High-quality texture real-time reconstruction method and device

Abstract

The invention discloses a high-quality texture real-time reconstruction method and device, and the method comprises the following steps: moving a single RGBD camera to shoot a plurality of angles of ato-be-measured object, and obtaining a color image sequence; screening out N key frames meeting a preset condition from the color image sequence; optimizing the N key frames through a deep neural network, and outputting color images after the N key frames are optimized; and back-projecting the optimized color images of the N key frames to the geometry of the object to be measured, and carrying out weighted average processing to obtain the high-quality texture of the shot part. According to the method, only a single RGBD camera is needed in the shooting process, and the high-quality texture ofthe static object can be reconstructed in real time.

Description

technical field [0001] The invention relates to the technical fields of computer vision and computer graphics, in particular to a texture reconstruction method and device for a three-dimensional model. Background technique [0002] The reconstruction of 3D objects has a wide range of applications in many fields such as games, animation, and movies. To obtain a 3D model of an object, it is necessary to obtain its corresponding geometry and texture. Some existing methods can use the depth image to obtain the geometry of the object. On this basis, if the captured color image is directly back-projected onto the object to obtain Textures are often affected by various errors, and the effect is poor. It will be affected by errors such as inaccurate camera parameters, image distortion, and geometric inaccuracy, resulting in texture reconstruction errors and blurring. There are methods to reconstruct high-quality textures by solving optimization problems, but these methods are often...

AI Technical Summary

Problems solved by technology

To obtain the 3D model of an object, it is necessary to obtain its corresponding geometry and texture. Some existing methods can use the depth image to obtain the geometry of the object. On this basis, if the captured color image is directly back-projected onto the object to obtain Textures are often affected by various errors, and the effect is poor. They will be affected by errors such as inaccurate camera parameters, image distortion, and geometric inaccuracies, resulting in texture reconstruction errors and blurring.

There are methods to reconstruct high-quality textures by solving optimization problems, but these methods are often time-consuming

Images