A neighborhood median weighted fuzzy c-means method for soil pore identification

Abstract Complex soil pore geometry and heterogeneity determines the ability of a soil to retain moisture and conduction. These soil properties are widely recognized as key factors of essential ecological functions and services. However, until recently, the existing pore identification methods have the problems of low identification accuracy and operating efficiency, which has restricted the development of soil science. The objective of this study was to propose a neighborhood median weighted fuzzy c-means method based on grayscale-gradient feature (NMFCM-G) to identify soil pore structure automatically and accurately. By combining three-dimensional (3D) printing technology with X-ray computed tomography technology, a 3D simulation model with known aperture (10-cm inner diameter) was adopted to evaluate the pore identification error rate of the NMFCM-G method quantitatively. Compared to the methods commonly applied in previous studies, the NMFCM-G method had the smallest average area relative error (2.98%), which was only 1/6 of that of the Image J method with the largest area relative error (18.46%). The NMFCM-G method also had the smallest average perimeter (5.46%), about 3/5 of that of the Image-Pro Plus method with the largest perimeter relative error (8.35%). Meanwhile, the NMFCM-G method was successfully tested on undisturbed cylindrical soil samples, providing encouraging results in terms of identifying irregular pore structure from the complex hierarchical organization of soil. The results show that the NMFCM-G method had the smallest distribution entropy (0.81), the smallest inter-class correlation (0.164 0), and the largest distribution coefficient (0.11), which proves that the NMFCM-G method performed the best in identifying different soil pore structures. Overall, the NMFCM-G method provides new insights into identifying pore structures and thus could provide an automatic and high-efficiency technique for studying the effects of tillage and freeze-thaw cycles on pore structure and soil quality in the future.

[1]  Alasdair N. Houston,et al.  Effect of scanning and image reconstruction settings in X-ray computed microtomography on quality and segmentation of 3D soil images , 2013 .

[2]  Dirk Mallants,et al.  Universal Spatial Correlation Functions for Describing and Reconstructing Soil Microstructure , 2015, PloS one.

[3]  J. Bezdek.  Cluster Validity with Fuzzy Sets , 1973 .

[4]  Simona M. Hapca,et al.  X-ray microtomography: A porosity-based thresholding method to improve soil pore network characterization? , 2014 .

[5]  Alasdair N. Houston,et al.  Quantification of the pore size distribution of soils: Assessment of existing software using tomographic and synthetic 3D images , 2017 .

[6]  Ana M. Tarquis,et al.  Detection of pore space in CT soil images using artificial neural networks , 2010 .

[7]  S. Mooney,et al.  Bimodal Soil Pore Structure Investigated by a Combined Soil Water Retention Curve and X-Ray Computed Tomography Approach , 2017 .

[8]  Tao Yang,et al.  A new method to characterize the spatial structure of soil macropore networks in effects of cultivation using computed tomography , 2014 .

[9]  Mark L. Rivers,et al.  Comparison of image segmentation methods in simulated 2D and 3D microtomographic images of soil aggregates , 2011 .

[10]  Y. Higo,et al.  Pore-scale investigations of partially water-saturated granular soil , 2018, Mechanics Research Communications.

[11]  Wei Wang,et al.  Observer-dependent variability of the thresholding step in the quantitative analysis of soil images and X-ray microtomography data , 2010 .

[12]  J. Koestel.  SoilJ: An ImageJ Plugin for the Semiautomatic Processing of Three‐Dimensional X‐ray Images of Soils , 2018 .

[13]  Wolfram Rühaak,et al.  Phase segmentation of X-ray computer tomography rock images usingmachine learning techniques: an accuracy and performancestudy , 2016 .

[14]  Jinhao Liu,et al.  Soil pore identification with the adaptive fuzzy C-means method based on computed tomography images , 2018, Journal of Forestry Research.

[15]  J. Bezdek,et al.  FCM: The fuzzy c-means clustering algorithm , 1984 .

[16]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[17]  J. Adrien,et al.  Assessment of X-ray Computed Tomography to characterize filtering media from Vertical Flow Treatment Wetlands at the pore scale. , 2019, The Science of the total environment.

[18]  Sheng Li,et al.  Tillage erosion and its effect on spatial variations of soil organic carbon in the black soil region of China , 2018 .

[19]  Alan Wee-Chung Liew,et al.  Spatial Possibilistic Fuzzy C-Mean Segmentation Algorithm Integrated with Brain Mid-sagittal Surface Information , 2017, Int. J. Fuzzy Syst..

[20]  Dorthe Wildenschild,et al.  Image processing of multiphase images obtained via X‐ray microtomography: A review , 2014 .

[21]  Dirk Mallants,et al.  Enhancing image resolution of soils by stochastic multiscale image fusion , 2018 .

[22]  J. C. Dunn,et al.  A Fuzzy Relative of the ISODATA Process and Its Use in Detecting Compact Well-Separated Clusters , 1973 .

[23]  T. Xiao,et al.  Quantification of soil aggregate microstructure on abandoned cropland during vegetative succession using synchrotron radiation-based micro-computed tomography , 2017 .

[24]  P. E. Rieke,et al.  Soil Bulk Density Analysis in Three Dimensions by Computed Tomographic Scanning , 1982 .

[25]  V ROYNA DAISY,et al.  Stability-integrated Fuzzy C means segmentation for spatial incorporated automation of number of clusters , 2018 .

[26]  Lei Liu,et al.  A Simplified Convolutional Network for Soil Pore Identification Based on Computed Tomography Imagery , 2019, Soil Science Society of America Journal.

[27]  Lianyou Liu,et al.  Soil Macropore Structure Characterized by X-Ray Computed Tomography Under Different Land Uses in the Qinghai Lake Watershed, Qinghai-Tibet Plateau , 2017, Pedosphere.

[28]  Ana M. Tarquis,et al.  Identification of pore spaces in 3D CT soil images using PFCM partitional clustering , 2014 .

[29]  H. E. Garrett,et al.  Influence of prairie restoration on CT-measured soil pore characteristics. , 2008, Journal of environmental quality.

[30]  A. McBratney,et al.  Application of fuzzy sets in soil science: fuzzy logic, fuzzy measurements and fuzzy decisions , 1997 .

[31]  Qiaoling Han,et al.  Model representation and quantitative analysis of pore three‐dimensional morphological structure based on soil computed tomography images , 2020, European Journal of Soil Science.

[32]  Richard J. Heck,et al.  Soil pore characteristics assessed from X-ray micro-CT derived images and correlations to soil friability , 2012 .

[33]  Ana M. Tarquis,et al.  Pore network complexity and thresholding of 3D soil images , 2009 .

[34]  E. Skvortsova,et al.  The Impact of Multiple Freeze–Thaw Cycles on the Microstructure of Aggregates from a Soddy-Podzolic Soil: A Microtomographic Analysis , 2018, Eurasian Soil Science.

[35]  D. Giménez,et al.  Pore‐Size Distributions of Soils Derived using a Geometrical Approach and Multiple Resolution MicroCT Images , 2017 .

[36]  Claude E. Shannon,et al.  The mathematical theory of communication , 1950 .