一种基于ViBe和改进LBP的目标跟踪算法

An Optimized Target Tracking Algorithm based on ViBe and Modified LBP

ViBe算法在跟踪视频场景中运动状态突然变化的物体时,效果较差,例如目标由静止状态到运动或者由运动状态到静止都会产生拖影区域(Ghost区域),而且Ghost区域在以后的跟踪中不会立即消除,这个问题大大降低了对物体跟踪的准确性.为了提高跟踪的正确率本文将ViBe算法和改进的局部二值算法(LBP)相结合使用,即针对ViBe算法提取出的前景图像,再使用改进的局部二值算法进行前景的二次提取,以达到快速消除Ghost区域的目的.仿真实验说明了该方法的有效性.

The use of ViBe algorithm will always generate a ghost area in the scene of target tracking video.The area will not be eliminated immediately,which seriously degrades the accuracy of the tracking.To solve this problem,this paper combines the Vibe algorithm and modified LBP algorithm to eliminate the ghost area rapidly.To be specific,extract the foreground image by using LBP algorithm right after the ViBe algorithm extracting.