ISAR Target Recognition Using Pix2pix Network Derived from cGAN

Inverse Synthetic Aperture Radar (ISAR) image processing has received much interest in recent years, due to its effectiveness in remote sensing and military use. Although ISAR can achieve all-time all-weather target detection, the quality of images is unstable due to many factors such as sea clutter, which will interfere with target recognition. Since, for sea objects, strong correlation exists between ISAR data and optical camera data, target information extraction accuracy and reliability can be improved by jointly processing the two types of data. In this paper, the pix2pix network derived from the conditional generative adversarial network (cGAN) is used to realize the translation of the ISAR images to the corresponding optical images. In order to remove the influence of lighting conditions on the color of the optical images, we use the grayscale images instead. We combine the generated and the ISAR images to train the CNN network for recognition. Experimental results demonstrate that the proposed method can effectively improve the recognition rate of target recognition based on the ISAR images.