Analysing Building Shapes Quality of Collaborative Mapping

Dr. Maythm Al-Bakri Lecturer Department of Surveying College of Engineering University of Baghdad Baghdad - Iraq

Abstract

The very fast developments of web and data collection technologies have enabled non- experts to collect and disseminate geospatial datasets through web applications. This new type of spatial data is usually known as collaborative mapping or volunteered geographic information VGI. There are various countries around the world could benefit from collaborative mapping data because it is cost free data, easy to access and it provides more customised data. However, there is a concern about its quality because the data collectors may lack the sufficient experience and training about geospatial data production. Most previous studies which have outlined and analysed VGI quality focused on positional and linear features. The current research has been conducted to investigate the quality of another feature type such as polygons (buildings) of collaborative mapping data. Two different VGI data sources have been tested: Google Maps and WikiMapia services. The VGI data was compared with reference data extracted from high resolution aerial image which was provided from General Directorate of Surveying. The suggested methodology based on applying several metrics and methods such as surface distance method, compactness, elongation, and ratio of areas computation. The polygon shape accuracy was analysed by comparing conventional statistical values such as mean, median, standard deviation, minimum, and maximum. The results indicated that there is no big difference between the shape similarities of collaborative mapping polygons. Hence, it can be used for several applications such as spatial data infrastructures (SDI) and urban planning.

Key words: google maps; wikimapia; shape accuracy; polygons; collaborative mapping

1. Introduction system. Crowdsource refers to the way that large number of distributed As a technology has evolved; the people can work on the same project Internet permits people not only to in a very powerful manner, creating gain information but also allows something where the whole is more users to create new content and new than sum of the parts. The online knowledge. This includes encyclopaedia Wikipedia is a good community based websites which is example of Crowdsource technique. usually called the crowdsource

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

Wikipedia is approach where in be due to the fact that in many principle anybody with a computer developing countries, there is a need can edit or submit content. In for free geospatial data because Wikipedia there is no peer review, authoritative data are not existing or and the material is subject to editing not complete. These countries lack by any user [3]. The concept of the financial or technical conditions to web as platform was also adopted by produce spatial data in digital geographers to establish the formats. Therefore VGI can provide volunteered geographic information a rich source for spatial which can (VGI) [9]. support a wide range of applications such as monitoring environment, The VGI is collaborative web-based disaster management, tourist efforts to collect, produce and services [7]. In addition, the disseminate free geospatial data authoritative data is expensive and provided voluntarily by individuals. map production is costly in most The VGI sites allow user to share countries around the world [4]. The their own contribution by defining VGI data can be captured using locations where certain features handheld GPS-devices or any smart exist. Nowadays, there are a wide phones, and it can be also digitised variety of programs exist to display from free aerial images. online maps on mobile phones. Furthermore, gaining benefits from Amongst supported devices are free software. It becomes possible to nearly all phones that can run Java use or download open source GIS for mobiles, as well as platforms software to produce and upload free such as Android, the iPhone, vector datasets. However, due to the Windows Mobile. The various VGI data are produced by volunteers programs distinguish themselves from different background, the according to key features like if they quality of VGI data is not use raster maps or vector maps, if standardised and needs to be they need an Internet connection or assessed [10]. can be used offline, if they support address search or advanced features Different approaches are being like routing, thus creating for a large developed by spatial researchers to number of different needs for assess VGI quality. For example, viewing maps. Some examples of [11] presented different techniques VGI phenomenon are WikiMapia, to measure the OSM building Google Maps Maker, completeness in Germany. The OpenStreetMap (OSM) [15]. comparison involved reference data from national mapping agencies. During the last decade, the amount Their results showed that the of VGI data sources continues to completeness of OSM data in grow on the Internet [6]. This may

