تعیین روش بهینه طبقه بندی و نقشه سازی کاربری/ پوشش اراضی با مقایسه الگوریتم های شبکه عصبی مصنوعی وماشین بردار پشتیبان با استفاده از داده های ماهواره ای (مطالعه موردی: تالاب بین المللی هامون)
محورهای موضوعی : هوش مصنوعی
امیرهوشنگ احسانی
1
(دانشیار، پردیس دانشکده‌های فنی، دانشکده محیط زیست، دانشگاه تهران، تهران، ایران. *(مسوول مکاتبات))
مجتبی شاکریاری
2
(کارشناس ارشد علوم محیط زیست، دانشکده محیط زیست، دانشگاه تهران، تهران، ایران.)
کلید واژه: شبکه عصبی مصنوعی, طبقه بندی پوشش اراضی, تالاب بین المللی هامون, داده های ماهواره ای,
چکیده مقاله :
زمینه و هدف: طبقه بندی تصاویر یکی از روش های مهم درتفسیرتصاویر ماهواره ای است که کاربرد زیادی در بررسی تغییرات زمین دارد. در این میان داده های ماهواره ای به دلیل ارایه اطلاعات به روز، ارزان بودن و تنوع اشکال بهترین وسیله برای آشکارسازی و ارزیابی تغییرات شناخته شده است. از طرفی دیگر در سال های اخیر روش های شبکه های عصبی مصنوعی به طور وسیع و گسترده جهت طبقه بندی داده های ماهواره ای استفاده می شود. هدف از این پژوهش مقایسه سه روش مختلف جهت طبقه بندی پوشش اراضی با استفاده از تصویر سنجده OLI سال 2014 طی یک دوره 26 ساله می باشد. روش بررسی: در این مقاله تصویر سنجنده OLI (1393) از لحاظ هندسی و اتمسفری در نرم افزار ENVI تصحیح شد. سپس جهت طبقه بندی تصویر به سه روش شبکه های عصبی مصنوعی آرتمپ فازی، شبکه عصبی مصنوعی پرسپترون چند لایه و روش ماشین بردار پشتیبان با استفاده از نرم افزار IDRIS Selva، نقشه پوشش اراضی به پنج کلاس آب، پوشش گیاهی، نیزار، اراضی بایر و اراضی شور طبقه بندی گردید. در نهایت به منظور ارزیابی صحت با استفاده از صحت کاربر، صحت تولید کننده، صحت کلی، ضریب کاپا و ماتریس خطا، نقشه ایجاد شده با نقشه واقعیت زمینی ایجاد شده توسط GPS و تصاویر گوگل ارث و بازدیدهای صحرایی مورد مقایسه قرار گرفت. بحث و نتیجه‌گیری: نتایج نشان دادند که روش آرتمپ فازی بیش ترین میزان دقت را با صحت کل 94.68 و ضریب کاپای91/. نسبت به دو روش شبکه عصبی مصنوعی پرسپترون چند لایه با صحت کل 92.99 و ضریب کاپای 89/. و ماشین بردار پشتیبان با صحت کل 90.93و ضریب کاپای 85/. در طبقه بندی داده های ماهواره ای دارد.
Background and Objective: Images classification is one of the important techniques for interpretation of satellite images that is widely used in survey of earth changes. In the meantime, satellite data has been recognized as the best tool for detection and evaluation of changes due to its update information, low costs and variety of forms. Therefore, land use/land cover map is one of the most important information required by the environmental managers and planners. On the other hand, in recent years, artificial neural network method has been used widely for the classification of satellite data. The aim of this study is to compare three different methods for land cover classification using 2014 OLI image over a 26-year period. Method: In this study, digital data of OLI (2014) sensor was used in order to optimize image classification method. Initially, the image was corrected in terms of geometry and radiometry in the ENVI software. Then IDRISI software was used for image classification using three different methods: fuzzy artmap, multilayer perceptron artificial neural networks and support vector machine. Finally, land cover maps were classified into five categories: water, vegetation, canebrake, barren lands and saline lands. To evaluate accuracy with the help of user accuracy, producer accuracy, overall accuracy, kappa coefficient and error matrix, the created map was compared with the ground reality map created by GPS, Google Earth images and field observations. Discussion and Conclusion: The results of image accuracy evaluation showed that among the applied methods the fuzzy artmap algorithm had the highest accuracy in classification of satellite data with an overall accuracy of 94.68 and kappa coefficient of 0.91 compared to both multilayer perceptron artificial algorithm with an overall accuracy of 92.99 and kappa coefficient of 0.89 and support vector machine with an overall accuracy of 90.93 and kappa coefficient of 0.85. This study showed that classification of fuzzy artmap artificial neural network algorithm has a high capability to create the land cover map with high accuracy.
