一种基于SIFT特征的自主目标跟踪方法
A Method Of Autonomous Target Tracking Based On The SIFT Feature
摘要
本文介绍了一种基于SIFT特征的自主目标跟踪方法。首先通过连续三帧图像差分检测运动目标并进行标记，提取图像中运动目标范围内的SIFT特征点，将这些特征点保留作为"模板"。在下一时刻，根据历史数据预测运动目标的位置，在预测位置周围一定范围内提取图像中SIFT特征点，并与"模板"进行匹配，计算运动目标的偏移量。将上一时刻对运动目标的标记位置按照该偏移量进行移动，从而实现目标跟踪。实验证明了该方法的有效性。Abstract
In this paper, a method of autonomous target tracking based on SIFT feature is introduced. At the beginning, it detects mobile target through the difference of three continuous frames and mark it. And then extract the SIFT points that belong to the mobile target and keep them as the template. At next time, predict next position of the mobile target according to history data and pick up SIFT points in the peripheral region of the position. Match the SIFT points with the template to calculate the offset, which indicates the movement of the mobile target from last frame. Mark the mobile target in the current frame, achieving the target tracking. The experiment proves that this method is effective. 关键词
模式识别/目标跟踪/SIFT/自主Key words
Pattern Recognition/Target Tracking/SIFT/Automation引用本文复制引用
赵昆,李宁,王雪刚,董良,王福.一种基于SIFT特征的自主目标跟踪方法[EB/OL].(2010-11-23)[2025-12-14].http://www.paper.edu.cn/releasepaper/content/201011-542.学科分类
自动化技术、自动化技术设备/计算技术、计算机技术
评论