摘要
在城市交通路口,车辆较多,并且车辆的停留时间比较长,对于这种情况,使用常用的帧差法和背景减除法都不能很好的分割出车辆目标。提出一种基于直方图的自适应阈值目标分割方法,无需更新背景,不依赖多帧图像的差值,只需对视频中的每帧图像进行路面区域分割即可得到汽车目标。实验证明此算法实时性好,能够较好的分割交通路口监控视频中的静止和运动车辆。
There are more vehicles in the traffic intersection of the city, and the vehicles will stay there for a longer period of time, in this case, the common methods of frame difference and background subtraction can not be a good algorithm for vehicle segmentation. A adaptive thresholding segmentation algorithm based on histogram. This new method no need to update the background, do not rely on the difference between multi-frame images, only need single frame image to divide vehicle area. Experiments show that this algorithm has better real-time performance and and vehicle segmentation performance.
出处
《微计算机信息》
2010年第21期214-215,共2页
Control & Automation
基金
基金申请人:赵勋杰
项目名称:视频鲁棒传输模型与算法研究
基金颁发部门:江苏省高校自然科学基金(BK2009116)
关键词
视频监控
图像分割
自适应阈值
直方图
Surveillance video
image segmentation
adaptive threshold
histogram
