Point Cloud Registration Algorithm Based on Extended Point Feature Histogram Feature

Hui Tang; Mingquan Zhou; Guohua Geng

doi:10.3788/LOP56.241503

Laser & Optoelectronics Progress, Volume. 56, Issue 24, 241503(2019)

Point Cloud Registration Algorithm Based on Extended Point Feature Histogram Feature

Hui Tang^1,2, Mingquan Zhou^1,3、*, and Guohua Geng¹

Author Affiliations

¹School of Information Science & Technology, Northwest University, Xi'an, Shaanxi 710127, China;

²Experimental Training Teaching Management Center, Xi'an University of Finance and Economics, Xi'an, Shaanxi 710127, China

³School of Arts & Communication, Beijing Normal University, Beijing 100875, China;

show less

Abstract Get PDF(in Chinese)

Figures & Tables(6)

Fig. 1. Effected region of PFH calculation for feature point p₀

Download full size

Fig. 2. Schematic of Darboux frame

Download full size

Fig. 3. Diagram of 3D spherical grid

Download full size

Fig. 4. Experimental results of the proposed method. (a)(d)(g)(j)(m) Original pieces; (b)(e)(h)(k)(n) results of rough matching based on SAC-IA algorithm; (c)(f)(i)(l)(o) results of fine matching based on improved k-d tree algorithm

Download full size

Table 1. Operating parameters of the proposed algorithm

View table

Table 1. Operating parameters of the proposed algorithm

Experimental data		Number of points onthe fracture surface	Accuracy /mm	Initial matchingtime /s	Precise matchingtime /s	Overall runningtime /s
Cake 1#,Cake 2#	Cake 1#	7035	0.0186	4.431	2.954	7.385
Cake 1#,Cake 2#	Cake 2#		0.0186	4.431	2.954	7942
Cake 3#,Cake 4#	Cake 3#	9688	0.0213	7.256	2.418	9.674
Cake 3#,Cake 4#	Cake 4#		0.0213	7.256	2.418	6354
G3-I-C	Fragment 1#	8643	0.0297	5.177	2.414	7.591
G3-I-C	Fragment 2#		0.0297	5.177	2.414	8937
G10-5	Fragment 3#	2543	0.0205	3.802	1.090	4.892
G10-5	Fragment 4#		0.0205	3.802	1.090	2711
G10-6	Fragment 5#	6041	0.0283	5.623	1.633	7.256
G10-6	Fragment 6#		0.0283	5.623	1.633	5486

Table 2. Comparison of operating parameters of different algorithms

View table

Table 2. Comparison of operating parameters of different algorithms

Experimental data		Number of points onthe fracture surface	Method			Overall runningtime /s
Cake 1#,Cake 2#	Cake 1#	7035	ICP	0.0336	12.364
	Cake 1#	7035	Method in Ref. [3]	0.0289			9.478
	Cake 2#	7942	Method in Ref. [7]	0.0243		10.752
	Cake 2#	7942	Our method	0.0186	7.385
Cake 3#,Cake 4#	Cake 3#	9688	ICP	0.0375	18.842
	Cake 3#	9688	Method in Ref. [3]	0.0248			13.621
	Cake 4#	6354	Method in Ref. [7]	0.0251		15.419
	Cake 4#	6354	Our method	0.0213	9.674
G3-I-C	Fragment 1#	8643	ICP	1.0346	13.254
	Fragment 1#	8643	Method in Ref. [3]	0.0325			8.027
	Fragment 2#	8937	Method in Ref. [7]	0.0318		8.361
	Fragment 2#	8937	Our method	0.0297	7.591
G10-5	Fragment 3#	2543	ICP	1.0510	6.468
	Fragment 3#	2543	Method in Ref. [3]	0.0304			5.124
	Fragment 4#	2711	Method in Ref. [7]	0.0212		5.867
	Fragment 4#	2711	Our method	0.0205	4.892
G10-6	Fragment 5#	6041	ICP	1.0397	10.412
	Fragment 5#	6041	Method in Ref. [3]	0.0341			7.983
	Fragment 6#	5486	Method in Ref. [7]	0.0365		8.025
	Fragment 6#	5486	Our method	0.0283	7.256

Tools

Get Citation

Copy Citation Text

Hui Tang, Mingquan Zhou, Guohua Geng. Point Cloud Registration Algorithm Based on Extended Point Feature Histogram Feature[J]. Laser & Optoelectronics Progress, 2019, 56(24): 241503

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Machine Vision

Received: Apr. 29, 2019

Accepted: Jun. 13, 2019

Published Online: Nov. 26, 2019

The Author Email: Zhou Mingquan (puss.tang@qq.com)

DOI:10.3788/LOP56.241503

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology