SAR IMAGE CHANGE DETECTION METHODS BASED ON GLCM TEXTURE FEATURES
Han Jing 1,2) ; Deng Kazhong 1,2) ; and Li Beicheng 1)
1)School of Environment Science and Spatial Informatics, CUMT, Xuzhou 221116;2)China University of Mining and Technology, Key Laboratory for Land Environment and Disaster Monitoring of SBSM, Xuzhou 221116
Abstract The authors found difference images based on the contrast can stand out changed information better using texture features extraction of SAR images based on gray level coocurrence matrix, to analyze the principle of the GLCM, feature vectors and the characteristic parameters determined,logarithmic ratio operator constructed difference images, we made the difference images based on the contrast as the base of change detection. As the images in accordance with the Gaussian mixture model, so we estimate the parameters of the Gaussian mixture model with expectation maximum (EM) algorithm, and then use Bayesian minimum error rate to extract change information,finally compare it with the change detection results based on the pixel grayscale value. The test proved that the change detection method based on GLCM texture features has the lower false alarm rate, the lower missing rate, the smaller overall error and better detection effect.