A point cloud registration method based on camera calibration and an ICP algorithm

Abstract

The embodiment of the invention discloses a point cloud registration method based on camera calibration and an ICP algorithm. The method comprises the following steps: respectively establishing a camera coordinate system and a world coordinate system on the depth camera and the calibration object, wherein the calibration object is a calibration plate with the same checkerboard on the front and back sides; using a plurality of depth cameras to shoot the calibration object simultaneously. the position or the angle of the calibration object is changed; shooting multiple sets of images, determining a conversion relationship between camera coordinate systems; a plurality of depth cameras are used to shoot a target object simultaneously. the method comprises the steps of obtaining point clouds under different viewpoints, executing pre-processing operation on the point clouds, accurately registering the point clouds by using an ICP algorithm after roughly registering the point clouds according to a calibration result, fusing the registered point clouds, filtering and smoothing the point clouds by using a moving least square method, adding RGB information, and outputting a three-dimensional color point cloud model. Through the scheme of the invention, the point cloud registration efficiency can be improved.

Description

technical field [0001] The invention belongs to the technical field of machine vision, and in particular relates to RGB and Depth calibration and registration methods of multiple depth cameras. Background technique [0002] In recent years, 3D depth perception equipment has begun to enter people's eyes. As a new medium for obtaining external information, high-precision 3D depth perception equipment is conducive to promoting the development of machine vision and helping robots understand the external world, making them more Intelligence also promotes the development of human-computer interaction. The application of 3D depth perception equipment is inseparable from 3D reconstruction, that is, to establish an appropriate mathematical model for 3D objects in the computer. [0003] Point cloud registration is the core link in 3D reconstruction, and the registration between point clouds of multiple depth cameras can be simplified as a pairwise registration problem. Pairwise regi...

AI Technical Summary

Problems solved by technology

However, the traditional ICP algorithm has requirements on the initial registration position of the two point clouds to be registered, and its robustness is poor, and it is easy to fall into a local optimal solution.

Images