车载激光点云与全景影像快速融合检校方法

Rapid Fusion Calibration Method of Vehicle Laser Point Cloud and Panoramic Image

针对车载激光点云与全景影像融合检校过程复杂的问题,通过分析检校的误差来源,量化误差影响因素,提出了一种快速检校方法。首先,构建更适用于全景球坐标计算的角度一致方程模型,求解全景相机检校参数;其次,通过分析检校模型,确定检校精度与参数初值和选点精度有关;然后,进一步分析选点精度与点云和全景像素密度差异、全景像片畸变的关系,得出密度差异用距离约束、像片畸变用立面点及对称选点约束的结论;最后,对同名特征点选取数量、距离及布设方式进行分析,确定最佳布点和选点方法并进行验证。实验结果表明,该检校方法简易可行,可实现点云与全景影像的快速检校,相比于传统方法具有明显优势。

Aiming at the problem of complex calibration process of vehicle laser point cloud and panoramic image fusion,a fast calibration method is proposed by analyzing the error source and quantifying the error influencing factors.Firstly,an angle-consistent equation model more suitable for the calculation of panoramic spherical coordinates is constructed to solve the calibration parameters of panoramic cameras.Secondly,by analyzing the calibration model,it is determined that the calibration accuracy is related to the initial value of parameters and the precision of point selection.And then,the relationship between point selection accuracy and pixel density difference between point cloud and panoramic image distortion is analyzed.It is concluded that density difference is constrained by distance,and image distortion is constrained by elevation point and symmetrical selection point.Finally,the number,distance and layout of the same feature points are analyzed to determine the optimal layout and selection method and verified.The experimental results show that the proposed method is simple and feasible.This method can realize the rapid calibration of point cloud and panorama,which has obvious advantages compared with the traditional methods.