Automated Sea Clutter Training Area Selection in Polarimetric SAR Imagery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting ship targets in synthetic aperture radar (SAR) imagery face challenges due to the complexity of sea clutter modeling, particularly in high-resolution data, where non-Gaussian distributions are more reliable but require varying window sizes for efficient detection, and often rely on amplitude or intensity components rather than the complete scattering vector or polarimetric covariance matrix.
Innovation Solution
An automated method for selecting training areas of sea clutter using multivariate complex Gaussian and non-Gaussian distributions, specifically formulating homogeneous and texture models based on squared radius and trace statistics for single- and multi-look polarimetric SAR data, allowing for reliable detection of small targets with low radar cross-sections without additional speckle filtering, and extending to near real-time hardware implementations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If non-Gaussian distributions (K-distribution, log-normal, Weibull) are used to model sea clutter in high-resolution SAR data, then the accuracy of sea clutter modeling is improved, but the complexity of the detection system increases due to requiring varying window sizes
Solution Approach 1:
The patent transforms the complex non-Gaussian sea clutter distribution into an equivalent Gaussian distribution by changing the statistical parameters. The method uses the equivalence between non-Gaussian and Gaussian distributions to simplify the detection algorithm while maintaining modeling accuracy, thereby resolving the contradiction between modeling precision and system complexity
Solution Approach 2:
The patent extracts and separates the sea clutter component from the composite signal (sea clutter + ship target) by using statistical properties. By isolating the sea clutter characteristics and modeling them independently, the method simplifies the overall detection process while maintaining high accuracy in representing sea clutter behavior
2Ease of operation
If amplitude or intensity components are used for ship detection, then the detection process is simplified, but the detection accuracy deteriorates due to not utilizing the complete scattering vector or polarimetric covariance matrix
Solution Approach 1:
The patent extracts the essential statistical characteristics (mean and covariance matrix) from the complete polarimetric scattering vector. By separating and utilizing only the necessary second-order statistical moments, the method achieves simplified processing while retaining the full information content needed for accurate ship detection in polarimetric SAR data
Solution Approach 2:
The patent transforms the complex polarimetric scattering vector into a simplified statistical representation using only the first and second moments (mean and covariance). This parameter transformation maintains detection accuracy by preserving the essential statistical properties while reducing computational complexity
3Reliability
If local processing approaches with guard rings and background rings are used, then the detection reliability is improved by excluding extended ship targets, but the productivity decreases due to requiring variable window sizes for different ship target sizes
Solution Approach 1:
The patent creates a universal detection algorithm that works effectively for ship targets of all sizes using a fixed processing window. By designing the algorithm to be size-independent and utilizing the statistical properties of sea clutter, the method eliminates the need to adjust window sizes for different target dimensions, thereby maintaining both reliability and productivity
Data Source
AI summary
Method for selecting a sea clutter training area in polarimetric synthetic aperture radar input data. A sea clutter reference distribution for a pixel magnitude value is provided. Based on the input data, one or more parameters of the reference distribution and a global covariance matrix are computed. The pixels are grouped into blocks. A block that minimizes a cost function is pre-selected, the cost function being derived from empirical moments of the block and moments of the reference distribution. A goodness-of-fit is computed for the pre-selected block with respect to the reference distribution. If the goodness-of-fit is sufficient, the block is selected as sea clutter training area. Otherwise, the steps of preselecting and computing a goodness of fit are repeated.