基于邻域搜索JADE的二维Otsu多阈值图像分割

doi:10.3969/j.issn.1001-506X.2020.10.03

摘要/Abstract

摘要：

为进一步提高分割精度并加快分割速度,提出了一种基于邻域搜索可选外部存档自适应差分进行算法(简称为JADE-GL)的二维Otsu多阈值图像分割方案。首先,针对原始JADE算法精英突变策略收敛速度慢、容易陷入局部最优等问题,提出了基于邻域半径搜索的改进突变策略,以提升算法的全局探索和开发能力。然后,将所提算法与现有分割方法以及其他JADE变种算法进行二维Otsu多阈值分割对比实验。最后,通过函数收敛曲线、分割距离测度、峰值信噪比(peak signal to noise ratio, PSNR)等指标定量分析算法的性能。实验结果表明,随着阈值数增加,提出的算法无论在收敛速度、分割精度还是分割图像效果上都有显著提升。

关键词: 图像分割, 二维Otsu, 多阈值, 邻域搜索, JADE算法

Abstract:

In order to further improve the segmentation accuracy and speed up the segmentation, a 2D Otsu multi-threshold image segmentation scheme based on the adaptive differential evolution algorithm with optional external archive, with neighborhood search (JADE-GL for short). Firstly, aiming at the problem of the original JADE algorithm such as slow convergence speed and easy to fall into local optimum, an improved mutation strategy based on neighborhood radius search is proposed to improve the global exploration and development ability of the algorithm. Then, the propesed algorithm is compared with existing segmentation methods and other JADE variant algorithms for 2D Otsu multi-threshold segmentation comparison experiments. Finally, the performance of the algorithm is quantitatively analyzed by function convergence curve, segmentation distance measure and peak signal to noise ratio (PSNR). The experimental results show that the proposed algorithm has more significant advantages in terms of convergence speed, segmentation accuracy and the effect of segmentation images as the number of thresholds increases.

Key words: image segmentation, 2D Otsu, multi-threshold, neighborhood search, adaptive differential evolution algorithm with optional external archive (JADE)

中图分类号:

TP391.41

罗钧, 刘建强, 庞亚男. 基于邻域搜索JADE的二维Otsu多阈值图像分割[J]. 系统工程与电子技术, 2020, 42(10): 2164-2171.

Jun LUO, Jianqiang LIU, Yanan PANG. Multi-threshold image segmentation of 2D Otsu based on neighborhood search JADE[J]. Systems Engineering and Electronics, 2020, 42(10): 2164-2171.

图/表 8

图1

图2

图3

表1

图4

图5

表2

5种算法的数据对比"