1- Badreldin N, Goossens R, 2013, Monitoring land use/land cover change using multi-temporal Landsat satellite images in an arid environment: a case study of El-Arish, Egypt. Arab J Geosci 7:1671– 1681. doi:10.1007/s12517-12013-10916-12513.
2- Gardi C, Panagos P, Van Liedekerke M, Bosco C, De Brogniez D, 2014, Land take and food security: assessment of land take on the agricultural production in Europe. J Environ Plann Manag 58:898–912.
3- Ghobadi Y, Pradhan B, Shafri H, Kabiri K ,2013, Assessment of spatial relationship between land surface temperature and landuse/cover retrieval from multi-temporal remote sensing data in South Karkheh Sub-basin, Iran. Arab J Geosci 8:525–537. doi:10.1007/ s12517-013-1244-3.
4- Matinfar H, Alavi Panah S, Zand F, Khodaei K, 2013, Detection of soil salinity changes and mapping land cover types based upon remotely sensed data. Arab J Geosci 6:913–919. doi:10.1007/s12517-011- 0384-6.
5- Arekhi S, Jafarzadeh A, 2014, Forecasting areas vulnerable to forest conversion using artificial neural network and GIS (case study: northern Ilam forests, Ilam province, Iran). Arab J Geosci 7:1073– 1085. doi:10.1007/s12517-012-0875-1.
6- Pradhan B, 2013, A comparative study on the predictive ability of the decision tree, support vector machine and neuro-fuzzy models in landslide susceptibility mapping using GIS. Comput Geosci 51: 350–365. doi: 10.1016/j.cageo.2012.08.023.
7- Alavi Panah, Seyyed Kazem, 2003. Application of Remote Sensing in Geosciences (Soil Sciences), First Edition, Tehran University Press, pp. 478. (In Persian)
8- Lu, D, and Weng, Q, 2007, A survey of image classification methods and techniques for improving classification performance. International Journal of Remote Sensing Vol.28, No. 5, 10 March 2007, 823–870.
9- Tien Bui D, Pradhan B, Lofman O, Revhaug I, Dick OB 2012 Landslide susceptibility assessment in the Hoa Binh province of Vietnam: a comparison of the Levenberg–Marquardt and Bayesian regularized neural networks. Geomorphology 171:12–19. doi:10.1016/j. geomorph.2012.1004.1023.
10- Joachims, T. 1999. Making large scale SVM learning practical. In Advances in
Kernel Methods-Support Vector Learning, pp.169-184.
11- Alavipanah, S.K., Masoudi, M., 2000, Land Use Mapping Using Landsat TM and Geographic Information System (GIS), a Case Study: Mouk Region of Fars Province, J. Agri. Sci. Natural Resources, Vol. 8, No. 1, pp. 65-75.
12- Borak, J.S. and Strahler, A.H. 1999. Feature selection and land cover classification of a MODIS-like dataset for a semiarid environment. International Journal of Remote Sensing. 20: 919-938.
13- Amiri, A., Chavooshi, H. and Amini, J. 2007. Comparison of Three Satellite Image Classification: Fuzzy, Neural Network and Minimum Distanc Geomatic Conference, National Cartographic Center, Tehran.
14- Mas, J.F., 2003, An Artificial Neural Networks Approach to Map Land Use/cover Using Landsat Imagery and Ancillary Data, Proceedings of the International Geosciences and Remote Sensing Symposium IEEE IGARSS 2003, Vol. VI, pp. 3498-3500, Toulouse, France.
15- Jianjun, J., Jie, Z., Hongan, W., Li, A., Hailing, Z., Li, Z., Jun, X., 2005, Land Cover Changes in the Rural-urban Interaction of Xian Region Using Landsat TM/ETM Data, Journal of Geographical Science, Vol. 15, No. 4, pp. 423-430.
16- Arekhi Saleh and Adibnejad. Mostafa, 2011, Evaluation of the efficiency of support vector machine algorithms for land use classification using Landsat + ETM satellite imagery (case study: Ilam Dam), Natural Resources Department, Agriculture Sciences Faculty, University of Ilam, Scientific Journal of Research Pasture and Desert of Iran, Volume 18, Number 3, Page 420-440. (In Persian)
17- Watts, D., 2001, Land Cover Mapping by Combinations of Multiple Artificial Neural Networks, MSc. Thesis, Department of Geomatics Engineering, University of Calgary.
