摘要
本文提出将聚类算法引入到ERT监测系统中,采用K均值(K-means)聚类、模糊C均值算法(FCM)以及混合高斯模型(GMM)3种常用聚类算法对ERT检测结果进行污染区域识别,通过一个数值模型分析了3种算法的识别效果.研究结果表明当污染区域与背景土壤的电阻率区分度较大时(电阻率差异性大于30%),采用3种聚类算法都可以识别出污染区域,K-means和FCM的识别效果优于GMM算法.最后,给出一个实际场地调查的应用案例.
Electrical resistivity tomography(ERT) has been used for pollution monitoring in contaminated sites in recent years because it is low cost and relatively fast. However, ERT monitoring data sets are usually analyzed and processed manually, as a result, a lot of manpower is required and the efficiency and accuracy of the ERT data identification are difficult to guarantee.This represents a strong limitation for the application of ERT monitoring system in the fields. To address this problem,clustering algorithms for ERT data analysis was introduced. A numerical model was used to research the effectiveness of contaminated areas recognition using K-means, fuzzy C-means(FCM), and Gaussian mixture model(GMM). The results showed that the three clustering algorithms identified the contaminated area effectively when the difference in resistivity values between the contaminated area and the background soil was larger than 30%. The recognition accuracy of the K-means and FCM algorithms was better than that of the GMM. Finally, a case of the clustering algorithms in ERT survey of a contaminated site was presented.
引文
[1]张宏凯,靳超,郭学茹.加油站泄漏污染物的迁移分布规律[J].中国环境科学,2018,38(4):1532-1539.Zhang H K,Jin C,Guo X R.The underground migration and distribution of petroleum contamination at a gas station[J].China Environmental Science,2018,38(4):1532-1539.
[2]李元杰,王森杰,张敏,等.土壤和地下水污染的监控自然衰减修复技术研究进展[J].中国环境科学,2018,38(3):1185-1193.Li Y J,Wang S J,Zhang M,et al.Research progress of monitored natural attenuation remediation technology for soil and groundwater pollution[J].China Environmental Science,2018,38(3):1185-1193.
[3]Masy T,Caterina D,Tromme O,et al.Electrical resistivity tomography to monitor enhanced biodegradation of hydrocarbons with Rhodococcuserythropolis T902.1at a pilot scale[J].Journal of Contaminant Hydrology,2016,184:1-13.
[4]Power C,Gerhard J I,Karaoulis M,et al.Evaluating four-dimensional time-lapse electrical resistivity tomography for monitoring DNAPLsource zone remediation[J].Journal of Contaminant Hydrology,2014,162-163:27-46.
[5]Lalaoui L,Mohamadi T,Djaalab A.New method for image segmentation[J].Procedia-Social and Behavioral Sciences,2015,195:1971-1980.
[6]Dubey Y K,Mushrif M M.FCM clustering algorithms for segmentation of brain MR images[J].Advances in Fuzzy System,2016,(3):1-14.
[7]Nayak J,Naik B,Behera H S.Computational intelligence in data mining[M].New Delhi:Springer India,2015,(2):133-149.
[8]Berkhin P.A survey of clustering data mining techniques[M].Heidelberg:Springer Berlin Heidelberg,2006:25-71.
[9]Likas A,Vlassis N,Verbeek J J.The global k-means clustering algorithm[J].Pattern Recognition,2003,36(2):451-461.
[10]Bradley P S,Fayyad U M.Refining initial points for k-means clustering[C]//Fifteenth international conference on machine learning,Morgan Kaufmann Publishers Inc,1998:91-99.
[11]Bagirov A M.Modified global k-means algorithm for minimum sum-of-squares clustering problems[J].Pattern Recognition,2008,41(10):3192-3199.
[12]Tzortzis G,Likas A,Tzortzis G.The MinMax k-means clustering algorithm[J].Pattern Recognition,2014,47(7):2505-2516.
[13]Dunn J C.A fuzzy relative of the ISODATA process and its use in detecting compact well-separated clusters[J].Journal of Cybernetics,1973,3(3):32-57.
[14]Wang X Y,Bu J.A fast and robust image segmentation using FCMwith spatial information[J].Digital Signal Processing,2010,20(4):1173-1182.
[15]Huang Z K,Chau K W.A new image thresholding method based on Gaussian mixture model[J].Applied Mathematics&Computation,2008,205(2):899-907.
[16]Baid U,Talbar S.Comparative study of k-means,gaussian mixture model,fuzzy c-means algorithms for brain tumor segmentation[C]//International Conference on Communication and Signal Processing,2017.
[17]Kannan S R,Ramathilagam S,Sathya A,et al.Effective fuzzy c-means based kernel function in segmenting medical images[J].Computers in Biology&Medicine,2010,40(6):572-579.
[18]Janssen R J,Jyl?nki P,Kessels R P C,et al.Probabilistic model-based functional parcellation reveals a robust,fine-grained subdivision of the striatum[J].Neuroimage,2015,119:398-405.
[19]徐立,白金牛,李磊民.基于脑部MR图像GMM特征决策分类的肿瘤诊[J].控制工程,2017,24(8):1718-1722.Xu L,Bai J N,Li L M.Diagnosis of tumor in brain MR images based on GMM features and decision tree classifier[J].Control Engineering,2017,24(8):1718-1722.
[20]Lalaoui L,Mohamadi T,Djaalab A.New method for image segmentation[J].Procedia-Social and Behavioral Sciences,2015,195:1971-1980.
[21]Li Y,Wang Q,Chen J,et al.K-means algorithm based on particle swarm optimization for the identification of rock discontinuity sets[J].Rock Mechanics&Rock Engineering,2015,48(1):375-385.
[22]Liu H,Yang H,Yi F.Experimental study of the complex resistivity and dielectric constant of chrome-contaminated soil[J].Journal of Applied Geophysics,2016,131,109-116.
[23]Carroll Z L,Oliver M A.Exploring the spatial relations between soil physical properties and apparent electrical conductivity[J].Geoderma,2005,128(3):354-374.
[24]王玉玲,能昌信,王彦文,等.重金属污染场地电阻率法探测数值模拟及应用研究[J].环境科学,2013,34(5):1908-1914.Wang Y L,Neng C X,Wang Y W,et al.Numerical simulation and application of electrical resistivity survey in heavy metal contaminated sites[J].Environmental Science,2013,34(5):1908-1914.
[25]王玉玲.重金属污染场地电阻率法检测数值模拟及反演研究[D].北京:中国矿业大学(北京),2013.Wang Y L.Study on numerical simulation and inversion of electrical resistivity method for heavy metal contaminated sites[D].Beijing:China University of Mining&Technology,Beijing,2013.