Combining top-hat, thresholding and watershed transformation for 3D Inverse Synthetic Aperture Radar Images segmentation

Abstract

Segmentation is a critical step, preliminary to High-level processing such as classification, detection and object extraction from radar images. It provides an easier way to extract information in automatic target recognition (ATR) fields. However, extracting the shape of the target from this kind of images is very difficult task, so this paper presents a new approach for Radar image segmentation and how to be able to extract target shape from 3D Inverse Synthetic Aperture Radar (ISAR) Images by combining mathematical morphology watershed and thresholding methods with top-hat transformation.