18- Gahegan, M., German, G. and West, G., 1999, Improving Neural Network Performance on the Classification of Complex Geographic Datasets, Journal of Geographical Systems, No. 1, pp. 3-22.
19- Wijaya, A., 2005, Application of Multi-Stage Classification to Detect Illegal Logging with the Use of Multi-Source Data, MSc. Thesis, ITC, Enschede, The Netherlands.
20- Liu, X.H., Skidmore, A.K. and Oosten, H.V., 2002, Integration of Classification Methods for Improvement of Land-cover Map Accuracy, ISPRS Journal of Photogrammetry & Remote Sensing, No.56, pp. 257-268.
21- Hung, C.C., Coleman, T.L. and Long, O., 2004, Supervised and Unsupervised Neural Models for Multispectral Image Classification, ISPRS, http://www.isprs.org/istanbul 2004/ Comm7/papers/20.pdf.
22- Vapnik, V .N., 1995.The Nature of Statistical Learning Theory (New York :Springer Verlag ).
23- Hsu, C.W. and Lin, C.K., 2002. Acomparison of methods for multiclass support vector machines. IEEE Transactions on Neutal Networks 13, 415-425.
24- Yousef, S., Tazeh, M., Mirezee, S., Moradi, H.R. and Tavanger, S.H. 2001. Comparison of different classification algorithms in satellite imagery to produce land use maps (Case study: Noor city), Journal of Applied RS & GIS Techniques in Natural Resource Science. 2(2): 15-25.
25- Dixon, B. and Candade, N. 2008. Multispectral land use classification using neural networks and support vector machines: one or the other, or both? International Journal of Remote Sensing. 29: 1185–1206.
26- Arekhi Saleh, 1393, Preparation of land use map of Abbas Ilam Plain using artificial neural network, backup vector machine and maximum probability, Journal of Rangeland, Issue 2: Pages 30-43. (In Persian)
27- Lizarazo, I., 2006, Urban Land Cover and Land Use Classification Using High Spatial Resolution Images and Spatial Metrics, Proceedings of the 2nd Workshop of the EARSeL SIG on Land Use and Land Cover, pp. 292-298.
1- Badreldin N, Goossens R, 2013, Monitoring land use/land cover change using multi-temporal Landsat satellite images in an arid environment: a case study of El-Arish, Egypt. Arab J Geosci 7:1671– 1681. doi:10.1007/s12517-12013-10916-12513.
2- Gardi C, Panagos P, Van Liedekerke M, Bosco C, De Brogniez D, 2014, Land take and food security: assessment of land take on the agricultural production in Europe. J Environ Plann Manag 58:898–912.
3- Ghobadi Y, Pradhan B, Shafri H, Kabiri K ,2013, Assessment of spatial relationship between land surface temperature and landuse/cover retrieval from multi-temporal remote sensing data in South Karkheh Sub-basin, Iran. Arab J Geosci 8:525–537. doi:10.1007/ s12517-013-1244-3.
4- Matinfar H, Alavi Panah S, Zand F, Khodaei K, 2013, Detection of soil salinity changes and mapping land cover types based upon remotely sensed data. Arab J Geosci 6:913–919. doi:10.1007/s12517-011- 0384-6.
5- Arekhi S, Jafarzadeh A, 2014, Forecasting areas vulnerable to forest conversion using artificial neural network and GIS (case study: northern Ilam forests, Ilam province, Iran). Arab J Geosci 7:1073– 1085. doi:10.1007/s12517-012-0875-1.
6- Pradhan B, 2013, A comparative study on the predictive ability of the decision tree, support vector machine and neuro-fuzzy models in landslide susceptibility mapping using GIS. Comput Geosci 51: 350–365. doi: 10.1016/j.cageo.2012.08.023.
7- Alavi Panah, Seyyed Kazem, 2003. Application of Remote Sensing in Geosciences (Soil Sciences), First Edition, Tehran University Press, pp. 478. (In Persian)
8- Lu, D, and Weng, Q, 2007, A survey of image classification methods and techniques for improving classification performance. International Journal of Remote Sensing Vol.28, No. 5, 10 March 2007, 823–870.
