基于特征点匹配的激光雷达与相机间外参标定方法  

Feature point matching based extrinsic parameter calibration between lidar and camera

作　　者：李新 王肖霞[1] 杨风暴[1] LI Xin;WANG Xiaoxia;YANG Fengbao(School of Information and Communication Engineering,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学信息与通信工程学院,山西太原030051

出　　处：《现代电子技术》2025年第3期19-25,共7页Modern Electronics Technique

摘　　要：激光雷达点云和相机图像融合常被应用在多个领域,准确的外参标定是融合两种信息的前提。现有的基于特征的标定算法,提取的3D⁃2D特征存在不匹配的问题,影响了整体标定的性能。为此提出一种基于特征点匹配的激光雷达与相机外参标定方法。首先利用圆心特征提取算法从点云和图像中分别获取标定板上4个圆心特征点的三维和二维坐标;然后建立三维点云数据和二维图像数据中点对间约束关系;最后通过非线性优化算法得到激光雷达和相机标定的外参,利用外参矩阵将激光雷达点云投影在对应的图像上。实验结果表明,文中所提方法平均重投影误差在2.6像素内,与基于手动匹配的标定方法相比,精度提升了42.2%;与基于两孔圆形标定板的标定方法相比,精度提升了27.8%。The fusion of lidar point cloud and camera images is often applied in multiple fields,and accurate extrinsic parameter calibration is a prerequisite for fusing the two types of information.The 3D⁃2D features extracted with the existing feature⁃based calibration algorithms do not match with each other,which affects the overall calibration performance.In view of the above,this article proposes a feature point matching based calibration method for extrinsic parameters between lidar and camera.The center feature extraction algorithm is utilized to obtain the 3D and 2D coordinates of the four center feature points on the calibration board from the point cloud and image,respectively.The constraint relationships between point pairs in 3D point cloud data and 2D image data are established.The extrinsic parameters for lidar and camera calibration are obtained by nonlinear optimization algorithms,and the extrinsic parameter matrix is utilized to project the lidar point cloud onto the corresponding image.The experimental results show that the average reprojection error of the proposed method is kept within 2.6 Pixels.Its accuracy has improved by 42.2%in comparison with the calibration method based on manual matching,and by 27.8%in comparison with the calibration method based on the circular calibration plate with two holes.

关 键 词：激光雷达 点云 相机图像融合 外参标定 特征点匹配 平均重投影误差 

分 类 号：TN249-34[电子电信—物理电子学]