Dual-mode Infrared Image Fusion Algorithm Selection Based on Possibility Information Quality Synthesis

Xiaoming GUO; Linna JI; Fengbao YANG

doi:10.3788/gzxb20215003.0310003

Journals >Acta Photonica Sinica >Volume 50 >Issue 3 >Page 167 > Article

Acta Photonica Sinica
Vol. 50, Issue 3, 167 (2021)

Dual-mode Infrared Image Fusion Algorithm Selection Based on Possibility Information Quality Synthesis

Xiaoming GUO, Linna JI, and Fengbao YANG

Author Affiliations

College of Information and Communications Engineering， North University of China， Taiyuan030051， China

show less

DOI: 10.3788/gzxb20215003.0310003 Cite this Article

Xiaoming GUO, Linna JI, Fengbao YANG. Dual-mode Infrared Image Fusion Algorithm Selection Based on Possibility Information Quality Synthesis[J]. Acta Photonica Sinica, 2021, 50(3): 167 Copy Citation Text

show less

Fig. 1. Flow chart of dual-mode infrared image fusion algorithm selection

Download full size | View in the Article

Fusion image results (based on fusion algorithm sets, from left to right are A1 to A11)

Fig. 3. Fusion image results (based on fusion algorithm sets, from left to right are A₁ to A₁₁)

Download full size | View in the Article

Fig. 4. Scattering distribution graphs of difference feature amplitude fusion effectiveness

Download full size | View in the Article

Fig. 5. The possibility distribution of the difference feature fusion effectiveness under multiple fusion algorithms

Download full size | View in the Article

Fig. 6. The credibility of the possibility distribution subsets of the difference feature fusion effectiveness

Download full size | View in the Article

Fig. 7. The separability of the possibility distribution subsets of the difference feature fusion effectiveness

Download full size | View in the Article

Fig. 8. The weight of the possibility distribution subsets of the difference feature fusion effectiveness

Download full size | View in the Article

Fig. 9. The information quantity of weighted composite subsets under each fusion algorithm

Download full size | View in the Article

Fig. 10. The credibility of the weighted composite subsets under each fusion algorithm

Download full size | View in the Article

Fig. 11. The score of weighted composite subsets under each fusion algorithm

Download full size | View in the Article

Fig. 12. The score of the non-dominated subsets under each fusion algorithm

Download full size | View in the Article

Fig. 13. S_w of different fusion algorithms of two groups of images

Download full size | View in the Article

Number of inclusive subsets	Weighted composite subset definition
1	$B_{1} = W_{1} \cdot π_{1}$
2	$B_{7} = (W_{1} \cdot π_{1} + W_{2} \cdot π_{2}) / 2$ …… $B_{21} = (W_{5} \cdot π_{5} + W_{6} \cdot π_{6}) / 2$
3	$B_{22} = (W_{1} \cdot π_{1} + W_{2} \cdot π_{2} + W_{3} \cdot π_{3}) / 3$ …… $B_{41} = (W_{4} \cdot π_{4} + W_{5} \cdot π_{5} + W_{6} \cdot π_{6}) / 3$
4	$B_{42} = (W_{1} \cdot π_{1} + W_{2} \cdot π_{2} + W_{3} \cdot π_{3} + W_{4} \cdot π_{4}) / 4$ …… $B_{56} = (W_{3} \cdot π_{3} + W_{4} \cdot π_{4} + W_{5} \cdot π_{5} + W_{6} \cdot π_{6}) / 4$
5	$B_{57} = (W_{1} \cdot π_{1} + W_{2} \cdot π_{2} + W_{3} \cdot π_{3} + W_{4} \cdot π_{4} + W_{5} \cdot π_{5}) / 5$ …… $B_{62} = (W_{2} \cdot π_{2} + W_{3} \cdot π_{3} + W_{4} \cdot π_{4} + W_{5} \cdot π_{5} + W_{6} \cdot π_{6}) / 5$
6	$B_{63} = (W_{1} \cdot π_{1} + W_{2} \cdot π_{2} + W_{3} \cdot π_{3} + W_{4} \cdot π_{4} + W_{5} \cdot π_{5} + W_{6} \cdot π_{6}) / 6$

