Deep-learning-based workflow for boundary and small target segmentation in digital rock images using UNet++ and IK-EBM
Abstract
We report three-dimensional (3D) X-ray micro-computed tomography (μCT) has been widely used in petroleum engineering because it can provide detailed pore structural information for a reservoir rock, which can be imported into a pore-scale numerical model to simulate the transport and distribution of multiple fluids in the pore space. The partial volume blurring (PVB) problem is a major challenge in segmenting raw μCT images of rock samples, which impacts boundaries and small targets near the resolution limit. We developed a deep-learning (DL)-based workflow for accurate and fast partial volume segmentation. The DL model's performance depends primarily on the training data quality and model architecture. This study employed the entropy-based-masking indicator kriging (IK-EBM) to segment 3D Berea sandstone images as training datasets. The comparison between IK-EBM and manual segmentation using a 3D synthetic sphere pack, which had a known ground truth, showed that IK-EBM had higher accuracy on partial volume segmentation. We then trained and tested the UNet++ model, a state-of-the-art supervised encoder-decoder model, for binary (i.e., void and solid) and four-class segmentation. We compared the UNet++ with the commonly used U-Net and wide U-Net models and showed that the UNet++ had the best performance in terms of pixel-wise and physics-basedmore »
- Authors:
-
- Virginia Polytechnic Inst. and State Univ. (Virginia Tech), Blacksburg, VA (United States)
- North Carolina State Univ., Raleigh, NC (United States)
- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- National Energy Technology Lab. (NETL), Morgantown, WV (United States)
- Environmental and Ocean Engineering, Hoboken, NJ (United States)
- Publication Date:
- Research Org.:
- National Energy Technology Laboratory (NETL), Pittsburgh, PA, Morgantown, WV, and Albany, OR (United States); Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- Sponsoring Org.:
- USDOE Office of Fossil Energy (FE)
- OSTI Identifier:
- 1875462
- Alternate Identifier(s):
- OSTI ID: 1870206
- Grant/Contract Number:
- FE0004000; FE0026825; AC05-00OR22725
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Journal of Petroleum Science and Engineering
- Additional Journal Information:
- Journal Volume: 215; Journal Issue: A; Journal ID: ISSN 0920-4105
- Publisher:
- Elsevier
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 02 PETROLEUM; digital rock physics; partial volume blurring; image segmentation; boundary and small targets; IK-EBM; supervised deep learning; UNet++
Citation Formats
Wang, Hongsheng, Dalton, Laura, Fan, Ming, Guo, Ruichang, McClure, James E., Crandall, Dustin, and Chen, Cheng. Deep-learning-based workflow for boundary and small target segmentation in digital rock images using UNet++ and IK-EBM. United States: N. p., 2022.
Web. doi:10.1016/j.petrol.2022.110596.
Wang, Hongsheng, Dalton, Laura, Fan, Ming, Guo, Ruichang, McClure, James E., Crandall, Dustin, & Chen, Cheng. Deep-learning-based workflow for boundary and small target segmentation in digital rock images using UNet++ and IK-EBM. United States. https://doi.org/10.1016/j.petrol.2022.110596
Wang, Hongsheng, Dalton, Laura, Fan, Ming, Guo, Ruichang, McClure, James E., Crandall, Dustin, and Chen, Cheng. Wed .
"Deep-learning-based workflow for boundary and small target segmentation in digital rock images using UNet++ and IK-EBM". United States. https://doi.org/10.1016/j.petrol.2022.110596. https://www.osti.gov/servlets/purl/1875462.
