Registration of 3D Point Cloud of Human Body Based on the Range Images and RGB Images

Abstract:

The existing 3D reconstruction techniques rarely can be easily used in people's daily life, and the traditional registration algorithms have the drawback of massive calculation. In this paper it presented a registration algorithm of body point cloud based on RGB images and Range images. First, it used kinect to obtain the RGB images and Range images from different perspectives. Then it extracted the pairs of 2D feature points on RGB images using scale-invariant feature transform (SIFT) and random sample consensus (RANSAC) algorithm application for the coarse registration and used the improved iterative closest point (ICP) algorithm for the fine registration. Second, it eliminated the background information and the noise points of the model edges. Finally it completed the registration process. Experimental results show that the algorithm can accurately accomplish the body point clouds registration using the low-cost instrument and has a relatively high efficiency.

• Periodical: Applied Mechanics and Materials (Volumes 738-739)
• Pages: 656-661
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.738-739.656
• Online since: March 2015
• Authors: Yan Wan, Jun Lu*, Ao Qiong Li
• Keywords: 3D Registration, Iterative Closest Point Algorithm, KINECT, Random Sample Consensus Algorithm, Scale-Invariant Feature Transform Algorithm

References

[1] Peter Henry, Michael Krainin, Evan Herbst, Xiaofeng Ren and Dieter Fox. RGB-D mapping: Using Kinect-style depth cameras for dense 3D modeling of indoor environments. In The International Journal of Robotics Research, 2012, 2(1): 1–17.

DOI: 10.1177/0278364911434148

[2] Ge Baozhen, Peng Bo, Tian Qingguo. Registration of Three-Dimensional Point-Cloud Data Based on Curvature Map[J]. Journal of TianJin University, 2013，46(2)：175 - 180.

[3] Wang Xin, Zhang Ming-ming, Yu Xiao, Zhang Ming-chao. Point cloud registration based on improved iterative closest point method[J]. Optics and Precision Engineering, 2012, 20(9): 2068-(2077).

DOI: 10.3788/ope.20122009.2068

[4] CHUA C S, JARVIS R. Point signatures: A new representation for 3d object recognition[J]. International Journal of Computer Vision, 1997, 25(1): 63-85.

[5] R.J. Ong, J.T. Dawley and P.G. Clem: submitted to Journal of Materials Research (2003).

[5] Lu Qiang, Wang Xiaolong, Liu Feng, Ni Peipei. Feature Extraction and the Description of Point Cloud Image Based on S-NARF Algorithm[J]. Science & Technology Review, 2013, 31(13: 45-48.

[6] Luo Wenchao, Liu Guodong, Yang Haiyan. Application of SIFT and advanced RANSAC algorithm on image registration. Computer Engineering and Applications, 2013, 49(15: 147-153).

[7] Chen Y，Medioni G．Object Modeling by Registration of Multiple Range Images[J]．Image and Vision Computing, 1992, 10(3): 145—155.

DOI: 10.1016/0262-8856(92)90066-c

[8] EZRA E, SHARIR M, EFRAT A. On the performance of the ICP algorithm[J]. Computational Geometry. 2008, 41(1): 77-93.

DOI: 10.1016/j.comgeo.2007.10.007

[9] Yuan Liang. Study on 3D point clouds registration algorithms in the pipeline of 3D model reconstruction[D]. Xidian university, (2010).

[10] EGGERT D W, LORUSSO A, FISHER R B. Estimating 3-D rigid body transformations: a comparison of four major algorithms[J]. Machine Vision and Applications, 1997, 9(5, 6): 272-290.

DOI: 10.1007/s001380050048

[11] CHEN Y, MEDIONI G. Object modeling by registration of multiple range images[J]. Image and vision computing, 1992, 10(3): 145-155.

DOI: 10.1016/0262-8856(92)90066-c

[12] Besl P J, McKay N D. A method for registration of 3D shape[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239-256.

DOI: 10.1109/34.121791