基于FMCW微波成像的非合作目标精细识别与形状提取方法

The imaging quality degeneration caused by high-speed movement of FMCW SAR imaging platform is a bottleneck problem that restricts precise extraction of target information. This project proposes a high-resolution imaging and precise information extraction method for non-collaborative targets based on spatial-temporal compressive receiving mode. First, the sparse representation model is built for both scenes and non-collaborative targets and then high resolution spatial-temporal sparse imaging algorithm is designed on the basis of joint spatial-temporal compressive transmitting and receiving model to balance between the high resolution and miniaturization requirements during imaging process. Secondly, aiming at the target imaging problem caused by complex moving trajectories, a transmitting wave optimization based automatic error correction algorithm is proposed to obtain the exact shape information of targets. Also, the impact of high-speed movement on imaging quality can be effectively addressed by improved FOCUSS sparse reconstruction algorithm. Furthermore, based on the spatial-temporal sparse sampling, target shape detection and recognition algorithm will be designed via saliency filtering to automatically extract target information in a rapid and precise manner. Based on this, a prototype FMCW SAR imaging verification system will be designed from the spatial-temporal compressive transmitting and receiving imaging model and the effectiveness of the proposed method will be checked on real data, so as to lay the foundations for FMCW micro imaging based fine target processing.

弹载FMCW平台的高速机动导致成像质量下降是制约精细化目标信息提取的瓶颈问题。针对该问题，本项目提出了基于空时二维压缩接收模型的面向非合作目标高分辨成像与精细化信息提取方法。首先，建立场景和非合作目标的稀疏表征模型，研究基于空时二维联合压缩成像的高分辨空时稀疏成像算法。其次，针对复杂运动轨迹带来的目标成像难题，提出基于发射波形优化的自动误差矫正算法，以获得准确的目标形状信息，改进FOCUSS稀疏重构成像算法，有效解决高速运动对成像质量的影响。再次，在空时稀疏采样基础上设计基于显著滤波的目标形状检测与识别算法，快速精确提取目标信息。最后，拟设计基于空时二维压缩收发成像模型的FMCW雷达成像原型验证系统，兼顾弹载成像中高分辨与小型化的需求。该项目为弹载高速运动平台下FMCW SAR的高分辨对地观测成像和精确目标信息获取提供有力理论和实践支撑。

弹载FMCW平台的高速机动导致成像质量下降是制约精细化目标信息提取的瓶颈问题。针对该问题，本项目提出了基于时空二维压缩接收模型的面向非合作目标高分辨成像与精细化提取方法。传统的ISAR成像为了获得方位像上的高分辨，需要对目标进行长时间观测，对高速运动的非合作目标，会导致严重的散焦，成像效果严重恶化。本项目提出了基于稀疏理论的ISAR成像和高分辨MIMO成像技术，获取准确的目标形状。首先，大多数ISAR成像数据是稀疏的；其次，相对于成像背景，ISAR成像数据也具有低秩特性，利用稀疏低秩等先验信息，可解决目标的在复杂运动场景下的高分辨成像问题。MIMO雷达成像相对于传统的ISAR成像，可以在很大程度上减小合成阵列时间，可通过少量的快拍矩阵进行成像。特别地，当信噪比足够高的情况下，可通过单快拍方式进行成像。另外，利用雷达图像和高光谱图像的相似性，研究了高光谱图像的精细化目标提取方法。本项目为弹载高速运动平台下的高分辨对地观测成像和精确目标信息获取提供了有力理论和实践支撑。