摘要
针对地面激光雷达点云和数码光学影像非同源异质数据自动配准困难的问题,本文提出了基于互信息的两种数据同名特征高精度自动提取的方法。首先,把点云数据生成中心平面投影的反射强度图像和基于RGB信息的彩色图像,应用点云彩色图像和数码光学影像的匹配,确定点云与影像的粗配准参数;然后,对反射强度图像进行特征提取,应用粗配准参数确定其在数码光学影像上的初始位置,应用互信息实现非同源数据的高精度匹配;最后,应用罗德里格矩阵和选权迭代方法计算高精度配准参数,生成三维彩色模型。试验证明,本文方法可以解决地面激光点云和数码光学影像非同源异质数据的配准问题,具有一定的研究和应用价值。
In order to solve the problem of automatic registration of non-homologous and heterogeneous data of ground LiDAR and digital optical image, this paper proposes a method of high-precision automatic extraction of homonymous features based on mutual information. Firstly, the reflection intensity image and the color image based on RGB information of the central plane projection are generated from the point cloud data, and the coarse registration parameters of the point cloud and the image are determined by the matching of the color image of the point cloud and the digital optical image. Then, the feature of the reflection intensity image is extracted, and its initial position on the digital optical image is determined by coarse registration parameters, and the high-precision matching of non-homologous data is realized by mutual information. Finally, Rodrigues matrix and weight selection iterative method are used to calculate the high-precision registration parameters and generate the three-dimensional color model. The experimental results show that this method can solve the registration problem of non-homologous and heterogeneous data for the terrestrial LiDAR point cloud and digital optical image, which has certain research and application value.
作者
胡春梅
夏国芳
刘喜
HU Chunmei;XIA Guofang;LIU Xi(Beijing University of Civil Engineering and Architecture,Beijing 100044,China;China Cultural Relics Information Consulting Center,Beijing 100029,China;Heilongjiang Surveying and Mapping Measuring Instrument Verification Station,Haribin 150081,China)
出处
《测绘通报》
CSCD
北大核心
2021年第10期83-87,共5页
Bulletin of Surveying and Mapping
基金
国家自然科学基金(41401536)
北京建筑大学市属高校基本科研业务费专项资金(X18230)。
关键词
地面激光点云
光学影像
匹配
互信息
配准
terrestrial LiDAR point cloud
optical image
matching
mutual information
registration
