基于高光谱与无人机LiDAR的路面健康状况监测方法研究

With the fast increase of the highway and freeway kilometers in China in recent years, road pavement maintenance and management has posed a pressing requirement on the road pavement health monitoring and assessment. Unmanned Aerial Vehicle (UAV) based LiDAR and hyperspectral imaging systems will be integrated and used to acquire high-resolution LiDAR point cloud and hyperspectral imagery data of the highway pavements. The pavement material and structural information (e.g. e.g. aging, wearing and tearing) is to be derived from the UAV-based and ground measurements of spectral features, while the road pavement geometric indictors (e.g. planeness, uplift) are calculated from the 3D geospatial data generated from the UAV-based LiDAR point cloud data. Afterwards a pavement health monitoring and diagnostic model is created based on the derivatives from the hyperspectral and LiDAR data, which can offer a new technique for the highway maintenance and management departments. The main research tasks of this project include identification of the spectral response characteristics of the typical highway road pavements such as asphalt and cement concrete, building spectral database for the road pavements with various health statuses, ages in use and environments, developing algorithms for retrieving the indices of pavement health conditions from the hyperspectral data, and algorithm for the extraction of road pavement planeness index from LiDAR data, and generating diagnostic and assessment model for the pavement health conditions. The expected achievement of this project can provide a new technique for Chinese highway pavement health monitoring and assessment, and extend and strengthen the research and application in the field of remote sensing for transportation.

随着我国公路尤其是高速公路的快速增长，路面养护与管理部门对如何快速获取大范围路面健康状况信息提出了迫切需求。利用无人机平台获取高精度公路路面LiDAR与高光谱数据，通过测量路面的光谱信息与几何特性，分别反演道路路面材质结构与路面平整度变化（如老化、磨损、塌陷、壅包等）的特征参量，构建路面健康状况遥感监测与诊断模型，为公路养护决策提供新的技术手段。本课题的研究内容包括：通过地面测量和特征提取研究沥青、水泥等典型路面材料的光谱特征，研发不同材质不同健康状况路面的光谱特征数据库，构建沥青路面健康状况表征参量的遥感反演算法与识别模型，基于LiDAR数据研发路面平整度检测的算法，并构建路面健康状况诊断与评价模型，为我国等级公路路面健康状况的快速监测评估提供新的途径，拓展与深化交通遥感研究与应用的水平。

随着我国公路里程的快速增长，传统依靠地面测量方式难以满足路面养护部门对快速获取大范围路面健康状况信息的迫切需求。本项目利用卫星高分辨率遥感影像数据，以及无人机平台获取的沥青路面高密度LiDAR点云与高分辨率光学影像数据，研究了适合大范围沥青路面老化状况的遥感诊断模型，以及对重点路段的病害目标提取与平整度评估方法。完成的主要研究内容包括：（1）依靠地面反射光谱测量和现场调查与分析，弄清了沥青路面老化过程及其光谱响应特征，构建了我国首个公路路面光谱数据库；（2）构建了基于亚米级卫星遥感影像的路面老化状况多端元光谱分解模型，可快速提取老化初期、中期和后期路面的亚像元丰度信息；（3）研发了一种基于激光强度影像且不依赖地面控制点或人工特征目标的无人机激光点云条带快速拼接算法，实现了多个飞行条带的快速配准；（4）融合高分辨率多光谱影像的光谱特征和激光点云的几何特征，建立了沥青路面激光点云快速分类与提取模型，并通过空间差值快速生成路面数字表面模型。.本项目建立了协同高分卫星遥感和低空无人机遥感的路面健康状况协同监测技术体系，为道路养护部门提供了路面健康状况遥感监测新的技术手段。已发表学术论文11篇，完成研究生学位论文7篇，授权发明专利3项，软著1项，部分成果获中国地理信息协会2018年度科技进步一等奖。本项目的实施建立了首个公路路面光谱数据库，积累了2套新疆沙湾县蘑菇湖附近一条沥青道路高密度无人机激光点云数据集以及高空间分辨率多光谱数据集，部分成果已被中国交通管理部门采用。项目的实施极大地促进了我国交通遥感领域的发展，而且必将拓展我国高分对地观测系统在交通领域的综合应用。