9- Tien Bui D, Pradhan B, Lofman O, Revhaug I, Dick OB 2012 Landslide susceptibility assessment in the Hoa Binh province of Vietnam: a comparison of the Levenberg–Marquardt and Bayesian regularized neural networks. Geomorphology 171:12–19. doi:10.1016/j. geomorph.2012.1004.1023.
10- Joachims, T. 1999. Making large scale SVM learning practical. In Advances in
Kernel Methods-Support Vector Learning, pp.169-184.
11- Alavipanah, S.K., Masoudi, M., 2000, Land Use Mapping Using Landsat TM and Geographic Information System (GIS), a Case Study: Mouk Region of Fars Province, J. Agri. Sci. Natural Resources, Vol. 8, No. 1, pp. 65-75.
12- Borak, J.S. and Strahler, A.H. 1999. Feature selection and land cover classification of a MODIS-like dataset for a semiarid environment. International Journal of Remote Sensing. 20: 919-938.
13- Amiri, A., Chavooshi, H. and Amini, J. 2007. Comparison of Three Satellite Image Classification: Fuzzy, Neural Network and Minimum Distanc Geomatic Conference, National Cartographic Center, Tehran.
14- Mas, J.F., 2003, An Artificial Neural Networks Approach to Map Land Use/cover Using Landsat Imagery and Ancillary Data, Proceedings of the International Geosciences and Remote Sensing Symposium IEEE IGARSS 2003, Vol. VI, pp. 3498-3500, Toulouse, France.
15- Jianjun, J., Jie, Z., Hongan, W., Li, A., Hailing, Z., Li, Z., Jun, X., 2005, Land Cover Changes in the Rural-urban Interaction of Xian Region Using Landsat TM/ETM Data, Journal of Geographical Science, Vol. 15, No. 4, pp. 423-430.
16- Arekhi Saleh and Adibnejad. Mostafa, 2011, Evaluation of the efficiency of support vector machine algorithms for land use classification using Landsat + ETM satellite imagery (case study: Ilam Dam), Natural Resources Department, Agriculture Sciences Faculty, University of Ilam, Scientific Journal of Research Pasture and Desert of Iran, Volume 18, Number 3, Page 420-440. (In Persian)
17- Watts, D., 2001, Land Cover Mapping by Combinations of Multiple Artificial Neural Networks, MSc. Thesis, Department of Geomatics Engineering, University of Calgary.
18- Gahegan, M., German, G. and West, G., 1999, Improving Neural Network Performance on the Classification of Complex Geographic Datasets, Journal of Geographical Systems, No. 1, pp. 3-22.
19- Wijaya, A., 2005, Application of Multi-Stage Classification to Detect Illegal Logging with the Use of Multi-Source Data, MSc. Thesis, ITC, Enschede, The Netherlands.
20- Liu, X.H., Skidmore, A.K. and Oosten, H.V., 2002, Integration of Classification Methods for Improvement of Land-cover Map Accuracy, ISPRS Journal of Photogrammetry & Remote Sensing, No.56, pp. 257-268.
21- Hung, C.C., Coleman, T.L. and Long, O., 2004, Supervised and Unsupervised Neural Models for Multispectral Image Classification, ISPRS, http://www.isprs.org/istanbul 2004/ Comm7/papers/20.pdf.
22- Vapnik, V .N., 1995.The Nature of Statistical Learning Theory (New York :Springer Verlag ).
23- Hsu, C.W. and Lin, C.K., 2002. Acomparison of methods for multiclass support vector machines. IEEE Transactions on Neutal Networks 13, 415-425.
24- Yousef, S., Tazeh, M., Mirezee, S., Moradi, H.R. and Tavanger, S.H. 2001. Comparison of different classification algorithms in satellite imagery to produce land use maps (Case study: Noor city), Journal of Applied RS & GIS Techniques in Natural Resource Science. 2(2): 15-25.
25- Dixon, B. and Candade, N. 2008. Multispectral land use classification using neural networks and support vector machines: one or the other, or both? International Journal of Remote Sensing. 29: 1185–1206.
26- Arekhi Saleh, 1393, Preparation of land use map of Abbas Ilam Plain using artificial neural network, backup vector machine and maximum probability, Journal of Rangeland, Issue 2: Pages 30-43. (In Persian)
27- Lizarazo, I., 2006, Urban Land Cover and Land Use Classification Using High Spatial Resolution Images and Spatial Metrics, Proceedings of the 2nd Workshop of the EARSeL SIG on Land Use and Land Cover, pp. 292-298.
