利用中低分辨率合成孔径雷达(SAR)影像,通过灰度共生矩阵提取不同纹理窗口大小的纹理特征来构造差异影像,并结合Otsu阈值分割方法来获取变化图像。实验结果表明,当检测地物单一、变化较明显的区域时,通过选用均值纹理特征并结合相应纹理窗口,中低分辨率SAR影像能够满足变化检测精度的要求。 The medium and low resolution synthetic aperture radar (SAR) images were used to construct the differential images by extracting the texture features of different texture windows through the gray level co-occurrence matrix. The Otsu threshold segmentation method was used to obtain the change images. The experimental results show that the medium and low resolution SAR images can meet the requirements of change detection accuracy by using the mean texture features and the corresponding texture windows when detecting the single and obvious changes of the features.