基于增量PCA的目标跟踪算法

Target Tracking Algorithm Based on the Incremental PCA

在空间遥操作系统中为了获取良好的视觉信息使操作员有种身临其境的感觉,在基于奇异值分解的PCA算法上采用增量PCA跟踪算法,避免了提取大量图片的问题;将该算法与PTZ摄像机相结合,设计了一套基于主动视觉的目标跟踪系统。为机械臂提供灵敏的感知系统,在不接触目标的情况下将相关信息及时传递给地面操作员,便于操作员自然、有效地解决问题。在机械臂动作之前执行规划,在动作过程中不断地反馈信息,适合于对环境变化有适应性要求的作业。

In the space of Tele-Operation systems, the visual telepresence has been widely recognized. In order to obtain good visual information and make operators have immersive feeling, this paper use an incremental PCA （principal component analysis） and the singular value decomposition,and avoid extracting a lot of pictures; with the combination of the algorithm and PTZ （Pan-Tilt-Zoom） camera, we can design a set of target tracking system based on active vision for providing a sensitive perception system of manipulators, in the case of not reaching target, the relevant information is passed to the operator on the ground timely and effectively, so the operator can solve the problem naturally and effectively. The plan execution before the manipulator action, the information is feedback in the process of action constantly, so it is suitable for the task which has the adaptability to environmental change.