测量路面三维纹理的激光约束双目鲁棒算法  

Laser-Constrained Binocular Robust Algorithm for Measuring 3D Texture of Pavement

作　　者：王元元 李仁杰 刘燕燕 张亨通 刘德政 WANG Yuanyuan;LI Renjie;LIU Yanyan;ZHANG Hengtong;LIU Dezheng(Hubei key Laboratory of Power System Design and Test for Electrical Vehicle,Hubei University of Arts and Science,Xiangyang 441053,Hubei,China;School of Mechanical Engineering,Hubei University of Arts and Science,Xiangyang 441053,Hubei,China;School of Materials Science and Engineering,Chongqing Jiaotong University,Chongqing 400074,China)

机构地区：[1]湖北文理学院纯电动汽车动力系统设计与测试湖北省重点实验室,湖北襄阳441053 [2]湖北文理学院机械工程学院,湖北襄阳441053 [3]重庆交通大学材料科学与工程学院,重庆400074

出　　处：《重庆交通大学学报（自然科学版）》2024年第2期24-29,56,共7页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：国家自然科学基金项目(52178422,51808084)。

摘　　要：为提升激光约束双目重构算法测量路面三维纹理的普适性,增强算法的抗光干扰能力,解决激光约束目标在变化光照下的鲁棒提取问题,基于Faster-RCNN(faster-region based convolutional neural network)构建激光约束识别模型,提取5~1050 lux光照范围内人为构造的激光约束目标,对待测路面的约束目标开展子区域分割,进而实现子区域分割立体匹配下双目三维重构。结果表明:改进算法在1050 lux光照强度下的平均构造深度测量值与铺砂法的拟合系数为0.987,优于原激光约束双目重构算法的0.887;在5~1050 lux光照范围内,改进算法最大波动误差、平均波动误差分别为0.141、0.084 mm,较原激光约束双目重构算法分别降低70.20%和76.83%。In order to improve the universality of the laser-constrained binocular reconstruction algorithm to measure the three-dimensional texture of the road surface,enhance the anti-light interference ability of the algorithm,and solve the problem of robust extraction under the changing illumination of the laser-constrained target,a laser-constraint recognition model was established based on Faster-RCNN(faster-region based convolutional neural network).The artificially constructed laser-constraint targets were extracted within the illumination range of 5~1050 lux,the sub-region segmentation of the constrained target to be tested was carried out,and the binocular 3D reconstruction under the stereo matching of sub-region segmentation was realized.The results show that the fitting coefficient of the mean texture depth measurement values of the improved algorithm and the sand paving method under the illumination intensity of 1050 lux is 0.987,which is better than 0.887 of the original laser-constrained binocular reconstruction algorithm.The maximum fluctuation error and the average fluctuation error of the improved algorithm in the illumination range of 5~1050 lux are 0.141 mm and 0.084 mm respectively,which are 70.20%and 76.83%lower than the original laser-constrained binocular reconstruction algorithm respectively.

关 键 词：道路工程 路面三维纹理 激光约束 双目重构 Faster-RCNN 

分 类 号：U416.217[交通运输工程—道路与铁道工程]