基于压缩感知关联成像算法的激光跟踪机理研究

Laser tracking system by active illumination is important for the application in the field of early warning of space junk and optical communication. The complexity of the system is increased by using the high resolution detector to obtain high resolution target information. The tracking precision and the reliability of the tracking system are decreased by the high complexity of the system. In this project, the theoretical model of laser imaging tracking of uncooperative object in the space based on the Algorithm of Compressive sensing computational ghost imaging is set up. The mechanism of imaging and laser tracking is researched. The Algorithm of Compressive sensing computational ghost imaging is applied to the laser imaging tracking system to increase the resolution limit and the tracking precision. A single pixel bucket detector is used to detect the echo and obtain the super-resolution target information. The recovery Algorithm of Compressed Sensing of Block-Sparse Signals based on subspace is used to increase the imaging resolution and the imaging speed. The sparse reconstruction algorithm based on total variational constraint is used to suppress speckle noise. A new algorithm of target detection in complex background is proposed combining fractal area measurement with fractal fitting error, the algorithm can detect extended target in complex background correctly and reserve the shape details of the target. The method of particle filter fused motion feature and fractal feature is proposed to reliably tracking. The research on mechanism of position recognition and imaging tracking of uncooperative object in the space based on the Algorithm of Compressive sensing computational ghost imaging can improve the tracking precision. The research on this project is not just for laser tracking, but also can be extended and applied to other imaging detection system. It has important practical value and theoretical significance.

空间激光主动照明跟踪系统在空间垃圾预警、轨道逼近、空间光通信等领域具有重要的应用；目前通过提高探测器的分辨率获取高分辨率的目标信息增加了系统的复杂度，降低了系统的可靠性，进而降低了系统跟踪精度。本项目建立基于压缩感知关联成像算法的激光主动照明跟踪系统的理论模型，进行成像跟踪机理的研究。将计算关联算法应用到成像跟踪系统中，通过单像素桶探测器获取超分辨率的目标信息，提高成像分辨率的同时降低系统复杂度；采用基于子空间的块稀疏信号的压缩感知关联成像算法，提高成像分辨率和成像速度；采用基于全变分约束的稀疏重构算法抑制散斑噪声；将面积度量和拟合误差两种分形特征相结合，实现目标的完整检测，增强目标检测的准确性和可靠性；在粒子滤波框架下融合目标的运动特征和分形特征，实现目标的稳定跟踪，提高系统的跟踪精度。项目的研究不仅为激光跟踪所用，还可进一步推广和应用到其他成像探测系统中，具有重要的实用价值和理论意义。

空间激光主动照明跟踪系统在空间垃圾预警、轨道逼近、空间光通信等领域具有重要的应用；目前通过提高探测器的分辨率获取高分辨率的目标信息增加了系统的复杂度，降低了系统的可靠性，进而降低了系统跟踪精度。.本项目建立了基于压缩感知关联成像算法的激光主动照明跟踪系统的理论模型，进行成像跟踪机理的研究。将计算关联算法应用到成像跟踪系统中，通过单像素桶探测器获取超分辨率的目标信息，提高成像分辨率的同时降低系统复杂度；采用基于子空间的块稀疏信号的压缩感知关联成像算法，提高成像分辨率和成像速度；采用基于全变分约束的稀疏重构算法抑制散斑噪声；将面积度量和拟合误差两种分形特征相结合，实现目标的完整检测，增强目标检测的准确性和可靠性；在粒子滤波框架下融合目标的运动特征和分形特征，实现目标的稳定跟踪，提高系统的跟踪精度。.在原理研究的基础上提出了一种基于玫瑰扫描的关联成像方案，把玫瑰扫描中的瞬时视场作为关联成像中的调制光场，以小范围的采样代替整体采样，减少了能量损失，在减少采样次数和采样时间的同时提高了重构图像质量。.提出了一种基于背景差分法的压缩鬼成像跟踪方案，利用关联成像采集图像信息，运用背景差分法获得目标图像的测量值，将通过压缩感知重构出目标图像进行投影定位获取形心坐标，利用多项式拟合的方法构建运动轨迹进行物体跟踪。在低采样的情况下仍能够准确跟踪到物体，大大减少了重构的测量次数，提高了跟踪效率。.项目的研究不仅为激光跟踪所用，还可进一步推广和应用到其他成像探测系统中，具有重要的实用价值和理论意义。