ISSN 2444-5002

Volume 5, Issue 16 – July – December – 2019

Journal of Computational

Systems and ICTs

ECORFAN-Spain Journal of Computational Systems and Chief Editor ICTs, Volume 5, Issue 19, of July - December MIRANDA - TORRADO, Fernando. PhD 2019, is a journal edited sixmonthly by

ECORFAN. 38 Matacerquillas street, Postcode: Executive Director 28411. Moralzarzal – Madrid WEB: RAMOS-ESCAMILLA, María. PhD www.ecorfan.org/spain, [email protected].

Editor in Chief: MIRANDA - TORRADO, Editorial Director Fernando. PhD, ISSN On line: 2444-5002. PERALTA-CASTRO, Enrique. MsC Responsible for the latest update of this number

ECORFAN Computer Unit. ESCAMILLA- Web Designer BOUCHÁN, Imelda. PhD, LUNA-SOTO, ESCAMILLA-BOUCHAN, Imelda. PhD Vladimir. PhD, 38 Matacerquillas street,

Postcode: 28411. Moralzarzal – Madrid, last Web Diagrammer updated December 31, 2019. LUNA-SOTO, Vladimir. PhD

The opinions expressed by the authors do not Editorial Assistant necessarily reflect the views of the editor of the SORIANO-VELASCO, Jesús. BsC publication.

Translator It is strictly forbidden to reproduce any part of DÍAZ-OCAMPO, Javier. BsC the contents and images of the publication

without permission of the Centro Español de Philologist Ciencia y Tecnología. RAMOS-ARANCIBIA, Alejandra. BsC

Journal of Computational Systems and ICTs

Definition of Journal

Scientific Objectives

Support the international scientific community in its written production Science, Technology and Innovation in the Field of Engineering and Technology, in Subdisciplines of software analysis, assistance control, data acquisition, information systems, design and components, mobile devices, network systems, free software, virtual environments, multiprocessor systems, algorithm analysis, web application, network infrastructure and telecommunication.

ECORFAN- SC is a Scientific and Technological Company in contribution to the Human Resource training focused on the continuity in the critical analysis of International Research and is attached to CONACYT-RENIECYT number 1702902, its commitment is to disseminate research and contributions of the International Scientific Community, academic institutions, agencies and entities of the public and private sectors and contribute to the linking of researchers who carry out scientific activities, technological developments and training of specialized human resources with governments, companies and social organizations.

Encourage the interlocution of the International Scientific Community with other Study Centers in Mexico and abroad and promote a wide incorporation of academics, specialists and researchers to the publication in Science Structures of Autonomous Universities - State Public Universities - Federal IES - Polytechnic Universities - Technological Universities - Federal Technological Institutes - Normal Schools - Decentralized Technological Institutes - Intercultural Universities - S & T Councils - CONACYT Research Centers.

Scope, Coverage and Audience

Journal of Computational Systems and ICTs is a Journal edited by ECORFAN-Mexico S.C in its Holding with repository in Spain, is a scientific publication arbitrated and indexed with semester periods. It supports a wide range of contents that are evaluated by academic peers by the Double-Blind method, around subjects related to the theory and practice of software analysis, assistance control, data acquisition, information systems, design and components, mobile devices, network systems, free software, virtual environments, multiprocessor systems, algorithm analysis, web application, network infrastructure and telecommunication with diverse approaches and perspectives, That contribute to the diffusion of the development of Science Technology and Innovation that allow the arguments related to the decision making and influence in the formulation of international policies in the Field of Engineering and Technology. The editorial horizon of ECORFAN-Mexico® extends beyond the academy and integrates other segments of research and analysis outside the scope, as long as they meet the requirements of rigorous argumentative and scientific, as well as addressing issues of general and current interest of the International Scientific Society.

Editorial Board

HERNÁNDEZ - PRIETO, María de Lourdes. PhD Universidad Gestalt

TIRADO - RAMOS, Alfredo. PhD University of Amsterdam

VALERDI, Ricardo. PhD Universidad de Arizona

LÓPEZ - BONILLA, Oscar Roberto. PhD State University of New York at Stony Brook

CASTILLO - LÓPEZ, Oscar. PhD Academia de Ciencias de Polonia

ROBLEDO - VEGA, Isidro. PhD University of South Florida

CENDEJAS - VALDEZ, José Luis. PhD Universidad Politécnica de Madrid

RODRIGUEZ - ROBLEDO, Gricelda. PhD Universidad Santander

MARTINEZ - ALVARADO, Luis. PhD Universidad Politécnica de Cataluña

MAYORGA - ORTIZ, Pedro. PhD Institut National Polytechnique de Grenoble

Arbitration Committee

GONZÁLEZ - SILVA, Marco Antonio. PhD Universidad Nacional Autónoma de México

TZILI - CRUZ, María Patricia. PhD Universidad Politécnica del Valle de México

ORTEGA - CORRAL, César. PhD Universidad Autónoma de Baja California

SÁNCHEZ - HERRERA, Mauricio Alonso. PhD Instituto Tecnológico de Tijuana

PALAFOX - MAESTRE, Luis Enrique. PhD Centro de Investigación Científica y de Educación Superior de Ensenada

ORANTES - JIMÉNEZ, Sandra Dinorah. PhD Centro de Investigación en Computación

TORRES, Sandra. PhD Universidad Tecnologica Fidel Velazquez

VALDEZ - ACOSTA, Fevrier Adolfo. PhD Universidad Autónoma de Baja California

RODRÍGUEZ - AGUILAR, Rosa María. PhD Universidad Autónoma Metropolitana

RODRIGUEZ - CARVAJAL, Ricardo. PhD Universidad de Guanajuato

SOLORZANO - SALGADO, Paulina. PhD Universidad Autónoma de Querétaro

Assignment of Rights

The sending of an Article to Journal of Computational Systems and ICTs emanates the commitment of the author not to submit it simultaneously to the consideration of other series publications for it must complement the Originality Format for its Article.

The authors sign the Authorization Format for their Article to be disseminated by means that ECORFAN- Mexico, S.C. In its Holding Spain considers pertinent for disclosure and diffusion of its Article its Rights of Work.

Declaration of Authorship

Indicate the Name of Author and Coauthors at most in the participation of the Article and indicate in extensive the Institutional Affiliation indicating the Department.

Identify the Name of Author and Coauthors at most with the CVU Scholarship Number-PNPC or SNI- CONACYT- Indicating the Researcher Level and their Google Scholar Profile to verify their Citation Level and H index.

Identify the Name of Author and Coauthors at most in the Science and Technology Profiles widely accepted by the International Scientific Community ORC ID - Researcher ID Thomson - arXiv Author ID - PubMed Author ID - Open ID respectively.

Indicate the contact for correspondence to the Author (Mail and Telephone) and indicate the Researcher who contributes as the first Author of the Article.

Plagiarism Detection

All Articles will be tested by plagiarism software PLAGSCAN if a plagiarism level is detected Positive will not be sent to arbitration and will be rescinded of the reception of the Article notifying the Authors responsible, claiming that academic plagiarism is criminalized in the Penal Code.

Arbitration Process

All Articles will be evaluated by academic peers by the Double Blind method, the Arbitration Approval is a requirement for the Editorial Board to make a final decision that will be final in all cases. MARVID® is a derivative brand of ECORFAN® specialized in providing the expert evaluators all of them with Doctorate degree and distinction of International Researchers in the respective Councils of Science and Technology the counterpart of CONACYT for the chapters of America-Europe-Asia- Africa and Oceania. The identification of the authorship should only appear on a first removable page, in order to ensure that the Arbitration process is anonymous and covers the following stages: Identification of the Journal with its author occupation rate - Identification of Authors and Coauthors - Detection of plagiarism PLAGSCAN - Review of Formats of Authorization and Originality-Allocation to the Editorial Board- Allocation of the pair of Expert Arbitrators-Notification of Arbitration -Declaration of observations to the Author-Verification of Article Modified for Editing-Publication.

Instructions for Scientific, Technological and Innovation Publication

Knowledge Area

The works must be unpublished and refer to topics of software analysis, assistance control, data acquisition, information systems, design and components, mobile devices, network systems, free software, virtual environments, multiprocessor systems, algorithm analysis, web application, network infrastructure and telecommunication and other topics related to Engineering and Technology.

Presentation of Content

In the first article we present, Seven frame phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing, by CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ-REBOLLAR, Alicia and BALTAZAR-FLORES, María del Rosario, with ascription in the Instituto Tecnológico de León and National Center for Research and Technological Development, as next article we present, Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of , by RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan, with ascription in the Universidad Politécnica del Estado de Guerrero, as next article we present, Design and verification of a testing platform for implementation of software defined radio communications systems, by GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ- VENEGAS, Jorge, with ascription in CINVESTAV del IPN, as next article we present, RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems, by YAÑEZ- VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and AGUILERA-GONZALEZ, Gabriel, with ascription in the Universidad Politécnica de Juventino Rosas.

Content

Article Page

Seven frame phase-shifting algorithms for the three-dimensional profilometry 1-8 using fringe projection and soft computing CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia and BALTAZAR-FLORES, María del Rosario Instituto Tecnológico de León National Center for Research and Technological Development

Augmented Reality App for tourism promotion in the Triangle of the Sun of the 9-13 State of Guerrero RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan Universidad Politécnica del Estado de Guerrero

Design and verification of a testing platform for implementation of software 14-19 defined radio communications systems GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ-VENEGAS, Jorge CINVESTAV del IPN

RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR 20-26 Imagery Systems YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and AGUILERA- GONZALEZ, Gabriel Universidad Politécnica de Juventino Rosas

1 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

Seven frame phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing

Algoritmos de desplazamiento de fase de siete cuadros para la perfilometría tridimensional usando proyección de franjas y cómputo suave

CRUZ-LANDEROS, Jose Juan1†*, MOSIÑO, Juan Francisco1, MARTÍNEZ-REBOLLAR, Alicia2, BALTAZAR-FLORES, María del Rosario1

1Instituto Tecnológico de León 2National Center for Research and Technological Development

ID 1st Author: Jose Juan, Cruz-Landeros / ORC ID: 0000-0002-9110-2007, CVU CONACYT ID: 859653

ID 1st Coauthor: Juan Francisco, Mosiño / ORC ID: 0000-0002-3925-3051, CVU CONACYT ID: 25513

ID 2nd Coauthor: Alicia, Martínez-Rebollar / ORC ID: 0000-0002-1071-8599, CVU CONACYT ID: 217272

ID 3st Coauthor: María del Rosario, Baltazar-Flores / ORC ID: 0000-0002-8847-8732

DOI: 10.35429/JCSI.2019.16.5.1.8 Received July 23, 2019; Accepted September 18, 2019

Abstract Resumen

In this paper, we present a set of seven-frame algorithms En este artículo se presenta una familia de algoritmos de for the phase shifthing profilometry and their application siete cuadros para la perfilometria de desplazamiento de for the three-dimensional surface imaging using soft fase y su aplicación para la obtención de perfiles computing. This set obtained generate the most common tridimensionales usando cómputo suave. Esta familia seven-frame algorithms in the state of art. The algorithms obtenida genera la mayoría de los algoritmos de siete are analized for the compensation of common errors such cuadros más comunes en el estado del arte. Los as phase shifthing, phase errors because variation of light algoritmos son analizados en la compensación de errores or bias. comunes como el desentonamiento de fase, errores de fase debido a la variación de iluminación o bias. Fringe analysis, Phase-shifting interferometry, Softcomputing Análisis de franjas, Interferometría de corrimiento de fase, Cómputo suave

Citation: CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ-REBOLLAR, Alicia, BALTAZAR- FLORES, María del Rosario. Seven frame phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019, 5-16: 1-8.

