光照变动条件下基于图切割算法的运动目标跟踪

Moving target tracking based on image cutting under illumination-variant condition

摘要:

为解决光照大范围变动条件下跟踪运动目标丢失问题,本文基于光照估计的算法建立光照模型,对跟踪视频中的光照进行估计,然后结合图切割算法计算出运动目标的光流向量,利用光流向量进行目标跟踪。实验表明,本文算法对光照的改变具有一定的适应性,可以准确地跟踪目标,提高了跟踪算法的鲁棒性。

Abstract:

Tracking moving targets under variant illumination in a large range is still a hard problem in computer vision.In order to solve it,the illumination model is established based on the outdoor light source estimation,which is used to estimate the illumiation in tracking vision.Then the optical flow vector of the moving target is computed combining the image cutting algorithm,so the moving target is tracked.Experimental results indicate that the proposed scheme is adaptive to illumination changes,can track the moving target accurately,and improves the algorithm′s robustness.