A CFAR intensity pattern detector for MMW SAR images

In this paper, we first demonstrate a novel CFAR intensity pattern matching detector, which is formulated based on the generalized likelihood ratio test (GLRT) for target detection in millimeter wave (MMW) SAR images. Both the widely used two-parameter CFAR and the /spl gamma/-CFAR are the special cases of the intensity matching detector with different intensity kernels to be matched. Secondly, to find the best intensity matching kernel, we propose to apply principal component analysis (PCA) to the radial intensity profiles of several targets. It can be experimentally shown for the MSTAR data set that the first eigenfunction of the radial intensity profile of targets can be well approximated by the first order Gamma kernel, which provides the optimal matching intensity kernel. Therefore, the /spl gamma/-CFAR detector approximately performs "normalized" maximum eigenfiltering, resulting in better performance when compared with the delta function stencil proposed by MIT/Lincoln Laboratories.