基于深度相机的碎石撒布质量检测方法  

A Method for Inspecting Quality of Gravel Spreading by Depth Camera

作　　者：金肃静 罗佳 顾海荣[2] 刘胜 韩帅 JIN Su-jing;LUO Jia;GU Hai-rong;LIU Sheng;HAN Shuai(Zhejiang Scientific Research Institute of Transport,Hangzhou Zhejiang 311305,China;National Engineering Laboratory for Highway Maintenance Equipment,Chang’an University,Xi’an Shaanxi 710064,China;Keylaboratory of Road and Bridge Inspection and Maintenance Technology Research of Zhejiang Province,Hangzhou Zhejiang 311112,China;Lishui Traffic Engineering Quality Supervision Station,Lishui Zhejiang 323050,China)

机构地区：[1]浙江省交通运输科学研究院,浙江杭州311305 [2]长安大学公路养护装备国家工程实验室,陕西西安710064 [3]浙江省道桥检测与养护技术研究重点实验室,浙江杭州311112 [4]丽水市交通工程质量监督站,浙江丽水323050

出　　处：《公路交通科技》2022年第5期50-56,共7页Journal of Highway and Transportation Research and Development

基　　金：浙江省交通运输厅科技项目(2021011);河南省重大科技专项(191110211500);陕西省青年科技新星项目(2020KJXX-044)。

摘　　要：在道路工程中,为了准确快速检测碎石封层施工中碎石的撒布质量,采用基于TOF原理的深度相机获取碎石封层表面的彩色图像和深度信息,基于最大类间方差法分析封层表面彩色图像,将图像分为前景背景两类,不断迭代计算类间方差,识别分离表面图像中的碎石颗粒与沥青,并基于形态学方法平滑处理碎石轮廓边缘,剔除小径颗粒,优化碎石轮廓,基于碎石轮廓的像素比计算碎石覆盖率。利用获取的深度信息直观地观察整个封层表面凹凸不平的起伏程度,对像素点深度信息进行卷积二重积分,构建碎石封层表面的空间曲面,计算碎石封层的路面构造深度。在实验室模拟直径9.5~19 mm的碎石颗粒撒布后的状态,搭建图像采集装置,分别在200 mm和300 mm高度获取碎石封层图像,对碎石封层图像进行识别,计算覆盖率,分析在不同高度下该方法的检测误差,并进行图像法深度信息检测与铺沙法检测结果的对比试验,验证路面构造深度检测方法的正确性和可行性。结果表明:利用深度相机采集碎石图像,基于最大类间方差法可准确地描绘识别出碎石轮廓,在不同高度下碎石覆盖率检测误差不大于2%;在同一路段,路面构造深度检测与铺沙法检测结果误差不超过3%,且可以保留封层表面的纹理与轮廓细节信息,能够快速检测分析碎石撒布质量。In road engineering,to accurately and quickly inspect the gravel spreading quality in the gravel seal,the color image and depth information of gravel seal surface are obtained by using the TOF principle based depth camera.The color image of the seal surface is analyzed based on the maximum inter-class variance method.The images are divided into foreground images and background images,the inter-class variance is iteratively calculated continuously,the gravel particles and asphalt mortar in the surface images are identified and separated.The edge of the gravel contour is smoothed based on morphological method,the small-diameter particles are removed,the gravel contour is optimized,and finally the gravel coverage is calculated according to pixel ratio of the gravel outline.The unevenness of the entire seal is visually observed with the obtained depth information,the double convolution integral is performed on the depth information of each pixel to construct the space surface of the gravel seal and calculate the pavement structure depth of the gravel seal.The spreaded state of gravel particles with diameter of 9.5-19 mm is simulated in laboratory,and an image acquisition device is set up.The images of gravel seal are obtained at 200 mm and 300 mm heights respectively,which are identified to calculate the coverage rate and analyze the detection errors of this method at different heights.The comparison test of the inspection results obtained by the image depth information method and sand spreading method is carried out,which verified the correctness and feasibility of the pavement structure depth inspection method.The result shows that(1)the gravel contour could be accurately described and identified based on the maximum inter-class variance method with the gravel images collected by depth camera,and the inspection error of the gravel coverage at different heights is not more than 2%;(2)the difference of inspection results between the 2 method is not more than 3%in the same road section,the texture and contour details

关 键 词：道路工程 碎石撒布质量检测 深度相机 碎石封层 覆盖率 构造深度 

分 类 号：U416.1[交通运输工程—道路与铁道工程]