摘要
针对激光雷达(LiDAR)点云与光学图像配准时提取的可用特征少的问题,该文提出了一种利用Zernike多项式的LiDAR点云与光学图像配准的方法。该方法能够在LiDAR点云数据和光学图像的基础上,同时提取Zernike多项式特征,用Zernike多项式特征构造配准描述子,实现LiDAR与光学图像的精确配准。实验结果表明,该文提出的配准方法能在多维度上对特征点进行表征,更突出目标特征,拟合精度更高,在几何角点多的场景中提高了LiDAR点云与光学图像的配准精度。
Aiming at the problem of low availability of features extracted from LiDAR point clouds when aligned with optical images,a method for the registration of LiDAR point clouds with optical images using Zernike polynomials was proposed in this paper.The Zernike polynomial features were extracted on the basis of LiDAR point cloud data and optical images simultaneously.Registration descriptors were constructed with Zernike polynomial features to achieve accurate registration of LiDAR and optical images.Experimental results showed that the registration method proposed in this paper could characterize feature points in multiple dimensions,highlight target features more prominently,optimize fitting accuracy,and improve alignment of LiDAR point clouds to optical images in scenes with many geometric corner points.
作者
金利强
黄桦
刘微微
JIN Liqiang;HUANG Hua;LIU Weiwei(Zhejiang Academy of Surveying&Mapping,Hangzhou 311100,China)
出处
《测绘科学》
CSCD
北大核心
2022年第10期124-131,共8页
Science of Surveying and Mapping
基金
浙江省自然资源厅2021年度科技项目(2021-55)
