Abstract:
This paper proposes a target tracking algorithm, which is based on multiple color information fusion and projection vector, and can describe the space information of target more than camshift. The method adopts the statistical projection of multiple colors, which includes Y, U, V and H etc, to fuse information, to describe the interesting moving target, and uses level and vertical histogram projection to describe the space information of target. Results show that the multiple color information fusion and projection vector can improve the tracking effect of moving target.
Key words:
target tracking,
multiple color space,
information fusion,
projection vector
摘要: 提出一种基于多颜色空间信息融合和投影向量的目标跟踪算法,用以解决传统camshift不能描述目标的空间信息等缺陷。该算法利用Y、U、V、H等多种颜色分量的统计直方图进行信息融合来描述目标颜色特征,通过水平和垂直投影直方图描述目标的空间位置信息。该算法可以反映跟踪目标的部分空间信息,并具备一定的形变适应性。实验表明,该算法能够较好地改进运动目标的跟踪效果。
关键词:
目标跟踪,
多颜色空间,
信息融合,
投影向量
CLC Number:
HAN Yan-Ban, LI Jin-Bing. Tracking Algorithm Based on Multiple Color Fusion and Projection Features[J]. Computer Engineering, 2011, 37(7): 190-192.
韩延彬, 李金屏. 基于多颜色融合及投影特征的跟踪算法[J]. 计算机工程, 2011, 37(7): 190-192.