Automatic target recognition of ISAR images based on Hausdorff distance

A method of ATR of ISAR images is addressed in this paper. In this method, the edge images of ISAR images are applied as features and the THD method is applied as classifier. The edge image has the advantage of less demand on storage space and computation time, which is superior in practical system. The recognition based on Hausdorff distance is invariant to rotation after the scale and the signal level of the images are normalized. The experimental results demonstrate that the method performs very well.