* Correspondence to Author (email: [email protected]) † Researcher contributing first author.

© ECORFAN - Spain www.ecorfan.org/spain 2 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

Introduction Phase shifting interferometry is one of the best known structured light projection The world around us is three-dimensional; techniques (Geng, 2011). In one of the stages of however, cameras and image sensors have the this technique, a phase shifting algorithm (PSA) ability to acquire two-dimensional (Gutiérrez García, et al., 2013), also known as a representations without depth information. But filter, is applied. The objective of this paper is in recent decades there have been significant to present the proposal to define filters using a advances regarding research, development and soft computation method known as differential commercialization of technologies capable of evolution, which is an algorithm inspired by capturing the depth of a surface. There are nature used to find values that define a treatable numerous applications of these technologies, solution to a problem (Simon, 2013). The filters some of these are facial recognition, creation of of the experiments performed are defined for exact dental models, simulation of plastic scenarios where there is no noise, simulated surgeries, reverse product engineering, accident noise of light variation, vibration in the scene investigation (Geng, 2011) and digital environment; and these filters are compared fossil conservation (Gutiérrez García, and with the most used for 7 images. others, 2015). Phase shifting interferometry The creation of three-dimensional images refers to techniques which are capable In this technique, a set of sinusoidal patterns of acquiring some properties of three- is projected on the surface of the object, dimensional objects. Nevertheless, surface obtaining a set of images (Geng, 2011). This acquisition techniques allow obtaining the technique consists of the following steps measurement of the coordinates (x, y, z) of the (Ordoñes Nogales, 2016). points on the surface of an object. The result of these techniques can be considered as a map of a) images are captured with a constant the depth as a function of the position (x, y). shifting angle. (Geng, 2011) b) A phase shifting algorithm of frames One of the methods for obtaining is chosen or designed to process the set surfaces is based on the use of structured light, of images and obtain the wrapped which is the active lighting of a scene with a phase. special two-dimensional design, varying the intensity patterns. This technique extracts the c) An unfolding algorithm is used or depth of the object in the scene based on the designed to recover the desired phase distortion information of the projected (the surface). structured light (Geng, 2011), usually captured using a camera and represented in an image (I). d) The phase is interpreted in physical A representation of this technique can be seen measurements. in Figure 1.

1. Data 2. Demodulation 3. Phase acquisition unwrapping

4. Physical interpretation of the phase

Figure 2 Stages of the phase shifting technique Source: (Ordoñes Nogales, 2016) Figure 1 Representation of the structured light projection technique A representation of these steps is seen in Source: (Zuo, Huang, Zhang, Chen, & Asundi, 2016) Figure 2.

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 3 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

In this technique, each image is captured Where and are vectors of from the surface with the projection of a shifted coefficients for the numerator and denominator sinusoidal pattern with a phase angle or of the filter, respectively. These values define carrier, obtaining a set of captured images the filter (Mosiño, Malacara Doblado, & that is represented in an image vector Malacara Hernández, 2009). Since the value (Equation 1): is contained in the function and it is discontinuous, then a phase development (1) algorithm corresponding to the third stage of this technique is used; this stage is beyond the In Figure 3, 4 shifted projection patterns scope of this work, and since we worked with are seen at a angle. simulated surfaces, we avoid the need for this technique in experiments.

Projection 1 흎ퟎ푡 흎ퟎ In (Mosiño, et al., 2011) an expression (Equation 4) was presented to obtain the and Projection 2 coefficients that define the filter by establishing the points at which the filter cuts Projection 3 the Fourier space. Where each set defines the frequency to be cut from the Fourier impulse response of the Projection 4 filter. The full explanation of Equation 4 is outside the scope of this work, but can be found Figure 3 4 shifted projection patterns are seen at a in (Mosiño, et al., 2011). angle Source: Prepared by the authors ∑ ( ) ∑ [∑ } ]} = (4) ∑ ∑ ∑ } Then, assuming a small noise value, the ( ) [ ]} set of images obtained can be represented as: Equation 1

( ) ( ) ( ) ( ) Then, the creation of a filter corresponds to the (2) choice of . In this work, the

expression of Equation 4 is used to obtain the In Equation 2, is the backlighting, is values that, under certain noise scenarios, the contrast of the patterns (amplitude), ( ) better recover the original surface. represents the phase (the z-value of pixel depth

( )) and is the carrier. Differential evolution

Once the images have been taken, a It is one of the most successful methods for filter is applied to the captured images. There continuous optimization (Simon, 2013). It is are numerous filters that are used in the classified as a soft computing technique the temporal phase shift (Mosiño, Malacara method of which has been inspired by the Doblado, & Malacara Hernández, 2009) (Zuo, observation of nature (Simon, 2013); and Huang, Zhang, Chen, & Asundi, 2016). But specifically to the algorithms inspired by the filters are designed for uniform spacing and process and mechanism of biological evolution. ideal cases; however, when they are applied to This process of evolution describes how by real problems, they are not uniform due to the means of natural selection (descendants with vibration effectors or the variation in the modifications) a population of a species adapts intensity of the projected light. Vibration and through generations to certain problems light effects are reflected in the values , and (Brownlee, 2011). captured by the images. Differential evolution involves the The tangent of the estimated phase of maintenance of a population of candidate any quadrature filter with an order (for solutions for recombination, evaluation and images) is given by: selection.

( ) ∑ [ ] ⃗ (3) ∑ [ ]

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 4 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

Recombination involves the creation of Then, the simulation of the images was new candidate components of the solution done by establishing each one according to the based on the weighted difference between two model of the Equation 2, where ( ) are the randomly selected members of the population peaks surface height values, and and then a third member of the population. The and where }. basic differential evolution algorithm is presented in Figure 4.

Input: Parameters: 퐹(scaling factor), 퐶푅 (crossing constant). Initialize the population (uniform random distribution); Repeat For (푖 ,푖 ≤ 푘,푖 ) Do /*Each individual*/ Mutate and recombine 푗푟푎푛푑 푖푛푡(푟푎푛푑푖 ∙ 퐷) ; For (푗 , 푗 ≤ 퐷,푗 ) Do If (푟푎푛푑푗 < 퐶푅) or (푗 푗푟푎푛푑) Then 푢푖푗 풗풊풋 풙풓ퟑ풋 푭(풙풓ퟏ풋 풙풓ퟐ풋) Else 푢푖푗 푥푖푗 Replace: 푢푖(푡 ) 푺풊 푓(푢푖 (푡 )) ≤ 푓(푥푖(푡)) 푥푖(푡 ) ) 푥푖(푡) 푬풍풔풆 End for Figure 5 Peaks function in Matlab Until the termination condition /*example: number of Source: Prepared by the authors generations*/ Output: Better population or solution. 1.2. Implementation of the differential Figure 4 Pseudocode of the differential evolution evolution algorithm algorithm The differential evolution algorithm shown in In this paper, a population of size 40 Figure 4 was coded. In this implementation, the was used, where each individual corresponds to individuals to be optimized correspond to 4 Fourier cuts of the filter. The filter is generated a vector that contains the Fourier cut-off points of Equation 4 (for these using Equation 4 and the coefficients that define the filter are obtained to test the profile 7-image experiments, and were kept constant to maintain the cyclic filter obtained with the real image. The function to be characteristic in Fourier). The values of minimized (Equation 4) is the mean square remained in the evolution between 0 and 2π. error between the surface obtained by the filter While the and values observed in the and the original profile. pseudocode were set at 0.8 and 0.9 respectively, these values being recommended in (Simon, ( ) ( ) (5) 2013), 1000 periods were used in the experimentation. Equation 2

Methodology 1.3. Evaluation of the coefficients obtained from the filters This section describes the experiments performed in this work. They are presented in 4 The defined coefficients and were then distinct sections: Both the algorithms and the used by differential evolution to test the filter in simulation environment were carried out in IDE a simulation of the peaks function using 11

Matlab in its R2017b (9.3.0.713579) version. different carriers from to and the

1.1. Definition of the simulation quadratic error was obtained between the phase generated by the obtained filter and the original The simulation was carried out using the profile. Matlab peaks function (Figure 5) as a surface to be recovered, defined as a 200x200 surface and the reduced height values to 0.15% of the original values of the function to avoid the phase unfolding step in the experiments.

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 5 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

1.4. Comparison of results with state of Average of the quadratic error of the the art filters differential evolution in the 11 carriers during the 1000 periods in Experiment 1 In this section, 7 new simulations of captured

10000 images of the peaks function were generated for 0,01 the 11 carriers and common filters of 7 images 1E-08 (shown in Figure 6) and those generated from 1E-14

the differential evolution were used to recover 1E-20 Quadratic Quadratic error

the surface. In the end, the quadratic error 1E-26

1

57

393 505 617 169 225 281 337 449 561 673 729 785 841 897 953 between the recovered surface and the 113 simulated surface with the peaks function was Epoch calculated. Three comparison experiments were carried out. Experiment 1, without adding noise Graph 1 Average of the quadratic error of the to the images. Experiment 2, adding random differential evolution in the 11 carriers during the 1000 noise of 0.01% to component ; and experiment periods in Experiment 1 3, adding random noise of 0.01% to component Source: Prepared by the authors

of the images. Average of the quadratic error of the In Figure 6 of the comparison filters, differential evolution in the 11 carriers corresponds to the carrier value for which they during the 1000 epochs in Experiment 2

were designed, the diagram shows the Fourier 100000 cut-off points of the filter (angles and the 1000 direction is clockwise as it is customary to 10 illustrate for this type of filters). 0,1

Quadratic error 0,001

Number 1

Diagram Coefficients 57

169 281 225 337 393 449 505 561 617 673 729 785 841 897 953 of filters 113 Epoch

1 Graph 2 Average of the quadratic error of the

differential evolution in the 11 carriers during the 1000 periods in Experiment 2 Source: Prepared by the authors

2

Average of the quadratic error of the

differential evolution in the 11 carriers during the 1000 periods in Experiment 3 √ 3 100000

10000 1000 100 10 √ 4 1 Quadratic error

0,1 1

57

337 617 897 113 169 225 281 393 449 505 561 673 729 785 841 953 Epoch Figure 6 Common filters of 7 images

Graph 3 Average of the quadratic error of the Results differential evolution in the 11 carriers during the 1000 periods in Experiment 3 In Graph 1, Graph 2, and Graph 3 the average Source: Prepared by the authors of the quadratic error of the differential evolution is shown for the 11 carriers evaluated Table 1 and Table 2 are the results of in the 3 experiments (the quadratic error is experiment 1. Table 1 shows the individuals expressed in base 10 logarithmic scale), a selected by the differential evolution algorithm similar behavior is observed, as well as the fact for each carrier of the experiment. These that the error starts too high converging early in individuals contain the , , , and that the period at a value close to 0. define the filter.

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 6 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

t Tables 3 and 4 show the results of

3.39341494 0.31415784 0.31416257 0.31416178 experiment 2. Adding a noise level of 0.01% to

0.56584501 0.56548641 0.56548579 3.68985352 component of the images. Table 3 shows the

values and obtained in the 0.81684604 0.81681588 3.75344502 0.81680483 differential evolution. It can be noted how the

3.44769663 1.06814547 1.06814117 1.06810668 values differ more from the carrier than in

1.31944629 1.31950415 1.31891933 1.31880409 experiment 1.

