基于OpenCV的运动目标检测与跟踪_尹俊超--20.pdfVIP

计算机工程与设计 Computer Engineering and Design 尹俊超，刘直芳：基于OpenCV 的运动目标检测与跟踪 2011, Vol.32, No.8 2817 基于OpenCV 的运动目标检测与跟踪 1 1,2 尹俊超 ， 刘直芳 (1. 四川大学计算机学院，四川成都610064 ； 2. 四川大学视觉合成图形图像技术国防重点学科实验室，四川成都610064) 摘 要：基于OpenCV 开源平台开发了一种多运动目标检测和跟踪的自动化智能方案。针对多个运动目标在复杂场景中的 情况，无法根据背景差和帧间差的方法来确定目标是否首次出现，根据新目标期望出现的位置范围和轮廓面积大小采用筛 选法来判断目标的性质。利用CAMShift 跟踪算法提取每个目标的颜色特征信息，实现对多个运动目标的跟踪，最后将运动 目标的轮廓和运动轨迹描述出来。实验结果表明，在 目标颜色特征明显的情况下文中方法具有很好的鲁棒性和精确性。 关键词：OpenCV; 检测; 跟踪; 帧间差; 背景差; 背景重构; CAMShift 中图法分类号：TP391.4 1 文献标识码：A 文章编号：1000-7024 (2011) 08-2817-04 Detection and tracking of moving obj ects based on OpenCV 1 1,2 YIN Jun-chao , LIU Zhi-fang (1. College of Computing, Science Sichuan University, Chengdu 610064, China; 2. State Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu 610064, China) Abstract ：Based on OpenCV open source platform, an automated and intelligent program for more moving targets’detection and tracking is developed. For multiple moving targets in complex scenes of the situation, can ’t be according to background subtraction and the frame difference method to determine whether the target the first time appears, by preparation, according to the area where a new target expected to appear to determine whether the target is new. CAMShift tracking algorithm can be adapted to extract the color feature information for each target to achieve the tracking of multiple moving targets. Finally, the outline and