Point Cloud Registration Method Based on Dual Quaternion Description of Line-Planar Feature Constraints

Raobo Li; Xiping Yuan; Shu Gan; Rui Bi; Sha Gao; Yan Guo

doi:10.3788/AOS202242.0214003

Acta Optica Sinica, Volume. 42, Issue 2, 0214003(2022)

Point Cloud Registration Method Based on Dual Quaternion Description of Line-Planar Feature Constraints

Raobo Li¹, Xiping Yuan^2,3, Shu Gan^1,2、*, Rui Bi¹, Sha Gao¹, and Yan Guo¹

Author Affiliations

¹Faculty of Land Resources and Engineering, Kunming University of Science and Technology, Kunming, Yunnan 650093, China

²Application Engineering Research Center of Spatial Information Surveying and Mapping Technology in Plateau and Mountainous Areas Set by Universities in Yunnan Province, Kunming, Yunnan 650093, China

³College of Geosciences and Engineering, West Yunnan University of Applied Sciences, Dali, Yunnan, 671009, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(9)

Fig. 1. Spatial geometric relation between straight line and plane. (a) Selected line-planar features in un-registered point cloud; (b) selected line-planar features in target point cloud

Download full size

Fig. 2. Flowchart of point cloud registration algorithm

Download full size

Fig. 3. Point cloud data of station 16 and station 17 before and after registration. (a) Before registration; (b) after registration

Download full size

Fig. 4. Influence of noise standard deviation on registration accuracy under different conditions. (a) Angle between lines; (b) angle between planes

Download full size

Fig. 5. Registration results of Ref. [7,14]. (a) Ref. [7]; (b) Ref. [14]

Download full size

Table 1. Real value of dual quaternion considering scale factor and calculated value of proposed method
View table
Table 1. Real value of dual quaternion considering scale factor and calculated value of proposed method
Value r₀ r₁ r₂ r₃ t₀ t₁ t₂ t₃ μ
Real value 0.9953 -0.0002 -0.0001 -0.0968 0.0139 -1.0697 -4.5558 0.1475 1.0000
Calculated value 0.9950 -0.0003 -0.0007 -0.0967 0.0107 -1.0690 -4.5549 0.1469 1.0002

Table 2. Number of iterations and running time of L-M algorithm and Gaussian-Newton algorithm under different initial values
View table
Table 2. Number of iterations and running time of L-M algorithm and Gaussian-Newton algorithm under different initial values
Algorithm Parameter 1 5 10 15 20 25
L-M Number of iterations 68 132 503 1720 4328 8641
Time /s 0.134 0.306 0.603 2.075 5.356 10.402
Gaussian-Newton Number of iterations 18 - - - - -
Time /s 0.028 - - - - -

Table 3. Spatial transformation parameters calculated by different methods

View table

Table 3. Spatial transformation parameters calculated by different methods

Method	R	T /m	u
Theoretical value	$[\begin{array}{l} 0.9813 & 0.1927 & - 0.0001 \\ - 0.1927 & 0.9813 & 0.0003 \\ 0.0001 & - 0.0003 & 1.0000 \end{array}]$	$[\begin{array}{l} - 3.0113 \\ - 8.8617 \\ 0.2976 \end{array}]$	1.0000
Proposed method	$[\begin{array}{l} 0.9795 & 0.1934 & - 0.0012 \\ - 0.1914 & 0.9803 & 0.0015 \\ 0.0017 & 0.0005 & 1.0000 \end{array}]$	$[\begin{array}{l} - 3.0102 \\ - 8.8634 \\ 0.2954 \end{array}]$	1.0002
Ref. [7]	$[\begin{array}{l} 0.9818 & 0.1935 & - 0.0013 \\ - 0.1955 & 0.9820 & - 0.0002 \\ - 0.0024 & 0.0002 & 1.0000 \end{array}]$	$[\begin{array}{l} - 3.0120 \\ - 8.8612 \\ 0.2961 \end{array}]$	0.9999
Ref. [14]	$[\begin{array}{l} 0.9830 & 0.1928 & 0.0019 \\ - 0.1922 & 0.9814 & 0.0021 \\ - 0.0019 & 0.0018 & 1.0000 \end{array}]$	$[\begin{array}{l} - 3.0123 \\ - 8.8649 \\ 0.2952 \end{array}]$	1.0005

Table 4. Angle between line and line, surface and surface and their errors

View table

Table 4. Angle between line and line, surface and surface and their errors

Method	α_L /(°)	Δα_L /(°)	α_P /(°)	Δα_P /(°)	Method	α_L /(°)	Δα_L /(°)	α_P /(°)	Δα_P /(°)
Theoretical value	0.0953	0.0805	0.0934	0.0789	Proposed method	0.1619	0.1148	0.1581	0.1131
	0.0837		0.0824			0.0492		0.0501
	0.0770		0.0760			0.1081		0.1081
	0.0578		0.0570			0.0947		0.0920
	0.0230		0.0181			0.0587		0.0595
Ref. [7]	0.1469	0.1266	0.1118	0.1240	Ref. [14]	0.1374	0.1286	0.1709	0.1253
	0.0787		0.0236			0.0760		0.0114
	0.0797		0.1172			0.0718		0.0964
	0.1162		0.1274			0.1412		0.1008
	0.1283		0.1358			0.1282		0.1183

Tools

Get Citation

Copy Citation Text

Raobo Li, Xiping Yuan, Shu Gan, Rui Bi, Sha Gao, Yan Guo. Point Cloud Registration Method Based on Dual Quaternion Description of Line-Planar Feature Constraints[J]. Acta Optica Sinica, 2022, 42(2): 0214003

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Lasers and Laser Optics

Received: Jun. 28, 2021

Accepted: Aug. 17, 2021

Published Online: Dec. 29, 2021

The Author Email: Gan Shu (n1480@qq.com)

DOI:10.3788/AOS202242.0214003

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Real value of dual quaternion considering scale factor and calculated value of proposed method

Table 1. Real value of dual quaternion considering scale factor and calculated value of proposed method

Table 2. Number of iterations and running time of L-M algorithm and Gaussian-Newton algorithm under different initial values

Table 2. Number of iterations and running time of L-M algorithm and Gaussian-Newton algorithm under different initial values

Table 3. Spatial transformation parameters calculated by different methods

Table 3. Spatial transformation parameters calculated by different methods

Table 4. Angle between line and line, surface and surface and their errors

Table 4. Angle between line and line, surface and surface and their errors