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

Germany in the federal states of Google Maps (GM) services. North Rhine-Westphalia were 25% WikiMapia is a web site enabling and 15% in Saxony in November collective annotation of geographic 2011. In a study undertaken by [5] satellite imagery, and is analysed the OSM building representative of similar efforts such footprints. The evaluation involved as Google Earth and mash-ups position, completeness, shape and created with web application semantic accuracy. The OSM data programming interface (API) to was compared to German Authority mapping services. It is allowing Topographic–Cartographic users to add information in the form Information System as official of a note to any location on Earth. datasets. The main results from the Although registration is not required paper indicated that the semantic to edit or add data to WikiMapia, accuracy and completeness have over 2,500,000 users from around high quality. Also, the shape of the world currently are registered OSM building footprints is very [16]. All content uploaded by users close to official data. However there becomes the intellectual property of is a difference four meters in WikiMapia and available for non- positional accuracy of OSM data. comercial use through WikiMapia Most recently [6] assessed the API [15]. completeness of OSM buildings data in UK. Three different study areas WikiMapia allows any contributor to were tested: London, Leeds and add a tag (placemark) to any location Sheffield / UK. Their findings by marking out a polygon around the proved that the quality of OSM location and then providing a default buildings is variable and inconsistent language, title, description and one within UK cities. or more categories. Location tagging is fully multi-lingual, meaning that In most of the existing studies, the there is no need to create separate OSM buildings were evaluated using tags for different languages. official dataset. The study reported WikiMapia also enables users to add here compares the polygons different map features such as (buildings) quality of other buildings, roads, railroads and rivers. collaborative mapping datasets such The interface allows specifying the as WikiMapia (Wiki) and Google size of features and providing a brief Maps (GM). description and coordinates about them, as explain in Fig. 1 [15]. In 2. Data Sources this study, the coordinates of tested features are displayed and recorded Two different datasets from manually from the main window of collaborative mapping projects have WikiMapia. been tested: WikiMapia (Wiki) and

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

Google map is another geospatial The overlay features are specified in dataset that has been investigated terms of their location within the and analysed in this research. Google Maps environment [1]. Google map is one of the webs mapping service which was Additional information is usually launched in 2005. The main basic displayed in an information window, interface of Google Maps is which is a sort of speech bubble that incredibly intuitive and is overlaid on the map. The straightforward. Users can select a information window usually appears map, move it around, and zoom in in response to the user selecting a and out to find particular area that, marker, either by clicking the marker as shown in Fig. 2. Google Maps in the map or by clicking a separate shows country, state, and county list of markers that is also displayed boundaries when available; regions on the page. Google Maps also offer and locations; country names. It also coordinates for places found in provides street-level and building Google map [1]. The positional data maps. The Google Maps enable you coordinates can be extracted directly to overlay lines and polygons objects from the interface of Google Map onto a map through Google Map service which is usually in World Maker service. This information can Geodetic System (WGS84). be used to provide additional detail about the location that looking for.

Fig 1. Screen shot of Wikimapia service, showing online information including map, place name, coordinates (http://wikimapia.org), Date: 20-07-2016.

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

Fig 2. Screen shot of Google Maps showing online information including map, place name, coordinates, photo of place (https://www.google.com), Date: 20-07-2016.

In order to estimate the shape approach, identifying the strength of accuracy of tested data (WikiMapia different datasets. and Google Maps), it was compared with self-generated dataset of higher 3. Methodology and Procedures accuracy. The digitized data will be used to create a definitive reference After presenting the main dataset (RD). The comparison will characteristics of the data sources in consist of an accurate vector dataset previous section, we now move to produced using high resolution (10 describe the overall procedures and cm) aerial photo from general methods which have been adopted in directorate of surveying. The this study. Many methods for compared datasets is composed of determining the similarity between fifty polygons (buildings) were polygons have been investigated and selected in the centre of Baghdad introduced. In the preliminary city. The intention is to compare the analysis, the function of the surface data quality of these datasets by distance operation was applied as applying the methods mentioned in follows [8]: section 3. Initially this can be done ( ) (1) with visual comparison of derived ( ) maps, to get a general picture but the Where: P and P are two different methods indicated below will be 1 2 polygons. applied to construct a quantitative

