背景约束的红外复杂背景下坦克目标分割方法

高敏; 李怀胜; 周玉龙; 方丹; 张宝全

doi:10.16383/j.aas.2016.c150492

背景约束的红外复杂背景下坦克目标分割方法

doi: 10.16383/j.aas.2016.c150492

高敏^1,,
李怀胜^2,,
周玉龙^1,3, ,,
方丹^1,,
张宝全^4,

1.
军械工程学院导弹工程系石家庄 050003
2.
66393部队装备部保定 071002
3.
66393部队博士后科研工作站保定 071002
4.
国网河北省电力公司检修分公司石家庄 050071

基金项目:

中国博士后科学基金 2014M562657

详细信息

作者简介:
高敏   军械工程学院导弹工程系教授, 66393部队博士后科研工作站合作导师.1992年获北京理工大学电子工程专业博士学位.主要研究方向为精确制导技术.E-mail:Gaomin@126.com

李怀胜   66393部队博士后科研工作站合作导师.主要研究方向为装备维修保障.E-mail:lihuaisheng11@126.com

方丹   军械工程学院导弹工程系讲师, 2013年获军械工程学院导弹工程系博士学位.主要研究方向为精确制导技术.E-mail:aq_fd@126.com

张宝全   国网河北省电力公司检修分公司工程师, 2008年获华北电力大学计算机科学与技术本科学历.主要研究方向为红外测温、红外成像检测.E-mail:18031168801@163.com

通讯作者:
周玉龙 66393部队博士后科研工作站在站博士后.2012年获军械工程学院光学与电子工程系博士学位.主要研究方向为红外成像制导技术中的目标分割、识别及跟踪.本文通信作者.E-mail:zyljq@126.com

计量
- 文章访问数: 2586
- HTML全文浏览量: 199
- PDF下载量: 763
- 被引次数: 0
出版历程
- 收稿日期: 2015-07-30
- 录用日期: 2015-12-03
- 刊出日期: 2016-03-20

Tank Segmentation Under Infrared Complex Background with Background Restriction

GAO Min^1
,,
LI Huai-Sheng^2
,,
ZHOU Yu-Long^{1,3
, ,},
FANG Dan^1
,,
ZHANG Bao-Quan^4
,

1.
Missile Engineering Department of Ordnance Engineering College, Shijiazhuang 050003
2.
66393 Troops Equipment Department, Baoding 071002
3.
66393 Troops Postdoctoral Science Research Workstation, Baoding 071002
4.
State Grid Hebei Electronic Power Company Maintenance Branch, Shijiazhuang 050071

Funds:

Postdoctoral Science Foundation of China 2014M562657

More Information

Author Bio:
  Professor at the Missile Engineering Department, Ordnance Engineering College. Postdoctoral tutor at the 66393 Troops Postdoctoral Science Research Workstation. He received his Ph. D. degree in electronic engineering from Beijing Institute of Technology in 1992. His main research interest is precise guidance technology.E-mail:

  Postdoctoral tutor at the 66393 Troops Postdoctoral Science Research Workstation. His main research interest is equipment maintenance.E-mail:

  Lecturer at the Missile Engineering Department, Ordnance Engineering College. He received his Ph. D. degree at the Missile Engineering Department, Ordnance Engineering College in 2013. His main research interest is precise guidance technology.E-mail:

   Engineer at the State Grid Hebei Electric Power Company. He received his bachelor degree in computer science and technology from North China Electric Power University in 2008. His research interest covers infrared temperature measurement, infrared imaging detection.E-mail:

Corresponding author: ZHOU Yu-Long Postdoctor at the 66393 Troops Postdoctoral Science and Research Workstation. He received his Ph. D. degree in optical engineering from Ordnance Engineering College in 2012. His research interest covers infrared target segmentation, recognition and tracking. Corresponding author of this paper.E-mail:zyljq@126.com

