基于机器视觉和姿态检测的循迹智能车  

A Tracking Smart Car Based on Machine Vision and Posture Detection

作　　者：王天河[1] 侯建军[1] 

机构地区：[1]北京交通大学电子信息工程学院,北京100044

出　　处：《兵工自动化》2012年第8期82-85,共4页Ordnance Industry Automation

摘　　要：针对未来汽车在一定条件下根据需求自主驾驶并安全到达目的地的趋势,阐述一种以数字摄像头作为道路传感器的智能循迹电动车制作方案。介绍智能车整车硬件结构与设计,采用图像二值化加速处理电路快速处理赛道信息,惯性导航元器件检测和计算车身姿态信息,给出智能车软件系统结构图,分析控制策略及优化。结果表明:通过优化控制策略和道路特征识别,可让电动模型车在赛道板内安全行驶不出界的情况下,以较快速度顺利通过各种弯道和适应不超过15°的上下坡道。Aiming at the trend for the future of the automobile autonomous driving and safe arriving at the destination, describe production scheme of an intelligent tracking electric vehicle with a digital camera. It has accelerated technology of image bi-naryzation to process track information; inertial navigation chip can detect and calculate car posture. The paper gives the smart car software system structure; analysis the control strategies and optimization. The results show that through optimal control strategy and the identification of road characteristics, it can run safely, rapidly and smoothly in the no-boundary track board. On the other hand, it also can safely drive through a ramp up and down within 15 degrees.

关 键 词：机器视觉 姿态检测 自动驾驶 循迹 MEMS 

分 类 号：TP273.3[自动化与计算机技术—检测技术与自动化装置]