Generative network-based optical-radar ISAR image conversion method and device

The invention relates to an optical-radar ISAR image conversion method and device based on a generative network. The method comprises the steps of introducing distribution characteristics of ISAR scattering points into a loss function of network training by taking a deep learning network as a main technical means based on target physical characteristic analysis and modeling of an ISAR image, and finally converting an optical image of a target into a corresponding ISAR image. Wherein the constraint condition that scattering point position matrixes extracted from an ISAR original image and an ISAR reconstructed image are equal is added into a loss function, so that the deep learning network can generate a more real ISAR image. 本申请涉及一种基于生成网络的光学-雷达ISAR图像转换方法及装置。所述方法包括：通过基于对ISAR图像的目标物理特性分析与建模，以深度学习网络为主要技术手段将ISAR散射点分布特点引入网络训练的损失函数中，最终实现将目标的光学图像转换至对应的ISAR图像。其中将ISAR原图像以及ISAR重构图像分别提取的散射点位置矩阵相等这一约束条件加入损失函数中，使得深度学习网络能够生成更为真实的ISAR图像。