自主移动机器人走廊视觉识别与跟踪方法研究

在基于图像的道路识别技术实现的自主驾驶问题研究中,图像识别和跟踪的速度是自主驾驶效果的关键。采用基于模型的走廊识别技术,通过模仿人类视觉过程,使用遗传算法,对实时图像进行目标识别和跟踪,实现了一套自主移动机器人系统;同时通过对识别模型匹配值的变化率分析,实现了对走廊拐角的识别。实验结果表明,该方法的走廊图像识别和跟踪速度能够满足系统实时性的要求,并具有较高的可靠性。

Research on Vision Recognition and Tracking for Corridor of Autonomous Driving Robot

The speed of image recognition and tracking is a key factor of the effect of autonomous driving in the research on problem of autonomous driving which is realized by image-based road recognition technology.The paper uses model-based corridor recognition technique and adopts genetic algorithm,which is used like the optical process of human being to recognize and track the target in the real-time image,and realizes an autonomous driving robot system.In addition,the system also realizes the recognition of the corridor corner by analyzing the variation ratio of the model matching value.The experimental results indicate that the high speed of corridor recognition and tracking,in this method can meet the real-time request of the system,and has high degree of reliability.