Description

SAR Image change detection using OpenCV-Deeplearning

SAR images are increasingly gaining importance in the field of remote sensing applications. The receiving SAR data are affected by multiplicative speckle noise which form granular patterns. Hence despeckling is done prior to object detection. Here a comparison of several transform domain techniques are carried out for despeckling SAR images. Performance parameters such as ENL, PSNR, SSIM etc. are used to evaluate the result. From the result, Curvelet is chosen as the optimum technique for despeckling. In this paper SAR image change detection is performed and further object detection is done using CA-CFAR.