复合高斯杂波中知识辅助的自适应检测算法

研究了在复合高斯杂波中利用先验知识自适应检测目标的问题，证明了基于双参数逆高斯分布纹理的复合高斯模型比传统的K分布、复t分布和单参数逆高斯分布纹理的复合高斯分布模型能够更好地拟合实际杂波。文中选择双参数逆高斯分布作为纹理分量的先验分布、基于广义似然比准则和贝叶斯方法设计得到了一种复合高斯杂波中的自适应检测器。理论分析和数值仿真表明，与自适应匹配滤波器和正则化自适应匹配滤波器相比，该检测器具有更好的检测性能

Knowledege-aided Adaptive Detection Algorithm in Compound Gaussian Clutter

The adaptive target detection problem is studiea in compound Gaussian clutter with the aid of prior knowledge. The compound Gaussian (CG) distribution with the two-parameter inverse Gaussian (2PIG) texture, called the two-parameter inverse Gaussian compound Gaussian (2PIG-CG) distribution, is validated to fit the real clutter data better than the traditional K distribution, complex t distribution and the compound Gaussian distribution with one parameter inverse Gaussian texture. In this paper, based on the Bayesian method and the generalized likelihood ratio criterion, using the 2PIG distribution as the prior distribution of the texture component, an adaptive detector in compound Gaussian clutter is proposed. Theoretical analysis and numerical simulations show that the proposed detector has better performance than the adaptive matched filter (AMF) and the normalized adaptive matched filter (NAMF).