Three-dimensional reconstruction method of single object, based on color depth camera

Abstract

The invention discloses a three-dimensional reconstruction method of single object, based on a color depth camera. The three-dimensional reconstruction method of single object, based on a color depthcamera includes the three steps: 1) during the scanning process, extracting the scanned object area; 2) according to the color depth data, performing camera tracking and local fusion of the depth data, and performing global non rigid body registration on the locally fused data, and gradually constructing an overall three-dimensional model and the accurate key frame camera position; and 3) performing grid extraction on the fusion model, according to the obtained key frame camera position and the key frame color image, calculating the texture map of the three-dimensional grid model. The method framework of the three-dimensional reconstruction method of single object, based on the color depth camera can guarantee that even if the proportion of the object in the image is relatively smaller, geometrical reconstruction and texture mapping with high quality can still be performed when single object is reconstructed. The three-dimensional reconstruction method of single object, based on a color depth camera has the advantages of being clear, being high in speed, being robust in the result, and being able to be applied to the virtual reality scene construction field and the like.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a method in the field of three-dimensional reconstruction and texture mapping of a single object using a color depth camera. Background technique [0002] 3D reconstruction technology is a method of reconstructing 3D objects based on 2D input. Traditional 3D reconstruction methods include StructureFromMotion (SfM), which mainly reconstructs 3D objects from a large number of color images. In recent years, with the development and commercialization of depth cameras such as Kinect, 3D reconstruction techniques based on color depth images have been extensively studied. The 3D reconstruction technology based on color depth images can be used to reconstruct dense and fine object models for the construction of virtual reality environments, combining current virtual reality and augmented reality hardware and software to provide a 3D interactive experience. [0003] 3D reconstruction algo...

AI Technical Summary

Problems solved by technology

However, due to the small proportion of a single object under ordinary camera settings, the total amount of its features is small, and the corresponding proportion of the total energy in the camera tracking and global optimization process is also small, which makes the existing reconstruction algorithm easy to converge to errors in the optimization process. The local optimal solution of

Images