摘要

摘要: 为实现自寻的反坦克导弹红外导引头对复杂背景下坦克目标的快速有效分割, 以红外导引头拍摄的坦克目标红外图像为研究对象, 采用计算简单的最大类间方差法, 对其分割效果进行了研究.根据实验结果, 揭示了最大类间方差法进行图像分割的有效性机理.在此基础上, 提出了对背景区域像素和灰度级别进行约束的思想, 以降低背景区域类内方差, 提高算法的分割精度, 并给出了具体的方法.首先利用坦克目标的先验信息, 根据光学成像原理, 推导了红外坦克目标图像的大小估计公式, 用来实现对背景像素的约束; 然后采用黄金分割法对背景灰度级别进行约束; 最后利用最大类间方差法实现了复杂背景下红外坦克目标的分割.实验表明, 本文方法的分割效果堪比手工分割效果, 且计算量较少, 算法耗时最大不超过1.44 ms, 完全满足对坦克目标图像分割的有效性和实时性需求.
- 自寻的反坦克导弹红外成像制导系统 /
- 坦克目标分割 /
- 最大类间方差法 /
- 阈值分析 /
- 背景约束
Abstract: To achieve fast and efficient tank segmentation under infrared complex background for infrared imaging guidance system of homing antitank missile, the maximum between-class variance method is studied using tank infrared images from the infrared imaging and guidance system, and the working mechanism of the maximum between-class variance method is revealed. Then the idea to restrict the background pixels and gray levels is proposed to lower the background within-class variance so as to improve the segmentation accuracy of the method. The detailed approach is as follows. Firstly, with the prior information of the tank, according to the optical imaging principle, the size equation of the tank image is deduced, by which background pixels are restricted; secondly, the golden section is used to restrict the background gray levels; lastly, with the maximum between-class variance method, segmentation of the tank target is achieved. Experiments show that the segmentation result of the proposed method is as good as manual segmentation's and the processing time is decreased to no more than 1.44 ms. So it can completely meet the needs of effectiveness and real-time processing for the tank segmentation under the infrared complex background of the homing antitank missile.
- Infrared imaging guidance system of the homing antitank missile /
- tank segmentation /
- the maximum between-class variance method /
- threshold analysis /
- background restriction

HTML全文

图 1 不同距离的坦克目标红外图像

Fig. 1 Infrared tank images of different distance

下载: 全尺寸图片幻灯片

图 2 灰度直方图

Fig. 2 Gray histograms

下载: 全尺寸图片幻灯片

图 3 Otsu法二值化分割图

Fig. 3 Two-valued results by Otsu method

下载: 全尺寸图片幻灯片

图 4 类内方差随灰度级的变化曲线

Fig. 4 The within-class variance curves changed with gray levels

下载: 全尺寸图片幻灯片

图 5 距离为1 455米处的白天坦克红外图像分割结果

Fig. 5 Segmentation results of tank infrared image taken in daytime with distance of 1 455 m

下载: 全尺寸图片幻灯片

图 6 距离为1 199米处的白天坦克红外图像分割结果

Fig. 6 Segmentation results of tank infrared image taken in daytime with distance of 1 199 m

下载: 全尺寸图片幻灯片

图 7 距离为936米处的白天坦克红外图像分割结果

Fig. 7 Segmentation results of tank infrared image taken in daytime with distance of 936 m

下载: 全尺寸图片幻灯片

图 8 距离为573米处的白天坦克红外图像分割结果

Fig. 8 Segmentation results of tank infrared image taken in daytime with distance of 573 m

下载: 全尺寸图片幻灯片

图 9 距离为446米处的白天坦克红外图像分割结果

Fig. 9 Segmentation results of tank infrared image taken in daytime with distance of 446 m

下载: 全尺寸图片幻灯片

图 10 距离为1 611米处夜间坦克红外图像分割结果

Fig. 10 Segmentation results of tank infrared image taken at night with distance of 1 611 m

下载: 全尺寸图片幻灯片

图 11 距离为1 251米处夜间坦克红外图像分割结果

Fig. 11 Segmentation results of tank infrared image taken at night with distance of 1 251 m

