摘要
根据图像序列中小目标自身特点,以及小目标在场景中运动轨迹特点,提出一种弱小目标检测的方法。采用多帧图像积累的方法得到小目标在较短时间内的运动轨迹,由改进后的最小二乘直线拟合法对运动轨迹进行拟合,再运用拟合直线判断目标点,最终检测出小目标。理论分析和实验结果都表明,该方法能有效检测出复杂背景下运动的小目标。
According to its own characteristics of small target image sequence,as well as the small target's trajectory characteristics in the scene ,this paper presents a small target detection method. It used the method for accumulating multi-frame images to get small target's trajectory characteristics within a short time,fiting the trajectory with improved least-squares linear fiting method ,used the line to determine the target-points ,and eventually found out small targets. Theoretical analysis and experimental results show that the method can effectively find out the moving-small targets under a complex background.
出处
《电子设计工程》
2010年第7期176-177,180,共3页
Electronic Design Engineering
基金
太原理工大学技术开发项目(203-24041221)
关键词
小目标检测
最小二乘直线拟合
图像积累
运动轨迹
small target
least-squares fitting a straight line
image accumulation
trajectories