基于线特征的鱼眼图像与地面激光雷达点云配准  

Registration of fisheyes panoramic image and terrestrial LiDAR date based on line feature

作　　者：熊友谊 乔纪纲 张文金 XIONG Youyi;QIAO Jigang;ZHANG Wenjin(Guangzhou Ouke Information Technology Corporation,Guangzhou 510663,China;School of Geography and Tourism,Guangdong University of Finance and Economics,Guangzhou 510320,China)

机构地区：[1]广州欧科信息技术股份有限公司,广东广州510663 [2]广东财经大学地理与旅游学院,广东广州510320

出　　处：《测绘通报》2021年第7期74-74,75-80,共7页Bulletin of Surveying and Mapping

基　　金：广东省科技计划(2019B010151001);广州市科技计划(201604010125)。

摘　　要：地面激光扫描数据(LiDAR)与全景图像进行联合分析在虚拟场景建设、文化遗迹保护等方面有较高的应用价值。将全景图像在其视点与LiDAR点云对齐是开展两种数据联合分析的基础,但需要克服数据维度差异转换和特征匹配的困难。本文提出了基于线特征的半自动配准方法纠正图像,将鱼眼图像和LiDAR点云投影为透视成像的柱面全景图像,采用Hough变换提取图像直线特征,并利用修正迭代Hough变换方法,实现在鱼眼全景图像视点约束下与离散激光点云的三维对齐。试验表明,该方法能在较少的人工干预下实现二维到三维数据对齐。Combining terrestrial laser point cloud and fisheye image are high efficiency on panoramic information analysis.Registration of fisheyes panoramic image and terrestrial LiDAR data is a prerequistite for its information complement each other.But it is necessary to overcome the difficulties of data dimensional difference transformation and feature matching.In this paper,a new semi-automatic method is brought up to registered these two data based on Hough algorithm,both fisheyes image and LiDAR points are reprojected to a same cylindrical surface.Line features are extracted from these two images based on Hough algorithm.And then,the modified Iterated Hough transform is employed to registered LiDAR range image and fisheye cylindrical surface image.Experimentation shows that the model has capability to register these two data in 3D screen with less manual intervention,and obtain high register precision within the area near fisheyes image viewing point.

关 键 词：鱼眼图像 LIDAR 线特征 MIHT 配准 

分 类 号：P237[天文地球—摄影测量与遥感]