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

ds: Surface distance value interval (0 and 1). The two polygons can be considered as similar, if the In this method, the surface distance value of the distance is close to zero, between polygons can be determined while the two polygons are disjoint, by dividing the intersection of if the distance is close to one. Fig. 3 polygons on union of them, as shows the basic elements of the shown in equation 1. The value of surface distance method [8]. surface distance is defined in the

P1 ds (P ,P ) = P2 1 2

(a) (b) Fig 3. The surface distance method

The comparison of polygonal ration between the area and similarity also involved compactness perimeter of the polygon, as follows: measurements. The compactness can (2) be defined as a numerical quantity ( ) representing the degree to which a shape is compact [12]. In GIScience, Where: C is compactness value. the shape compactness analysis has been a long standing for studying Elongation was also determined as urban sprawl, city planning and another metric of shape quality describing the hydrological measurement. It can be calculated by properties of drainage basins [2]. In dividing the width and the length of literature, several methods for the smallest rectangle that contain compactness measurement have the polygon, Fig. 4. The values of been introduced; see for example elongation are in the range of (0-1). [13]; [2]; [8]. The [13] method is The value close to zero means that adopted in this study because it can the shape is symmetric in all be considered as the most common directions such as square while method for measuring compactness. elongation value close to one The compactness is determined as a denotes that the shape is elongated in

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

one or more directions [14]. The A2 : the area of the same polygon in equation for elongation can be another datasets. expressed as: R : the ratio of the area of polygon

value. (3) The first step in the methodology Where: consisted in the extraction the W : is the width of the smallest coordinates of required polygons rectangle containing the polygon. from tested datasets. The extracted L : is the length of the smallest position information was imported rectangle containing the polygon. by ArcGIS to generate point layers E : is the elongation value. for showing building locations as

represented in the three datasets: GM, Wiki and RD. The resulting points map was subsequently

L connected and joined with stratified closed polygons. In the following step of the analysis, the geometric W properties of polygons, such as areas, perimeters, length of sides, and intersection of area, were Fig 4. Elongation concept determined. To be able to determine the shape similarity based on the Additional method for measuring geometric properties of tested similarity between polygons is the datasets, MatLab programming ratio of areas (R) computation. The language was employed to design a ratio of area is computed as a ratio of specialized program to perform this the area of polygon in first dataset task. The designed program has the and the area of the same polygon in ability to compute and compare the another dataset. The equation of this values of surface distance method, descriptor is given as follows: compactness, elongation, and ratio of areas computation through

(4) statistical analysis. The methodology

followed in this research can be Where: shown in Fig. 5. A1 : the area of polygon in one dataset.

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

100

Dataset #1 (GM) Dataset #2 (Wiki) Dataset #3 (RD)

Geospatial Data Extraction

Data Manipulation using ArcGIS

Geometric Properties Computations

Shape Accuracy Assessment

Statistical Analysis

Fig 5. Flow diagram of the shape similarity process

4. Quality Assessment of Figure 6 shows the distribution of Collaborative Mapping the output of the surface distance Buildings method. It is clear that the compare data are close to each other, although Using the surface distance method the frequency of GM/RD data is a described in section 3, the accuracy little more in the first value. The of tested buildings was computed. concentration peak is between 0.05

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

101

and 0.35, which corresponds small data (i.e. buildings mapped from difference. This is also confirmed WikiMapia and Google Maps) are from the descriptive statistics in less than 0.5 which means that the table 1. The mean values for the differences between output values surface distance method of all tested are small.

0.4

0.35

0.3

0.25 0.2 GM/RD Wiki/RD

Frequency 0.15 GM/Wiki 0.1 0.05

Surface distance data

Fig 6. The distribution of the surface distance output for the tested datasets

Table 1. Descriptive statistics of surface distance method

GM/RD Wiki/RD GM/Wiki

Max 0.9503 0.8124 0.9231 Min 0.0003 0.0020 0.0011 Mean 0.3320 0.2692 0.4800 Medean 0.1536 0.2211 0.3571 SD 0.3794 0.2471 0.5946

