A piece-wise polynomial fitting method to filter the interferogram phase noise [InSAR]

Among many important factors for the whole interferometric processing, the interferometric phase quality affects the phase unwrapping and the accuracy of the DEM directly. The noise, which exists in the single complex image, affects the judgment of the image. After the interferogram phase (IP) is taken out from the complex images, the distribution of the phase noise is complicated. To get the high quality IP, the IP noise, which always exists, must be eliminated. The IP noise should be distributed continuously in one period and disconnected from one period to another. The polynomial fitting method can eliminate the phase noise in the continuous interval, but it can also add noise in the discontinuous interval, especially in the whole phase interval. This paper brings forward the piece-wise polynomial fitting (PWPF) method that can eliminate the IP noise. The PWPF can be obtained from the range and azimuth director respectively in the whole IP image. The experiment shows this method has a better IP noise eliminating effect.