Dynamic scene real-time three-dimensional reconstruction method based on binary feature code matching

Abstract

The invention provides a dynamic scene real-time three-dimensional reconstruction method based on binary feature coding matching. The method comprises the steps: carrying out the shooting of a dynamic scene through employing an RGBD camera, obtaining a color depth image sequence, and converting a single-frame image in the color depth image sequence into a three-dimensional point cloud; matching the parameters through binary feature coding, matching point pairs between the three-dimensional point cloud and vertexes of a standard model are obtained, wherein the attitude parameters of the standard model are determined according to a first frame of color depth image in the color depth image sequence; determining an energy function corresponding to the set corresponding to the matching point pair, and constructing and updating non-rigid motion field parameters from the standard model to the three-dimensional point cloud according to the energy function; and updating the standard model according to the non-rigid motion field parameters to obtain a target model. Therefore, the real-time performance, the robustness and the accuracy of the reconstruction of the dynamic scene can be effectively improved.

Description

Technical field[0001]The present invention relates to the field of computer vision and computer graphics, and more particularly to a three-dimensional reconstruction method based on dynamic scene real-based dynamic scene based on binarized feature encoding.Background technique[0002]Dynamic scene 3D reconstruction is the basis of computer vision and computer graphics and important issues. High quality dynamic scenes 3D reconstruction in virtual reality, sports games, film and refreshures, etc. have extensive application prospects and important application value.[0003]In the relevant technique, the system based on laser scanner or multiphase array requires that the scanning process is required to be absolutely stationary, it is difficult to handle dynamic scenes, and this type of equipment is expensive, it is difficult to popularize the daily life of ordinary people. The reconstruction method of the existing RGBD camera is typically based on hypothetical hypothesis, it is difficult to...

AI Technical Summary

Problems solved by technology

[0003] In related technologies, systems based on laser scanners or multi-camera arrays require the object to be scanned to remain absolutely still during the scanning process, which makes it difficult to handle dynamic scenes, and such equipment is expensive and difficult to popularize in the daily lives of ordinary people.

Existing RGBD camera-based reconstruction methods are usually based on the assumption that the object's motion range is small, and it is difficult to complete high-precision reconstruction of fast-moving non-rigid objects

Images