2.793

2018影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于局部空间信息的可变类模糊阈值光学遥感图像分割

杨蕴 李玉 赵泉华

杨蕴, 李玉, 赵泉华. 基于局部空间信息的可变类模糊阈值光学遥感图像分割. 自动化学报, 2020, 46(x): 1−12 doi: 10.16383/j.aas.c190412
引用本文: 杨蕴, 李玉, 赵泉华. 基于局部空间信息的可变类模糊阈值光学遥感图像分割. 自动化学报, 2020, 46(x): 1−12 doi: 10.16383/j.aas.c190412
Yang Yun, Li Yu, Zhao Quan-Hua. Fuzzy threshold optical remote sensing image segmentation with variable class number based on local spatial information. Acta Automatica Sinica, 2020, 46(x): 1−12 doi: 10.16383/j.aas.c190412
Citation: Yang Yun, Li Yu, Zhao Quan-Hua. Fuzzy threshold optical remote sensing image segmentation with variable class number based on local spatial information. Acta Automatica Sinica, 2020, 46(x): 1−12 doi: 10.16383/j.aas.c190412

基于局部空间信息的可变类模糊阈值光学遥感图像分割

doi: 10.16383/j.aas.c190412
基金项目: 国家自然科学基金 (41301479, 41271435)资助
详细信息
    作者简介:

    杨蕴:辽宁工程技术大学测绘与地理科学学院博士研究生. 主要研究方向高分辨遥感图像的地物目标几何以及特征提取.E-mail: m13147945981@163.com

    李玉:辽宁工程技术大学测绘与地理科学学院教授. 主要研究方向为遥感数据处理理论与应用基础研究. 本文通信作者.E-mail: liyu@lntu.edu.cn

    赵泉华:辽宁工程技术大学测绘与地理科学学院教授. 主要研究方向为遥感图像建模与分析随机几何在遥感图像处理中的应用.E-mail:zhaoquanhua@lntu.edu.cn

Fuzzy Threshold Optical Remote Sensing Image Segmentation with Variable Class Number Based on Local Spatial Information

Funds: Supported by National Natural Science Foundation of China (41301479, 41271435)
  • 摘要: 阈值法分割在光学遥感图像分析中被得到广泛的应用, 然而传统阈值法也存在诸多局限性, 如对噪声敏感, 需人为设定类别数, 计算复杂度高等. 针对传统阈值法的局限性, 提出一种基于局部空间信息的可变类模糊阈值光学遥感图像分割方法. 首先, 以图像光谱的一阶矩为初始类中心, 利用二分法原理和区域间最大相似度准则来快速确定类别数及其中心. 然后, 通过岭形模糊隶属函数计算各像素点对不同类的隶属程度, 同时考虑到像素点的隶属度局部空间信息, 在隶属度域中定义一个模糊加权滤波器对各类的隶属度矩阵进行滤波, 以滤波后的隶属度集合为依据, 按照最大隶属原则确定图像的标号场. 最后, 对标号场中的局部异常标号进行替换, 将修正后的标号场由对应的类中心赋色得到分割图像. 视觉和统计分析评价结果表明, 与传统阈值法相比, 该方法能在减少计算时间的同时获得更好的分割结果, 可适用于光学遥感图像的多阈值分割.
  • 图  1  多级岭形隶属函数

    Fig.  1  Multilevel ridge membership function

    图  2  算法流程图

    Fig.  2  Algorithm flow chart

    图  3  模拟图像

    Fig.  3  Simulated images

    图  4  模拟图像分割结果

    Fig.  4  Simulated image segmentation results

    图  5  全色遥感图像和分割结果

    Fig.  5  Panchromatic remote sensing images and segmentation results

    图  6  多光谱遥感图像和分割结果

    Fig.  6  Multispectral remote sensing images and segmentation results

    表  1  各同质区域的高斯分布参数

    Table  1  Gaussian distribution parameters of homogeneous regions

    模拟图像 参数
    图3(b1) 均值 70 90 130 180 160
    方差 6 2 7 4 8
    图3(b2) 均值 20/120/40 70/80/200 120/160/80 150/60/160 200/200/110
    方差 5/7/4 7/5/3 4/2/7 3/4/5 5/6/2
    下载: 导出CSV

