基于目标CSAR回波模型的SAR自动目标识别算法

基于模板的SAR目标识别需要存储海量的目标模板，给识别系统的设计和算法效率的提高都造成了严重的困难，而基于模型的方法克服了上述问题，并已成为下一代目标识别算法研究的热点。该文提出的基于圆周SAR(CSAR)回波模型的识别算法，从目标的3维CAD模型出发，利用弹射线原理构建目标的CSAR回波，并通过在线实时预测目标聚束SAR图像来完成识别。同传统的基于散射中心模型的算法相比，利用CSAR回波的算法不仅预测结果准确，而且算法简单高效。仿真实验验证了算法的有效性，并比较了相关算法的优缺点。

SAR ATR Algorithm Based on CSAR Raw Echo Modeling

Template-based ATR algorithm needs the storage of massive templates of the targets, which cause a heavy burden of system design and algorithm optimization. On the contrary, model-based algorithm can overcome this drawback and become a promising solution of the next generation ATR system. A new algorithm based on CSAR raw echo of the targets is proposed by applying the SBR (Shooting and Bouncing Rays) method onto the 3D CAD model of the targets, then the recognition procedure is completed via online prediction of the hypothetic targets. The predicted SAR image of the target is more accurate and efficient comparing with the traditional scattering center approach. The simulate experiments validate the proposed method and compare it with the traditional methods.