Wavelet Domain InSAR Phase Filtering with Local Frequency Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for interferometric phase filtering in the wavelet domain fail to effectively remove noise while preserving details, leading to inaccuracies in interferometric phase measurements.
Innovation Solution
A method that transforms interferometric phase data into a complex field, performs local frequency estimation, and applies wavelet decomposition with threshold shrinkage processing to distinguish and filter noise from useful information, ensuring effective noise removal and detail preservation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If wavelet transform is applied to interferometric phase filtering to preserve detail information, then the resolution and detail preservation are improved, but noise information is not suppressed effectively resulting in poor noise removal effect
Solution Approach 1:
The patent applies different thresholding strategies to different wavelet coefficient sub-bands based on their frequency characteristics. Low-frequency sub-bands (approximating coefficients) use one thresholding approach while high-frequency sub-bands (detail coefficients) use another, allowing localized optimization of noise suppression versus detail preservation in different frequency regions
Solution Approach 2:
The patent dynamically adjusts thresholding parameters based on the statistical properties of wavelet coefficients and noise characteristics. By changing threshold values and thresholding methods according to local signal-to-noise ratio estimates, the filter adapts to preserve details in clean regions while suppressing noise in noisy regions
2Ease of manufacture
If conventional spatial domain filtering is used for interferometric phase filtering, then the implementation is simple, but phase details are easily broken and resolution is reduced when fringes are dense
Solution Approach 1:
The patent replaces conventional spatial domain filtering mechanisms with wavelet transform-based filtering in the frequency domain. This substitution allows filtering operations to be performed on wavelet coefficients rather than directly on pixel values, preserving spatial relationships while enabling more sophisticated noise suppression that maintains phase details
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
The present disclosure provides a method for InSAR interferometric phase filtering in a wavelet domain in conjunction with local frequency estimation. The method for InSAR interferometric phase filtering in a wavelet domain enables distinction of the useful information sub-bands from the noise sub-bands in the wavelet coefficients of the complex interferometric phase by using the local frequency estimation. With characteristics of the good noise removal performance for the general threshold shrinkage method and the strong detail maintenance capability for the neighborhood threshold shrinkage method, respectively, the neighborhood threshold shrinkage is performed on the wavelet coefficients of the sub-bands in which the useful information is located, while the general threshold shrinkage is performed on the wavelet coefficients of other sub-bands. In this way, the noise may be removed as much as possible, while the detail information of the interferometric fringes may be kept from being destroyed, thereby achieving a highly accurate interferometric phase filtering and providing foundations for the highly accurate interferometric measurement.