    表  2  模拟图像分割的定量评价结果

    Table  2  Quantitative evaluation results of simulated image segmentation

    图像 指标 区域 Kmeans FCM 文献[19] 本文方法
    图3(b1) 用户精度/% 69.8 72.1 96.9 99.9
    92.2 90.1 98.6 99.9
    44.7 81.7 98.1 99.9
    90.4 80.4 97.0 99.8
    58.8 69.7 94.5 99.3
    产品精度/% 56.3 66.5 99.6 99.7
    88.5 86.1 78.3 100
    33.9 75.6 98.5 99.5
    86.5 86.5 97.1 100
    63.3 71.7 90.4 99.6
    总精度/% 55.4 75.2 88.6 99.4
    Kappa系数/% 53.9 74.6 85.3 99.7
    图3(b2) 用户精度/% 42.4 96.5 97.5 99.5
    38.9 84.5 70.6 96.0
    63.2 96.1 96.6 99.5
    85.6 59.6 95.9 99.0
    88.4 86.7 88.3 97.2
    产品精度/% 55.3 69.7 90.4 98.7
    55.3 90.5 76.8 99.9
    48.7 81.7 88.4 93.7
    78.3 71.4 69.5 98.5
    90.1 73.8 90.3 99.2
    总精度/% 53.2 81.2 89.7 98.3
    Kappa系数/% 48.9 80.0 88.1 98.6
    下载: 导出CSV

    表  3  全色遥感图像分割质量评价指标

    Table  3  Quality evaluation of panchromatic remote sensing image segmentation

    图像 MV JM
    Kmeans FCM 文献[19] 本文方法 Kmeans FCM 文献[19] 本文方法
    图5(a1) 1.626 1.335 0.973 0.632 0.887 0.742 0.712 0.633
    图5(b1) 2.344 1.698 1.335 0.966 0.831 0.787 0.737 0.596
    图5(c1) 1.886 1.475 1.203 1.079 0.759 0.703 0.663 0.645
    图5(d1) 0.982 0.875 0.619 0.512 0.692 0.640 0.598 0.582
    下载: 导出CSV

    表  4  计算复杂度对比

    Table  4  Computational complexity comparison

    方法 计算复杂度
    Kmeans $O((K+M \times N / K) \times t)$
    FCM $O((M \times N \times K \times t)$
    文献[19] $O\left((1+M \times N) \times K \times t \times \omega^{2}\right)$
    本文方法 $O\left(M \times N \times K+2 \times M \times N \times \omega^{2}\right)$
    下载: 导出CSV

    表  5  全色图像分割时间对比(s)

    Table  5  Panchromatic images segmentation time comparison (s)

    方法 图5(a1) 图5(b1) 图5(c1) 图5(d1)
    Kmeans 1.49 1.62 4.17 2.38
    FCM 2.63 2.87 7.46 4.28
    文献[19] 38.04 41.66 110.31 60.48
    本文方法 1.66 1.80 3.87 2.29
    下载: 导出CSV

    表  6  多光谱遥感图像分割质量评价

    Table  6  Quality evaluation of multispectral remote sensing image segmentation

    指标 方法 图6(a1) 图6(b1) 图6(c1) 图6(d1)
    MV Kmeans 1.971 1.613 2.316 2.146
    FCM 1.813 1.404 1.833 1.799
    文献[19] 1.570 1.071 1.279 1.344
    本文方法 1.376 0.796 0.941 1.001
    JM Kmeans 0.832 0.797 0.774 0.808
    FCM 0.748 0.624 0.647 0.734
    文献[19] 0.662 0.588 0.541 0.631
    本文方法 0.575 0.534 0.532 0.565
    E Kmeans 0.671 0.572 0.607 0.632
    FCM 0.524 0.466 0.573 0.597
    文献[19] 0.456 0.403 0.434 0.463
    本文方法 0.347 0.332 0.293 0.306
    下载: 导出CSV

    表  7  多光谱图像分割计算时间对比(s)

    Table  7  Multispectral images segmentation time comparison (s)

