摘要
给出了一种新的视频运动目标检测算法.该算法采用差异积累的方法自适应更新背景模型,用背景差法进行运动检测,用Otsu法计算二值化阈值,给出了Otsu法类间方差简化公式的详细推导.在背景差图像阈值化的基础上,对视频帧进行网格区域划分,并定义网格矩阵,设定网格内前景点个数的阈值,对视频帧像素进行重新定义,再对网格矩阵进行连通区域合并和前景区域定位.采用不同的视频测试序列,从检测效果及耗时上研究了基于网格的视频运动目标检测算法的性能,并与区域生长法进行了比较.实验结果表明,该算法具有良好的检测效果和实时性能.
A novel detecting algorithm for moving objects in video sequences was proposed. Background was updated adaptively using difference accumulation information and movements were detected by background subtraction. Otsu method was used to threshold the background subtraction image and the covariance equation was inferentially reasoned to a much simpler style. Based on the binary result, the image was partitioned using grid. Grid matrix was defined, setting the threshold value for the number of foreground pixels and redefining the frame pixels. And adjacent regions were combined together and foreground regions were located. Different video sequences were used to test the performance of the proposed algorithm. And it was also compared with the region growing method. Experiment results showed that the proposed algorithm was effective and efficient.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2009年第3期442-447,476,共7页
Journal of Zhejiang University：Engineering Science
关键词
差异积累
OTSU
网格
检测
区域生长
differences accumulation
Otsu
grid
detection
region growing
