An adaptive scalloping suppression method for spaceborne ScanSAR images based on pre-estimation and weighted filtering

Abstract

Spaceborne scanning synthetic aperture radar (ScanSAR) is widely used in global observations due to its ability to perform wide-swath mapping in the range direction. However, its special working mode causes scalloping, which reduces the quality of images and affects subsequent applications. According to its characteristics, an adaptive method based on pre-estimation and weighted filtering is proposed in this paper to suppress scalloping in the image domain. First, the azimuth intensity distribution of the image after scalloping suppression is estimated, which is used for scene stationarity test. Then, the images that cannot meet the stationary standard are segmented into subimages using maximum entropy principle and mathematical morphology. Finally, an algorithm based on adaptive weighted filtering is introduced to suppress scalloping, and the suppressed subimages are fused to obtain final results. The performance of the proposed method is tested with real ScanSAR data from the Gaofen-3 (GF-3) satellite. The experimental results indicate that the effect of scalloping suppression is excellent, since the depth of scalloping can be suppressed to approximately 0.3 dB. Notably, the proposed system is efficient in processing large-area images, such as GF-3 ScanSAR images whose actual width is more than 300 km. The entire process requires no parameter adjustment, and the proposed method is suitable for various complex scenes. The effectiveness, high efficiency, adaptability and robustness of the proposed system are verified.