Effect Analysis and Spectral Weighting Optimization of Sidelobe Reduction on SAR Image Understanding

This paper focuses on effect analysis and spectral weighting optimization of sidelobe reduction on synthetic aperture radar (SAR) image understanding. First, a statistical model for each pixel in complex-amplitude distribution of the SAR image is derived to investigate how the sidelobes of the nearby scatterers change the speckle noise. Second, the phase error is analyzed for interferometry applications when high sidelobe is in the presence. The potential benefit of using a sidelobe reduction method for improving interferometric phase accuracy is also investigated. Third, to better understand how the sidelobes affect the quality of SAR images at different sidelobe levels, with the same imaging resolution, one novel type of weights is optimized to improve the performance of spectral weighting for SAR sidelobe reduction by using one bioinspired method. Intensive sidelobe reduction experiments are carried out, and the results are verified on the probability density function (PDF), the level of speckle noise, and interferometric phase accuracy of the SAR images. The experimental results indicate that the sidelobe reduction methods can change the speckle noise and the statistical distributions for the underlying SAR imagery. One of the interesting findings is that a better equivalent number of looks (ENL) and interferometric phase accuracy can be obtained by using the optimized spectral weighting. However, the spatially variant apodization damages the PDFs of the SAR data, decreases the ENL of the SAR images, and leads to significant phase distortion for SAR interferometry.