摘要
运动物体的实时跟踪是移动机器人视觉的关键技术之一.为了实现对目标快速有效的跟踪,本文提出了一种改进的移动机器人视觉跟踪算法,该算法在mean shift算法的基础上,利用颜色特征作为视觉跟踪依据,并且引入Kalman滤波进行迭代窗口的预测.实验仿真结果表明,本文算法一定程度上消除了光照条件的影响,而且很好的解决了当目标被遮挡时发生目标跟踪偏差或丢失的问题,具有实用价值.
Tracking a target in real - time is one of the key visual techniques on mobile robot. In order to track a moving object fast and effectively, an improved algorithm is proposed in this paper. This new method uses color features as the basis for the target＇ s tracking based on mean shift algorithm, and then combined with Kalman filter for iterative window forecast. Simulation results show that this new algorithm not only eliminates the impact of lighting conditions to some extent, but also is a good solution for tracking bias or missing when the target was blocked.
出处
《佳木斯大学学报（自然科学版）》
CAS
2009年第2期203-205,共3页
Journal of Jiamusi University：Natural Science Edition
