基于加权动态规划和航迹关联的小目标检测技术

Detection and Tracking of Dim Targets Based on Dynamic Programming and Track Matching

低信噪比下运动小目标的检测一直是成像目标检测中的一个热点问题。提出了一种新的小目标检测算法,采用方向加权的动态规划算法和二值航迹关联,克服了低信噪比下目标机动和传感器的不稳定对小目标检测的影响,实现了对深空背景下运动方式任意、速度最大达1像素/帧的弱小运动目标的有效检测。

The observation of lowSNR moving targets has been a hot spot for a long time. A dynamic programming (DP) and track matching algorithm has been developed for the detection of subpixelsized, lowSNR targets observed by mosaic imaging sensors. The primary advantages of this technique are its sensitivity to weak targets along with its robustness to target maneuvers and sensor instabilities. Enhancement is achieved by performing directional weight operation precede to DP and target data association. Experiments have been conducted and it is shown that lowSNR moving targets with maneuver and maximum speed of 1 pixel/frame can be detected with high probability.