基于全维状态观测器的视频球体跟踪

Video Spherical Target Tracking Using Full State Observer

视频目标跟踪是计算机领域的一核心内容,该文构造一种全维状态观测器以实现在静态无噪声基础上对运动单目标的跟踪。在MATLAB环境下,运用图像处理技术,将从图像中提取目标球体;对目标球体进行分析找出若干跟踪参数;基于找出的跟踪参数,建立状态空间方程;最后构建全维状态观测器,实现对球体目标的跟踪。实验表明,全维状态观测器能很好地实现运动目标的跟踪,其跟踪精度主要取决于极点配置的好坏。

Video target tracking is a core element of the computer field. In this paper, try to construct a full state observer to realize the single moving target tracking. Advantage of this approach lies with avoidance of hypothesis for the stochastic noise uncertainties existing in process＇ dynamics. In the MATLAB environment, using image processing technology, the target is extracted from the image; then analyzing to identify ball tracking parameters; then the state space equation is found; Finally, based on full state observer theory, the ball can be tracked. The result shows that full state observer can be good for moving target tracking, the tracking accuracy depends on the pole placement. Therefore Full State Observer is an alternative video tracking method.