1.56295155 1.57069825 1.57079604 1.56978077

1.82213 1.82210915 1.82213792 1.5884965 0.31415927 3.84595269 4.3097081 3.21293577

2.073 2.073454 2.07345115 0.01856206 0.56408641 3.47891916 3.47891919 3.47891914

2.32477803 0.00089924 2.32478016 2.32494418 0.81681818 0.68943062 0.81683439 4.6475956

2.57611004 0.00094596 2.57609418 2.57611317 0.42835963 0.42835958 0.42835953 0.97643749

2.82743408 0.20611196 2.82747649 2.82743618 4.34392235 1.31946884 0.24623126 1.31946627

2.88707736 2.88707727 1.58142495 2.88707767

Table 1 Values , , , and in Experiment 1 1.82214712 0.08621156 4.86019418 1.8220794 Source: Prepared by the authors

4.66441939 2.07345217 3.32705553 2.07345113

It is observed how many of the values 2.32477856 0.03785148 2.61020667 4.95782448 obtained by the differential evolution

2.37501618 2.37501623 5.39835857 2.37501571 correspond to the carrier value, others to values close to 0, and others to values that correspond 1.02587409 0.76972043 2.82743339 4.36500391 to harmonics of the carrier.

Table 3 Values , , , and in Experiment 2 Quadratic Quadratic Quadratic Quadratic Quadratic error Source: Prepared by the authors error error error error Filter Filter 1 Filter 2 Filter 3 Filter 4 obtained

2.0267E-26 1289.17545 709.193493 1011.83758 2608.92567 And Table 4 shows the quadratic error

6.0748E-28 495.490852 62.8168649 91.3724312 483.399975 between the original surface peaks and the

surface obtained by the filters generated in 3.1399E-28 149.023385 4.27249863 2.9615834 55.3055054

experiment 2. It is observed that it obtains 1.3602E-28 28.4029947 0.13122533 0.00014448 0.23972935 better results in general for 7 values of the

1.8345E-28 1.73799776 0.00044905 3.25728372 20.5986079 carrier, so this procedure reflects good results

1.8241E-28 2.4351E-28 1.7947E-28 38.0972638 33.5218508 against 0.01% noise in component of the

2.023E-28 1.73799776 0.00044905 167.890785 20.5986079 images.

3.9314E-28 28.4029947 0.13122533 509.098423 0.23972935 Quadratic Quadratic Quadratic Quadratic Quadratic error error error error error 1.1446E-27 149.023385 4.27249863 1303.08776 55.3055054 Filter Filter 1 Filter 2 Filter 3 Filter 4 obtained 8.6198E-27 495.490852 62.8168649 3139.45571 483.399975 11.3156004 1332.04539 829.635101 1052.93484 2682.98442

2.9105E-25 1289.17545 709.193493 7858.36742 2608.92567 1.85225274 536.444531 44.4973128 73.8519551 438.797349

2.06810974 145.131457 4.93457309 3.27576934 56.2346913

Table 2 Quadratic error between the original surface and 27.3841552 31.1390406 2.3100546 0.0915791 0.50409375 that obtained by the filters in Experiment 1

Source: Prepared by the authors 0.12627019 1.65109314 0.37787341 4.75278043 18.1970708

0.48040327 0.00288633 0.01147011 37.0954392 35.5608772

Table 2 contains the quadratic error between the simulated original surface of the 0.48816892 1.56377579 0.0331104 163.054158 23.6688965 peaks function and the surfaces recovered from 70.2612303 27.8119454 2.82438951 5753.21601 5556.94097

experiment 1. It is observed that the filter 0.59793114 154.792944 4.11593077 1228.03868 49.9164952 obtained for each carrier generally obtained 5.3790476 475.695833 73.4079079 2989.77769 449.976993 better results than the other filters compared 1.41064121 1262.87099 747.974768 8123.44365 2767.71522 here, the filter error value that was lower for each carrier is shown in bold (the same is shown for Table 4 and Table 6). Table 4 Quadratic error between the original surface and that obtained by the filters in Experiment 2 Source: Prepared by the authors

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 7 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

Tables 5 and 6 show the results of Conclusions experiment 3, where noise is added to the images in component . Table 5 contains the In this paper a different way to the state of the corresponding values and art was tested in the use of soft computing for obtained. Table 6shows the quadratic errors the technique of phase shifting interferometry. for experiment 3. Table 5 shows how the We looked for filters that are a treatable values obtained are varied, trying to solution at different noise levels in obtaining compensate for the error in . Table 6 shows image captures. Based on the results section, how the filter obtained generates bad the differential evolution seems to obtain approximations of the original surface with the satisfactory filters for noise of 0.01% in the filters created. Although the other filters also do component of the model of the images, while it not have considerably good results in this type does not obtain good results against noise of of noise. 0.01% in the component of the captured images. It was also observed that the periods in

which the differential evolution reaches values

3.32920506 1.30797008 0.18626818 4.53170912 close to 0 are early, so it is inferred that it has

1.04400711 0.41752438 1.48632945 3.59359454 been a problem not so complicated to solve for

the technique or that it has stagnated in local 2.41891002 4.67177629 0.84677996 4.77445194 optimums. This method could be improved and

3.78062461 1.34567746 1.7207E-15 1.34567746 analyzed by evaluating other noise scenarios or

6.09515811 1.30248931 4.52554356 2.21220134 using a larger filter search space, that is, with

more images. 1.13518769 5.09820149 1.13518767 1.13518751

1.79390759 1.03679378 3.38555486 4.90012928 References

1.87011461 3.53110598 1.83660486 5.26091338

Brownlee, J. (2011). Clever Algorithms Nature- 0.08522184 2.29078968 0.08522179 6.28318531 Inspired Programming Recipes.

1.31637597 1.31637598 2.51937157 1.31637604

0.32691101 0.21442545 2.82578976 5.31039657 Geng, J. (2011). Structured-light 3D surface imaging: a tutorial. (O. S. America, Ed.) Advances in Optics and Photonics, 128-160. Table 5 Values , , , and in Experiment 3 Source: Prepared by the authors Gutiérrez García, J. C., Gutiérrez García, T. A., Quadratic Quadratic Quadratic Quadratic Quadratic Mosiño, J. F., Vázquez Domínguez, E., error t error error error error Filter Martínez, A., & Arroyo, C. J. (2015). A novel Filter 1 Filter 2 Filter 3 Filter 4 obtained application of the white light/fringe projection 372.645432 2615.21151 1367.18855 1313.2444 2966.18916

duo: recovering high precision 3-D images 31.7460575 726.451404 42.8653768 59.8989512 409.622363 from fossils for the digital preservation of

82.319874 224.439255 22.1017502 13.3538321 36.6120774 morphology. Paleontología Electronica , 1-13.

25.9686161 36.6630661 25.8437282 36.6350729 37.7581843

35.3537115 2.3875674 4.09725675 21.6537373 56.5979154 Gutiérrez García, J. C., Mosiño, J. F., Martínez,

22.3487838 5.15276423 1.42260927 109.567887 5.87751501 A., Gutiérrez García, T. A., Vázquez

Dominguez, E., & Arroyo Cabrales, J. (2013).

101.036962 7.01629465 0.58385936 287.474212 35.3712173 Practical eight-frame algorithms for fringe

28.6485437 26.1559882 10.0363416 42613.6722 30350.6701 projection profilometry. Optics Express, 21(1),

29.7515912 109.889506 9.05457366 1824.03349 693.502067 903-917.

7.123764 607.816808 44.8672903 6808.00038 2894.14073

Mosiño, J. F., Gutiérrez-García, J. C., 41.6824643 2035.08558 459.75393 10437.1562 4114.31313 Gutiérrez-García, T. A., Castillo, F., García- González, M. A., & Gutiérrez-García, V. A. Table 6 Quadratic error between the original surface and (2011). Algorithm for phase extraction from a that obtained by the filters in Experiment 3 Source: Prepared by the authors set of interferograms with arbitrary phase shifts. Optics Express, 19(18), 4908-4923.

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 8 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 1-8

Mosiño, J. F., Malacara Doblado, D., & Malacara Hernández, D. (2009). Calculus of exact detuning phase shift error in temporal phase shifting algorithms. Optics Express, 17(18), 15766-15771.

Ordoñes Nogales, S. (2016). Algoritmo Iterativo Eficiente para el Análisis de Interferogramas con Corrimiento de Fase Aleatorio. Universidad de Guadalajara, Centro Universitario de Ciencias Exactas e Ingenierías, Guadalajara.

Simon, D. (2013). Evolutionary Optimization Algorithms. New Jersey: John Wiley & Sons, Inc.

Zuo, C., Huang, L., Zhang, M., Chen, Q., & Asundi, A. (6 de Mayo de 2016). Temporal phase unwrapping algorithms for fringe projection profilometry: A comparative review. Optics and Lasers in Engineering, 84-103.

ISSN 2444-5002 CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ- REBOLLAR, Alicia, BALTAZAR-FLORES, María del Rosario. Seven frame ECORFAN® All rights reserved phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing. Journal of Computational Systems and ICTs. 2019. 9 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 9-13

Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero

Aplicación de Realidad Aumentada para promoción turística en el Triángulo del Sol del Estado de Guerrero

RÍOS, David†*, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan

Universidad Politécnica del Estado de Guerrero

ID 1st Author: David, Ríos / ORC ID: 0000-0002-0494-7045, CVU CONACYT ID: 739442

ID 1st Coauthor: Daniel, Urueta / ORC ID: 0000-0002-8741-6978

ID 2nd Coauthor: Reynaldo, Alanís / ORC ID: 0000-0001-5397-7016

ID 3rd Coauthor: Juan, Mota / ORC ID: 0000-0001-6099-9408, CVU CONACYT ID: 240517

DOI: 10.35429/JCSI.2019.16.5.9.13 Received July 07, 2019; Accepted September 25, 2019

Abstract Resumen

To show information about the three main tourist Mostrar información sobre los tres destinos turísticos destinations in the state of Guerrero: “ de principales del estado de Guerrero que son: Acapulco de Juárez”, “Ixtatapa-” and “ de Alarcón” Juárez, Ixtatapa-Zihuatanejo y Taxco de Alarcón, hacia a to users using an application for devices with Android los usuarios haciendo uso de una aplicación para Operating System with Augmented Reality, a dispositivos con Sistema Operativo Android con promotional video reproduced by recognizing the target Realidad Aumentada, un video promocional que se of the city and showing information like activities or reproduzca al reconocer el target de la cuidad y muestre attractions of each of these destinations; these images información como actividades o atracciones de cada uno will be on a poster located at different terminals within de estos destinos, estas imágenes estarán en un cartel the Mexican Republic and the state of Guerrero, thus ubicado en las diferentes terminales dentro de la achieving greater impact and interest in national users. República Mexicana y del estado de Guerrero, y así lograr un mayor impacto e interés en los usuarios mayor Augmented Reality, Triangle of the Sun, Tourism atracción turística por parte de usuarios Nacionales.

Realidad Aumentada, Triangulo del Sol, Turismo

Citation: RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan. Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero. Journal of Computational Systems and ICTs. 2019, 5-16: 9-13

* Correspondence to Author (email: [email protected]) † Researcher contributing first author.