下载: 全尺寸图片幻灯片

图 12 类距离为980米处夜间坦克红外图像分割结果

Fig. 12 Segmentation results of tank infrared image taken at night with distance of 980 m

下载: 全尺寸图片幻灯片

图 13 距离为740米处夜间坦克红外图像分割结果

Fig. 13 egmentation results of tank infrared image taken at night with distance of 740 m

下载: 全尺寸图片幻灯片

图 14 距离为353米处夜间坦克红外图像分割结果

Fig. 14 Segmentation results of tank infrared image taken at night with distance of 353 m

下载: 全尺寸图片幻灯片

表 1 坦克目标距离及所占最大像素数

Table 1 The target distance and its maximum pixels

	Daytime images					Night images
	Fig. 5(a)	Fig. 6(a)	Fig. 7(a)	Fig. 8(a)	Fig. 9(a)	Fig. 10(a)	Fig. 11(a)	Fig. 12(a)	Fig. 13(a)	Fig. 14(a)
Target distance (m)	1455	1199	936	573	446	1 611	1 251	980	740	353
Target pixels	526	775	1 271	3 390	5 601	429	712	1 159	2 032	8 924

下载: 导出CSV

表 2 算法分割精度对比(%)

Table 2 The segmentation accuracy comparison of different methods (%)

	Images	Standard Otsu method	2-D maximum entropy	KFCM	Proposed method
Daytime	Fig. 5(a)	0.25	76.57	85.93	96.01
	Fig. 6(a)	0.26	0.28	98.19	91.54
	Fig. 7(a)	0.90	0.91	97.47	99.69
	Fig. 8(a)	62.08	98.08	94.09	97.08
	Fig. 9(a)	78.68	96.17	89.40	98.51
Night	Fig. 10(a)	0.09	0.09	95.26	99.57
	Fig. 11(a)	71.84	73.26	66.81	86.23
	Fig. 12(a)	0.59	79.35	0.81	99.91
	Fig. 13(a)	0.68	0.67	0.76	99.89
	Fig. 14(a)	7.10	54.93	99.92	96.98
	Average	22.25	48.03	72.86	96.54

下载: 导出CSV

表 3 算法耗时对比

Table 3 The consuming time comparison of different methods

	Images	Standard Otsu method (ms)	2-D maximum entropy (ms)	KFCM (ms)	Proposed method (ms)
Daytime	Fig. 5(a)	1.42	4 260.79	27597.60	1.40
	Fig. 6(a)	1.32	4 247.61	21 529.17	1.27
	Fig. 7(a)	1.55	4 237.53	41 853.46	1.41
	Fig. 8(a)	1.47	4 031.80	54170.71	1.39
	Fig. 9(a)	1.36	4 158.24	52 666.22	1.29
Night	Fig. 10(a)	1.44	4 375.39	21 598.83	1.40
	Fig. 11(a)	1.64	4 521.92	26614.01	1.40
	Fig. 12(a)	1.43	4153.51	27991.33	1.34
	Fig. 13(a)	1.52	3 839.10	43 789.32	1.44
	Fig. 14(a)	1.40	4 671.52	45 569.53	1.32
	Average	1.45	4 249.74	36 338.02	1.37

下载: 导出CSV

参考文献(31)