    方法 图6(a1) 图6(b1) 图6(c1) 图6(d1)
    Kmeans 4.15 2.75 3.01 5.39
    FCM 7.33 4.99 5.34 9.54
    文献[19] 106.02 66.65 77.49 138.51
    本文方法 3.56 3.15 3.37 4.68
    下载: 导出CSV
  • [1] Min Wang, Jiru Huang, Dongping Ming. Region-line association constraints for high-resolution image segmentation. IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing, 2016, 10(2): 1−10
    [2] Andrés Troya-Galvis, Gancarski P, Berti-Equille L. Remote sensing image analysis by aggregation of segmentation-classification collaborative agents. Pattern Recognition, 2018, 73: 259−274 doi: 10.1016/j.patcog.2017.08.030
    [3] 李修霞, 荆林海, 李慧, 唐韵玮, 戈文艳. 参考1维光谱差异的区域生长种子点选取方法. 中国图象图形学报, 2016, 21(9): 1256−1264 doi: 10.11834/jig.20160915

    Li Xiu-Xia, Jing Lin-Hai, Li Hui, Tang Yun-Wei, Ge Wen-Yan. Seed extraction method for seeded region growing based on one-dimensional spectral differences. Journal of Image and Graphics, 2016, 21(9): 1256−1264 doi: 10.11834/jig.20160915
    [4] 游江, 唐力伟, 邓士杰, 苏续军. 完全基于边缘信息的目标靶快速分割算法. 激光与红外, 2017, 47(3): 372−378 doi: 10.3969/j.issn.1001-5078.2017.03.023

    You Jiang, Tang Li-Wei, Deng Shi-Jie, Su Xue-jun. Fast target segmentation algorithm fully based on edge information. Laser & Infrared, 2017, 47(3): 372−378 doi: 10.3969/j.issn.1001-5078.2017.03.023
    [5] Funke J, Tschopp F D, Grisaitis W, Sheridan A, Singh C, Saalfeld Se. Large scale image segmentation with structured loss based deep learning for connectome reconstruction. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2018, 41(7): 1669−1680
    [6] 李擎, 唐欢, 迟健男, 邢永跃, 李华通. 基于改进最大类间方差法的手势分割方法研究. 自动化学报, 2017, 43(4): 528−537

    Li Qing, Tang Huan, Chi Jian-Nan, Xing Yong-Yue, Li Hua-Tong. Gesture segmentation with improved maximum between-cluster variance algorithm. Acta Automatica Sinica, 2017, 43(4): 528−537
    [7] Zeggada A, Melgani, Farid, Ba zi, Yakoub. A deep learning approach to UAV image multilabeling. IEEE Geoscience & Remote Sensing Letters, 2017, 14(5): 694−698
    [8] 田娟秀, 刘国才, 谷珊珊, 鞠忠建, 刘劲光, 顾冬冬. 医学图像分析深度学习方法研究与挑战. 自动化学报, 2018, 44(3): 401−424

    Tian Juan-Xiu, Liu Guo-Cai, Gu Shan-Shan, Ju Zhong-Jian, Liu Jin-Guang, GU Dong-Dong. Deep learning in medical image analysis and its challenges. Acta Automatica Sinica, 2018, 44(3): 401−424
    [9] Fredo A R J, Abilash R S, Kumar C S. Segmentation and analysis of damages in composite images using multi-level threshold methods and geometrical features. Measurement, 2017, 100(100): 270−278
    [10] Yamini B, Sabitha R. Image steganalysis: adaptive color image segmentation using Otsu’s method. Journal of Computational & Theoretical Nanoscience, 2017, 14(9): 4502−4507
    [11] 黄扬, 郭立君, 张荣. 融合全局和局部相关熵的图像分割. 中国图象图形学报, 2018, 20(12): 1619−1628

    Huang Yang, Guo Li-Jun, Zhang Rong. Integration of global and local correntropy image segmentation algorithm. Journal of Image and Graphics, 2018, 20(12): 1619−1628
    [12] Choy S K, Shu Y L, Yu K W, Lee W Y, Leung K T. Fuzzy model-based clustering and its application in image segmentation. Pattern Recognition, 2017, 100(68): 141−157
    [13] Kurt B, Nabiyev V V, Turhan K. A novel automatic suspicious mass regions identification using Havrda & Charvat entropy and Otsu's N thresholding. Computer Methods & Programs in Biomedicine, 2014, 114(3): 349−360
    [14] Mousavirad S J, Ebrahimpour-Komleh H. Multilevel image thresholding using entropy of histogram and recently developed population-based metaheuristic algorithms. Evolutionary Intelligence, 2017, 10(1): 45−75
    [15] 产思贤, 周小龙, 张卓, 陈胜勇. 一种基于超像素的肿瘤自动攻击交互式分割算法. 自动化学报, 2017, 43(10): 1829−1840