© ECORFAN - Spain www.ecorfan.org/spain 10 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 9-13

Introduction 1.1.2 Levels of augmented reality

The relationship which existed in the twentieth There are different levels of augmented reality, century between education and technology has numbered from 0 to 3; these are (Fitzgerald, become close since the twenty-first century; 2009): making it easier for students to learn by using today's technologies, such as laptops, wireless • Level 0: Hyperlinks in the physical transmission devices, like GoogleCast, world. By reading a barcode or QR, smartphones, and even tablets. hyperlinks are made to other content such as websites, there is no 3D In addition, the development of mobile registration or bookmark tracking. applications aimed at various educational areas, such as: Sexual and Reproductive Education, • Level 1: Augmented reality based on Languages, Mathematics, Economics and markers. In this level, the activators are Tourism, etc., are aimed at promoting and markers, these are black and white improving cognitive learning through the use of images, generally square, with simple augmented reality; even cultural heritage sites and asymmetrical drawings that portrays have made use of this tool (Urueta, 2017). a virtual object which overlaps with the real image when scanned. The problem that exists when visiting certain sites is bad promotion in diverse means • Level 2: Augmented reality without of information like social networks. This markers. The activators are images, research proposes an application of augmented objects or GPS locations by which POI reality in the area of Economy and Tourism to (points of interest) are superimposed on promote the tourist attractions of the Sun real-world images. Triangle of the state of Guerrero as an alternative to the different conventional forms Methodology of tourism promotion. 2.1 Application development 1.1 Concepts and definitions Through a mobile device with Android 1.1.1 Augmented Reality operating system version 4.4+, the user can view a 30-minute promotional short video of a Augmented reality (AR) is a variation of virtual place within the Sun Triangle, allowing and environments (VE) or virtual reality as it is encouraging visitors to go to that place, through commonly called. the interest caused by the video. The general operation is presented in the following Figure: VE technologies completely immerse the user in a synthetic environment; while Augmented reality immersed, the user cannot see the real world android application around them. In contrast, AR allows the user to see the real world with virtual objects that are superimposed or in composition with an image of the real world, obtained through a screen, Pattern search using 3D models or other computer generated information. That is, augmented reality supplements reality rather than completely replacing it (Azuma, 1997). Identified pattern

According to the aforementioned, AR possesses the following three characteristics (Espinosa, 2015): Promotional site display 1. Combines the real and the virtual 2. Interactive and real time 3. Registered in 3D Figure 1 General operation

RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan. ISSN 2444-5002 ECORFAN® All rights reserved Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero. Journal of Computational Systems and ICTs. 2019. 11 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 9-13

To carry out the visualization of the video, three images that are unique to said tourist place were edited as shown in Figure 2.A and Figure 3.A, to make use of them in the application. The edited images were subjected to an editing process to be able to attract more public attention as shown in Figure 2.B.

Figure 3.B Image with detected patterns

A challenge faced in the development of the application was the selection of the massive tourist information on these selected destinations; however, videos of shorter duration and with more striking content were

selected, as shown in Figure 4. Another Figure 2.A Raw Image challenge is the lack of platforms as official repositories of tourism material, which is why we opted to embed the video within the application, greatly increasing the size of it.

Figure 2.B Edited image of the tourist place

From the edited images, with the Vuforia tool, they were subjected to a process Figure 4 Application execution for pattern identification, obtaining the "target" as shown in Figure 3-B. 2.2 App Evaluation

After the creation of the application, 30 people were randomly selected in the Palmas terminal in Cuernava and 30 people in the Iguala terminal in Guerrero, addressing the issue of activities that can be carried out in these tourist spots through Augmented Reality, in addition to having a mobile device with Android 4.4+.

Half of them had the application installed (Group A) and the other half, not (Group B) and had a verbal presentation. Following the use of the application or verbal explanation, a brief questionnaire was made Figure 3.A Image of the tourist place (see Appendix). The results are shown below.

RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan. ISSN 2444-5002 ECORFAN® All rights reserved Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero. Journal of Computational Systems and ICTs. 2019. 12 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 9-13

20,5 There was a greater interest in traveling 20 20 and knowing the activities and places shown in 20 the video after using the application. 19,5 19 It is also possible to conclude that, with 18,5 the help of digital technology, tourism 18 promotion can be improved, this proposal can 18 be expanded by creating an official state 17,5 repository with this information. 17 Future Work

Compatible devices will be extended so that the

Question 1 Question Question 2 Question 3 Question application can also work with an iOS system.

Figure 5 Questions about the application The development of new tourism Additionally, from question 4, points promotion materials in the Sun Triangle will could be obtained to improve the application; continue; it is also expected to reduce the size these include the following: being able to use it of the application through the use of video with iOS devices and reducing the size of the streaming, with which it is expected to focus application. more attention on the promotion and therefore increase of tourism within the state of Guerrero. In the case of question 5, the holiday destination of each user after having used the References application or listened to the talk is shown in Figure 6. Azuma, R. T. (1997). A Survey of Augmented Reality. Presence: Teleoperators and Virtual Environments, 6(4), 355–385. 25 23 20 20 17 Espinosa, Carlos Prendes. "Realidad aumentada y educación: análisis de experiencias prácticas." 15 Pixel-Bit. Revista de Medios y Educación 46 10 (2015): 187-203. 5 Urueta Hinojosa, D., Ríos Mendoza, D., 0 Vázquez Calderón, H., "Realidad aumentada e infografías como medio para educar sobre el

Taxco embarazo no deseado", procc., 17 Congreso de - Acapulco Investigación en Salud Pública; Cuernavaca, Zihuatanejo Morelos, 1-3 March 2017.

Figure 6 Holiday destination chosen by respondents

Results

In this paper, an augmented reality application was developed for Android devices which show activities and places in a tourist destination within the Sun Triangle in the state of Guerreo.

Conclusions

The application had a high level of acceptance by the users; however, it can improve substantially due to the comments provided by users.

RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan. ISSN 2444-5002 ECORFAN® All rights reserved Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero. Journal of Computational Systems and ICTs. 2019. 13 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 9-13

APPENDIX

Questionnaire used in the investigation

Instructions and considerations.

• Read carefully before choosing an answer. • If the application was not installed on your device, answer only question 5

1. Do you think the application is engaging?

a) Yes b) No

2. Do you think that the application can influence the visit of said tourist place?

a) Yes b) No

3. Would you recommend the use of the application?

a) Yes b) No

4. What would you like to include in the application?

5. Select one place you would like to visit on your next holiday.

a) Acapulco b) Ixtapa – Zihuatanejo c) Taxco

RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan. ISSN 2444-5002 ECORFAN® All rights reserved Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero. Journal of Computational Systems and ICTs. 2019. 14 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 14-19

Design and verification of a testing platform for implementation of software defined radio communications systems

Diseño y verificación de una plataforma de prueba para la implementación de sistemas de comunicación basados en radio definido por software

GARCÍA-DZUL, Rogelio†*, PARRA-MICHEL, Ramón and SÁNCHEZ-VENEGAS, Jorge

CINVESTAV del IPN, Unidad Guadalajara, Laboratorio de Telecomunicaciones

ID 1st Author: Rogelio, García-Dzul / ORC ID: 0000-0002-2334-3060

ID 1st Coauthor: Ramón, Parra-Michel / ORC ID: 0000-0003-2327-2482

ID 2nd Coauthor: Jorge, Sanchez-Venegas / ORC ID: 0000-0001-5342-3400

DOI: 10.35429/JCSI.2019.16.5.14.19 Received July 25, 2019; Accepted September 05, 2019

Abstract Resumen

The software-defined radio platforms that currently exist Las plataformas de radio definido por software que are mainly focused on the development of software existen actualmente se centran principalmente en el communication blocks, leaving aside the hardware desarrollo de bloques de comunicaciones mediante development, so the need arises to have a testing platform software dejando de lado el desarrollo de hardware, por for hardware blocks which also allows us to interact with lo que surge la necesidad de contar con una plataforma de a radio frequency stage to perform blocks tests of prueba para bloques de hardware la cual además nos complete communication systems in the air such as WiFi, permita la interacción con una etapa de radio frecuencia Bluethoot, Zig Bee among others. This article describes para realizar pruebas de bloques de sistemas de the design of a platform based on a Field-Programmable comunicaciones completos en el aire tales como WiFi, Gate Array (FPGA) technology and microprocessor, Bluethoot, Zig Bee entre otros. En este artículo se which allows having multiple blocks and algorithms in describe el diseño de una plataforma basada en la hardware and software and interacting between them. In tecnología de matriz de puertas programables (FPGA, por the same way there is a significant number of general sus siglas en inglés) y microprocesador, la cual permite purpose interfaces for testing the signals, as well as tener multiples bloques y algoritmos en hardware y having specific interfaces which allow us to interconnect software e interactuar entre estos. De igual manera se with different radio frequency platforms and cuenta con un número significativo de interfaces de communication cards with the computer. This allows you propósito general para prueba de las señales, al igual que to work with different radio technologies according to tiene interfaces específicas las cuales nos permiten la your needs. A use case is presented in which a Bluethoot interconexión con diferentes plataformas de radio transmitter is tested with the use of the platform. frecuencia y tarjetas de comunicación con la computadora. Esto permite trabajar con diferentes Software defined radio, FPGA platform tecnologías de radio según sean las necesidades. Se presenta un caso de uso en el cual se prueba un transmisor de Bluethoot con el uso de la plataforma.

Radio definido por software, plataforma FPGA

Citation: GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ-VENEGAS, Jorge. Design and verification of a testing platform for implementation of software defined radio communications systems. Journal of Computational Systems and ICTs. 2019, 5-16: 14-19.

* Correspondence to Author (email: [email protected]) † Researcher contributing first author.

© ECORFAN - Spain www.ecorfan.org/spain 15 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 14-19

1. Introduction In the same way, there is a significant number of general purpose interfaces for testing Communication systems are made up of the signals, as well as having specific interfaces different stages such as coding, modulation, that allow us to interconnect with different demodulation, and decoding. From that base radio frequency platforms and communication have emerged diferents wireless standards such cards with the computer. This allows working as WiFi, Bluetooth, ZigBee, LTE and others, with different radio technologies according to each of these include different architectures and the needs of the final user. algorithms. In the same way, a case of use of the The need for implementation and testing platform is presented, where it is used to carry of different communication blocks to be able to out the tests of a Bluetooth transmitter through have specific communication systems such as the implementation of a modulator and a MAC those mentioned above has given rise to the layer of the standard. emergence of new tools and technologies such as software defined radio (SDR). 2. Design

The SDR technology allows us to make As mentioned in the previous section, the significant changes in the hardware through platform should allow the user flexibility and changes through the software, this gives a accessibility to be able to make different significant advantage in terms of costs and ease configurations in each of the elements, that is of operation and implementation of different why the platform should integrate the sections wireless communication systems. mentioned below. Figure 1 shows the blocks that integrate the platform, which are described The implementation of algorithms, below. modulators, complete standards and in general blocks of communications can be done through 2.1. FPGA software and/or hardware, the latter through hardware description languages such as The integration of an FPGA in the platform is Verilog. of vital importance since it will allow us to implement algorithms and communication The software-defined radio platforms blocks through hardware description language. that currently exist focus mainly on the With this element, it will be possible to development of communication blocks in configure external elements such as the clock software and in the case of having some tools generator and the RF section of which will be for implementation in hardware such as a discussed later. The FPGA included in the FPGA, the end user has limited access to this. platform will be from Intel.

