Cyber-Physical System for Environmental Monitoring Based on Deep Learning

Cyber-Physical System for Environmental Monitoring Based on Deep Learning

sensors Article Cyber-Physical System for Environmental Monitoring Based on Deep Learning Íñigo Monedero 1 , Julio Barbancho 1,* , Rafael Márquez 2 and Juan F. Beltrán 3 1 Tecnología Electrónica, Escuela Politéncia Superior, Universidad de Sevilla, Calle Virgen de África 7, 41012 Sevilla, Spain; [email protected] 2 Fonoteca Zoológica, Departamento de Biodiversidad y Biología Evolutiva, Museo Nacional de Ciencias Naturales (CSIC), Calle José Gutiérrez Abascal, 2, 28006 Madrid, Spain; [email protected] 3 Departamento de Zoología, Facultad de Biología, Universidad de Sevilla, Avenida de la Reina Mercedes, s/n, 41012 Sevilla, Spain; [email protected] * Correspondence: [email protected]; Tel.: +34-955-55-28-38 Abstract: Cyber-physical systems (CPS) constitute a promising paradigm that could fit various applications. Monitoring based on the Internet of Things (IoT) has become a research area with new challenges in which to extract valuable information. This paper proposes a deep learning classification sound system for execution over CPS. This system is based on convolutional neural networks (CNNs) and is focused on the different types of vocalization of two species of anurans. CNNs, in conjunction with the use of mel-spectrograms for sounds, are shown to be an adequate tool for the classification of environmental sounds. The classification results obtained are excellent (97.53% overall accuracy) and can be considered a very promising use of the system for classifying other biological acoustic targets as well as analyzing biodiversity indices in the natural environment. The paper concludes by observing that the execution of this type of CNN, involving low-cost and reduced computing resources, are feasible for monitoring extensive natural areas. The use of CPS enables Citation: Monedero, Í.; Barbancho, J.; Márquez, R.; Beltrán, J.F. flexible and dynamic configuration and deployment of new CNN updates over remote IoT nodes. Cyber-Physical System for Environmental Monitoring Based on Keywords: convolutional neural network; deep learning; machine learning; cyber-physical systems; Deep Learning. Sensors 2021, 21, 3655. passive active monitoring; Internet of Things https://doi.org/10.3390/s21113655 Academic Editors: Alessandro Gardi and Roberto Sabatini 1. Introduction 1.1. Environmental Sound Classification Received: 30 March 2021 Recently, international institutions such as the United Nations (UN) have paid grow- Accepted: 13 May 2021 ing attention to global strategies to coordinate the interactions among different agents Published: 24 May 2021 (countries, companies ... ) around the world. These strategies [1] are focused on the development of sustainable development goals (SDGs). Special interest has been shown Publisher’s Note: MDPI stays neutral regarding goal 11, “sustainable cities and communities”; goal 13, “climate action”; goal 14, with regard to jurisdictional claims in published maps and institutional affil- “life below water”; and goal 15, “life on land”. Local and regional institutions are designing iations. roadmaps to reach these goals. This is the case for the Doñana Biological Station (DBS), a public Research Institute belonging to the Spanish Council for Scientific Research CSIC in the area of Natural Resources located in the South of Spain that focuses its research mainly on the Doñana National Park (DNP), a 135 km2 natural park composed of several ecosystems. The DNP is a key environment where several species of birds stop on their Copyright: © 2021 by the authors. migration paths between Europe and Africa. These features make this space a unique Licensee MDPI, Basel, Switzerland. enclave for analyzing several aspects of the interactions among human beings and nature. This article is an open access article distributed under the terms and This paper describes a project named SCENA-RBD, developed by the University conditions of the Creative Commons of Seville in collaboration with the DBS. The main aim of the project is to design and Attribution (CC BY) license (https:// implement a cyber-physical system (CPS) that allows distributed data processing for creativecommons.org/licenses/by/ biological information. This CPS will constitute a useful tool for different governments 4.0/). (local, regional, and national) to evaluate the sustainability of an area. Several metrics [2] Sensors 2021, 21, 3655. https://doi.org/10.3390/s21113655 https://www.mdpi.com/journal/sensors Sensors 2021, 21, x FOR PEER REVIEW 2 of 18 This paper describes a project named SCENA-RBD, developed by the University of Sensors 2021, 21, 3655 Seville in collaboration with the DBS. The main aim of the project is to design and imple-2 of 19 ment a cyber-physical system (CPS) that allows distributed data processing for biological