ISAR Target Reconstruction by Inference Rule Based Systems

The Inverse Synthetic Aperture Radar (ISAR) is a useful technique to identify moving targets by observing high resolution images. ISAR generates two dimensional maps of the target by processing the return echoes from different angles of the target. Such a high resolution images may help the radar operator to easily identify and to classify the targets. This paper describes a non linear processing for ships ISAR image reconstruction, based on two Inference Fuzzy rule filters and one morphological filter. The reconstruction is basically performed in two main steps. At first the target (ship) is extracted from the noisy background of the image, and then the orientation and size of the target are estimated. The second step is done by using two binary templates, one oriented along the above axes and the other vertical, which supply the inputs to the fuzzy filter. The final output consists in a binary image representing a reconstructed ship silhouette of the corresponding ISAR images with low distortion.