Neutral beam infrared image distortion correction based on vanishing point detection

Liping Chen; Yongjian Xu; Zichen Yu; Rixin Wang; Xufeng Peng; Yizhen Xu; Ling Yu

doi:10.11884/HPLPB202335.220109

Journals >High Power Laser and Particle Beams >Volume 35 >Issue 2 >Page 029001 > Article

High Power Laser and Particle Beams
Vol. 35, Issue 2, 029001 (2023)

Neutral beam infrared image distortion correction based on vanishing point detection

Liping Chen^1、2, Yongjian Xu^2、*, Zichen Yu^1、2, Rixin Wang^2、3, Xufeng Peng^2、3, Yizhen Xu^2、3, and Ling Yu²

Author Affiliations

¹Institute of Physical Science and Information Technology, Anhui University, Hefei 230601, China

²Institute of Plasma Physics, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei 230031, China

³University of Science and Technology of China, Hefei 230026, China

show less

DOI: 10.11884/HPLPB202335.220109 Cite this Article

Liping Chen, Yongjian Xu, Zichen Yu, Rixin Wang, Xufeng Peng, Yizhen Xu, Ling Yu. Neutral beam infrared image distortion correction based on vanishing point detection[J]. High Power Laser and Particle Beams, 2023, 35(2): 029001 Copy Citation Text

show less

Fig. 1. Section of experimental platform

Download full size | View in the Article

Fig. 2. Diagnostic target and experimental thermal infrared image

Download full size | View in the Article

Fig. 3. Flowchart of image rectification

Download full size | View in the Article

Fig. 4. Results of line segment detection after Sobel convolution

Download full size | View in the Article

Fig. 5. Diagram of line segment relations

Download full size | View in the Article

Fig. 6. Result of image rectification

Download full size | View in the Article

Fig. 7. Changes in temperature

Download full size | View in the Article

Fig. 8. Comparison of detection results of each line segment detection algorithm

Download full size | View in the Article

algorithm	line segment clustering algorithm based on line segment relation
input	the sample set D = {L1,L2,…,Lm}, L means coordinate values of two endpoints of a line segment
output	coordinates of vanishing points
1:	R = $ \varnothing $
2:	fori = 1,2,…,mdo
3	C= $ \varnothing $
4:	forj = 1,2….,mdo
5:	calculate $ h $ betweenLi and Lj
6:	calculate $ \Delta \theta $ between Li and Lj
7:	if$ h $＜threshold and $ \Delta \theta $＜threshold then
8:	C=C$ \cup $Lj
9:	remove Lj from D
10:	R=R$ \cup $C
11:	remove Li from D
12:	take the two sets from R with the largest number of line segments.
13:	use the least square method to fit the two lines in Step 12.
14:	calculate the intersection of two lines. That is the vanishing point.

Table 1. Line segment clustering algorithm

	the degree of continuous	the number of lines detected in x-direction	the effective number of lines detected in x-direction	the number of lines detected in y-direction	the effective number of lines detected in y-direction	run time/s
Hough transform	1	3	3	0	0	0.021
probabilistic Hough transform	0	105	98	11	0	0.014
Cannylines	0	53	44	31	24	3.08
our algorithm	1	5	5	5	5	4.61

Table 2. Comparison of linear detection methods

Download Citation

Save the article for my favorites

Paper Information