MAXIMUM-ENTROPY REGULARIZATION IN INVERSE SYNTHETIC APERTURE RADAR IMAGERY

The method of maximum entropy is applied to the regularization of inve rse synthetic aperture radar (ISAR) image reconstructions. This is acc omplished by considering an ensemble of images with associated "allowe d" probability density functions. Instead of directly considering the "solution" to be an image, we take it to be the a posteriori probabili ty density found by minimizing a regularization functional composed of the usual "least squares" term and a Kullback (cross-entropy) informa tion difference term. The desired image is then found as the expectati on of this density. The basic model of this approach is similar to tha t used in usual maximum a posteriori analysis and allows for a more ge neral relationship between the image and its "configuration entropy" t han is usually employed. In addition, it eliminates the need for inapp ropriate nonnegativity constraints on the (generally complex-valued) i mage.