| In recent years,3D motion reconstruction used widely becomes a research hotspot in the field of computer vision. How to describe the character of 3D motion reconstruction, and to reconstruct the motion, and how to improve the accuracy of 3D motion skeleton reconstruction, which would play an important role for solving the problems about 3D motion reconstruction.Based on existing research, a new method for three dimensional motion skeleton reconstruction is proposed in this paper by studying the algorithm and technology of 3D motion reconstruction. The RGB color and depth image are obtained by Kinect, and the background of the RGB color image is removed, and body joint points are positioned, and the three-dimensional skeleton is extracted by combining with the depth information. In general, the specific research work in this paper is expanded from the following aspects:(1) The research direction is proposed, by studying the development situation at home and abroad and analyzing the advantages and disadvantages of different methods about 3 D motion reconstruction.(2) Getting the RGB color and depth image by Kinect at the same time is proposed, and they are aligned, based on summarizing acquisition method for the RGB color and depth image. At the same time, algorithm about removing RGB color image background is proposed based on background difference methods.(3) By analyzing exciting skeleton extraction algorithm and comparing the advantages and disadvantages of various algorithms, an algorithm of positioning body joint points is put forward, and body joint points are found out.(4) Because each pixel of the depth image is three-dimensional, a method to transforming two-dimensional coordinate for the three dimensional coordinate is put forward, then get the 3D body skeleton.A large number of experimental results and data showed that the test system can accurately extract the 3D skeleton, by using the proposed new three-dimensional movement skeleton reconstruction method. |
PDF Full Text Request