Table 1. Weighted composition of multiple vector subsets

View in the Article

Fusion algorithm	Non-dominated subsets
A₁	{B₃，B₂₄，B₂₇，B₃₃，B₃₇，B₄₀，B₄₃，B₄₇，B₅₀，B₅₉，B₆₃}
A₂	{B₅，B₆，B₃₅，B₃₇，B₄₀，B₅₆，B₅₇，B₅₉，B₆₁，B₆₂，B₆₃}
A₃	{B₅，B₃₃，B₃₇，B₄₀，B₄₇，B₅₀，B₅₇，B₅₉，B₆₁}
A₄	{B₂，B₃，B₂₅，B₃₃，B₃₄，B₄₀，B₄₃，B₄₄，B₅₀，B₅₉，B₆₀，B₆₁，B₆₃}
A₅	{B₅，B₃₁，B₃₃，B₄₀，B₄₃，B₅₀，B₅₄，B₅₅，B₅₈，B₆₃}
A₆	{B₂，B₃，B₂₄，B₂₇，B₃₃，B₃₇，B₄₀，B₄₃，B₄₇，B₅₀，B₅₄，B₅₉，B₆₃}
A₇	{B₂，B₆，B₁₅，B₂₁，B₃₇，B₅₉，B₆₁，B₆₃}
A₈	{B₅，B₃₃，B₃₇，B₃₈，B₄₀，B₄₁，B₅₀，B₅₄，B₅₅，B₅₆，B₅₇，B₅₉，B₆₁，B₆₃}
A₉	{B₅，B₁₅，B₂₁，B₃₇，B₄₀，B₄₁，B₅₆，B₆₁，B₆₂}
A₁₀	{B₆，B₃₇，B₄₀，B₅₆，B₅₉，B₆₁，B₆₃}
A₁₁	{B₆， B₁₅，B₃₅，B₄₀，B₄₁，B₅₅，B₅₆，B₆₁，B₆₃}

Table 2. Non-dominated subsets of the first group image

View in the Article

Fusion algorithm	Non-dominated subsets
A₁	{B₃，B₂₄，B₂₅，B₂₇，B₃₁，B₃₃，B₃₇，B₄₃，B₄₇，B₅₀，B₅₄，B₅₉，B₆₃}
A₂	{B₅，B₆，B₄₀，B₅₉，B₆₂，B₆₃}
A₃	{B₃，B₂₄，B₂₇，B₃₃，B₃₇，B₄₀，B₄₃，B₅₀，B₅₄，B₆₃}
A₄	{B₂，B₃，B₁₂，B₂₂，B₂₄，B₂₅，B₃₃，B₃₄，B₄₃，B₄₄，B₅₀，B₅₄，B₆₃}
A₅	{B₃，B₂₈，B₃₃，B₃₄，B₃₇，B₄₀，B₄₃，B₄₇，B₅₀，B₅₄，B₅₉，B₆₃}
A₆	{B₅，B₆，B₁₉，B₂₀，B₃₁，B₃₃，B₃₇，B₄₀，B₄₇，B₅₀，B₅₄，B₆₃}
A₇	{B₆，B₂₇，B₂₈，B₃₁，B₃₇，B₄₀，B₄₇，B₅₀，B₆₀，B₆₃}
A₈	{B₃，B₅，B₂₇，B₄₀，B₄₇，B₅₀，B₆₃}
A₉	{B₅，B₂₇，B₃₃，B₄₀，B₅₀，B₅₄，B₅₉，B₆₂，B₆₃}
A₁₀	{B₃，B₄₀，B₅₀，B₅₅，B₆₁，B₆₃}
A₁₁	{B₅，B₆， B₂₈，B₃₇，B₄₀，B₄₇，B₅₀，B₅₉，B₆₂，B₆₃}

