摘要
为了改善运动目标在跟踪中的实时性问题,提出了一种基于SURF算子的目标跟踪算法。通过对SURF特征点集进行描述快速确定目标位置,实验表明该方法是一种简洁有效的目标跟踪识别方法。同时针对目标运动过程中短时间的遮挡问题,提出了一种目标遮挡检测机制,为目标遮挡的处理提供了一种途径。
In order to improve the real-time object track, a new object tracking algorithm based on SURF operator was pro- posed. Through the description of SURF feature point sets, the target position was quickly determined. The experimental results show that the method is a kind of concise and effective target tracking recognition. At the same time, aiming at the short time occlusion issues, a kind of object shelter detection system is put forward, which provides a kind of way to deal with occlusion.
出处
《江南大学学报(自然科学版)》
CAS
2012年第5期515-518,共4页
Joural of Jiangnan University (Natural Science Edition)
基金
教育部自主科研计划项目(JUSRP211A17)
关键词
目标跟踪
遮挡
目标识别
特征匹配
object tracking, occlusion, object recognition, feature matching