Panoramic Texture Mapping Algorithm Based on Multi-image Pose Estimation

Abstract

Texture mapping technology, which is an effective means to obtain true color point clouds with rich texture information, has been widely used in many fields with its unique advantages. It has a very broad application prospect. We study a method for solving multiple image poses based on joint calibration of 3D laser scanners and external digital cameras, and obtain a method for obtaining panoramic true color point clouds. The basic idea is to use the inherent relative position and posture of the camera and the laser scanner to calibrate the first image obtained by rotating the camera. After obtaining its position and posture, the geometrical characteristics of the camera's space rotation are used to determine the position and orientation of the rest of the images in terms of the pose of first image. Then the texture mapping of multiple images can be accomplished to obtain a panoramic color point cloud. Compared with the current mainstream panoramic image texture mapping algorithm, this algorithm has certain improvement in accuracy and efficiency. By performing texture mapping experiments on a variety of point cloud data, the results show that the method can quickly and accurately obtain true three-dimensional panoramic color point clouds, providing a data foundation for three-dimensional refined modeling.