Research on 3D Point Cloud Model Reconstruction Method Based on Multi-Kinects

摘要

3D reconstruction technology has always been a challenging research topic in the field of computer vision, a new 3D reconstruction system based on the new generation Kinect V2 is proposed in this paper, use two Kinects to collect object images from different views on the same horizontal line, through the calibration, denoising, registration and image mosaic, the 3D point cloud model of the object is obtained, the experimental results show that the method has better practicability and feasibility than kinectv1, which can be widely used in 3D reconstruction.