计算机与数字工程2025，Vol.53Issue(1)：257-262,6.DOI:10.3969/j.issn.1672-9722.2025.01.046
基于多线结构光的车载路面凹陷检测方法
Vehicle-mounted Pavement Sag Detection Method Based on Multi-line Structured Light
摘要
Abstract
Aiming at the application scenario of pavement depression detection,a vehicle pavement depression detection method based on multi-line structured light is proposed.The projection equipment is used to project multi-line structured light to the road in front of the vehicle,and the pavement depression at different angles is perceived according to the pixel density after the light bar is deformed.The skeleton-based orientation template extracts the center of the line structured light fringe.The light plane is calibrated according to the points at different positions on the coplanar reference target.Finally,the distance between the system and the depression is studied according to the light bar deformation information for the vehicle to decelerate and avoid,which is a key step to obtain the relationship between the average longitudinal distance difference and its included angle in the multi-line struc-tured light on both sides of the sag and on the flat road.In order to verify the feasibility,four sets of experiments are used to detect the characteristic depression of every 20° angle,and then linear fitting is used to solve the accuracy between every 10°.Different from the traditional single-line stripes,the experimental image has 16 multi-line light strips,and the accuracy of the distance error from the system to the depression is verified to be within 3%according to the pixel density.The experimental results show that the deformation under multi-line structured light has higher accuracy and better real-time effect for judging whether there is a dent and estimating the distance between the dent and the vehicle.关键词
多线结构光/标定/方向模板/中心提取/凹陷检测Key words
multi-line structured light/calibration/direction template/center extraction/depression detection分类
信息技术与安全科学引用本文复制引用
孙晗笑,李锋,李桂玲..基于多线结构光的车载路面凹陷检测方法[J].计算机与数字工程,2025,53(1):257-262,6.基金项目
国家自然科学基金项目(编号:61671221)资助. （编号:61671221）
