SAR Cross-Range Streak Detector for Accurate Pose Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cross-range streaking in synthetic aperture radar (SAR) images caused by rotating radar arrays leads to inaccurate target position and pose estimates, resulting in poor automatic target recognition (ATR) scores and increased software run times due to longer search processes and larger database requirements.
Innovation Solution
A SAR chip level cross-range streak detector identifies and filters cross-range streaks by defining regions, thresholding, forming profiles, applying low-pass filters, and excluding affected pixels from bounding box calculations, allowing for improved pose and position estimation without modifying the SAR chip data passed to the ATR algorithm.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If cross-range streaks are detected and processed by searching all possible pose values, then target identification accuracy may improve, but software run time increases and false target identification rate increases
Solution Approach 1:
The patent applies preliminary action by detecting and removing cross-range streaks from the SAR image before the target identification process begins. The streak detection algorithm processes the image in advance to identify and eliminate problematic pixels, creating a cleaned image that reduces the search space for subsequent pose estimation. This preliminary processing prevents streaks from causing false identifications during the main recognition process, thereby reducing both time and false alarm rates.
2Measurement precision
If cross-range streaks are modeled into ATR models/templates, then target identification accuracy may improve, but software run time increases and database size increases
Solution Approach 1:
The patent extracts and removes cross-range streaks from the SAR image data before it enters the ATR processing pipeline. By separating and eliminating the harmful streak components from the target image data, the system avoids the need to incorporate complex streak models into the ATR database. This extraction approach simplifies the database requirements while maintaining identification accuracy, as the cleaned image data can be processed using standard ATR algorithms without needing extended or modified templates.
3Productivity
If cross-range streaks are not addressed, then software performance remains efficient, but target position and pose estimation accuracy decreases
Solution Approach 1:
The patent converts the harmful effect of cross-range streaks into a beneficial process by implementing a detection and removal algorithm that specifically targets these streaks. The algorithm identifies streak characteristics (such as linear patterns extending across the image) and systematically removes them, transforming the problematic streak data into cleaned, usable image data. This conversion allows the system to maintain efficient processing while achieving accurate target estimation, as the streaks no longer interfere with pose and position calculations.
Data Source
AI summary
A method of reducing cross-range streaking in a radar image includes determining a number of on-pixels in each line of at least a portion of the radar image, determining which lines have a determined number of on-pixels that exceeds a threshold number, and removing the on-pixels of lines having the determined number of on-pixels exceeding the threshold number.