Urban Change Detection in TerraSAR Image Using the Difference Method and SAR Coherence Coefficient

Zhang Xuedong, Beijin g University of Civil Engineering and Architecture , Beijing 10 2616, China, Liu Wenxi, He Shuguang, Key Laboratory of modern urban surveying and mapping, State Bureau of surveying, mapping and Geographical Information Bureau, Beijing 102616, China, Beijin g University of Civil Engineering and Architecture , Beijing 10 2616, China, Key Laboratory of modern urban surveying and mapping, State Bureau of surveying, mapping and Geographical Information Bureau, Beijing 102616, China, China Association of Constructio n Education, Beijing 100037, China
2018 Journal of Engineering Science and Technology Review  
At present, most methods of urban change detection with synthetic aperture radar (SAR) images are based on image amplitude information. SAR images of the same region over different periods are usually compared and analyzed by using the difference method, which often neglects the SAR image interference information, exhibiting sensitivity to changes in ground scatterers. Therefore, this study presented a novel method combining the difference method and the SAR coherence coefficient to accurately
more » ... nd comprehensively detect the characteristics of urban change information. In this method, the amplitude difference between two images was obtained by the image difference method, and the suspected change areas in the urban area were determined. The suspected change areas of the image were further analyzed by using the coherence coefficient of the images and were regarded as genuinely changed areas when the results of the difference method and the coherence coefficient both changed. Finally, the proposed method and the difference method were employed to conduct a comparative study at Huilongguan Village in Changping District of Beijing city, China. Results indicate that the difference method can only reflect urban changes caused by urban construction (e.g., demolition and urban expansion), whereas the proposed method can not only detect urban changes due to urban construction but also remove urban pseudo-changes attributed to season change, vegetation, and tree growth. Thus, the proposed method can effectively improve the accuracy of change detection. Comparative analysis and field verification of the selected region of interest showed that the result accuracy of the proposed method reaches 84.5%, which is much higher than that of the single difference method (61.4%). This finding proves that the proposed method is feasible for urban change detection, and the conclusions can also provide a reference for urban planning and construction.
doi:10.25103/jestr.113.03 fatcat:jkujwfccsbapfl4mcpd7oxd66u