[1]	Gao C, Zhou D G, Guo Y C. Automatic iterative algorithm for image segmentation using a modified pulse-coupled neural network. Neurocomputing, 2013, 119:332-338 doi: 10.1016/j.neucom.2013.03.025
[2]	Wei S, Hong Q, Hou M S. Automatic image segmentation based on PCNN with adaptive threshold time constant. Neurocomputing, 2011, 74(9):1485-1491 doi: 10.1016/j.neucom.2011.01.005
[3]	汪海洋, 潘德炉, 夏德深.二维Otsu自适应阈值选取算法的快速实现.自动化学报, 2007, 33(9):968-972 doi: 10.1360/aas-007-0968 Wang Hai-Yang, Pan De-Lu, Xia De-Shen. A fast algorithm for two-dimensional Otsu adaptive threshold algorithm. Acta Automatica Sinica, 2007, 33(9):968-972 doi: 10.1360/aas-007-0968
[4]	Kim B G, Park D J. Novel target segmentation and tracking based on fuzzy membership distribution for vision-based target tracking system. Image and Vision Computing, 2006, 24(12):1319-1331 doi: 10.1016/j.imavis.2006.04.008
[5]	范朝冬, 张英杰, 欧阳红林, 肖乐意.基于改进斜分Otsu法的回转窑火焰图像分割.自动化学报, 2014, 40(11):2480-2489 http://www.aas.net.cn/CN/Y2014/V40/I11/2480 Fan Chao-Dong, Zhang Ying-Jie, Ouyang Hong-Lin, Xiao Le-Yi. Improved Otsu method based on histogram oblique segmentation for segmentation of rotary kiln flame image. Acta Automatica Sinica, 2014, 40(11):2480-2489 http://www.aas.net.cn/CN/Y2014/V40/I11/2480
[6]	聂守平, 王鸣, 刘峰.低对比度图像分割算法研究.中国激光, 2004, 31(1):89-91 http://www.cnki.com.cn/Article/CJFDTOTAL-JJZZ200401026.htm Nie Shou-Ping, Wang Ming, Liu Feng. Image segmentation algorithm study for low contrast image. Chinese Journal of Lasers, 2004, 31(1):89-91 http://www.cnki.com.cn/Article/CJFDTOTAL-JJZZ200401026.htm
[7]	郭海涛, 田坦, 王连玉, 张春田.利用二维属性直方图的最大熵的图像分割方法.光学学报, 2006, 26(4):506-509 http://www.cnki.com.cn/Article/CJFDTOTAL-GXXB200604005.htm Guo Hai-Tao, Tian Tan, Wang Lian-Yu, Zhang Chun-Tian. Image segmentation using the maximum entropy of the two-dimensional bound histogram. Acta Optica Sinica, 2006, 26(4):506-509 http://www.cnki.com.cn/Article/CJFDTOTAL-GXXB200604005.htm
[8]	龙建武, 申铉京, 陈海鹏.自适应最小误差阈值分割算法.自动化学报, 2012, 38(7):1134-1144 doi: 10.3724/SP.J.1004.2012.01134 Long Jian-Wu, Shen Xuan-Jing, Chen Hai-Peng. Adaptive minimum error thresholding algorithm. Acta Automatica Sinica, 2012, 38(7):1134-1144 doi: 10.3724/SP.J.1004.2012.01134
[9]	刘金, 金炜东. 3维自适应最小误差阈值分割法.中国图象图形学报, 2013, 18(11):1416-1424 http://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201311004.htm Liu Jin, Jin Wei-Dong. Three-dimensional adaptive minimum error thresholding segmentation algorithm. Journal of Image and Graphics, 2013, 18(11):1416-1424 http://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201311004.htm
[10]	范九伦, 雷博.二维直线型最小误差阈值分割法.电子与信息学报, 2009, 31(8):1801-1806 http://www.cnki.com.cn/Article/CJFDTOTAL-DZYX200908006.htm Fan Jiu-Lun, Lei Bo. Two-dimensional linear-type minimum error threshold segmentation method. Journal of Electronics and Information Technology, 2009, 31(8):1801-1806 http://www.cnki.com.cn/Article/CJFDTOTAL-DZYX200908006.htm
