The Study Of Urban Extraction Method Based On SAR And Optical Remote Sensing Satellite Images

In the global land resources developed and utilized environment,urbanization has been an important social and economic phenomenon taking place in an unprecedented scale and rate all over the world.But the uncontrollable urbanization has negative impacts on the society,ecology,and economy.Therefore,it is of great importance to detect urban area in order to prevent these negative consequences.Nowadays,with the successful launch of high-resolution Synthetic Aperture Radar(SAR),it is becoming more and more convenient to obtain high-resolution SAR images,and radar imagery is at an absolute advantage in areas where environmental conditions impede the acquisition of optical image data.So extraction of urban based on SAR data has become a research focus all over the world.In this paper,Through the analysis of existing approaches of extraction urban area based SAR images,it is found that these methods are only applicable to the extraction of the built-up area of the mature urban construction.In the high-speed constructing urban area where frequent occurrence of the demolition area existed,leakage phenomenon was very serious,resulting in poor extraction accuracy.The demolition area in the context of large-scale regional research,especially for urban rapid expansion and transformation of the region,the demolition area cannot be ignored.To solve this problem,we propose a method to extract the urban area combined SAR image and optical satellite image to improve the accuracy of extraction urban in large scale in this study.The idea of the approach starts from the analysis of the local speckle characteristics and intensity information of SAR image to monitor spatial extent of the built-up area and non-built-up area.Then the maximum likelihood method and post-classification are used to extract the demolition area within cities based on the optical image.The final extraction of urban area includes the above built-up area and demolition area.In order to evaluate the applicability and accuracy of the improved method,we take different resolution SAR images(high-resolution single-polarized TerraSAR-X and ALOS-PALSAR image)as the data source,and select the old city of Changzhou as the study area.The Erdas modeling method is used extract the BAs from TerraSAR-X and ALOS-PALSAR image,and the maximum likelihood method is used to extract BAs of Rapid Eye image,then comparison of the classification results and evaluation of the accuracy are made.The results of accuracy analysis indicate that the method based on TerraSAR-X image is better than the other two images and the method based on Erdas modeling is more suitable for high-resolution SAR images.Then,the whole urban area including both build-up area and non-build-up area based on TerraSAR-X image and optical satellite image is extracted.The results show that the overall classification accuracy is higher than the previous methods,indicating that this method is feasible.