摘要
由于实际场景复杂多变,目标在运动过程中往往会出现形变、遮挡等问题,增加了跟踪的难度。为了解决上述问题,提出一种基于特征点匹配的自适应目标跟踪算法。算法初始化时在选定的目标区域内提取特征点,跟踪过程中通过对前后两帧的特征点进行匹配,计算出目标的位置、尺度和旋转变化,进而实现对目标的跟踪。同时通过对特征点的不断更新,可以使算法具有一定的抗遮挡能力。实验表明,该方法在实际应用中效果很好。
Due to the complexity of the actual scene, problems of deformation or occlusion always occur during the motion of objects and make tracking more difficult. For solving these problems, an adaptive object tracking algorithm based on feature-points matching is proposed. In initialization, we extract feature points of objects in the selected object area. Then we match the feature points between the two successive frames, and evaluate the change in position, scale and rotation of objects for tracking. Moreover constantly updating feature points can make the method robust while objects are occluded. The experiments prove that the proposed method is effective in actual scene.
出处
《微型机与应用》
2015年第8期17-19,共3页
Microcomputer & Its Applications
关键词
特征点匹配
自适应目标跟踪
尺度变化
旋转
feature points matching
adaptive object tracking
scale variation
rotation