information. This CPS will constitute a useful tool for different governments (local, re- gional, and national) to evaluate the sustainability of an area. Several metrics can be con- cansidered be considered for this purpose. for this The purpose. Acoustic The Entropy Acoustic Index Entropy (H), Index which (H), is calculated which is calculated according according to the Shannon information theory, indicates the richness of a specific habitat. to the Shannon information theory, indicates the richness of a specific habitat. The higher The higher the value of a habitat’s H, the richer it is. Another common metric is the the value of a habitat’s H, the richer it is. Another common metric is the Acoustic Com- Acoustic Complexity Index (ACI), which considers the soundscape of an environment by plexity Index (ACI), which considers the soundscape of an environment by measuring the measuring the variability of its natural audio signals in relation to audio signals with a variability of its natural audio signals in relation to audio signals with a human origin. human origin. Aligned with this idea, the Normalised Difference Soundscape Index (NDSI) Aligned with this idea, the Normalized Difference Soundscape Index (NDSI) is a metric is a metric that estimates the influence of human acoustic presence (anthrophony) over that estimates the influence of human acoustic presence (anthrophony) over the biological the biological soundscape (biophony). Based on this, a new paradigm is rising: passive soundscape (biophony). Based on this, a new paradigm is rising: passive acoustic moni- acoustic monitoring (PAM). PAM is a promising research area that has recently increased toring (PAM). PAM is a promising research area that has recently increased in importance in importance due to the development of low-cost wireless electronic sensors. due to the development of low-cost wireless electronic sensors. The CPS of the present paper consists of two stages (Figure1). The first stage is focusedThe on CPS the of identification the present ofpaper acoustic consists patterns. of two Each stages pattern (Figure is identified 1). The first according stage is to fo- a specificcused on algorithm. the identification There could of acoustic be several patterns. algorithms Each running pattern at is the identified same time according depending to a onspecific the target algorithm. under There study—biological could be several acoustic algorithms targets running such as at animals the same or time climate depending events oron even the target environmental under study—biol targets suchogical as acoustic engines, targets gun shots, such humanas animals voices, or climate etc., could events, be consideredor even environmental in the classification. targets In such this as stage engine a classifications, gun shots, procedure human voices, is executed etc., basedcould on be previouslyconsidered supervised in the classification. training. TheIn this second stage, stage a classification correlates the procedure classification is executed donein based the previouson previously stage withsupervised the purpose training. to develop The second biodiversity stage correlates indices. the classification carried out in the previous stage with the purpose to develop biodiversity indices. Figure 1. Stages in the CPS of SCENA-RBD project. Figure 1. Stages in the CPS of the SCENA-RBD project. This paper is focused on the first stage of the CPS. The first step is the audio recording. This paper is focused on the first stage of the CPS. The first step is the audio record- Industrial devices such as the Song Meter [3] have become solid platforms for recording ing. Industrial devices such as the Song Meter [1] have become solid platforms for record- wildlife sounds in real-world conditions with a high quality of recording. These devices ing wildlife sounds in real-world conditions with a high quality of recording. These de- usually offer complementary software that enables a post-analysis of the recorded data. However,vices usually this sortoffer of complementary products has features software that th mustat enables be considered a post-analysis in an extensive of the recorded sensor devicesdata. However, deployment. this sort First, of the product price ofhas the featur deviceses that is quitemust high.be considered If the area in to an be extensive studied issensor large, devices a significant deployment. number First, of elements the price must of the be devices spread is out. quite This high. could If the increase area to the be coststudied of the is large, project a significant considerably. number Although of elem someents projectsmust be arespread based out. on This open could hardware consid- (e.g.,erably AudioMoth increase the [4 cost]) that of the reduce project. the Although cost of fabrication,

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    19 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us