基于改进的高斯混合模型和图割模型的水体图像提取算法

doi:10.7523/j.ucas.2023.028

摘要/Abstract

摘要： 合成孔径雷达具有全天时、全天候成像，观测范围广、成图周期短等特点，使其在水体提取中的应用具有明显优势。但是在湖泊提取中，已有算法易受湖泊周围环境噪声的干扰，运行效率低。针对此，提出一种将改进的高斯混合模型(GMM)与图割模型相结合的检测方法。先是利用两级Otsu阈值方法以获取湖泊的初始分割图，将对其计算得到的参数集作为GMM初始参数，其后应用最大期望算法(EM)迭代求取GMM的最佳参数，最后基于图割模型精准定位水体位置。实验结果表明初始参数越精确，水体轮廓越清晰，引入两级Otsu算法不仅可大幅度减少EM算法的迭代次数，且与预处理中的下采样相结合有效地提高了算法的运行速度。此外，改进传统图割模型的能量函数使得无需后处理即可得到精确的湖泊边界。

关键词: 合成孔径雷达, 高斯混合模型, 图割模型, 二级Otsu阈值

Abstract: Synthetic aperture radar (SAR) has the characteristics of all-day and all-weather imaging, wide observation range, and short mapping period, which make it highly advantageous in water extraction. However, existing algorithms for lake extraction are easily affected by the surrounding environment of lakes and noise interference, resulting in low operational efficiency. Therefore, this paper proposes a detection method that combines an improved Gaussian mixture model (GMM) with graph cut model (GCM). First, the two-level Otsu threshold method is used to obtain the initial segmentation map of the lake, and the calculated parameter set is used as the initial parameter of the GMM. The expectation maximum algorithm (EM) is employed to obtain the optimal parameters of the GMM iteratively. The experimental results demonstrate that the more accurate the initial parameters, the clearer the outline of the water body. The introduction of the two-level Otsu algorithm not only greatly reduces the times of iterations of the EM algorithm, but also effectively enhances the running speed of the algorithm in combination with downsampling in preprocessing. In addition, the energy function of the graph cut model enables accurate lake boundaries to be obtained without requiring any post-processing.

Key words: synthetic aperture radar, Gaussian mixture model, graph cut model, two-level Otsu threshold

中图分类号:

TP753

包立男, 吕孝雷. 基于改进的高斯混合模型和图割模型的水体图像提取算法[J]. 中国科学院大学学报, 2024, 41(6): 794-802.

BAO Linan, LYU Xiaolei. Water image extraction algorithm based on improved Gaussian mixture model and graph cut model[J]. Journal of University of Chinese Academy of Sciences, 2024, 41(6): 794-802.