算法		Lena(512×512)			Car(512×512)			Plane(512×512)
算法		双阈值	三阈值	四阈值	双阈值	三阈值	四阈值	双阈值	三阈值	四阈值
JADE	距离测度	4 842.77	4 906.97	4 924.38	6 775.94	6 822.67	6 820.92	4 237.65	4 298.51	4 302.95
	PSNR/dB	11.79	14.25	16.02	10.95	14.90	15.43	13.68	15.81	17.76
	最佳阈值	(89, 149) (193, 195)	(79, 81) (124, 152) (181, 182)	(71, 73) (116, 121) (141, 148) (183, 186)	(78, 78) (106, 156)	(78, 79) (105, 123) (161, 161)	(75, 76) (101, 102) (110, 122) (157, 158)	(93, 141) (184, 185)	(91, 91) (100, 135) (171, 171)	(31, 36) (86, 88) (114, 127) (169, 175)
L-SHADE	距离测度	4 862.06	4 958.56	4 979.32	6 775.94	6 823.60	6 824.28	4 242.17	4 308.93	4 346.97
	PSNR/dB	12.93	15.17	16.23	10.97	15.01	15.62	13.78	14.73	18.28
	最佳阈值	(86, 148) (193, 193)	(83, 83) (118, 150) (188, 189)	(81, 81) (118, 124) (139, 148) (194, 194)	(77, 78) (106, 156)	(76, 77) (106, 127) (169, 169)	(74, 75) (93, 95) (111, 128) (170, 171)	(91, 135) (180, 180)	(36, 41) (93, 130) (172, 172)	(87, 89) (105, 126) (158, 159) (170, 170)
CJADE	距离测度	4 863.53	4 977.39	4 999.63	6 775.87	6 820.06	6 828.37	4 244.89	4 314.66	4 352.11
	PSNR/dB	13.93	15.64	16.25	11.00	14.89	15.87	14.10	16.52	18.99
	最佳阈值	(87, 149) (194, 194)	(77, 78) (113, 149) (194, 194)	(78, 80) (117, 139) (156, 156) (199, 199)	(75, 78) (106, 155)	(76, 76) (104, 121) (157, 157)	(75, 78) ((106, 117) ((128, 128) ((161, 162)	(92, 131) (177, 177)	(87, 88) (104, 127) (171, 172)	(81, 81) (105, 124) (138, 142) (170, 170)
WOA	距离测度	4 846.78	4 936.64	4 944.23	6 724.17	6 810.15	6 801.26	4 201.23	4 264.65	4 228.01
	PSNR/dB	12.85	15.15	16.13	7.60	14.85	15.00	9.11	13.72	16.24
	最佳阈值	(85, 157) (192, 193)	(75, 76) (130, 142) (195, 195)	(65, 65) (118, 121) (145, 157) (192, 192)	(75, 155) (255, 255)	(77, 93) (104, 129) (164, 164)	(75, 83) (108, 116) (142, 151) (255, 255)	(76, 76) (97, 158)	(120, 44) (91, 136) (182, 182)	(13, 42) (73, 73) (104, 122) (136, 157)
JADE-GL	距离测度	4 863.53	4 977.70	5 001.16	6 771.43	6 825.41	6 834.79	4 245.88	4 325.61	4 385.06
	PSNR/dB	14.02	16.23	17.90	13.51	15.19	16.15	14.22	17.03	19.49
	最佳阈值	(87, 149) (194, 194)	(79, 79) (114, 149) (194, 194)	(78, 78) (114, 136) (155, 155) (194, 194)	(74, 125) (160, 161)	(76, 78) (108, 129) (170, 170)	(74, 78) (106, 117) (129, 129) (172, 172)	(93, 133) (177, 177)	(84, 84) (106, 131) (166, 166)	(81, 81) (106, 121) (144, 144) (185, 185)

