InSAR Phase Optimization Using L-Looks Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current InSAR time-series phase optimization methods are inefficient and lack precision, struggling to meet the demands of timely and accurate deformation monitoring due to limitations in spatial resolution and computational efficiency.
Innovation Solution
The proposed optimization method involves obtaining a time-series SAR data set, performing L-looks processing, and estimating a covariance matrix based on a land cover image and differential interferometric data set, using a maximum likelihood estimator to achieve optimized time-series phase estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SqueeSAR technology uses pixel-by-pixel search to meet statistical model assumptions, then image details are preserved, but processing time increases significantly
Solution Approach 1:
The patent performs preliminary L-looks processing on the SAR data before phase optimization to pre-aggregate pixel information. This preliminary action reduces the data dimension and creates coarser resolution data that can be processed more efficiently while still preserving essential deformation signals, thereby reducing the computational burden of subsequent processing without completely sacrificing detail information
Solution Approach 2:
The patent applies partial action by selectively processing only certain interferograms or data subsets rather than performing exhaustive pixel-by-pixel search on all data. This approach achieves acceptable phase optimization accuracy for the most critical deformation monitoring areas while reducing overall processing time for large-scale time-series analysis
2Productivity
If small baseline subsets (SBAS) technology filters interferograms one by one, then processing efficiency is improved, but spatial resolution degrades
Solution Approach 1:
The patent introduces L-looks processing as an additional dimension of data aggregation that processes data both spatially (through looks) and temporally (through interferogram filtering). This multi-dimensional approach allows efficient batch processing while maintaining spatial resolution by preserving the structure of deformation signals across different look dimensions
Solution Approach 2:
The patent merges L-looks processing with small baseline subset filtering by combining spatial aggregation (looks) with temporal filtering (baseline subsets). This merging allows the method to process data in batches efficiently while the combined spatial-temporal filtering preserves deformation signal characteristics, achieving both efficiency and resolution
3Productivity
If L-looks processing is applied to reduce data volume, then computational efficiency improves, but data detail information is lost
Solution Approach 1:
The patent changes the resolution parameter by applying L-looks processing that aggregates pixels into larger super-pixels. This parameter change reduces data volume and improves computational efficiency while the patent compensates for information loss by using appropriate look numbers and preserving phase information through the processing chain
Data Source
AI summary
Disclosed are an optimization method and apparatus for an Interferometric Synthetic Aperture Radar (InSAR) time-series phase. The optimization method includes: obtaining a time-series SAR data set, and performing registration and L-looks processing on the time-series SAR data set to obtain an L-looks intensity data set and an interferometric data set respectively; taking the L-looks intensity data set as a reference, obtaining a preset digital elevation model (DEM) and a preset land cover image, performing registration and geocoding on the preset DEM to obtain a digital elevation in a SAR image coordinate system, and performing registration and geocoding on the preset land cover image to obtain a land cover image in the SAR image coordinate system; performing a differential operation on the interferometric data set to obtain a differential interferometric data set; and estimating a covariance matrix at each spatial pixel position, and estimating and obtaining an optimized time-series phase.