Group Intelligent Hybrid Optimization Algorithm for Image Segmentation of Deep Space Exploration

Qiying Nie; Zhencai Zhu; Yonghe Zhang; Yamin Wang

doi:10.3788/LOP202158.0210002

Journals >Laser & Optoelectronics Progress >Volume 58 >Issue 2 >Page 0210002 > Article

Laser & Optoelectronics Progress
Vol. 58, Issue 2, 0210002 (2021)

Group Intelligent Hybrid Optimization Algorithm for Image Segmentation of Deep Space Exploration

Qiying Nie^1、2、3, Zhencai Zhu^1、3、*, Yonghe Zhang^1、3, and Yamin Wang^1、3

Author Affiliations

¹Innovation Academy for Microsatellites of CAS, Shanghai 201203, China

²University of Chinese Academy of Sciences, Beijing 100049, China

³Key Laboratory of Microsatellites, Shanghai 201203, China

show less

DOI: 10.3788/LOP202158.0210002 Cite this Article Set citation alerts

Qiying Nie, Zhencai Zhu, Yonghe Zhang, Yamin Wang. Group Intelligent Hybrid Optimization Algorithm for Image Segmentation of Deep Space Exploration[J]. Laser & Optoelectronics Progress, 2021, 58(2): 0210002 Copy Citation Text

show less

Fig. 1. Wolves algorithm hierarchy model

Download full size

Fig. 2. Segmentation process of proposed algorithm

Download full size

Fig. 3. Segmentation results of single threshold segmentation method. (a) Original images; (b) segmentation results

Download full size

Fig. 4. Segmentation results of Otsu algorithm. (a) Original images; (b) segmentation results; (c) gray histograms

Download full size

Fig. 5. Segmentation results of proposed algorithm. (a) Image 1; (b) image 2; (c) image 3

Download full size

Fig. 6. Segmentation detail diagram of different algorithms. (a) Original image; (b) single threshold segmentation; (c) Otsu algorithm; (d) proposed algorithm

Download full size

Name	Function	n	Range
F1	f₁(x)= $\overset{n}{\sum_{i = 1}} x_{i}^{2}$	30	[-100,100]
F2	f₂(x)= $\overset{n}{\sum_{i = 1}}$ \|x_i\|+ $\overset{n}{\prod_{i = 1}}$ \|x_i\|	30	[-10,10]
F3	f₃(x)= $\max_{i} [\|x_{i}\|, 1 \leq i \leq n]$	30	[-100,100]

Table 1. Unimodal benchmark functions

Name	Function	n	Range
F4	$\begin{matrix} f_{4} (x) = - 20 \exp (- 0.2 \sqrt[]{\frac{1}{n} \overset{n}{\sum_{i = 1}} x_{i}^{2}}) - \exp [\frac{1}{n} \overset{n}{\sum_{i = 1}} \cos (2 π x_{i})] + 20 + e \end{matrix}$	30	[-32,32]
F5	f₅ $(x)$ = $\frac{1}{4000} \overset{n}{\sum_{i = 1}} x_{i}^{2}$ - $\overset{n}{\prod_{i = 1}}$ cos $(x_{i} / \sqrt[]{i})$ +1	30	[-600,600]
F6	f₆ $(x)$ = $\overset{n}{\sum_{i = 1}} x_{i}^{2}$ -10cos $(2 π x_{i})$ +10	30	[-5.12,5.12]

Table 2. Multimodal benchmark functions

Name	Function	n	Range	Min
F7	f₇ $(x)$ = $\overset{11}{\sum_{c = 1}} {[a_{c} - \frac{x_{1} (b_{c}^{2} + b_{c} x_{2})}{b_{c}^{2} + b_{c} x_{3} + x_{4}}]}^{2}$	4	[-5,5]	0.00030
F8	f₈ $(x)$ = ${(x_{2} - \frac{5.1}{4 π^{2}} x_{1}^{2} + \frac{5}{π} x_{1} - 6)}^{2}$ +10 $(1 - \frac{1}{8 π})$ cos x₁+10	2	[-5,5]	0.398
F9	$\begin{array}{l} f_{9} (x) = [1 + {(x_{1} + x_{2} + 1)}^{2} (19 - 14 x_{1} + 3 x_{1}^{2} - 14 x_{2} + 6 x_{1} x_{2} + 3 x_{2}^{2})] \times \\ [30 + {(2 x_{1} - 3 x_{2})}^{2} \times (18 - 32 x_{1} + 12 x_{1}^{2} + 48 x_{2} - 36 x_{1} x_{2} + 27 x_{2}^{2})] \end{array}$	2	[-2,2]	3

Table 3. Fixed-dimension multimodal benchmark functions

Functions	Proposed algorithm	GWO	PSO	DE	GSA
F1	5.18887×10^-49	2.80179×10^-27	0.000237219	0.00314	2.37×10^-16
F2	2.7456×10^-25	9.93897×10^-17	0.04260275	0.00339	1.08782
F3	1.9321×10^-15	6.53673×10^-7	1.120836	0.00334	43.30114
F4	4.3668×10^-15	1.10667×10^-13	0.07106	0.000275933	8.71582
F5	0	0.003175	0.016024	0.1462	335.2185875
F6	0	0.293711	66.0182	786.4320	13.13346
F7	0.0014554	0.010344514	0.000938607	1.50073×10^-7	0.00536
F8	0.398652	0.397891	0.3979	0.39771	0.3979
F9	3.00071	3.00733	3	3	3

Table 4. Average of different algorithms under different functions

Functions	Proposed method	GWO	PSO	DE	GSA
F1	2.83492×10^-49	4.09404×10^-27	0.000207836	0.000685079	9.74×10^-17
F2	2.18425×10^-25	8.07479×10^-17	0.067624678	0.000860814	0.980801962
F3	6.10984×10^-15	4.7406×10^-7	0.278588176	0.001478137	1.147027104
F4	4.3668×10^-15	6.19657×10^-16	0.20023266	5.5731×10^-5	1.479301656
F5	0	0	0.012184864	0.08498613	180.0696952
F6	0	49.4631	20.8949	0	3.24273875
F7	3.40825×10^-5	0.01056044	0.00012533	1.44066×10^-7	0.004790778
F8	0.001191216	3.16228×10^-6	0	9.21×10^-8	0
F9	0.001747029	0.008511176	0	0	0

Table 5. Standard deviation of different algorithm under different functions

Image	Single threshold	Otsu	Proposed algorithm
Image	Single threshold	Otsu	Logarithmic entropy	Exponential entropy	T entropy
1	110	65, 109	76, 155	74, 148	78, 149
2	127	80, 120	74, 154	78, 153	78, 152
3	135	80, 123	74, 150	74, 153	71, 149

Table 6. Threshold comparison of different algorithms

Image	Number of thresholds is 2	Number of thresholds is 3	Number of thresholds is 4	Number of thresholds is 5
1	75, 153	74, 123, 211	67, 74, 153, 186	74, 77, 153, 207, 243
2	78, 149	74, 92, 153	74, 153, 202, 230	74, 82, 114, 153, 188
3	72, 154	74, 131, 153	74, 119, 127, 153	74, 127, 153, 185, 208

Table 7. Results of multi-level threshold search

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information