SAR moving-target detection using difference between two looks

A novel scheme is presented to detect moving targets in SAR images. In the two looks, the images of a stationary target are similar, but those of a moving target are different. So, moving targets can be detected using the difference between the two looks. First, the complex image is compensated to remove the shifting and the blurring of the stationary ground due to the velocity error of the radar. This will reduce the loss of the similarity between the two looks of the stationary ground. Then, the Doppler spectrum is halved, each half is used to obtain a look, and the two looks are despeckled to reduce the loss of their similarity. Finally, the normalized difference between the two looks is calculated, smoothed, absolutized, and thresholded. This scheme is effective for targets moving in range, azimuth, or both. In addition, it is easy to implement and efficient in computation.