一种简易的激光雷达点云与光学影像自动配准方法

将激光雷达点云与光学影像进行配准融合,在数字城市建模、自动驾驶和导航避障等方面有重要的应用价值。当前多数自动配准方法需要专门设计特殊的场景结构,操作复杂。根据激光雷达对玻璃具有穿透的特性,提出利用透明的门窗等室内建筑物场景来实现对激光雷达点云与光学影像的自动配准。首先,将激光点云量化为图像,提取光学影像与量化图像的门窗角点特征,并利用相关系数法实现特征点的自动匹配;然后,对激光雷达与相机之间的配准转换参数进行求解;最后,对两种数据进行配准融合。实验结果表明:利用室内门窗这种规则的半透明场景,可以轻松实现激光点云与光学影像的自动化配准,简单实用,操作性强。

A simple automatic registration method for Lidar point cloud and optical image

The registration and fusion of lidar point clouds and optical images have an essential application value in digital city modelling,autonomous driving and navigation obstacle avoidance. Most current automatic registration methods require a unique design of a particular scene structure and complex operations. The penetration characteristics of lidar to glass proposes transparent doors and windows and other indoor building scenes to realize automatic registration of lidar point clouds and optical images. First,quantify the laser point cloud into an image,extract the door and window corner features of the optical image and the quantized image,and use the correlation coefficient method to achieve automatic matching of feature points;then,solve the registration conversion parameters between the lidar and the camera;Finally,the two data are registered and fused. The experimental results show that regular translucent scenes,such as indoor doors and windows,can quickly realize the automatic registration of laser point clouds and optical images,which is practical and straightforward and has strong operability.