Camera Calibration Method with Multi-condition Constraints Based on Laser Points

Cailin LI; Benya SU; Baoyun GUO; Zhiyong WANG; Fuhong ZHU

doi:10.3788/gzxb20215004.0415001

Journals >Acta Photonica Sinica >Volume 50 >Issue 4 >Page 172 > Article

Acta Photonica Sinica
Vol. 50, Issue 4, 172 (2021)

Camera Calibration Method with Multi-condition Constraints Based on Laser Points

Cailin LI, Benya SU^*, Baoyun GUO, Zhiyong WANG, and Fuhong ZHU

Author Affiliations

Institute of Architecture and Engineering， Shandong University of Technology， Zibo， Shandong255000China

show less

DOI: 10.3788/gzxb20215004.0415001 Cite this Article

Cailin LI, Benya SU, Baoyun GUO, Zhiyong WANG, Fuhong ZHU. Camera Calibration Method with Multi-condition Constraints Based on Laser Points[J]. Acta Photonica Sinica, 2021, 50(4): 172 Copy Citation Text

show less

Fig. 1. Flow of overall algorithm

Download full size | View in the Article

Fig. 2. 3D control field with different postures

Download full size | View in the Article

Fig. 3. Initial homonymy points selection of Confucius statue and stone

Download full size | View in the Article

Fig. 4. Three-dimensional reconstruction of Confucius statue before and after camera calibration

Download full size | View in the Article

Fig. 5. Three-dimensional reconstruction of stone before and after camera calibration

Download full size | View in the Article

Fig. 6. Reflection error

Download full size | View in the Article

Fig. 7. Effect of laser point cloud reflection of Confucius and stone statue

Download full size | View in the Article

Fig. 8. The error distribution of the total unit weight of each iteration in the adjustment process

Download full size | View in the Article

Date	Number of images	Image size（pixel×pixel）	Number of image 3D points	Number of laser points
Confucius statue	62	3 168×2 112	8 375	713 209
Stone	39	3 168×2 112	14 297	313 937

Table 1. Basic information of experiment data

View in the Article

	3D control field	Proposed method
		Confucius statue		Stone
		Initial parameters	Parameters after calibration	Initial parameters	Parameters after calibration
$f$ (pixel)	4 642.07	4 599.21	4 608.82	4 607.99	4 605.99
$x_{0}$ (pixel)	1 581.02	1 570.44	1 579.97	1 585.44	1 586.42
$y_{0}$ (pixel)	1 041.89	1 060.56	1 051.27	1 057.41	1 051.27
$k_{1}$	-7.11×10^-9	-2.92×10^-13	-3.38×10^-9	-6.63×10^-9	-6.5×10^-9
$k_{2}$	1.59×10^-16	-1.46×10^-25	1.64×10^-17	1.65×10^-16	5.41×10^-17
$k_{3}$	2.54×10^-23	1.91×10^-38	6.28×10^-24	1.91×10^-23	4.54×10^-23
$p_{1}$	1.06×10^-7	2.12×10^-7	-5.59×10^-8	1.58×10^-7	1.62×10^-7
$p_{2}$	5.48×10^-8	-9.56×10^-8	1.49×10^-8	6.22×10^-8	9.37×10^-8
$σ_{0}$	0.224	\	0.363	\	0.260
$σ_{I}$ (pixel)	0.257	\	0.356	\	0.242

Table 2. Comparison of camera calibration results

View in the Article

3D model

accuracy

Confucius statue

Stone

Initial parameters

Parameters after calibration

Initial parameters

Parameters after calibration

Accuracy

19.8

14.8

17.9

12.4

Integrity

21.3

15.4

16.2

9.8

Table 3. 3D reconstruction model accuracy statistics

View in the Article

		5%	20%	35%	50%	65%	80%	95%	100%
Confucius statue	Number of laser points	35 665	142 661	249 658	356 604	463 650	570 647	677 544	713 209
	Density of laser points/ $m^{- 2}$	2 822	1 1291	19 763	28 228	36 702	45 172	53 634	56 457
	$σ_{0}$	0.487	0.394	0.378	0.369	0.368	0.368	0.369	0.363
	$σ_{I}$ /pixel	0.409	0.368	0.359	0.359	0.358	0.359	0.357	0.356
Stone	Number of laser points	15 696	62 787	109 877	156 968	204 059	251 149	298 240	313 937
	Density of laser points/ $m^{- 2}$	1 986	7 947	13 908	19 869	25 830	31 791	37 752	39 739
	$σ_{0}$	0.389	0.274	0.263	0.263	0.262	0.263	0.264	0.260
	$σ_{I}$ /pixel	0.366	0.259	0.246	0.244	0.250	0.253	0.243	0.242

Table 4. Effect of different sampling proportions of laser point cloud on the accuracy of camera inspection

Cailin LI, Benya SU, Baoyun GUO, Zhiyong WANG, Fuhong ZHU. Camera Calibration Method with Multi-condition Constraints Based on Laser Points[J]. Acta Photonica Sinica, 2021, 50(4): 172

Download Citation

Tools

Save the article for my favorites

Paper Information