Current Issue Cover
基于全向视觉和前向视觉的足球机器人目标识别

卢惠民1, 王祥科1, 刘斐1, 季秀才1, 郑志强1(国防科技大学机电工程与自动化学院,长沙 410073)

摘 要
为了快速正确地进行机器人的目标识别,首先介绍了RoboCup中型组足球机器人NuBot使用的全向视觉系统及其目标识别方法;然后通过引入一套简单的前向视觉系统来弥补全向视觉的不足,以提高机器人对其前方目标球的感知识别精度,同时给出了一种简单有效的前向视觉系统标定方法;最后介绍两套视觉系统共同工作的实现机制,以提高机器人的目标识别处理速度。实验结果表明,两套视觉系统能够很好地共同完成机器人的目标识别。
关键词
Omni-Vision and Front Vision Based Object Recognition for Soccer Robots

()

Abstract
The omni-vision system for RoboCup middle-size league soccer robot-NuBot and its method for object recognition are introduced firstly in the paper,and then a simple front vision system is used to improve the accuracy of recognition of the ball in front,and meantime an easy and effective calibration method for front vision system is presented,and finally the realization mechanism for the two vision systems working together is introduced,in which the processing speed of robot's object recognition can be accelerated.The experimental result shows that the object recognition can be done successlly with the function of both vision systems.
Keywords

订阅号|日报