低空背景下红外目标提取跟踪算法研究

为解决低空复杂背景下红外目标的提取和跟踪,提出了地平线检测排除地物背景的方法,然后在地平线之上的区域,采用边缘检测、阈值分割、边界搜索算法并结合多帧图像的相关性以及目标形态和运动参数的连续性来提取目标。实验表明,本文的算法计算量小、实时性强,可以有效检测出地平线,排除低空背景和假目标的干扰,实现稳定的跟踪,并最终完成了DSP的硬件实现。

Research on IR targets extraction and tracking under low altitude background

In order to solve the extraction and tracking problems of IR targets under complicated low altitude background,a horizon detection algorithm is proposed to exclude background of ground objects.Then above the horizon region,uses the edge detection,the threshold segmentation,the boundary search algorithm combined with the multi-frame image correlation as well as the continuity of target shape and motion to extract the target.Experiments show that the algorithm of this article is small amount of calculation,real-time,horizon can be effectively detected to remove low altitude background and the false target disturbance,may realize stable tracking and eventually completed a DSP hardware implementation.