The compactness vales are the low mean of 0.2564, 0.2515, and determined for each polygon of 0.2601 for the RD, GM, and Wiki tested dataset. More than fifty respectively. The compactness percent of RD data has compactness standard deviations were computed value less than 0.30, and it is almost for each dataset with a range of similar to the distribution of GM and 0.0975 and 0.1105. Therefore, in Wiki compactness values as shown general all tested datasets have a in Fig. 7. The computed very similar compactness values compactness is almost consistent because the buildings have regular among tested datasets. The shapes and compact. uniformity is presented in table 2 by

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

102

0.3

0.25

0.2 RD 0.15 GM

Frequency 0.1 Wiki

0.05

0 0.05-0.10 0.1-0.15 0.15-0.20 0.20-0.25 0.25-0.30 0.30-0.35 0.35-0.40 0.40-0.45 0.45-0.50 0.50-0.55 Compactness Data

Fig 7. The distribution of the compactness output for the tested datasets

Table 2. Descriptive statistics of compactness method

RD GM Wiki Max 0.4686 0.4877 0.5358 Min 0.1072 0.0954 0.0847 Mean 0.2564 0.2515 0.2601 Medean 0.2287 0.2207 0.2346 SD 0.0975 0.0985 0.1105

As stated before, the elongation for with lowest elongation mean is RD, each polygon was also determined and the highest is Wiki (0.4039 and and investigated. The elongation 0.4340, respectively). The results of RD, Wiki and GM datasets differences between GM and Wiki are presented in table 3 and Fig. 8. elongation are relatively small. The The elongation standard deviations RD overall elongation values are of the full sample (50 buildings) are slightly higher, which indicates that 0.3013, 0.2784, and 0.2725 for RD, the RD polygons are a little GM and Wiki respectively, with a elongated. range of 0.0169 to 0.9888. The data

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

103

0.2 0.18 0.16 0.14 0.12 0.1 RD

Frequency 0.08 GM 0.06 Wiki 0.04 0.02 0

Elongation Data

Fig 8. The distribution of the elongation output for the tested datasets

Table 3. Descriptive statistics of elongation method

RD GM Wiki Max 0.9564 0.9888 0.9658 Min 0.0427 0.0346 0.0169 Mean 0.4039 0.4112 0.4340 Medean 0.3902 0.3577 0.3568 SD 0.3013 0.2784 0.2725

Table. 4 and Fig. 9 compare the differences of standard deviations of output results of the ratio of area 0.5007, 0.3463, and 0.7665 for method. It is clear from Fig. 9 that GM/RD, Wiki/RD, and GM/Wiki most ratios of area data are respectively. Hence, it can be said distributed between 0 and 1.8. This that in general the areas of tested variety in values is reflected by the polygons tend to be different.

0.3

0.25

0.2

GM/RD 0.15

Wiki/RD Frequency

0.1 GM/Wiki

0.05

0 0-0.2 0.2-0.4 0.4-0.6 0.6-0.8 0.8-1 1-1.2 1.2-1.4 1.4-1.6 1.6-1.8 2-2.2 2.2-2.4 2.4-2.6 2.6-2.8 2.8-3 >3

Ratio of Area Data

Fig 9. The distribution of the ratio of areas output for the tested datasets

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

104

Table 4. Descriptive statistics of the ratio of areas method

GM/RD Wiki/RD GM/Wiki Max 2.4503 1.5319 3.7131 Min 0.0676 0.1676 0.2355 Mean 0.9222 0.8774 1.0460 Medean 0.9900 0.9547 0.9564 SD 0.5007 0.3463 0.7665

