Abstract:

Inverse synthetic aperture radar (ISAR) images, consisting of many sparse scatters, usually vary with different view angles, which will greatly impact on the feature extraction of the ISAR image and decrease the performance of target recognition via the image. An available preprocessing approach to ISAR images is presented to enhance the stability and recognition performance of ISAR images. It contains suppressing the speckle noise and interferential stripes, connecting the scatters with morphological processing, smoothing images with neighbor average and normalizing the mass center, scale and view angle. Simulation results show that the stability of ISAR images is enhanced greatly after the preprocessing, which is propitious to the latter feature extraction and recognition.