Power Equipment Infrared and Visible Images Registration Based on Cultural Wolf Pack Algorithm

Fig. 2. Saliency gradient of infrared image. (a) Original image; (b) image after saliency detection; (c) infrared image after division; (d) image after enhancing the saliency area; (e) image of saliency gradient

Download full size

Fig. 3. Saliency gradient of visible image. (a) Visible image; (b) image of saliency gradient

Download full size

Fig. 4. Schematic diagram of CWPA

Download full size

Fig. 5. Experimental sample of standard registration data set. (a) Visible image; (b) infrared image

Download full size

Fig. 6. Comparison results of different algorithms. (a) SMI; (b) GWW-NMI; (c) SGNMI

Download full size

Fig. 7. Part of the standard registration test image set. (a) Visible image; (b) infrared image

Download full size

Fig. 8. Test results of blurred images. (a) MAE; (b) RMSE

Download full size

Fig. 9. Visible image set and infrared image set. (a) Visible image; (b) infrared image

Download full size

Fig. 10. Test results of the actual data set. (a) Registration time; (b) MAE; (c) RMSE

Download full size

Fig. 11. Experimental results of actual data set. (a) σ_TRE; (b) registration time

Download full size

Table 1. Solution space of registration parameters
View table
Table 1. Solution space of registration parameters
Registration parameter h v q r
Solution space [-1000,1000] [-1000,1000] [-10,10] [0,360]

Table 2. Parameters of optimization algorithm

View table

Table 2. Parameters of optimization algorithm

Algorithm	Parameter
CPSO	N=100, i_ter=2000, inertia weight ω=0.7, learning factor c₁=c₂=1.5, individual speed limit [-0.5, 0.5]
WPA	N=100, ferocious wolves∶scout wolves=1∶1, i_ter=2000, T_max=10, step factor S=0.1; judging distance d=0.08, update scale factor β=3
CWPA	N=100, ferocious wolves∶scout wolves=1∶1, T_max=10, threshold parameter ε=0.5, update scale factor β=3

Table 3. Registration result of standard test image set

View table

Table 3. Registration result of standard test image set

Sample	MAE			RMSE			Registration time /s
Sample	GWW-NMI	SMI	SGNMI	GWW-NMI	SMI	SGNMI	GWW-NMI	SMI	SGNMI
1	0.897	1.435	0.931	1.213	2.241	1.391	0.734	4.231	1.032
2	1.293	1.692	1.125	1.479	2.693	1.592	0.823	3.328	1.143
3	0.736	1.613	0.962	0.986	2.861	1.242	0.672	4.054	1.097
4	1.043	1.973	0.947	1.435	3.173	1.374	0.743	4.426	0.969

Table 4. Mean value of registration results of 50 sets of standard test image sets
View table
Table 4. Mean value of registration results of 50 sets of standard test image sets
Mean MAE Mean RMSE Mean registration time /s
GWW-NMI SMI SGNMI GWW-NMI SMI SGNMI GWW-NMI SMI SGNMI
1.010 1.673 1.040 1.387 2.490 1.324 0.926 3.847 1.239

Table 5. Parameters of infrared camera
View table
Table 5. Parameters of infrared camera
Parameter Value
Resolution /pixel×pixel 384×288
Scene temperature range /℃ 0--200
Temperature accuracy /% ±2
Wavelength /μm 7--13
Focus range /m >0.6
Frame rate /Hz 8.7

Table 6. Standard test functions

View table

Table 6. Standard test functions

Function	Expression	Feature	Solution space
Sphere	$f_{1} = \overset{D}{\sum_{i = 1}} x_{i}^{2}$	L/U	[-10,10]²
Sumsquares	$f_{2} = \overset{D}{\sum_{i = 1}} i x_{i}^{2}$	L/U	[-10,10]¹⁰⁰
Booth	$f_{3} = (x_{1} + 2 x_{2} {- 7)}^{2} + (2 x_{1} + x_{2} {- 5)}^{2}$	L/M	[-10,10]²
Quadric	$f_{4} = \overset{D}{\sum_{i = 1}} {(\overset{i}{\sum_{k = 1}} x_{k})}^{2}$	L/M	[-30,30]¹⁰⁰
Powersum	$f_{5} = \overset{D}{\sum_{i = 1}} {[(\overset{D}{\sum_{j = 1}} x_{j}^{i}) - b_{i}]}^{2}$	H/U	[-10,10]²
Zakharov	$f_{6} = \overset{D}{\sum_{i = 1}} x_{i}^{2} + {(\overset{D}{\sum_{i = 1}} 0.5 i x_{i})}^{2} + {(\overset{D}{\sum_{i = 1}} 0.5 i x_{i})}^{4}$	H/U	[-10,10]¹⁰⁰
Griewank	$f_{7} = \frac{1}{4000} \overset{D}{\sum_{i = 1}} x_{i}^{2} - \overset{D}{\prod_{i = 1}} \cos (\frac{x_{i}}{\sqrt[]{i}}) + 1$	H/M	[-600,600]¹⁰⁰
Ackley	$\begin{matrix} f_{8} = - 20 \exp (- 0.2 \sqrt[]{\frac{1}{D} \overset{D}{\sum_{i = 1}} x_{i}^{2}}) - \exp [\frac{1}{D} \overset{D}{\sum_{i = 1}} \cos (2 π x_{i})] + 20 + e \end{matrix}$	H/M	[-32,32]¹⁰⁰

Table 7. Performance comparison of optimization algorithms

View table

Table 7. Performance comparison of optimization algorithms

Function	Algorithm	MEAN	STD	SR /%	AEN
Sphere	WPA	6.31×10^-91	8.13×10^-87	100	111.30
	CPSO	7.8×10^-188	2.1×10^-185	100	34.70
	CWPA	2.45×10^-91	8.97×10^-88	100	102.10
Sumsquares	WPA	2.16×10^-96	8.57×10^-96	100	88.46
	CPSO	1.03×10^-8	1.05×10^-8	100	33.64
	CWPA	4.84×10^-96	2.17×10^-95	100	77.56
Booth	WPA	1.32×10^-6	1.7×10^-6	100	145.50
	CPSO	0	0	100	31.24
	CWPA	1.07×10^-9	1.25×10^-9	100	87.42
Quadric	WPA	7.60×10^-84	2.21×10^-89	100	326.52
	CPSO	6.88×10⁺²	1.46×10⁺²	0	2000.00
	CWPA	6.82×10^-90	3.12×10^-90	100	226.60
Powersum	WPA	6.63×10^-95	2.30×10^-95	100	113.30
	CPSO	1.09×10^-6	1.01×10^-6	100	1172.00
	CWPA	4.84×10^-183	1.27×10^-183	100	38.42
Zakharov	WPA	3.18×10^-2	2.98×10^-1	100	391.80
	CPSO	1.20	2.20	0	2000.00
	CWPA	4.15×10^-16	2.06×10^-16	100	237.98
Griewank	WPA	1.44×10^-89	5.56×10^-89	100	243.52
	CPSO	1.03×10⁺³	2.97×10⁺²	0	2000.00
	CWPA	6.63×10^-90	1.88×10^-89	100	221.94
Ackley	WPA	9.33×10^-1	1.19	60	991.04
	CPSO	9.37×10⁺¹	1.37	0	2000.00
	CWPA	4.62×10^-10	3.29×10^-12	100	193.51

Tools

Get Citation

Copy Citation Text

Hongshan Zhao, Zeyan Zhang. Power Equipment Infrared and Visible Images Registration Based on Cultural Wolf Pack Algorithm[J]. Acta Optica Sinica, 2020, 40(16): 1610003

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites