Speckle Reduction for SAR Superpixel Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Synthetic aperture radar (SAR) images often contain speckle, which complicates the application of superpixel segmentation algorithms and reduces their effectiveness in accurately segmenting and categorizing terrain and structures.
Innovation Solution
A speckle reduction process, such as sub-aperture multilook, mean coherent change detection, or median radar cross section, is applied to SAR images before superpixel segmentation, resulting in a speckle-reduced image product that is then segmented into superpixels, which are registered back to the original image for accurate labeling and classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If speckle reduction processes (such as sub-aperture multilook, mean coherent change detection, or median radar cross section) are applied to SAR images before superpixel segmentation, then the segmentation accuracy is improved, but the spatial resolution is reduced
Solution Approach 1:
The patent applies speckle reduction processes to the SAR image before performing superpixel segmentation. This preliminary action removes the harmful speckle noise that would otherwise interfere with the segmentation algorithm, allowing for more accurate segmentation while accepting the trade-off of reduced spatial resolution in the processed image
Solution Approach 2:
The patent introduces an intermediary speckle reduction process between the original SAR image and the superpixel segmentation algorithm. This intermediary step creates a modified image version that has reduced speckle noise, which then serves as the input for segmentation, effectively mediating between the noisy original image and the segmentation requirements
Data Source
AI summary
Various embodiments presented herein relate to assigning labels to segments of a synthetic aperture radar (SAR) image, where the segments are based upon a speckle-reduced SAR image product. A plurality of SAR images of a scene are co-registered to form a registered stack of SAR images. A speckle-reduced SAR image product is generated based upon at least one registered SAR image in the registered stack of SAR images. The speckle-reduced SAR image product is segmented into a plurality of superpixels, and boundaries of the superpixels are applied to the at least one registered SAR image to form a segmented SAR image. A segment of the SAR image is then labeled as including or not including a feature, wherein the label is assigned based upon values of pixels in the segment.