5. Conclusion and Outlook The findings showed that the polygons quality of different The VGI data or collaborative collaborative mapping datasets is mapping is where volunteers, usually almost similar, although there is a untrained, and they have different slight difference between statistical background and expertises, create values of tested datasets. For spatial data based on web platform. instance, the mean values of Due to the lacking of compactness are 0.2564, 0.2515, and standardization, the VGI quality is 0.2601 for RD, GM and Wiki data vary and heterogeneous from respectively. On the other hands, the different data sources. Therefore, mean values of surface distance there is a need for better method are 0.3320, 0.2692 and understanding of the quality of free 0.4800 for GM/RD, Wiki/RD and geospatial data, in particular their GM/Wiki comparison respectively. accuracy. To the best of our Therefore, this kind of datasets can knowledge, the majority of be used to enrich and improve each published research assessed the other. In addition, free geospatial quality of OpenStreetMap (OSM) data can also aid the completion of data as a common VGI data source. geospatial datasets in some areas Here, this paper proposed a around the world, especially in methodology for evaluating the countries lacking full authoritative quality of buildings on a further VGI geospatial datasets. Collaborative services such as WikiMapia and mapping data can be utilised to Google Maps. Fifty polygons develop spatial data infrastructure (buildings) were selected and (SDI) or urban planning. Thus, examined in centre of Baghdad. The policy and institutional issues could evaluation process involved several be addressed to demonstrate the methods and techniques such as advantages of the approach surface distance method, developed. compactness, elongation, and ratio of area. For further studies, the above methodology can be transferred to

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

105

multi data sources such as Bing Sons Ltd, West Sussex, United Maps, Yahoo Maps, etc. This can be Kingdom. done by the development of a prototypical application that allows [4] Elwood, S., Goodchild, M. and for location, selection and assessed Sui, D, (2012), Researching shape accuracy in the web Volunteered Geographic environment. Thus, web services Information: Spatial Data, need to be included in the data Geographic Research, and New flowline. Subsequently, refinements Social Practice. Annals of the and feedback will be undertaken to Association of Geographers 102(3). the framework. Another possible p.p.571-590. area of future research would be to investigate the creation of a quality [5] Fan, H., Zipf, A., Fu, Q., and evaluation index which may assist to Neis, P., (2014), Quality assessment determine the extent to which dataset for building footprints data on can be used for a range of purposes. OpenStreetMap, International Issues of the creation of a universal Journal of Geographical Information mathematically derived index need Science, 28 (4), 700-719. to be addressed as well. [6] Fram, C, Chistopoulou, K, and References Ellul, C, (2015), Assessing the quality of OpenStreetMap building [1] Brown, M., (2006), Hacking data and searching for a proxy Google Maps and Google Earth, variable to estimate OSM building Wiley Publishing, Inc., Indianapolis, data completeness, GISRUK 2015 Indiana, USA. conference, Leeds, UK, p.p. 1-11.

[2] Chandra, S., Chhetri, P. and [7] Genovese & Roche, (2010), Corcoran, J., (2009). Spatial patterns Potential of VGI as a resource for of urban compactness in Melbourne: SDIs in the North/South context, an urban myth or a reality. In: Geomatica, 64 (4), p.p. 439-450. Ostendorf, B., Baldock, P., Bruce, D., Burdett, M. and Corcoran, P., [8] Girres & Touya (2010). Quality eds. Proceedings of the Surveying assessment of the French and Spatial Sciences Institute OpenStreetMap dataset. Biennial International Conference, Transactions in GIS 14(4). p.p. 435- Adelaide, Australia, p.p. 231-242. 459. [3] Crampton, J., (2010), Mapping A critical Introduction to [9] Goodchild, M. F., (2007), Cartography and GIS, John Wily & Citizens as Sensors: The World of

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

106

Volunteered Geography, Geographical Information Science GeoJournal, 69, 211–221. 27(6), p.p. 1227–1250.

[10] Haklay, M., (2010), How Good [13] Maceachren (1985). is Volunteered Geographical Compactness of geographic shape: Information? A Comparative Study Comparison and evaluation of of OpenStreetMap and Ordnance measures. Geografiska Annaler Survey Datasets, Environment and 67(1). p.p 53-67. Planning B: Planning and Design, 37 (4), PP. 682–703. [14] Mingqiang, Kidiyo & Joseph (2008). A survey of shape feature [11] Hecht R, Kunze C, Hahmann extraction techniques. Pattern S,(2013), Measuring completeness of Recognition p.p.43-90. building footprints in OpenStreetMap over space and time, [15] Surhone, L., Tennoe, M., ISPRS Int J Geo-Information 2, Henssonow, S., 2010, Volunteered p.p.1066-1091. Geographic Information, VDM [12] Li, W., M. F. Goodchild, and R. Publishing house Ltd., Beau Bassin, L. Church. (2013). An Efficient Mauritius. Measure of Compactness for Two- Dimensional Shapes and Its [16] Wikimapia Statistics, (2016), Application in Regionalization available at Problems. International Journal of http://wikimapia.org/wiki/Wikimapi a_Statistics [Accessed 10 April 2016]

