摘要
光照变化对传统的背景减除算法有很大的影响,文中提出一种新的基于矢量线性相关性分析的背景减除算法,可以适应光照的变化。先对图像序列建模,对比分析当前图像与背景,把当前图像分为无变化、光照变化和内容有变化的三类区域,然后对内容变化区域进行各个像素矢量线性相关性分析,判断像素属于前景还是背景,得到最终的结果。实验结果显示,这种方法对光照变化具有很好的适应性,改进了一般算法在光照变化时的性能。
Varying illumination influences the performance of traditional background subtraction algorithms. A new background subtraction algorithm based on analyzing vector linear dependence is proposed in this paper. It can tolerate varying illumination. Firstly, the paper constructs a background model. Then, it analyzes and compares current image with background model, classifies all the pixels into three classes which are invariant, varying illumination and varying content. Next, it analyzes the vector linear dependence of every pixel which belongs to varying content class. Finally, it gets the target which is linear independent. Experiment results indicate that this algorithm has better tolerance to varying illumination and improves the performance of traditional background subtraction algorithms.
出处
《计算机仿真》
CSCD
2007年第9期214-217,共4页
Computer Simulation
关键词
光照变化
相关性
景减除
Varying illumination
Vector linear dependence
Background subtraction
