摘要
针对传统的帧差分算法在运动目标检测过程中目标存在较大空洞和丢失边沿轮廓的不足,提出了一种新的运动目标检测算法。该算法对视频序列进行预处理,进行传统的帧差分运算,引入新定义的邻域综合因子来更新像素的前景概率,通过数学形态学分割处理与边缘检测结果进行逻辑"或"运算,提取出运动目标。实验结果表明,该算法能较大程度上克服了以上不足,而且识别率更高,实时性更好。
Aiming at the deficiencies of existing bigger voids and losing edge in the moving object by the conventional frame difference algorithm in the processing of the moving object detection, a new moving object detection algorithm is proposed. Firstly, this algorithm preprocesses the video sequence, moreover, does conventional frame difference operation, and then, updates pixel foreground probability by the new defined contiguous area factors; Finally, making use of math morphological segmentation and edge detection to computer with or logic, and extract the moving object. The experimental results show that the new algorithm is able to overcome the shortage of discussed above, the recognition rate performs higher and the real-time is better.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第12期4331-4335,共5页
Computer Engineering and Design
基金
教育部重点科研基金项目(208098)
湖南省科技计划基金项目(2012FJ30052)
关键词
帧差分
邻域综合因子
自适应阈值
数学形态学
边缘检测
运动目标检测
frame difference
contiguous area integrated factors
adapted threshold
math morphological
edge detection
moving object detection