The implementation of blocks in Flash hardware is essential if you want, for example, PCinterface CLK memory generator to have an integrated circuit (IC) as the final SPI producer. This is why there is a need among hardware designers to have a test platform for connector RF hardware blocks which provides the user JTAG flexibility of operation and interaction between FPGA hardware and software, in addition to having SRAM general purpose pins for data analysis and in a similar way to perform wireless communication tests. PinHeader Micro controller

This paper presents the design of a test PinHeader platform for the implementation of USBOTG SWD communication systems in hardware and software and interacting among them. Figure 1 Platform block diagram

GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ- ISSN 2444-5002 ECORFAN® All rights reserved VENEGAS, Jorge. Design and verification of a testing platform for implementation of software defined radio communications systems. Journal of Computational Systems and ICTs. 2019. 16 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 14-19

To be able to use the FPGA it is 2.4. Interfaces necessary to have at least one way of programming it, that is why together with the The platform has input/output pins that are FPGA two interfaces are included, one of directly connected to the FPGA, these allow us programming through JTAG and the other with to have a visualization of desired signals, data programming by means of a programming entry, connection with other devices, as well as memory, which allows having a pre-loaded the interconnection between two or more test program when the platform is turned on. platforms.

To give the user an additional memory One of the fundamental requirements in to the one contained in the FPGA, an SRAM having an interface for the connection of a memory was integrated, which is completely commercial RF card, that is why an RF accessible through this same element. connector was included which is adapted to the Myriad commercial card, which includes an 2.2. Microcontroller LMS6002D transceiver. This transceiver performs its configuration by writing data in its An STM microcontroller is included in the registers through the Serial Peripheral Interface platform which will allow us to implement (SPI) protocol. The SPI block will be algorithms in programming languages, just as it implemented within the FPGA and will be will have complete communication with the discussed later. FPGA through general purpose input/output pins; this allows us to have interaction between 3. Development the algorithms that are implemented in the two different elements of the platform. It is necessary to be able to configure the clock generator and the RF stage, which is why the In a similar way to the FPGA it is I2C and SPI blocks were implemented, which necessary to have a programming interface for allow us to receive the data from the computer the microcontroller, additionally, it has a USB and send them to the corresponding OTG connection, which allows the interaction configuration registers. of the microcontroller with the computer. 3.1. IP-SPI 2.3. Clock generator The IP-SPI block has connections for a The radio frequency (RF) card to which the communication interface with the computer, as platform is adapted requires different clock well as the connections that go to the radio frequencies for the operation of the transceiver, interface which correspond to an SPI as well as to be able to make use of different connection. sampling rates according to the need of the user. For this, a clock generator is included in This block will allow us to write the the platform, which, from an input clock, desired configuration in the LMS6002D chip generates 8 independent clock outputs that can registers in order to obtain, for example, the be adjusted between a range of 8 KHz to 160 desired transmission and/or reception MHz. frequency, bandwidth, gains of the amplifiers, among other configurations. In the same way, it To adjust the output frequencies, it is allows us to read the values of the registers to necessary to configure the clock generator, be able to see the current configurations and which is done by sending values to the internal perform the necessary calibrations within the registers of the chip through the Inter-Integrated transceiver. Circuit (I2C) protocol. This configuration can be done through the FPGA or the The configuration registers of the microcontroller, with the FPGA it is necessary LMS6002D are divided into 8 logical blocks. to have a block in hardware that allows to The write/read are made through 16 cycles of receive the configuration data and send them SPI clock with which the same number of bits through the protocol which will be discussed is sent, the most significant bit corresponds to a later. read or writes command indicating it with a 0 or 1.

GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ- ISSN 2444-5002 ECORFAN® All rights reserved VENEGAS, Jorge. Design and verification of a testing platform for implementation of software defined radio communications systems. Journal of Computational Systems and ICTs. 2019. 17 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 14-19

Of the remaining 15 bits, the 3 more significant is used to indicate which logical block to configure and the next 4 bits indicate a particular address within the block.

The remaining 8 bits are the data to be written in the register chosen in the case of the writings and in the readings only the clock cycles are used to read the register value and send it to the SPI block. Table 1 shows the logical blocks of the LMS6002D.

Address Description 000:XXXX Top level 001:XXXX TX PLL 010:XXXX RX PLL 011:XXXX TX LPF Figure 2 SW-HW Testing platform 100:XXXX TX RF 101:XXXX RX LPF, DACs and ADCs 110:XXXX RX VGA2 Given the connections between the 111:XXXX RX RF FPGA and the microcontroller, it is important to mention that the configuration of the Table 1 Memory map SPI LMS6002D aforementioned elements is not exclusive of the FPGA since they can be done in the same way 3.2. IP-I2C by means of the microcontroller.

As mentioned above, this block will allow us to 4. Case of use configure the Si5351C clock generator of Silabs in order to give the desired frequency to each of The blocks used to perform the tests are shown the elements of the platform. Similar to the IP- in Figure 3, it is noted that in addition to the SPI, this block includes a connection to a Bluetooth block the I2C and SPI configuration communication interface with the computer, but blocks are necessary. with the diference in the connection on the other side of the block corresponds to an I2C SPI interface. SEN SCLK SPI & SDIO I2C SDO Clock The Si5351C supports read and write SCL SDA generator operations for its 255 registers, this through the Interface I2C CLK Tx_CLK CLK_IN corresponding I2C commands. This IC supports Tx_EN Tx_IQ_SEL LMS6002D writes and reads of a single record, as well as Bluetooth Data_Tx 12 reading of the record in burst mode, that is, TX reading or writing in a certain number of FPGA consecutive records, since the address auto- Platform increment. Figure 3 Block diagram for the BLE transmitter tests The configurations available for this IC through this block are read the status indicator, A use case is presented, which consists set the multiplication and division values to in the use of the platform to test a physical obtain the desired frequency, enable or disable architecture and MAC layer of a reconfigurable each of the clock outputs, and compensation of Bluetooth transmitter developed by JM the exit phase. Sánchez-Venegas whose detailed architecture can be seen in (Sánchez-Venegas, 2018).

Figure 4 shows the blocks that make up the architecture of Bluetooth transmitter.

GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ- ISSN 2444-5002 ECORFAN® All rights reserved VENEGAS, Jorge. Design and verification of a testing platform for implementation of software defined radio communications systems. Journal of Computational Systems and ICTs. 2019. 18 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 14-19

To carry out the tests, a commercial Bluetooth dongle receiver was used, which is configured in the listening mode of the Bluetooth LE advertising packages; to make use of this device it is necessary to have a computer with Linux operating system to be able to execute the corresponding commands. Figure 6 shows the results of the tests performed.

Figure 4 Architecture overview of PHY and MAC layers of all-digital Bluetooth transmitter Source: (Sánches-Venegas, 2018)

The physical layer of the modulator corresponds to the architecture of a GFSK modulator, which is completely digital and is based on a CPM modulator. This transmitter is capable of generating any CPM scheme with an arbitrary index, but for this example, it is adapting to comply with the Bluetooth 4.0 Figure 6 Decoded advertising LE 1M PHY packet on commercial Bluetooth dongle. standard. Source: (Sánches-Venegas, 2018)

The architecture of the MAC layer has 5. Conclusions two interfaces, the first to communicate with a microprocessor, through which the module is In this work a platform was developed that configured and sends the payload of the allows the easy prototyping and testing of packets. The second interface is for blocks of hardware and software oriented to communication with the wireless communication systems and modulator/demodulator. The MAC layer is communications in general. responsible for sending a bit stream of a defined packet to the modulator, which contains the Being able to make the configurations of device information such as the MAC address the elements that make up the platform through and the device name. Similarly, it receives the the FPGA or the microcontroller gives the user bits from the demodulator and processes them. versatility and likewise the opportunity to have The use of the platform in this test is shows in available the full capacity in available logic the Figure 5. elements of the FPGA, which is a problem in the SDR platforms from the market.

Another advantage with respect to commercial SDR platforms is the possibility of being able to carry out tests with different transceivers given the connection with the platform, which allows you to test your own RF card designs.

6. Acknowledgements

The present work was supported by ANR- CONACYT TOLTECA project No. 273562.

Figure 5 Transmitter testing GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ- ISSN 2444-5002 ECORFAN® All rights reserved VENEGAS, Jorge. Design and verification of a testing platform for implementation of software defined radio communications systems. Journal of Computational Systems and ICTs. 2019. 19 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 14-19

7. References

Burns, P. (2003). Software Defined Radio for 3G. Artech House, Inc.

Harada, H., & Prasad, R. (2002). Simulation and software radio for mobile communications. Artech House.

Mazover, V. K. (2015). Fundamentos de comunicaciones digitales. Limusa.

Microsystems, L. (2019). LMS6002D Datasheet. Retrieved from Lime Microsystems: www.limemicro.com

Sánchez-Venegas, J. M., Ramírez-Pérez, A., Jaramillo-Ramírez, R., & Parra-Michel, R. (2018). An all-digital physical and MAC layer architectutes for a reconfigurable Bluetooth transmitter. 2018 IEEE 9th Latin American Symposium on Circuits & Systems (LASCAS).

Silabs, S. (2019). Silicon Labs. Retrieved from https://www.silabs.com/

GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ- ISSN 2444-5002 ECORFAN® All rights reserved VENEGAS, Jorge. Design and verification of a testing platform for implementation of software defined radio communications systems. Journal of Computational Systems and ICTs. 2019. 20 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems

Fusión de imágenes RASF y DEDR-POCS a través de perceptron multicapa en sistemas SAR

YAÑEZ-VARGAS, Israel†*, QUINTANILLA-DOMÍNGUEZ, Joel and AGUILERA-GONZALEZ, Gabriel

Universidad Politécnica de Juventino Rosas, Departamento de Ingeniería Telemática, Hidalgo 102, Comunidad de Valencia, Santa Cruz de Juventino Rosas, Gto.

ID 1st Author: Israel, Yañez-Vargas / ORC ID: 0000-0001-5749-8442, CVU CONACYT ID: 295711

ID 1st Coauthor: Joel, Quintanilla-Domínguez / ORC ID 0000-0003-2442-2032

ID 2nd Coauthor: Gabriel, Aguilera-Gonzalez / ORC ID: 0000-0002-4160-448X

DOI: 10.35429/JCSI.2019.16.5.20.26 Received July 18, 2019; Accepted September 11, 2019

Abstract Resumen

This paper presents a novel multi-layer perceptron (MLP) Este artículo presenta una novedosa técnica de fusión de based image fusion technique, which fuses two synthetic imágenes basada en perceptrón multicapa (MLP), que aperture radar (SAR) images, obtained from the same fusiona dos imágenes de radar de apertura sintética spatial reflectivity map, acquired with a conventional (SAR), obtenidas a partir del mapa de reflectividad low-cost fractional synthetic aperture radar (Fr-SAR) espacial, adquiridas con un radar de apertura sintética system, enhanced via two different methodologies. The fraccional de bajo costo convencional (Fr- Sistema SAR), first image is enhanced using the traditional descriptive mejorado a través de dos metodologías diferentes. La experiment design regularization (DEDR) framework primera imagen se mejora utilizando el marco tradicional through the projection onto convex solution sets (POCS) del diseño descriptivo de experimentos de regularización method; the second image is enhanced with the DEDR (DEDR) a través del método de proyección en conjuntos framework by incorporating the robust adaptive spatial de soluciones convexas (POCS); la segunda imagen se filtering (RASF) solution operator. This work describes a mejora con el marco DEDR incorporando el operador de MLP based technique applied to the pixel level multi- la solución robusta de filtrado espacial adaptativo focus fusion problem characterized by the use of image (RASF). Este trabajo describe una técnica basada en windows with the idea of reducing noise and determining MLP aplicada al problema de fusión de enfoque múltiple which pixel is clearer between the two images. a nivel de píxeles caracterizada por el uso de ventanas de Experimental results show that the proposed novel imagen con la idea de reducir el ruido y determinar qué method outperforms the discrete wavelet transform based píxel es más claro entre las dos imágenes. Los resultados most competing approach. experimentales muestran que el nuevo método propuesto tiene un mayor enfoque en la mejora de las imágenes Robust Adaptive Filtering, Descriptive Experiment basado en la transformación discreta de wavelet. Desing Regularization, Multi-Layer Perceptron FIltrado Robusto Adaptivo, Diseño Descriptivo de Experimentos de Regularización, Perceptrón Multicapa

Citation: YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019, 5-16: 20-26.

