非结构化道路条件下单目视觉实时测距研究
Research on Monocular Vision Real-time Ranging Under Unstructured Road Conditions
沈平 1杨国平1
作者信息
- 1. 上海工程技术大学机械与汽车工程学院 上海 201620
- 折叠
摘要
为了自动驾驶汽车能够在非道路条件下进行有效的防碰撞,提高智能汽车安全性.论文通过对传感系统中的单目摄像头对前方物体的测距方法进行研究,来提高车辆在非结构化道路上的安全性能.对单目摄像头的标定采用了基于两个灭点的标定方法的数学模型,使用了针孔透视模型来成像以及单帧静态图像测距模型来计算前方物体与本车距离,加强图像的稳定性使用了RANSAC算法,Prescan和Simulink进行仿真试验,最后进行实车在非道路条件下的实时测距试验.试验结果表明:汽车在非结构化道路化的条件下行驶摄像头能进行实时测距.
Abstract
In order to improve the safety of intelligent vehicles,autonomous driving vehicles can effectively prevent collision in off-road conditions.In this paper,the distance measurement method of the front object by the monocular camera in the sensing system is studied to improve the safety performance of vehicles on unstructured roads.For the calibration of the monocular camera,the mathematical model based on the calibration method of two vanishing points is adopted,the pinhole perspective model is used to image,and the single frame static image ranging model is used to calculate the distance between the front object and the car,and the RANSAC algorithm,Prescan and Simulink are used to enhance the stability of the image for simulation tests.Finally,the real-time distance measurement test of real-vehicle in non-road condition is carried out.The test results show that the camera can carry out real-time ranging under the condition of non-road driving.
关键词
智能驾驶/单目摄像/非结构化道路/实时测距Key words
intelligent driving/monocular camera/unstructured road/real-time ranging引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据