空间移动小目标的检测算法

空间序列图像中移动小目标检测与轨迹提取是空间目标监视、跟踪及编目的关键技术之一.为了及时有效地处理星载可见光相机输出的多维海量数据,提出了一种基于四邻域二值量化方法和改进概率Hough变换的目标检测及轨迹提取算法.相较于美国空间目标在轨检测(MTI,moving target indicator)算法,改进算法虚警噪声点及候选路径数目明显减少,计算复杂度降低,运行效率提高.实验结果表明,算法可有效检测序列图像中快速移动的弱小目标并提取其直线运动轨迹.

Detection Algorithm of Space Small Moving Targets

Detection of space small moving targets in time sequence of space images is a key technique for space - based surveillance, tracking and catalog. To process the mass of multidimensional data from an optical telescope in a timely and computationally efficient fashion, an algorithm is introduced for targets detection and trace acquisition based on neighborhood binary and advanced PHT （probability hough transformation） method. Compared with MTI （moving target indicator） algorithm, the number of positive false and candidate tracks is reduced obviously. Simulation results show that the algorithm is useful for optically detecting moving targets which move in a straight path quickly across many frames of data.