    Chan Si-Xian, Zhou Xiao-Long, Zhang Zhuo, Chen Sheng-Yong. Interactive multi-label image segmentation with multi-layer tumors automata. Acta Automatica Sinica, 2017, 43(10): 1829−1840
    [16] 姜枫, 顾庆, 郝慧珍, 李娜, 郭延文, 陈道蓄. 基于内容的图像分割方法综述. 软件学报, 2017, 28(1): 160−183

    Jiang Feng, Gu Qing, Hao Hui-Zhen, Li Na, Guo Yan-Wen, Chen Dao-Xu. Survey on content-based image segmentation methods. Journal of Software, 2017, 28(1): 160−183
    [17] 肖满生, 文志诚, 张居武, 汪新凡. 一种改进隶属度函数的FCM聚类算法. 控制与决策, 2015, 30(12): 2270−2274

    Xiao Man-Sheng, Wen Zhi-Cheng, Zhang Ju-Wu, Wan Xin-Fan. An FCM clustering algorithm with improved membership function. Control and Decision, 2015, 30(12): 2270−2274
    [18] Meena Prakash R, Shantha S K R. Fuzzy C means integrated with spatial information and contrast enhancement for segmentation of MR brain images. International Journal of Imaging Systems & Technology, 2016, 26(2): 116−123
    [19] 赵雪梅, 李玉, 赵泉华. 参数自适应的可变类FLICM灰度图像分割算法. 控制与决策, 2017, 32(2): 262−268

    Zhao Xue-Mei, Li Yu, Zhao Quan-Hua. Self-adaptive FLICM algorithm for gray image segmentation with unknown number of clusters. Control and Decision, 2017, 32(2): 262−268
    [20] Ning J, Zhang L, Zhang D, Chengke W. Interactive image segmentation by maximal similarity based region merging. Pattern Recognition, 2010, 43(2): 445−456 doi: 10.1016/j.patcog.2009.03.004
    [21] 杜茂康, 王忠思, 宋强. 基于Bhattacharyya系数的改进相似度度量方法. 重庆邮电大学学报(自然科学版), 2018, 30(05): 115−120

    Du Mao-Kang, Wang Zhong-Si, Song Qiang. Research of improving similarity measure based on Bhattacharyya coefficient. Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition), 2018, 30(05): 115−120
    [22] Choi S H, Jung H Y, Kim H. Ridge fuzzy regression model. International Journal of Fuzzy Systems, 2019, 21(7): 2077−2090 doi: 10.1007/s40815-019-00692-0
    [23] Mansoori G, Zolghadri J, Katebi D. A weighting function for improving fuzzy classification systems performance. Fuzzy Sets & Systems, 2007, 158(5): 583−591
    [24] Wang Y, Qi Q, Liu Y. Unsupervised segmentation evaluation using area-weighted variance and Jeffries-Matusita distance for remote sensing images. Remote Sensing, 2018, 10(8): 1193−1217 doi: 10.3390/rs10081193
    [25] Chen H C, Wang S J. Visible colour difference-based quantitative evaluation of colour segmentation. IEE Proceedings-Vision, Image and Signal Processing, 2006, 153(5): 598−609 doi: 10.1049/ip-vis:20045221
    [26] 李玉, 徐艳, 赵雪梅, 赵泉华. 利用高斯混合模型的多光谱图像模糊聚类分割. 光学精密工程, 2017, 25(2): 509−518 doi: 10.3788/OPE.20172502.0509

    Li Yu, Xu Yan, Zhao Xue-Mei, Zhao Quan-Hua. Multispectral image segmentation by fuzzy clustering algorithm used Gaussian mixture model. Optics and Precision Engineering, 2017, 25(2): 509−518 doi: 10.3788/OPE.20172502.0509
    [27] Yellamraju T, Boutin M. Clusterability and clustering of images and other “real” high-dimensional data. IEEE Transactions on Image Processing, 2018, 27(4): 1927−1938 doi: 10.1109/TIP.2017.2789327
  • 加载中
计量
  • 文章访问数:  125
  • HTML全文浏览量:  38
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-05-27
  • 录用日期:  2019-12-02

目录

    /

    返回文章
    返回