[11]	吴一全, 占必超, 吴加明.基于类内绝对差和混沌粒子群的红外图像分割.光学学报, 2010, 30(1):79-85 doi: 10.3788/AOS Wu Yi-Quan, Zhan Bi-Chao, Wu Jia-Ming. An infrared image segmentation method based on within-class absolute difference and chaotic particle swarm optimization. Acta Optica Sinica, 2010, 30(1):79-85 doi: 10.3788/AOS
[12]	李昱川, 田铮.基于图的加权核K均值的图像多尺度分割.光学学报, 2009, 29(10):2762-2767 doi: 10.3788/AOS Li Yu-Chuan, Tian Zheng. Multiscale image segmentation based on graph weighted kernel K-means. Acta Optica Sinica, 2009, 29(10):2762-2767 doi: 10.3788/AOS
[13]	乔立永, 徐立新, 高敏.红外成像制导二维斜分最大熵分割的快速实现.红外与激光工程, 2013, 42(7):1691-1699 http://www.cnki.com.cn/Article/CJFDTOTAL-HWYJ201307009.htm Qiao Li-Yong, Xu Li-Xin, Gao Min. Fast maximum entropy thresholding based on two-dimensional histogram oblique segmentation in infrared imaging guidance. Infrared and Laser Engineering, 2013, 42(7):1691-1699 http://www.cnki.com.cn/Article/CJFDTOTAL-HWYJ201307009.htm
[14]	胡敏, 李梅, 汪荣贵.改进的Otsu算法在图像分割中的应用.电子测量与仪器学报, 2010, 24(5):443-449 doi: 10.3724/SP.J.1187.2010.00443 Hu Min, Li Mei, Wang Rong-Gui. Application of an improved Otsu algorithm in image segmentation. Journal of Electronic Measurement and Instrument, 2010, 24(5):443-449 doi: 10.3724/SP.J.1187.2010.00443
[15]	张天序, 赵广州, 王飞, 朱光喜.一种快速递归红外舰船图像分割新算法.红外与毫米波学报, 2006, 25(4):225-230 http://www.cnki.com.cn/Article/CJFDTOTAL-HWYH200604012.htm Zhang Tian-Xu, Zhao Guang-Zhou, Wang Fei, Zhu Guang-Xi. Fast recursive algorithm for infrared ship image segmentation. Journal of Infrared and Millimeter Waves, 2006, 25(4):225-230 http://www.cnki.com.cn/Article/CJFDTOTAL-HWYH200604012.htm
[16]	Otsu N. A threshold selection method from gray-level histograms. IEEE Transactions on System Man and Cybernetic, 1979, 9(1):62-66 doi: 10.1109/TSMC.1979.4310076
[17]	许向阳, 宋恩民, 金良海. Otsu准则的阈值性质分析.电子学报, 2009, 37(12):2716-2719 http://www.cnki.com.cn/Article/CJFDTOTAL-DZXU200912019.htm Xu Xiang-Yang, Song En-Min, Jin Liang-Hai. Characteristic analysis of threshold based on Otsu criterion. Acta Electronica Sinica, 2009, 37(12):2716-2719 http://www.cnki.com.cn/Article/CJFDTOTAL-DZXU200912019.htm
[18]	华罗庚.优选法.北京:科学出版社, 1981. Hua Luo-Geng. Optimum Selection Method. Beijing:Science Press, 1981.
[19]	解永春, 吴宏鑫.黄金分割在自适应鲁棒控制器设计中的应用.自动化学报, 1992, 18(2):177-185 http://www.aas.net.cn/CN/Y1992/V18/I02/177 Xie Yong-Chun, Wu Hong-Xin. The application of the golden section in adaptive robust controller design. Acta Automatica Sinica, 1992, 18(2):177-185 http://www.aas.net.cn/CN/Y1992/V18/I02/177
[20]	Sen S K, Ravi P A. Golden ratio in science, as random sequence source, its computation and beyond. Computers and Mathematics with Applications, 2008, 56(2):469-468 doi: 10.1016/j.camwa.2007.06.030
[21]	Lu Y. A golden section approach to optimization of automotive friction materials. Journal of Materials Science, 2003, 38(5):1081-1085 doi: 10.1023/A:1022362217043
