To improve the performance of autofocus in SAR images with an azimuth adaptive filter

In SAR azimuth signal processing, it is very common to subsample the raw data with a filter, which may reduce the Signal Noise Ratio(SNR) of the raw image. Since autofocus is to estimate the Doppler parameters from the raw data, with the reduced SNR, the result of autofocus may be incorrect. And images processed with the incorrect parameters will be worsened. This paper analyses the effect of SNR on autofocus and proposes an adaptive filter in SAR azimuth subsampling to keep SNR as high as possible, so that the performance of autofocus can be improved and higher image quality achieved. Finally, an image processed with the adaptive filter is contrasted with the result of a traditional filter, which proves the effectiveness of the adaptive filter on autofocus and the final image.