Interferometric phase filtering method based on difference chart and compensation filtering

Abstract

The invention provides an interferometric phase filtering method based on a difference chart and compensation filtering. The interferometric phase filtering method comprises the steps of: firstly, extracting principal components of interferometric phases from an input interferometric phase image through frequency domain filtering based on the differential chart, so as to reduce density of interferometric fringes in the residual interferometric phase image and facilitate subsequent processing; and secondly, performing compensation filtering on the residual interferometric phase image, wherein compensation filtering adopts BM3D filtering taking a phase derivative variogram of the residual interferometric phase image as self-adaptive filtering parameters, and can better extract detail components of the interferometric phases. The interferometric phase filtering method comprises frequency domain filtering and space domain filtering, is suitable for processing the interferometric phases containing dense interferometric fringes, and can effectively filter out noise without destroying the continuity of the dense interference fringes.

Description

technical field [0001] The invention belongs to the intersecting technical field of remote sensing and signal processing, in particular to an interference phase filtering method suitable for dense interference fringes. Background technique [0002] The interferometric phase is obtained by conjugate multiplication of two SAR images and taking the phase of the multiplication result. Interferometric phase is an important physical quantity in SAR interferometry, and its quality will determine the accuracy of the final product—digital elevation model or terrain deformation. However, due to decorrelation factors, there is always serious space-varying noise in the interferometric phase. Space-variant noise will not only introduce residual points but also destroy the distribution of the interferometric phase, which increases the difficulty of subsequent phase unwrapping and eventually leads to a decrease in product accuracy, so it must be filtered out through the interferometric ph...

AI Technical Summary

Problems solved by technology

[0004] However, there are two problems in applying BM3D filtering algorithm to interferometric phase filtering: (1) For dense interference fringes, since the filtering window of BM3D filtering algorithm may span multiple interference fringes, the continuity of interference fringes will be destroyed; (2) The BM3D filtering algorithm needs adaptive parameters that can represent the variance of the image noise as the filtering input parameters. Different filtering input parameters will lead to different filtering results, but the existing BM3D filtering algorithm does not provide a calculation method for the filtering input parameters.

Images