* Correspondence to Author (email: [email protected]) † Researcher contributing first author.

© ECORFAN - Spain www.ecorfan.org/spain 21 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

1. Introduction We propose a pixel level multi-focus image fusion method based on the use of image Image fusion is generally performed at different windows and a MLP neural network, with three levels of information representation, namely, principal features (visibility, spatial frequency pixel level, feature level, and decision level and edge detection) for classifying the (Shutao Li & Bin Yang, 2010) . Up to now, information, reducing the speckle noise and many image fusion techniques have been determining which pixel is clearer in each developed. Basically, the fusion technique can image even when the image has noise and mis- be categorized into spatial domain fusion and registration. The rest of the paper is organized transform domain fusion (Li, Yang & Hu, as follows. The imaging problem formalism 2011). The spatial domain-based methods select will be described in section two, in section regions from source images in the spatial three we will describe the DEDR-POCS domain to construct the fused image (Pajares & phenomenology, RASF formalism will be Manuel de la Cruz, 2004) and (Hui Li, presented in section four, the proposed fusion Manjunath & Mitra, 1995). method and scheme will be developed in section five, and finally in section six and seven The basic concept of the transformed we will describe some experiments and domain-based methods is to perform certain concluding remarks respectively. multi-resolution decomposition on each source image, then integrate all these decompositions 2. Imaging Problem Formalism to obtain one combined representation according to some fusion rules, and finally Consider the numerical approximation of the reconstruct the fused image by performing the imaging remote sensing (RS) system equation inverse transformation to the combined of observation (EO) that is represented in the representation. Discrete Wavelet Transform vector form as (Henderson & Lewis, 1998), (DWT) provide directional information and (Shkvarko, 2010) and (Shkvarko et al., 2011): without carrying redundant information across different resolutions. Moreover, DWT has good = ̃v+n=( )v+n (1) locality of time frequency (Zheng et al., 2007) and (Yang et al., 2010). However, these The matrix ̃=( ) represents the methods based on multi-scale transforms are discrete-form approximation of the uncertain shift-variant; namely, their performance will signal formation operator (SFO) function kernel quickly deteriorate when there is with 〈 ̃〉 representing the regular SFO misregistration of the source images. and representing the zero mean random SFO perturbation term, respectively. Here, v, n, u This work focuses on fusing DEDR- are random zero-mean vectors composed of the POCS and DEDR-RASF enhanced images. The decomposition coefficients } } DEDR-POCS method centers on the idea of and } , respectively (Li et al., 2011) spatial spectrum pattern (SSP) reconstruction characterized by the correlation matrices, that allows adaptive high-resolution radar Rv=D(b)=diag(b), Rn=N0I, and image formation with speckle reduction and 〈 ̃ ̃ 〉 , correspondingly, where <⋅> dynamic scene image enhancement, in the real- defines the averaging over the randomness of world uncertain operational scenarios. The the SFO, superscript + stands for Hermitian RASF method enhances the image and reduces conjugate, and N0 is the power of the white the blurring through an image filter. observation noise vector n. Vector b is referred to as a vector-form representation of the SSP. These aspects are formalized by SSP The vector b represents the lexicographical estimation techniques, minimizing noise ordering of the sensed scene through an improved balance between } ,, over the K K pixel-framed 2-D scene resolution and noise suppression, unified by x y {k = 1,…, K ; k = 1,…, K ; k = 1,…, K = regularization tools. The idea is to compensate x x y y K K } (Henderson & Lewis, 1998), (Shkvarko, the different attributes that each method x y 2010) and (Shkvarko et al., 2011). produces into a new fused image, balancing the performance that each method and algorithm provides.

YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and ISSN 2444-5002 ECORFAN® All rights reserved AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019 22 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

The nonlinear inverse problem for ( )is the discrete-form Laplacian, recovery of the SSP vector b from the available defines a hard threshholding operator ̂ data recordings u, i.e., }, that at each iteration clips off all ̂ depends on the employed estimation strategy. entries of lower than the user specified In the basic DEDR-related framework nonnegative sparsity preserving tolerance (Henderson & Lewis, 1998), the original low threshold level  and represents the ̂ resolution (LR) estimation of the SSP MSF solution-dependent point spread function (PSF) MSF } is performed employing the operator standard matched spatial filtering (MSF) processing in the positive convex cone solution ̂ ̂ ( ( ) ) ( ( ) ) set ( )in the Euclidian space with the metric (5) structure induced by the scalar product which does not involve the image type total Constructed from the diagonal loaded variation (TV) norm (Shkvarko et al., 2011) by factor N composition of the regular MSF and (Torres et al., 2014). Thus, the MSF response matrix  = S+S and the solution- estimate ̂ is formed as a square detected ̂ MSF dependent ( ) at the current ith iteration output step where symbol defines the element wise matrix product. In practical RS scenarios, the ̂ MSF (SQ-D T }) ( } ) (2) regularization factor N can be evaluated empirically from the speckle corrupted LR Of the complex coherent MSF imaging image following one of the local statistics + system q = S u post processed by the techniques exemplified in (Henderson & Lewis, despeckling filter defined by the POCS operator 1998). The adaptive iterative DEDR-POCS . procedure (3) incorporates two regularization factors and that balance the 3. DEDR-POCS Phenomenology corresponding discrepancy terms.

The LR image (2) serves as an input for any HR reconstructive post processing techniques, i.e., ̂ ̂ as a zero step iteration MSF. The general DEDR-POCS inspired adaptive iterative HR image enhancing approach featured in (Henderson & Lewis, 1998) and

(Shkvarko, 2010) is based on the DEDR- restructured minimum variance distortion-less Figure 1 Schematic diagram of the proposed fusion response (MVDR) method (Shkvarko, 2010) method that implies the nonlinear adaptive iterative ̂ processing (Henderson & Lewis, 1998), The iterative process is initialized with (Shkvarko, 2010) and (Shkvarko et al., 2011) ̂ ̂ MSF and is terminated at for which for HR SSP recovery in the solution space with the user specified -norm convergence the balanced type image and its gradient map tolerance level TL is attained at some i = I. In metrics that yields the implicit iterative the simulations, we specified  = 0.03 and technique (Shkvarko, 2010) and (Shkvarko et TL treated different feasible assignments to and al., 2011) (Shkvarko, 2010) and (Shkvarko et al.,

2011). ̂ ̂ ̂ ̂ (3)

̂ ̂ ( ) ( ( )) } 4. Robust Adaptive Spatial Filtering

Initialized by the despeckled MSF The DEDR optimal SSP estimate b, is acquired image (2). Here, as a solution to the nonlinear equation

̂ } ̂ ( ̂ ) ( ̂ ) , (4) diag (6)

YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and ISSN 2444-5002 ECORFAN® All rights reserved AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019 23 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

Where the solution operator (SO) 2. Extract three features that reflect its represents the DEDR operator in matrix-form clarity. Indicate the feature vectors for and indicates a POCS operator. The SO Ai and Bi by (SFAi, VIAi, EGAi) and estimator considered in this work is the robust (SFBi, VIBi, EGBi), respectively. adaptive spatial filtering (RASF), at the same time the solution component is given by: 3. Design and train a MLP to determine whether Ai or Bi is clearer. The

̂ difference vector (SFAi, SFBi, VIAi, VIBi, D DR RASF ( ) (7) EGAi, EGBi) is used as input, and the With ( ) ( ) and output is labeled according to if A is clearer than considering ̂ diag( ̂ ). target { other ise.

The term in (6) is composed of 3 4. Validate the trained MLP on all image operators as follows. window pairs obtained in step 1. The ith

window Zi, of the fused image is then (8) if out created as { other ise, Where is a desplecking operator. , expresses the SAR system Where outi is the MLP output using the impulse response function separability. Finally, ith image window pair as input. represents the convergence guaranteed projector onto the nonnegative convex solution 5. Verify the fusion result obtained in step set, specified as the positivity operator 4. Specifically, if the MLP decides that , that has the effect of suppressing all a particular window is to come from A negative values (Torres et al., 2014). but with the majority of its surrounding windows taken from B, this block might 5. Multi-layer Perceptron Based be switched to come from B. Multifocus Image Fusion 5.2. Feature vector extraction Figure 1 shows a schematic diagram of the MLP based multi-focus image fusion proposed The two M x N windows Ai and Bi can be fed method. Here, we consider two source images directly into a MLP for discrimination (Li et al., gotten through the RASF and DEDR-POCS 2002) and (Kulkarni, 2001). techniques. It is important to remark that the algorithm can be extended straightforwardly to In this paper, we extract three features handle more than two source images. from each divided image window to represent Moreover, these images are assumed to have its clarity. These are the spatial frequency, been registered. visibility and an edge feature.

5.1. Algorithm 5.3. Spatial Frequency

This method first divides the initial images into Spatial frequency is used to measure the overall windows. Given two of these windows (one activity level of an image. For an M x N image from each initial image) a MLP is trained to F, with the gray value at Pixel position (m,n) determine which one is clearer. Then it denoted by F(m,n), its spatial frequency is proceeds by selecting the clearer window and defined as constructing the final image (Li et al., 2002). √ (9) The full method is summarized as follows: Where RF and CF are the row frequency and column frequency respectively, the 1. Divide the two initial images A and B equations (10-11) represents the information for into M x N windows. Denote the ith extracting information pixel by pixel and with image window pair by Ai and Bi this create a feature vector (Eskicioglu & respectively. Fisher, 1995).

YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and ISSN 2444-5002 ECORFAN® All rights reserved AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019 24 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

iii. Compute the activation of each neuron. √ ∑ ∑ ( ( ) ( )) (10) The activation of neuron in layer

is expressed by

√ ∑ ∑ ( ( ) ( )) (11)

(∑ ) (14) 5.5. Visibility (VI) Where summation is done over all This feature is inspired from the human visual neurons in the( )layer, and (⋅)is the system, and is defined as sigmoid activation function. The weight for the connection from the neuron in the ( ) layer ( ) ∑ ∑ (12) to the neuron in the layer is denoted by ;

the bias of the neuron in layer is represented

Where μ is the mean intensity value of by . the image, and α is a visual constant ranging from 0.5 to 0.8 (Eskicioglu & Fisher, 1995). iv. Find the error between the layer output vector and the target vector via

5.5. Edge Features (EF) ∑ ( )

This feature is based on the number of edges The calculated error is then propagated extracted from the image. Intuitively, for backward in order to obtain the change in the images of comparable complexity, a clearer synaptic weights between layers } image will have more edges. Here, we first through where is a training apply the Canny edge detector (Eskicioglu & Fisher, 1995) to each decomposed window. The rate coefficient [12]. total number of one’s in the resultant binary image window is then taken as the edge feature. v. Update the synaptic weights using

; (16) 5.6. Neural Network Classification

Many neural networks (NN) models have been Where }are the iterations proposed for a diverse range of problems, required for achieving the desired NN output. including pattern classification. The fusion problem examined here isconsidered as a 6. Experiments classification problem. For this paper we proposed the next idea: A feed-forward back- In this section, we first experimentally propagation sigmoid NN is chosen as classifier; demonstrate the effectiveness of the three a type of supervised NN, since a desired output features, proposed in the previous section is required in order to be trained. The NN (namely SF; VI and EF), in representing the model consists of a multi-layer perceptron clarity level of an image. An image block of (MLP) which uses a back-propagation learning size 64x 64 is extracted from the ‘‘Lena’’ methodology and a sigmoid activation function. image Figure 1. The figure shows two degraded The back-propagation learning method is versions of the same image, the first one with appropriate for non-linearly-separable inputs; it Gaussian noise in the up part and second one is is summarized as follows (Li et al., 2002) and degraded with speckle noise in the down part, (Eskicioglu & Fisher, 1995): this process represents the additive and multiplicative noise in RS imagery. After i. Select an input vector extracting and classifying the features, the algorithm provides an image without noise in ( ) and a desired output Figure 1 (final image estimate). This suggests

vector ( ) ; subscript that all three features can be used to represent stands for transpose. image clarity and classify the features in order to reduce the noise in the fused image. ii. Initialize the synaptic weights between layers } with small random values.

YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and ISSN 2444-5002 ECORFAN® All rights reserved AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019 25 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

7. Simulations and Concluding Remarks The MLP contain three input units and one output unit, with 90 hidden units. For An experiment is performed on an image of comparison purposes, we also perform fusion size 512x512. Image windows of size 8x8 are using the DWT-based method mentioned used. Two regions, each containing 2048 image earlier. The “daubechies” avelet basis is used. window pairs, are selected from the two Thus, the only way to analyze and compare the DEDR-POCS and RASF images. A training set noise suppression and the pixel clarity is to with a total of 4096 image window pairs is thus perform the simulation experiments. formed. The three features, SF, VI and EF, are extracted ( ith α=0.5 in (12)) and normalized We must simulate the fractional SAR to the range [0, 1] before feeding into the MLP. (F-SAR) imaging RS system with operational

 parameters similar to the comparative studies. The simulation results of enhancement of such F-SAR imagery applying different DEDR- POCS and RASF methods are reported in Figure 2. The test 512x512-pixel HR image Figure 2(a) borrowed from the real-world SAR imagey (Skymed, s.f.) relates to the hypothetical full focused SAR imaging mode. a) b) The 512x512-pixel LR speckle corrupted image of the same scene presented in Figure 2(b) corresponds to the single look F-SAR mode (quick look modality (2)) for the typical F-SAR operational scenario specifications, like the comparative previous studies. Figure 2(c) and 2(d) report the feature-enhanced radar imaging results obtained via the DEDR-POCS and RASF techniques. c) d) In order to evaluate the quality of the high-resolution enhanced images, two metrics have been adopted: the peak signal-to-noise ratio (PSNR) as an extension of the mean squared error (MSE) and the structural similarity index (SSIM) (Hore & Ziou, 2010):

( ̂ ) ∑ ∑ ( ̂ ) (13)

e) f) Where represents the original SSP ̂ Figure 2 (a) Original 512x512-pixels test scene frame and is the fusion image approximation borrowed from the real-world high resolution SAR obtained through DWT fusion or MLP fusion. imagery ; (b) LR 512x512-pixels speckle corrupted radar image of the same scene formed with a simulated LR F- { ̂ } (14) SAR system; model system parameters: triangular range ( ) PSF (the width at ½ of the peak value, y = 10 pixels); Gaussian bell azimuth PSF (the width at ½ of the peak ( )( ) ̂ ̂ value, x = 15 pixels); single-look scenario with the fully ( ̂ ) (15) developed speckle, SNR = 0 dB; (c) 512x512-pixels ( )( ) ̂ ̂ image enhanced using the DEDR framework with the solution operator RASF, (3) with 1 = 1, 2 = 0 (convergence at I = 15, PSNR = 17.7208 dB); (d) Where is the mean intensity, is 512x512-pixels image enhanced with the most prominent standard deviation and is the covariance competing structured DEDR-POCS technique (3) with 1 between the involved images. = 2 = 1 (convergence at I = 10, PSNR = 22.6092 dB); (e) Fused image using DWT (Daubechies wavelet); (f) Fused image using MLP (with backpropation, three input neurons and one for output)

YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and ISSN 2444-5002 ECORFAN® All rights reserved AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019 26 Article Journal of Computational Systems and ICTs December 2019 Vol.5 No.16 20-26

The comparative results are presented in Pajares, G., & Manuel de la Cruz, J. (2004). A Table 1. These results verify that the best wavelet-based image fusion tutorial. Pattern perceptual fusion image performances Recognition, 37(9), 1855–1872. ̂ ̂ regarding the ( ) and ( ) metrics are reached by the MLP fusion method. Shkvarko, Y. (2010). Unifying Experiment Design and Convex Regularization Techniques Test Case Comparative Metrics (average) for Enhanced Imaging With Uncertain Remote DWT (Daubechies) MLP Sensing Data—Part I: Theory. IEEE PSNR 17.8571 dB 21.1645 dB Transactions on Geoscience and Remote SSIM O.6225 0.7998 Sensing, 48(1), 82–95. https://doi.org/10.1109/tgrs.2009.2027695 Table 1 Comparative Metrics

In general, if the window size is too Shkvarko, Y., Tuxpan, J., & Santos, S. (2011). large, a particular window may contain two or Dynamic Experiment Design Regularization more objects or information at the different Approach to Adaptive Imaging with Array distances and consequently will lead to a less Radar/SAR Sensor Systems. Sensors, 11(5), clear image. On the other hand, using small 4483–4511. window size may lead to the saw-tooth effect. Shutao Li, & Bin Yang, (2010). Hybrid Multiresolution Method for Multisensor References Multimodal Image Fusion. IEEE Sensors Eskicioglu, A., & Fisher, P. (1995). Image Journal, 10(9), 1519–1526. quality measures and their performance. IEEE Transactions on Communications, 43(12), Skymed, C. O. S. M. O. (s.f.). COSMO- 2959–2965. https://doi.org/10.1109/26.477498 SkyMed / CUGS. Recuperado 5 mayo, 2017, de http://www.cosmo-skymed.it. Henderson, F. M., & Lewis, A. V. (1998). Principles and Applications of Imaging Radar, Torres, D. G., Tuxpan, J., & Yanez, I. (2014). Manual of Remote Sensing (3ª ed.). N.Y, USA: Spatial spectrum pattern estimation based on Willey. the DEDR-TV extended approach for fractional SAR imagery enhancement. 2014 International Hore, A., & Ziou, D. (2010). Image Quality Conference on Electronics, Communications Metrics: PSNR vs. SSIM. 2010 20th and Computers (CONIELECOMP), . International Conference on Pattern https://doi.org/10.1109/conielecomp.2014.6808 Recognition, . 560 https://doi.org/10.1109/icpr.2010.579 Yang, Y., Park, D. S., Huang, S., & Rao, N. Hui Li, Manjunath, B., & Mitra, S. (1995). (2010). Medical Image Fusion via an Effective Multi-sensor image fusion using the wavelet Wavelet-Based Approach. EURASIP Journal transform. Proceedings of 1st International on Advances in Signal Processing, 2010(1). Conference on Image Processing, https://doi.org/10.1155/2010/579341

Kulkarni, A. (2001). Computer Vision and Zheng, Y., Essock, E. A., Hansen, B. C., & Fuzzy-Neural Systems (Ed. rev.). N.J., USA: Haun, A. M. (2007). A new metric based on Prentice Hall. extended spatial frequency and its application to DWT based fusion algorithms. Information Li, S., Kwok, J. T., & Wang, Y. (2002). Fusion, 8(2), 177–192. Multifocus image fusion using artificial neural networks. Pattern Recognition Letters, 23(8), 985–997.

Li, S., Yang, B., & Hu, J. (2011). Performance comparison of different multi-resolution transforms for image fusion. Information Fusion, 12(2), 74–84. https://doi.org/10.1016/j.inffus.2010.03.002

YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and ISSN 2444-5002 ECORFAN® All rights reserved AGUILERA-GONZALEZ, Gabriel. RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems. Journal of Computational Systems and ICTs. 2019 Article Journal of Computational Systems and ICTs Month, Year Vol.1 No.1 1-15-[Using ECORFAN]

[Title in Times New Roman and Bold No. 14 in English and Spanish]

Surname (IN UPPERCASE), Name 1st Author†*, Surname (IN UPPERCASE), Name 1st Coauthor, Surname (IN UPPERCASE), Name 2nd Coauthor and Surname (IN UPPERCASE), Name 3rd Coauthor

Institutional Affiliation of Author including Dependency (No.10 Times New Roman and Italic)

International Identification of Science - Technology and Innovation

ID 1st Author: (ORC ID - Researcher ID Thomson, arXiv Author ID - PubMed Author ID - Open ID) and CVU 1st author: (Scholar-PNPC or SNI-CONACYT) (No.10 Times New Roman)

ID 1st Coauthor: (ORC ID - Researcher ID Thomson, arXiv Author ID - PubMed Author ID - Open ID) and CVU 1st coauthor: (Scholar or SNI) (No.10 Times New Roman)

ID 2nd Coauthor: (ORC ID - Researcher ID Thomson, arXiv Author ID - PubMed Author ID - Open ID) and CVU 2nd coauthor: (Scholar or SNI) (No.10 Times New Roman)

ID 3rd Coauthor: (ORC ID - Researcher ID Thomson, arXiv Author ID - PubMed Author ID - Open ID) and CVU 3rd coauthor: (Scholar or SNI) (No.10 Times New Roman)

(Report Submission Date: Month, Day, and Year); Accepted (Insert date of Acceptance: Use Only ECORFAN)

Abstract (In English, 150-200 words) Abstract (In Spanish, 150-200 words)

Objectives Objectives Methodology Methodology Contribution Contribution

Keywords (In English) Keywords (In Spanish)

Indicate 3 keywords in Times New Roman and Bold No. Indicate 3 keywords in Times New Roman and Bold No. 10 10

Citation: Surname (IN UPPERCASE), Name 1st Author, Surname (IN UPPERCASE), Name 1st Coauthor, Surname (IN UPPERCASE), Name 2nd Coauthor and Surname (IN UPPERCASE), Name 3rd Coauthor. Paper Title. Journal of Computational Systems and ICTs. Year 1-1: 1-11 [Times New Roman No.10]

* Correspondence to Author ([email protected]) † Researcher contributing as first author.

© ECORFAN-Spain www.ecorfan.org/spain

Article Journal of Computational Systems and ICTs Month, Year Vol.1 No.1 1-15-[Using ECORFAN]

Introduction

Text in Times New Roman No.12, single space.

General explanation of the subject and explain why it is important.

What is your added value with respect to other techniques?

Clearly focus each of its features Figure 1 Title and Source (in italics)

Clearly explain the problem to be solved and the Should not be images-everything must be editable. central hypothesis.

Explanation of sections Article.