Table 3. Non-dominated subsets of the second group image

View in the Article

Evaluation index	Fusion algorithm
Evaluation index	A₁	A₂	A₃	A₄	A₅	A₆	A₇	A₈	A₉	A₁₀	A₁₁
X₁	0.400 7	0.573 4	0.369 8	0.451 1	0.310 4	0.433 1	0.579 1	0.405 6	0.577 2	0.427 8	0.558 0
X₂	0.493 1	0.411 8	0.380 7	0.524 1	0.389 3	0.533 5	0.535 6	0.377 7	0.424 2	0.446 3	0.416 1
X₃	0.564 8	0.706 0	0.559 1	0.544 1	0.530 1	0.284 8	0.711 3	0.565 1	0.713 3	0.645 9	0.723 6
X₄	0.217 3	0.287 8	0.202 2	0.258 2	0.188 1	0.143 7	0.325 1	0.190 5	0.311 4	0.248 7	0.286 9
X₅	0.463 3	0.388 6	0.344 5	0.519 7	0.345 8	0.528 7	0.511 6	0.343 2	0.401 2	0.419 9	0.396 7
X₆	0.754 2	0.463 5	0.617 4	0.545 1	0.559 9	0.493 6	0.688 6	0.432 7	0.597 0	0.570 4	0.602 8
X₇	4.085 4	2.821 3	3.350 7	7.412 1	3.685 3	7.850 6	7.101 1	3.117 3	2.988 9	3.165 3	2.845 1
X₈	0.041 8	0.055 6	0.063 7	0.045 1	0.049 5	0.028 2	0.067 4	0.037 6	0.066 9	0.048 0	0.068 9
X₉	15.051 7	18.009 7	22.937 8	16.231 6	17.832 7	10.144 8	19.227 7	13.540 3	18.093 7	17.284 8	14.810 2
X₁₀	31.485 5	26.205 1	30.853 5	29.999 8	30.081 9	29.951 2	30.628 9	29.502 2	30.387 8	29.862 1	30.361 0

Table 4. Evaluation index of different fusion algorithms of the first group image

View in the Article

Evaluation index	Fusion algorithm
Evaluation index	A₁	A₂	A₃	A₄	A₅	A₆	A₇	A₈	A₉	A₁₀	A₁₁
X₁	0.284 3	0.489 5	0.453 6	0.502 8	0.439 8	0.492 1	0.484 9	0.431 9	0.466 3	0.444 2	0.437 7
X₂	0.052 4	0.354 9	0.385 3	0.466 5	0.384 6	0.436 8	0.377 9	0.342 1	0.388 3	0.370 4	0.366 6
X₃	0.065 2	0.648 8	0.538 6	0.535 9	0.553 8	0.640 6	0.657 0	0.563 9	0.675 3	0.637 5	0.681 3
X₄	0.024 4	0.345 8	0.371 9	0.471 1	0.367 6	0.485 1	0.374 0	0.330 5	0.382 4	0.361 9	0.354 6
X₅	0.025 5	0.167 8	0.169 1	0.169 9	0.182 9	0.298 9	0.310 0	0.159 0	0.268 0	0.282 3	0.270 8
X₆	4.662 4	1.565 4	4.500 3	6.329 7	4.438 2	8.063 4	1.907 7	2.757 9	1.923 6	2.263 8	1.675 4
X₇	0.035 4	0.045 9	0.060 5	0.048 0	0.065 4	0.065 7	0.053 8	0.043 0	0.066 0	0.050 4	0.065 9
X₈	0.020 3	0.251 7	0.125 9	0.150 5	0.117 5	0.335 1	0.298 6	0.156 8	0.291 5	0.227 0	0.261 3
X₉	12.757 5	16.521 0	21.761 7	17.265 1	23.551 5	23.442 9	19.368 9	15.492 7	23.754 6	18.150 7	23.732 3
X₁₀	52.973 2	25.410 9	45.088 1	44.423 6	44.891 7	61.023 9	38.592 1	34.108 8	35.717 3	33.491 7	36.720 0

Table 5. Evaluation index of different fusion algorithms of the second group image

View in the Article

Fusion algorithm	A₁	A₂	A₃	A₄	A₅	A₆	A₇	A₈	A₉	A₁₀	A₁₁
S_w	5.727 2	4.577 6	4.627 0	6.496 5	3.317 2	4.342 2	9.037 6	2.430 7	6.409 8	4.920 5	5.928 0
Rank	5	8	7	2	10	9	1	11	3	6	4

Table 6. S_w of different fusion algorithms of the first group image

View in the Article

Fusion algorithm	A₁	A₂	A₃	A₄	A₅	A₆	A₇	A₈	A₉	A₁₀	A₁₁
S_w	1.250 5	5.235 1	6.584 8	6.743 5	6.865 0	9.736 1	6.932 9	4.676 7	7.469 8	6.035 5	7.134 5
Rank	11	9	7	6	5	1	4	10	2	8	3

Table 7. S_w of different fusion algorithms of the second group image

Xiaoming GUO, Linna JI, Fengbao YANG. Dual-mode Infrared Image Fusion Algorithm Selection Based on Possibility Information Quality Synthesis[J]. Acta Photonica Sinica, 2021, 50(3): 167

Download Citation

Tools

Save the article for my favorites

Paper Information