[22]	Cai J M, Han D, Chen C X, Chen S Y. Application of the golden section search algorithm in the nonlinear isoconversional calculations to the determination of the activation energy from nonisothermal kinetic conversion data. Solid State Sciences, 2010, 12(5):829-833 doi: 10.1016/j.solidstatesciences.2010.02.010
[23]	Tsai C H, Kolibal J, Li M. The golden section search algorithm for finding a good shape parameter for meshless collocation methods. Engineering Analysis with Boundary Elements, 2010, 34(8):738-746 doi: 10.1016/j.enganabound.2010.03.003
[24]	Stakhov A P. The Generalized Principle of the Golden Section and its applications in mathematics, science, and engineering. Chaos, Solitons and Fractals, 2005, 26(2):263-289 doi: 10.1016/j.chaos.2005.01.038
[25]	周祥, 何小荣, 陈丙珍.基于最优变异因子的遗传算法在ANN训练中的应用.清华大学学报(自然科学版), 2002, 42(5):619-621 http://www.cnki.com.cn/Article/CJFDTOTAL-QHXB200205013.htm Zhou Xiang, He Xiao-Rong, Chen Bing-Zhen. Best mutation coefficient genetic algorithm for ANN training. Journal of Tsinghua University (Science and Technology), 2002, 42(5):619-621 http://www.cnki.com.cn/Article/CJFDTOTAL-QHXB200205013.htm
[26]	李道平, 姚小兰, 伍清河, 曲亦直.自适应小波去噪算法及其在偏心补偿中的应用.北京理工大学学报, 2010, 30(2):179-182, 196 http://www.cnki.com.cn/Article/CJFDTOTAL-BJLG201002014.htm Li Dao-Ping, Yao Xiao-Lan, Wu Qing-He, Qu Yi-Zhi. Self-adaptive wavelet threshold de-noising method and its application in the compensation of roll eccentricity signal. Transactions of Beijing Institute of Technology, 2010, 30(2):179-182, 196 http://www.cnki.com.cn/Article/CJFDTOTAL-BJLG201002014.htm
[27]	魏雪峰, 刘晓.基于2维最大熵最佳阈值算法的图像分割研究.激光技术, 2013, 37(4):519-522 http://www.cnki.com.cn/Article/CJFDTOTAL-JGJS201304024.htm Wei Xue-Feng, Liu Xiao. Research of image segmentation based on 2-D maximum entropy optimal threshold. Laser Technology, 2013, 37(4):519-522 http://www.cnki.com.cn/Article/CJFDTOTAL-JGJS201304024.htm
[28]	洪霞, 周牧, 田增山, 董会宁.基于二维最大熵阈值分割的SIFT图像匹配算法.半导体光电, 2013, 34(4):689-693, 705 http://www.cnki.com.cn/Article/CJFDTOTAL-BDTG201304036.htm Hong Xia, Zhou Mu, Tian Zeng-Shan, Dong Hui-Ning. SIFT image matching algorithm based on two-dimensional maximum entropy-aided threshold segmentation. Semiconductor Optoelectronics, 2013, 34(4):689-693, 705 http://www.cnki.com.cn/Article/CJFDTOTAL-BDTG201304036.htm
[29]	康家银, 纪志成, 龚成龙.一种核模糊C均值聚类算法及其应用.仪器仪表学报, 2010, 31(7):1657-1663 http://www.cnki.com.cn/Article/CJFDTOTAL-YQXB201007038.htm Kang Jia-Yin, Ji Zhi-Cheng, Gong Cheng-Long. Kernelized fuzzy C-menas clustering algorithm and its application. Chinese Journal of Scientific Instrument, 2010, 31(7):1657-1663 http://www.cnki.com.cn/Article/CJFDTOTAL-YQXB201007038.htm
[30]	Ji Z X, Xia Y, Chen Q, Sun Q S, Xia D S, Feng D D. Fuzzy c-means clustering with weighted image patch for image segmentation. Applied Soft Computing, 2012, 12(6):1659-1667 doi: 10.1016/j.asoc.2012.02.010
[31]	Wang H, Zhang H, Ray N. Adaptive shape prior in graph cut image segmentation. Pattern Recognition, 2013, 46(5):1409-1414 doi: 10.1016/j.patcog.2012.11.002