Shape detection by mathematical morphology techniques for radar target classification

This paper presents a process of knowledge Discovery from Data (KDD) applied on 2D Inverse Synthetic Aperture Radar (ISAR) images. This process is based on four crucial steps which are data acquisition, data pre-processing, data representation and data classification to make the final decision. We propose a new method for data representation based on combining mathematical morphology top-Hat, Thresholding and Watershed transformation for target shape extraction from ISAR images. Then, the feature vector is achieved by Fourier descriptors of each extracted target shape. And finally the classification results are obtained using Support Vector Machine (SVM) classifier. In the last section, experimental results are provided and discussed.