摘要
针对像素域运动对象的提取,较系统地分析比较了现有主要算法各自的原理、提取效果及时间特性,为研究者在各种不同的场合选择不同的提取算法提供了依据。为使算法具有可比性,实验在同一Hall_Monitor序列上进行。然后,提出一种改进的、同时进行帧间差分和减背景的运动对象提取算法,其鲁棒性在于能对光照变化、运动物体的暂停,显露部分的背景区域等复杂情况作出正确判断处理。
In view of extraction process, the principles, the extracted effects and extracted time attributions of the main algorithms about the moving object extraction （ MOE ） from video pixel domain are analyzed and compared. The contribulions lie in the selection rules of the appropriation algorithm for own uses according to different conditions. Simulation results are demonstrated, which are conducted on Hall-Monitor sequence. Then, a improved MOE algorithm is presented, which is based on frame difference and background subtraction simultaneously. Its robustness can deal with some complicated situations such as light change, the paused moving object in background and the uncovered background.
出处
《湖南工业大学学报》
2008年第6期50-54,共5页
Journal of Hunan University of Technology
基金
国家自然科学基金资助项目(60572127)
湖南省自然科学基金资助项目(05JJ30113)
关键词
运动对象
提取
视频处理
像素域
moving object
extraction
video processing
pixel domain
