CAJ | 학술논문

针对很多高精度目标跟踪算法不能满足实时性要求的问题，提出使用Harris角点检测方法提取目标的少量特征；使用高精度的光流法对提取的部分特征点进行匹配；使用重心算法计算分散特征点的中心作为目标的脱靶量。该算法在保持精度的同时，提高了速度。设计的系统已经应用到实际工程中，对目标旋转、光照变换、变形等具有较高的鲁棒性，并且能够在20ms内完成目标位置的计算，满足实时目标跟踪的要求。
침대흔다고정도목표근종산법불능만족실시성요구적문제，제출사용Harris각점검측방법제취목표적소량특정；사용고정도적광류법대제취적부분특정점진행필배；사용중심산법계산분산특정점적중심작위목표적탈파량。해산법재보지정도적동시，제고료속도。설계적계통이경응용도실제공정중，대목표선전、광조변환、변형등구유교고적로봉성，병차능구재20ms내완성목표위치적계산，만족실시목표근종적요구。
Many new high precision methods can't meet real-time requirement of target tracking system. In order to overcome this problem, a new method is given which extracts little part of target corner feature by Harris corner extracting algorithm and finds those features in the following video frame. In the end, bary algorithm is used to compute the bary of those matched feature points as displacement of target. The proposed method is already used in actually projects and it has high robustness to target rotating, illumination change and target transform. Moreover, this method can get target location in less than 20ms and meets the real-time requirement.