Landslide Detection Based on GLCM Using SAR Images

Detecting the change of texture information is very important for landslide detection. In polarimetric SAR imagery, the landslides not only have typical polarimetric features but also have rich texture features. In this study, the texture feature extraction technique based on gray scale co-occurrence matrix was applied to process two SLC Sentinel-1A C-band SAR images and two COSMO-SkyMed X-band SAR images acquired before and after landslip for the landslide detection leading to the Jomda landslide event in 2018, Tibet of China and the Qinglong landslide in 2018, Guizhou of China. Several important characteristic parameters of the gray scale co-occurrence matrix were studied, and the optimal sliding window size and the optimal step size for landslide detection were obtained through rigorous analysis and comparison. The experimental results show that the analysis of the complex texture information can effectively detect the landslide in the mountain area with high accuracy.