Target extraction from clutter images using wavelet packet analysis

We present a wavelet shrinkage method that yields high clutter suppression by using a best-tree wavelet packet analysis. An integrated automatic target recognition processor utilizing a wavelet packet transform and a shape extraction method is introduced to demonstrate the feasibility of using wavelet packet analysis for automatic target extraction from synthetic aperture radar (SAR) images. Two analysis procedures are processed independently and the two outputs from each process are combined to increase the detection performance. Experimental demonstrations of target extraction are also provided. The preliminary experiments show that target extraction using wavelet packet analysis has high detection performance as well as low false detection performance.