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017

107

حتليل جودة أشلال البنايات يف اخلرائط التعاونية

د. ميثم مطصر شرقي مدرس قسم هندسة املساحة / كلية اهلندسة - جامعة بغداد بغداد – العراق

اخلالصة لكد وكنت التطٌزات الطسٍعْ يف تكنَات احلٌاضَب ًاالنرتنَت الغري املدتصني ون مجع ًنشس البَانات اجلغسافَْ عمٓ االنرتنَت بكن ضوٌلْ ًبدًن تعكَد. ان هرا النٌع اجلدٍد ون البَانات عادّ واٍعسف باخلسائط التعاًنَْ اً املعمٌوات املكانَْ اليت ٍنتجوا املتطٌعٌن. هناك العدٍد ون البمدان حٌه العامل ممكن ان تطتفاد ون هري البَانات اً املعمٌوات ًذلك لكٌنوا دلانَْ ًتٌفس املصٍد ون البَانات املكانَْ ًحطب الطمب ، خصٌصا يف تمك الدًه اليت التتٌفس فَوا خسائظ زمسَْ وعتىدّ ًونتجْ ون املٌضطات املدتصْ. وع ذلك فان هناك قمل وتصاٍد حٌه نٌعَْ هري املعمٌوات ًدقتوا لكٌن االشداص املتطٌعني جلىع هري املعمٌوات ًنشسها قد تعٌشهي اخلةرّ ًاالختصاصيف هرا اجملاه. زكصت وعظي الدزاضات الطابكْ عمٓ دزاضْ ًحتمَن دقْ العٌازض النكطَْ ًاخلطَْ هلرا النٌع ون البَانات. تودف الدزاضْ احلالَْ اىل تكََي جٌدّ نٌع اخس ون العٌازض ًهُ املطمعات )املبانُ( لنٌعني ون بَانات اخلسائط التعاًنَْ Google Maps , WikiMapia. لكد متت وكازنْ هري البَانات وع بَانات وسجعَْ وطتدمصْ ون صٌز جٌٍْ عالَْ الٌضٌح دلوصّون اهلَاّ العاوْ لمىطاحْ . ان املنوجَْ املكرتحْ هلرا البحث تعتىد بشكن اضاضُ عمٓ تطبَل عدد ون املٌدٍالت ًاالضالَب السٍاضَْ وثن ratio of areas ,elongation ,compactness ,surface distance method. لغسض تكََي ًحتمَن دقْ املطمعات املطمٌبْ فكد مت حطاب ًوكازنْ الكَي االحصائَْ التكمَدٍْ وثن الٌضط احلطابُ ، الٌضَط ، االحنساف املعَازِ ، احلد االدنٓ ، احلد االعمٓ لمنتائج املطتحصمْ ون تطبَل املٌدٍالت السٍاضَْ الطابكْ. أشازت نتائج هري الدزاضْ اىل ًجٌد فسًقات بطَطْ بني دقْ اشكاه املطمعات )البناٍات( املكازنْ ًعمَى ميكن تٌظَف هري البَانات يف تطٌٍس البنٓ التحتَْ لمىعمٌوات املكانَْ اً تطبَكات التدطَط احلطسِ ًاالقمَىُ.

الكمىات املفتاحَى: خسائط الكٌكن ، دقْ االشكاه ، املطمعات ، اخلسائط التعاًنَْ .

Maythm Al-Bakri Association of Arab Universities Journal of Engineering Sciences NO. 2 Volume. 24 Year. 2017