Development of headings and subheadings of the article with subsequent numbers Table 1 Title and Source (in italics)

[Title No.12 in Times New Roman, single Should not be images-everything must be editable. spaced and bold] Each article shall present separately in 3 folders: Products in development No.12 Times New a) Figures, b) Charts and c) Tables in .JPG Roman, single spaced. format, indicating the number and sequential Bold Title. Including graphs, figures and tables-Editable For the use of equations, noted as follows: In the article content any graphic, table and r figure should be editable formats that can change Yij = α + ∑h=1 βhXhij + uj + eij (1) size, type and number of letter, for the purposes of edition, these must be high quality, not Must be editable and number aligned on the right pixelated and should be noticeable even side. reducing image scale. Methodology [Indicating the title at the bottom with No.10 and Times New Roman Bold] Develop give the meaning of the variables in linear writing and important is the comparison of the used criteria.

Results

The results shall be by section of the article.

Annexes

Tables and adequate sources

Graphic 1 Title and Source (in italics) Thanks

Should not be images-everything must be editable. Indicate if they were financed by any institution, University or company.

Conclusions

Explain clearly the results and possibilities of improvement.

ISSN 2444-5002 Surname (IN UPPERCASE), Name 1st Author, Surname (IN ECORFAN® All rights reserved UPPERCASE), Name 1st Coauthor, Surname (IN UPPERCASE), Name 2nd Coauthor and Surname (IN UPPERCASE), Name 3rd Coauthor. Paper Title Journal of Computational Systems and ICTs. Year [Times New Roman No. 8] Article Journal of Computational Systems and ICTs Month, Year Vol.1 No.1 1-15-[Using ECORFAN]

References

Use APA system. Should not be numbered, nor with bullets, however if necessary numbering will be because reference or mention is made somewhere in the Article.

Use Roman Alphabet, all references you have used must be in the Roman Alphabet, even if you have quoted an Article, book in any of the official languages of the United Nations (English, French, German, Chinese, Russian, Portuguese, Italian, Spanish, Arabic), you must write the reference in Roman script and not in any of the official languages.

Technical Specifications

Each article must submit your dates into a Word document (.docx):

Journal Name Article title Abstract Keywords Article sections, for example:

1. Introduction

2. Description of the method 3. Analysis from the regression demand curve 4. Results 5. Thanks 6. Conclusions 7. References

Author Name (s) Email Correspondence to Author References

Intellectual Property Requirements for editing:

-Authentic Signature in Color of Originality Format Author and Coauthors

-Authentic Signature in Color of the Acceptance Format of Author and Coauthors

ISSN 2444-5002 Surname (IN UPPERCASE), Name 1st Author, Surname (IN ECORFAN® All rights reserved UPPERCASE), Name 1st Coauthor, Surname (IN UPPERCASE), Name 2nd Coauthor and Surname (IN UPPERCASE), Name 3rd Coauthor. Paper Title Journal of Computational Systems and ICTs. Year [Times New Roman No. 8]

Reservation to Editorial Policy

Journal of Computational Systems and ICTs reserves the right to make editorial changes required to adapt the Articles to the Editorial Policy of the Journal. Once the Article is accepted in its final version, the Journal will send the author the proofs for review. ECORFAN® will only accept the correction of errata and errors or omissions arising from the editing process of the Journal, reserving in full the copyrights and content dissemination. No deletions, substitutions or additions that alter the formation of the Article will be accepted.

Code of Ethics - Good Practices and Declaration of Solution to Editorial Conflicts

Declaration of Originality and unpublished character of the Article, of Authors, on the obtaining of data and interpretation of results, Acknowledgments, Conflict of interests, Assignment of rights and Distribution

The ECORFAN-Mexico, S.C Management claims to Authors of Articles that its content must be original, unpublished and of Scientific, Technological and Innovation content to be submitted for evaluation.

The Authors signing the Article must be the same that have contributed to its conception, realization and development, as well as obtaining the data, interpreting the results, drafting and reviewing it. The Corresponding Author of the proposed Article will request the form that follows.

Article title:

– The sending of an Article to Journal of Computational Systems and ICTs emanates the commitment of the author not to submit it simultaneously to the consideration of other series publications for it must complement the Format of Originality for its Article, unless it is rejected by the Arbitration Committee, it may be withdrawn.

– None of the data presented in this article has been plagiarized or invented. The original data are clearly distinguished from those already published. And it is known of the test in PLAGSCAN if a level of plagiarism is detected Positive will not proceed to arbitrate.

– References are cited on which the information contained in the Article is based, as well as theories and data from other previously published Articles.

– The authors sign the Format of Authorization for their Article to be disseminated by means that ECORFAN-Mexico, S.C. In its Holding Spain considers pertinent for disclosure and diffusion of its Article its Rights of Work.

– Consent has been obtained from those who have contributed unpublished data obtained through verbal or written communication, and such communication and Authorship are adequately identified.

– The Author and Co-Authors who sign this work have participated in its planning, design and execution, as well as in the interpretation of the results. They also critically reviewed the paper, approved its final version and agreed with its publication.

– No signature responsible for the work has been omitted and the criteria of Scientific Authorization are satisfied.

– The results of this Article have been interpreted objectively. Any results contrary to the point of view of those who sign are exposed and discussed in the Article.

Copyright and Access

The publication of this Article supposes the transfer of the copyright to ECORFAN-Mexico, SC in its Holding Spain for its Journal of Computational Systems and ICTs, which reserves the right to distribute on the Web the published version of the Article and the making available of the Article in This format supposes for its Authors the fulfilment of what is established in the Law of Science and Technology of the United Mexican States, regarding the obligation to allow access to the results of Scientific Research.

Article Title:

Name and Surnames of the Contact Author and the Coauthors Signature 1. 2. 3. 4.

Principles of Ethics and Declaration of Solution to Editorial Conflicts

Editor Responsibilities

The Publisher undertakes to guarantee the confidentiality of the evaluation process, it may not disclose to the Arbitrators the identity of the Authors, nor may it reveal the identity of the Arbitrators at any time.

The Editor assumes the responsibility to properly inform the Author of the stage of the editorial process in which the text is sent, as well as the resolutions of Double-Blind Review.

The Editor should evaluate manuscripts and their intellectual content without distinction of race, gender, sexual orientation, religious beliefs, ethnicity, nationality, or the political philosophy of the Authors.

The Editor and his editing team of ECORFAN® Holdings will not disclose any information about Articles submitted to anyone other than the corresponding Author.

The Editor should make fair and impartial decisions and ensure a fair Double-Blind Review.

Responsibilities of the Editorial Board

The description of the peer review processes is made known by the Editorial Board in order that the Authors know what the evaluation criteria are and will always be willing to justify any controversy in the evaluation process. In case of Plagiarism Detection to the Article the Committee notifies the Authors for Violation to the Right of Scientific, Technological and Innovation Authorization.

Responsibilities of the Arbitration Committee

The Arbitrators undertake to notify about any unethical conduct by the Authors and to indicate all the information that may be reason to reject the publication of the Articles. In addition, they must undertake to keep confidential information related to the Articles they evaluate.

Any manuscript received for your arbitration must be treated as confidential, should not be displayed or discussed with other experts, except with the permission of the Editor.

The Arbitrators must be conducted objectively, any personal criticism of the Author is inappropriate.

The Arbitrators must express their points of view with clarity and with valid arguments that contribute to the Scientific, Technological and Innovation of the Author.

The Arbitrators should not evaluate manuscripts in which they have conflicts of interest and have been notified to the Editor before submitting the Article for Double-Blind Review.

Responsibilities of the Authors

Authors must guarantee that their articles are the product of their original work and that the data has been obtained ethically.

Authors must ensure that they have not been previously published or that they are not considered in another serial publication.

Authors must strictly follow the rules for the publication of Defined Articles by the Editorial Board.

The authors have requested that the text in all its forms be an unethical editorial behavior and is unacceptable, consequently, any manuscript that incurs in plagiarism is eliminated and not considered for publication.

Authors should cite publications that have been influential in the nature of the Article submitted to arbitration.

Information services

Indexation - Bases and Repositories

LATINDEX (Scientific Journals of , Spain and Portugal) RESEARCH GATE (Germany) GOOGLE SCHOLAR (Citation indices-Google) REDIB (Ibero-American Network of Innovation and Scientific Knowledge- CSIC) MENDELEY (Bibliographic References Manager)

Publishing Services

Citation and Index Identification H Management of Originality Format and Authorization Testing Article with PLAGSCAN Article Evaluation Certificate of Double-Blind Review Article Edition Web layout Indexing and Repository ArticleTranslation Article Publication Certificate of Article Service Billing

Editorial Policy and Management

38 Matacerquillas, CP-28411. Moralzarzal –Madrid-España. Phones: +52 1 55 6159 2296, +52 1 55 1260 0355, +52 1 55 6034 9181; Email: [email protected] www.ecorfan.org

ECORFAN®

Chief Editor MIRANDA - TORRADO, Fernando. PhD

Executive Director RAMOS-ESCAMILLA, María. PhD

Editorial Director PERALTA-CASTRO, Enrique. MsC

Web Designer ESCAMILLA-BOUCHAN, Imelda. PhD

Web Diagrammer LUNA-SOTO, Vladimir. PhD

Editorial Assistant SORIANO-VELASCO, Jesús. BsC

Translator DÍAZ-OCAMPO, Javier. BsC

Philologist RAMOS-ARANCIBIA, Alejandra. BsC

Advertising & Sponsorship

(ECORFAN® Spain), [email protected]

Site Licences

03-2010-032610094200-01-For printed material ,03-2010-031613323600-01-For Electronic material,03-2010-032610105200-01-For Photographic material,03-2010-032610115700-14-For the facts Compilation,04-2010-031613323600-01-For its Web page,19502-For the Iberoamerican and Caribbean Indexation,20-281 HB9-For its indexation in Latin-American in Social Sciences and Humanities,671-For its indexing in Electronic Scientific Journals Spanish and Latin-America,7045008- For its divulgation and edition in the Ministry of Education and Culture-Spain,25409-For its repository in the Biblioteca Universitaria-Madrid,16258-For its indexing in the Dialnet,20589-For its indexing in the edited Journals in the countries of Iberian-America and the Caribbean, 15048-For the international registration of Congress and Colloquiums. [email protected]

Management Offices

38 Matacerquillas, CP-28411. Moralzarzal –Madrid-España.

Journal of Computational Systems and ICTs

“Seven frame phase-shifting algorithms for the three-dimensional profilometry using fringe projection and soft computing” CRUZ-LANDEROS, Jose Juan, MOSIÑO, Juan Francisco, MARTÍNEZ-REBOLLAR, Alicia and BALTAZAR-FLORES, María del Rosario Instituto Tecnológico de León National Center for Research and Technological Development

“Augmented Reality App for tourism promotion in the Triangle of the Sun of the State of Guerrero” RÍOS, David, URUETA, Daniel, ALANÍS, Reynaldo and MOTA, Juan Universidad Politécnica del Estado de Guerrero

“Design and verification of a testing platform for implementation of software defined radio communications systems” GARCÍA-DZUL, Rogelio, PARRA-MICHEL, Ramón and SÁNCHEZ-VENEGAS, Jorge CINVESTAV del IPN

“RASF and DEDR-POCS Image Fusion through Multi-layer Perceptron in SAR Imagery Systems” YAÑEZ-VARGAS, Israel, QUINTANILLA-DOMÍNGUEZ, Joel and AGUILERA-GONZALEZ, Gabriel Universidad Politécnica de Juventino Rosas