基于协方差矩阵的极化干涉SAR图像海面慢动目标检测新方法研究

The detection of slow moving target in complex marine environments is a difficult problem in the field of radar detection. It is an effective method to solve this problem by establishing the accurate statistical model of polarimetric SAR interferometry in image domain, estimating the model parameters accurately and applying the optimal moving target detection algorithm. .This project fully considers the influence of complex marine environments, the physical characteristics of slow motion and the polarization of ship targets, and develops a new area of slow target detection via along track interferometry based on Polarimetric SAR in image domain. .First the slow moving target scattering models are established based on the different representation of polarimetric SAR interferometry by scattering matrix and covariance matrix, and then the novel slow moving target detection algorithm is put forward using general likelihood ratio test(GLRT) via the above full information covariance matrix models..Second the performance of the methods is improved by developing image registration and polarimetric whitening filtering of polarization SAR interferometry to suppress sea clutter influence, which can also reduce the dimension of the covariance matrix; and then the optimal detection variable will be found by the covariance matrix decomposition and key parameter recombination when physical differences between moving target and sea clutter are fully discriminated and analyzed..Last the physical and electromagnetic differences of ship wake and ocean current velocity, especially the velocity distribution, is examined carefully. The new method of ship detection is put up based on the tail characteristics of the flow, and then the slow moving target detection theory system is fully established for polarimetric along track interferometric SAR.

复杂海洋环境下舰船等慢动目标的检测是雷达探测领域的难点问题。建立精确的极化干涉SAR图像统计模型、进行准确的参数估计并采取最优的运动目标检测算法是解决此难题的有效手段。.本项目充分考虑复杂海洋环境的影响，以海面舰船等目标的慢动特征和极化特征为物理基础，开创极化沿迹干涉SAR图像基于协方差矩阵进行海面慢动目标检测的研究先河。建立基于散射矩阵和协方差矩阵表征下的极化干涉SAR慢动目标散射理论模型，提出基于协方差矩阵和干涉矩阵的慢动目标检测方法；通过极化SAR图像配准新方法以及极化白化滤波等海杂波抑制手段，提高慢动目标检测性能，同时降低协方差矩阵维数；充分挖掘慢动目标和海杂波的物理特性差异，寻求基于矩阵分解和参量组合的海上慢动目标最优检测量；通过舰船尾流和海洋洋流速度和物理分布等特性差异，提出基于极化沿迹干涉SAR尾流特性的舰船检测新方法，建立较为完整极化干涉SAR图像海面慢动目标检测理论体系。

本项目充分考虑复杂海洋环境对目标检测的影响，以海面舰船等目标的慢速运动特性为物理基础，开辟了极化干涉SAR基于图像域进行海面慢动目标检测的新方向。拓展极化SAR-GMTI的数学模型，建立极化干涉SAR动目标散射基本理论模型，提出基于多通道沿迹干涉SAR的全极化高维协方差矩阵和干涉矩阵的慢动目标检测新方法；通过极化SAR图像配准新方法以及极化白化滤波等海杂波抑制手段，提高慢动目标检测性能，同时降低极化协方差矩阵维数；充分挖掘慢动目标和海杂波的物理特性差异，构建基于协方差矩阵分解和参量组合变换的海上慢动目标最优检测量，建立了极化沿迹干涉SAR图像海面慢动目标检测理论体系。在现有体制无法满足干涉要求的情形下，提出了基于极化子孔径分解的虚拟孔径干涉检测方法，取得了较好的效果。研究成果对慢小目标检测、隐身目标检测提供了有力工具，有效提升了远海预警探测能力。