参考文献

[1] Ian G Cumming，Frank H Wong.合成孔径雷达成像:算法与实现[M]. 洪文，胡东辉, 等译. 北京：电子工业出版社，2007.
[2] 邵媛媛, 周军伟, 母锐敏, 等. 中国城市发展与湿地保护研究[J]. 生态环境学报, 2018, 27(2): 381-388. DOI：10.16258/j.cnki.1674-5906.2018.02.024.
[3] Li L I, Zhao J, Xue X, et al. Comprehensive improvement and water quality simulation of Nanming River in Guizhou Province[J]. Acta Scientiae Circumstantiae, 2018. DOI: 10.13671/j.hjkxxb.2018.0029.
[4] Aristizabal F, Judge J. Mapping fluvial inundation extents with graph signal filtering of river depths determined from unsupervised clustering of synthetic aperture radar imagery[C]//2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS. July 11-16, 2021, Brussels, Belgium. IEEE, 2021: 6124-6127. DOI: 10.1109/IGARSS47720.2021.9553575.
[5] Gasnier N, Denis L, Fjørtoft R, et al. Narrow River extraction from SAR images using exogenous information[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 5720-5734. DOI: 10.1109/JSTARS.2021.3083413.
[6] 方贺, 蔡菊珍, 何月, 等.基于RADARSAT-2四极化SAR影像的海面风速反演[J].海洋预报, 2021, 38(2):42-54. DOI：10.11737/j.issn.1003-0239.2021.02.005.
[7] 张帆, 陆圣涛, 项德良, 等.一种改进的高分辨率SAR图像超像素CFAR舰船检测算法[J].雷达学报, 2022, 11:1-20. DOI：10.12000/JR22067.
[8] 李发森, 李显巨, 陈伟涛, 等.基于深度特征的双极化SAR遥感图像岩性自动分类[J].地球科学, 2022, 47(11): 4267-4279.DOI: 10.3799/dqkx.2022.129.
[9] Otsu N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62-66. DOI: 10.1109/TSMC.1979.4310076.
[10] 韩思奇，王蕾. 图像分割的阈值法综述[J]. 系统工程与电子技术, 2002, 24(6):91-94, 102.DOI：10.3321/j.issn:1001-506X. 2002.06.027.
[11] 田佳宁, 张果荣, 宋宝.基于阈值分割算法的湖泊边缘特征提取研究[J].现代信息科技, 2021, 5(4):67-71, 76. DOI：10.19850/j.cnki.2096-4706.2021.04.017.
[12] 韩纪普, 段先华, 常振.基于SLIC和区域生长的目标分割算法[J].计算机工程与应用, 2021, 57(1): 213-218.DOI：10.3778/j.issn.1002-8331.1911-0254.
[13] Guo Z S, Wu L, Huang Y B, et al. Water-body segmentation for SAR images: past, current, and future[J]. Remote Sensing, 2022, 14(7): 1752. DOI: 10.3390/rs14071752.
[14] 冷英, 刘忠玲, 张衡, 等. 一种改进的ACM算法及其在鄱阳湖水域监测中的应用[J]. 电子与信息学报, 2017, 39(5): 1064-1070. DOI：10.11999/JEIT160870.
[15] Xu K J. Expectation-maximization algorithm[M]//Encyclopedia of Systems Biology. New York, NY: Springer New York, 2013: 699-699. DOI: 10.1007/978-1-4419-9863-7_449.
[16] Xiang D L, Ban Y F, Wang W, et al. Edge detector for polarimetric SAR images using SIRV model and Gauss-shaped filter[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(11): 1661-1665. DOI: 10.1109/LGRS.2016.2600704.
[17] Liang J Y, Liu D S. A local thresholding approach to flood water delineation using Sentinel-1 SAR imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 159:53-62. DOI：10.1016/j.isprsjprs.2019.10.017.
[18] Awange J L, Palancz B, Lewis R, et al. An algebraic solution of maximum likelihood function in case of Gaussian mixture distribution[J]. Australian Journal of Earth Sciences, 2016, 63(2): 193-203. DOI:10.1080/08120099.2016.1143876.
[19] Xiao P F, Yuan M, Zhang X L, et al. Cosegmentation for object-based building change detection from high-resolution remotely sensed images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(3): 1587-1603. DOI: 10.1109/TGRS.2016.2627638.
[20] Boykov Y, Veksler O. Graph cuts in vision and graphics: theories and applications[M]//Handbook of Mathematical Models in Computer Vision. New York: Springer-Verlag, 2005: 79-96. DOI: 10.1007/0-387-28831-7_5.
[21] Boykov Y, Kolmogorov V. An experimental comparison of Min-cut/max-flow algorithms for energy minimization in vision[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(9): 1124-1137. DOI: 10.1109/TPAMI.2004.60.
[22] Boykov Y Y, Jolly M P. Interactive graph cuts for optimal boundary & region segmentation of objects in N-D images[C]//Proceedings Eighth IEEE International Conference on Computer Vision. ICCV. July 7-14, 2001, Vancouver, BC, Canada. IEEE, 2002: 105-112. DOI:10.1109/ICCV.2001.937505.
[23] Zhang K Y, Fu X K, Lv X L, et al. Unsupervised multitemporal building change detection framework based on cosegmentation using time-series SAR[J]. Remote Sensing, 2021, 13(3): 471. DOI: 10.3390/rs13030471.
[24] 张雯. 太湖湖滨带生态现状及其健康评价[D]. 上海:华东师范大学, 2017.
[25] Li F, Huang J H, Zeng G M, et al. Spatial risk assessment and sources identification of heavy metals in surface sediments from the Dongting Lake, Middle China[J]. Journal of Geochemical Exploration, 2013, 132:75-83. DOI：10.1016/j.gexplo.2013.05.007.
[26] Zhang J Q, Xu K Q, Yang Y H, et al. Measuring water storage fluctuations in Lake Dongting, China, by topex/Poseidon satellite altimetry[J]. Environmental Monitoring and Assessment, 2006, 115(1): 23-37. DOI: 10.1007/s10661-006-5233-9.
[27] Plank S. Rapid damage assessment by means of multi-temporal SAR-a comprehensive review and outlook to Sentinel-1[J]. Remote Sensing, 2014, 6(6): 4870-4906. DOI: 10.3390/rs6064870.
[28] Nagler T, Rott H, Ripper E, et al. Advancements for snowmelt monitoring by means of Sentinel-1 SAR[J]. Remote Sensing, 2016, 8(4):348. DOI：10.3390/rs8040348.
[29] Bouman C A, Shapiro M. A multiscale random field model for Bayesian image segmentation[J]. IEEE Transactions on Image Processing: a Publication of the IEEE Signal Processing Society, 1994, 3(2): 162-177. DOI:10.1109/83.277898.
[30] 杜钰娇，徐嘉，刘盛英杰. 基于改进图割的遥感图像水体自提取算法[J].计算机工程与设计, 2019, 40(5):1413-1417, 1423. DOI:10.16208/j.issn1000-7024.2019. 05.039.
[31] 董忠言, 蒋理兴, 王俊亚, 等.基于图像复杂度的一维Otsu改进算法[J].计算机科学, 2015, 42(S1):171-174. DOI:CNKI:SUN:JSJA.0.2015-S1-041.