一种基于傅里叶变换的改进ViBe算法

针对Vi Be算法在第1帧图像中含有运动目标时容易引入Ghost区域以及不能很好地去除阴影等问题.我们基于块的背景建模结合傅里叶变换,对Vi Be算法进行了改进.该算法对每一帧图像进行分块,在块内使用快速傅里叶变换,利用第1帧图像中每一块的直流分量建立背景模型,以后各帧与背景模型比较,检测出运动物体,最后通过扫描图像计算每一列前景像素的个数与设置的阈值进行比较,来消除阴影.实验结果表明,改进的算法可以快速的去掉Ghost区域以及很好地去除阴影.

An improved Vi Be algorithm based on Fourier transform

As the Vi Be algorithm will introduce an artifact commonly called a ghost when there are moving objects in the first frame and can't remove the shadow,an improved Vi Be algorithm based on block and Fourier transform is proposed. Every frame image is divided into several block and each block carries out fast Fourier transform,then the value of dc component of every block in the first frame is used to build background model. Then,starting from the second frame,the dc component of current block is compared with the corresponding sample set to detect the moving objects. Lastly,the shadow is removed by comparing the foreground numbers in every column of image with the threshold value. The experimental results show that this algorithm can quickly get rid of the ghost area,and it can be good remove the shadow region.