Body scanning oriented 3D object modelling method based on multi-view depth camera

Abstract

To implement the human body scanning process, in this paper, we propose a novel 3D object modelling method based on multi-view depth camera. The main innovations of this paper lie in that we model 3D object via fusing several depth images captured from depth cameras. Different from the 3D scanner-based 3D object modelling, this proposed method is much cheaper and easier to operate. To construct the multiple depth camera system, a camera is located at the centre point, and others are positioned at four directions with the same intervals. Afterwards, a cost function is defined to fuse the multiple views of depth images based on depth map of the surface points. Experimental results demonstrate that the proposed method can effectively 3D scan the human body, and our method is slightly inferior to 3D scanner. However, the proposed depth camera based method is much cheaper and easier to master.