面向运动目标检测的ViBe算法改进

背景差分法是静态背景下运动目标检测的常用方法,ViBe算法是它的主要建模方法之一.针对ViBe算法对鬼影消除缓慢的问题,提出了结合帧间差分技术的ViBe改进算法,使用帧间差分技术通过记录相关像素值的时域变化来判断鬼影像素,提高消除鬼影的速度.针对ViBe算法的固定阈值不能反映每个像素具体情况的问题,提出了一种自适应阈值的方法,可根据像素值的变化为每个像素设定阈值,提高前景检测的准确度.实验结果表明,结合帧间差分技术的ViBe算法能够较快地消除检测结果中的鬼影,应用自适应阈值的ViBe算法能够更准确地进行前景检测.

An ImproVed ViBe Algorithm for MoVing Object Detection

Background differencing is the commonly used method for the detection of moving objects in the static background， and the ViBe algorithm is the main modeling approach. In order to solve the problem about low rate of ghost elimination caused by the execution of ViBe algorithm， an improved ViBe algorithm combining with frame difference method is proposed. By using the frame difference method， the ghost pixel is judged according to the changes in time domain for related pixel value， which can improve the rate of ghost elimination. Since the specific condition of each pixel cannot be reflected with the fixed threshold， a method with self-adaptive threshold is proposed. The threshold of each pixel is set according to the change of the pixel value， which can improve the accuracy of foreground detection. The experimental results show that the ViBe algorithm combining with frame difference technology can be used to eliminate the ghost in the detection results more quickly， and the foreground can be detected more accurately using the ViBe algorithm with self-adaptive threshold.