表2

图6

参考文献 26

1	WIHARTO , ESTI S , MURDOKO S . The hybrid method of SOM artificial neural network and median thresholding for segmentation of blood vessels in the retina image fundus[J]. International Journal of Fuzzy Logic and Intelligent Systems, 2019, 19 (4): 323- 331. doi: 10.5391/IJFIS.2019.19.4.323
2	GAO B , LI X Q , WOO W L , et al. Physics-based image segmentation using first order statistical properties and genetic algorithm for inductive thermography imaging[J]. IEEE Trans.on Image Processing, 2018, 27 (5): 2160- 2175. doi: 10.1109/TIP.2017.2783627
3	SINGH N, GOYAL S.Determination and segmentation of brain tumor using threshold segmentation with morphological operations[M].Soft Computing: Theories and Applications.Singapore: Springer, 2018: 715-726.
4	PAUL D, DAW N, ROY N D, et al.An automated dual threshold band-based approach for malaria parasite segmentation from thick blood smear[M].Emerging Technology in Modelling and Graphics.Singapore: Springer, 2020: 485-500.
5	PARE S , BHANDARI A K , KUMAR A , et al. Backtracking search algorithm for color image multilevel thresholding[J]. Signal Image Video Process, 2018, 12 (2): 385- 392. doi: 10.1007/s11760-017-1170-z
6	OTSU N . A threshold selection method from gray-level histograms[J]. IEEE trans.on Systems, Man, and Cybernetics, 1979, 9 (1): 62- 66. doi: 10.1109/TSMC.1979.4310076
7	QIN J , SHEN X J , MEI F , et al. An Otsu multi-thresholds segmentation algorithm based on improved ACO[J]. The Journal of Supercomputing, 2019, 75 (2): 955- 967. doi: 10.1007/s11227-018-2622-0
8	TRUONG M T N , KIM S . Automatic image thresholding using Otsu's method and entropy weighting scheme for surface defect detection[J]. Soft Computing, 2017, 22 (13): 4197- 4203.
9	PUN T . A new method for grey-level picture thresholding using the entropy of the histogram[J]. Signal Processing, 1980, 2 (3): 223- 237. doi: 10.1016/0165-1684(80)90020-1
10	MISHRA S , PANDA M . Bat algorithm for multilevel colour image segmentation using entropy-based thresholding[J]. Arabian Journal for Science & Engineering, 2018, 43 (6): 7285- 7314.
11	刘健庄, 栗文青. 灰度图象的二维Otsu自动阈值分割法[J]. 自动化学报, 1993, 19 (1): 101- 105.
	LIU J Z , LI W Q . Two-dimensional Otsu automatic threshold segmentation method for gray image[J]. Acta Automatica Sinica, 1993, 19 (1): 101- 105.
12	宋文青, 王英华, 卢红喜, 等. 基于幂次变换的SAR图像Otsu分割法[J]. 系统工程与电子技术, 2015, 37 (7): 1504- 1511.
	SONG W Q , WANG Y H , LU H X , et al. Otsu segmentation algorithm for SAR images based on power transformation[J]. Systems Engineering and Electronics, 2015, 37 (7): 1504- 1511.
13	高飞彪, 陈南南. 二维Otsu图像快速分割方法的改进[J]. 黑河学院学报, 2019, 10 (10): 216- 220.
	GAO F B , CHENG N N . Improvement of 2D Otsu image fast segmentation method[J]. Journal of Heihe University, 2019, 10 (10): 216- 220.
14	MAL S, KUMAR A.Heuristic approach for finding threshold value in image segmentation[M].Emerging Technology in Modelling and Graphics.Singapore: Springer, 2020: 45-53.
15	胡加鑫, 贾鹤鸣, 邢致恺. 基于鲸鱼算法的森林火灾图像多阈值分割[J]. 森林工程, 2018, 34 (4): 70- 74, 95.
	HU J X , JIA H M , XING Z K . Multi-threshold segmentation of forest fire image based on whale algorithm[J]. Forest Engineering, 2018, 34 (4): 70- 74, 95.
16	赵汝海, 孙凡, 朱广. 基于模糊OTSU与布谷鸟寻优的火灾图像多阈值分割算法[J]. 安徽建筑大学学报, 2019, 27 (4): 65- 70, 97.
	ZHAO R H , SUN F , ZHU G . Multi-threshold segmentation algorithm for fire images based on fuzzy OTSU and cuckoo search[J]. Journal of Anhui University of Architecture, 2019, 27 (4): 65- 70, 97.
17	SATAPATHY S C , SRI M R N , RAJINIKANTH V , et al. Multi-level image thresholding using Otsu and chaotic bat algorithm[J]. Neural Computing and Applications, 2018, 29 (12): 1285- 1307. doi: 10.1007/s00521-016-2645-5
18	ZHANG J Q , SANDERSON A C . JADE:adaptive differential evolution with optional external archive[J]. IEEE Trans.on Evolutionary Computation, 2009, 13 (5): 945- 958. doi: 10.1109/TEVC.2009.2014613
19	TANABE R, FUKUNAGA A S.Improving the search perfor-mance of SHADE using linear population size reduction[C]//Proc.of the IEEE Congress on Evolutionary Computation, 2014: 1658-1665.
20	罗钧, 杨永松, 侍宝玉. 基于改进的自适应差分演化算法的二维Otsu多阈值图像分割[J]. 电子与信息学报, 2019, 41 (8): 2017- 2024.
	LUO J , YANG Y S , SHI B Y . Two-dimensional Otsu multi-threshold image segmentation based on improved adaptive differential evolution algorithm[J]. Journal of Electronics & Information Technology, 2019, 41 (8): 2017- 2024.
21	STORN R , PRICE K . Differential evolution-a simple and efficient heuristic for global optimization over continuous spaces[J]. Journal of Global Optimization, 1997, 11 (4): 341- 359. doi: 10.1023/A:1008202821328
22	LI H W , LIU J Y , CHEN L , et al. Chaos-enhanced moth-flame optimization algorithm for global optimization[J]. Journal of Systems Engineering and Electronics, 2019, 30 (6): 1144- 1159. doi: 10.21629/JSEE.2019.06.10
23	WANG X Y , LI J , WANG T B , et al. Satellite constellation design with genetic algorithms based on system performance[J]. Journal of Systems Engineering and Electronics, 2016, 27 (2): 105- 111.
24	TRIVEDI A, SANYAL K, VERMA P, et al.A unified differential evolution algorithm for constrained optimization problems[C]//Proc.of the IEEE Congress on Evolutionary Computation, 2017: 1231-1238.
25	LUO J , WANG Q . A method for axis straightness error evaluation based on improved artificial bee colony algorithm[J]. The International Journal of Advanced Manufacturing Technology, 2014, 71 (5/8): 1501- 1509.
26	ELSHAZLY E , ABDELWAHAB S , ABOUZAID R , et al. A secure image steganography algorithm based on least significant bit and integer wavelet transform[J]. Journal of Systems Engineering and Electronics, 2018, 29 (3): 639- 649. doi: 10.21629/JSEE.2018.03.21

编号	特征	函数表达式	上下界	维度
F₁	US	$f_{1}(x)=\sum_{i=1}^{n} x_{i}^{2}$	[-100, 100]	30
F₂	UN	$f_{2}(x)=\sum\limits_{i=1}^{n-1}\left[100\left(x_{i+1}-x_{i}^{2}\right)+\left(x_{i}-1\right)^{2}\right]$	[-30, 30]	30
F₃	MS	$\begin{array}{*{20}{c}}{f_{3}(x)=0.1\left\{\sin ^{2}\left(3 \pi x_{1}\right)+\sum\limits_{i=1}^{n}\left(x_{i}-1\right)^{2}\left[1+\sin ^{2}\left(3 \pi x_{i}+1\right)\right]+\right.}\\{\left.\left(x_{n}-1\right)^{2}\left[1+\sin ^{2}\left(2 \pi x_{n}\right)\right]\right\}+\sum\limits_{i=1}^{n} u\left(x_{i}, 5, 100, 4\right)}\end{array}$	[-50, 50]	30
F₄	MN	$f_{4}(x)=-\sum\limits_{i=1}^{n} \sin x_{i} \sin ^{20} \frac{i x_{i}^{2}}{\pi}$	[0, π]	10

[1]	吴涛, 王伦文, 朱敬成. 基于迁移学习和注意力机制的伪装图像分割[J]. 系统工程与电子技术, 2022, 44(2): 376-384.
[2]	吕鑫, 慕晓冬, 张钧. 基于改进麻雀搜索算法的多阈值图像分割[J]. 系统工程与电子技术, 2021, 43(2): 318-327.
[3]	韩子硕, 王春平. 基于改进FCM与MRF的SAR图像分割[J]. 系统工程与电子技术, 2019, 41(8): 1726-1734.
[4]	彭书娟, 曲长文, 李健伟, 邵嘉琦, 骆卉子. 基于ROEWA算子局部活动轮廓的SAR图像分割算法[J]. 系统工程与电子技术, 2019, 41(2): 280-290.
[5]	刘松涛, 刘振兴, 姜康辉. 基于模糊Renyi熵和区域增长的图像目标分割方法[J]. 系统工程与电子技术, 2018, 40(8): 1693-1701.
[6]	蔡青, 刘慧英, 孙景峰, 周三平, 李靖. 基于自适应分割与偏移场估计的活动轮廓模型[J]. 系统工程与电子技术, 2018, 40(5): 1148-1154.
[7]	罗忠涛, 吴太锋, 何子述, 陈绪元. 基于图像分割的高频雷达射频干扰提取算法[J]. 系统工程与电子技术, 2018, 40(4): 776-781.
[8]	雷俊, 王立辉, 何芸倩, 张智. 适用于机器人视觉的图像分割方法[J]. 系统工程与电子技术, 2017, 39(7): 1653-1659.
[9]	刘艳玲, 胡绍海. 运动模糊图像检测分割盲复原[J]. 系统工程与电子技术, 2017, 39(3): 662-667.
[10]	岳温川, 王卫卫, 李小平. 基于加权稀疏子空间聚类多特征融合图像分割[J]. 系统工程与电子技术, 2016, 38(9): 2184-2191.
[11]	杜琨, 王威, 年丰, 陈炜, 胡锋杰. 主动毫米波图像的人体携带危险物检测研究[J]. 系统工程与电子技术, 2016, 38(6): 1462-1469.
[12]	李守荣, 周秋, 周三平, 郝建红. 基于局部与全局信息的活动轮廓图像分割模型[J]. 系统工程与电子技术, 2016, 38(5): 1189-.
[13]	毕凯, 王晓丹, 邢雅琼. 基于改进BPSO的聚类选择性集成[J]. 系统工程与电子技术, 2016, 38(3): 685-691.
[14]	张梦梦1, 张泾周1, 周三平2, 张永涛1. 基于局部熵的边界与区域水平集图像分割模型[J]. 系统工程与电子技术, 2016, 38(12): 2884-2888.
[15]	涂松, 李禹, 粟毅. 基于主动轮廓模型的SAR图像分割方法综述[J]. 系统工程与电子技术, 2015, 37(8): 1754-1766.