@article{osti_1875462,
title = {Deep-learning-based workflow for boundary and small target segmentation in digital rock images using UNet++ and IK-EBM},
author = {Wang, Hongsheng and Dalton, Laura and Fan, Ming and Guo, Ruichang and McClure, James E. and Crandall, Dustin and Chen, Cheng},
abstractNote = {We report three-dimensional (3D) X-ray micro-computed tomography (μCT) has been widely used in petroleum engineering because it can provide detailed pore structural information for a reservoir rock, which can be imported into a pore-scale numerical model to simulate the transport and distribution of multiple fluids in the pore space. The partial volume blurring (PVB) problem is a major challenge in segmenting raw μCT images of rock samples, which impacts boundaries and small targets near the resolution limit. We developed a deep-learning (DL)-based workflow for accurate and fast partial volume segmentation. The DL model's performance depends primarily on the training data quality and model architecture. This study employed the entropy-based-masking indicator kriging (IK-EBM) to segment 3D Berea sandstone images as training datasets. The comparison between IK-EBM and manual segmentation using a 3D synthetic sphere pack, which had a known ground truth, showed that IK-EBM had higher accuracy on partial volume segmentation. We then trained and tested the UNet++ model, a state-of-the-art supervised encoder-decoder model, for binary (i.e., void and solid) and four-class segmentation. We compared the UNet++ with the commonly used U-Net and wide U-Net models and showed that the UNet++ had the best performance in terms of pixel-wise and physics-based evaluation metrics. Specifically, boundary-scaled accuracy demonstrated that the UNet++ architecture outperformed the regular U-Net architecture in the segmentation of pixels near boundaries and small targets, which were subjected to the PVB effect. Feature map visualization illustrated that the UNet++ bridged the semantic gaps between the feature maps extracted at different depths of the network, thereby enabling faster convergence and more accurate extraction of fine-scale features. The developed workflow significantly enhances the performance of supervised encoder-decoder models in partial volume segmentation, which has extensive applications in fundamental studies of subsurface energy, water, and environmental systems.},
doi = {10.1016/j.petrol.2022.110596},
journal = {Journal of Petroleum Science and Engineering},
number = A,
volume = 215,
place = {United States},
year = {Wed May 04 00:00:00 EDT 2022},
month = {Wed May 04 00:00:00 EDT 2022}
}
Works referenced in this record:
Image thresholding by indicator kriging
journal, July 1999
- Wonho Oh, ; Lindquist, B.
- IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 21, Issue 7
Temporal evolution of pore geometry, fluid flow, and solute transport resulting from colloid deposition: TEMPORAL EVOLUTION OF PORE GEOMETRY FROM COLLOID DEPOSITION
journal, June 2009
- Chen, Cheng; Lau, Boris L. T.; Gaillard, Jean-François
- Water Resources Research, Vol. 45, Issue 6
Pore scale investigation of hydrogen injection in sandstone via X-ray micro-tomography
journal, October 2021
- Jha, Nilesh Kumar; Al-Yaseri, Ahmed; Ghasemi, Mohsen
- International Journal of Hydrogen Energy, Vol. 46, Issue 70
Digital rock physics benchmarks—part II: Computing effective properties
journal, January 2013
- Andrä, Heiko; Combaret, Nicolas; Dvorkin, Jack
- Computers & Geosciences, Vol. 50
Segmentation of digital rock images using deep convolutional autoencoder networks
journal, May 2019
- Karimpouli, Sadegh; Tahmasebi, Pejman
- Computers & Geosciences, Vol. 126
Benchmarking conventional and machine learning segmentation techniques for digital rock physics analysis of fractured rocks
journal, January 2022
- Reinhardt, Marcel; Jacob, Arne; Sadeghnejad, Saeid
- Environmental Earth Sciences, Vol. 81, Issue 3
Active contours without edges
journal, January 2001
- Chan, T. F.; Vese, L. A.
- IEEE Transactions on Image Processing, Vol. 10, Issue 2
Unsupervised Learning of Image Segmentation Based on Differentiable Feature Clustering
journal, January 2020
- Kim, Wonjik; Kanezaki, Asako; Tanaka, Masayuki
- IEEE Transactions on Image Processing, Vol. 29
CNN Explainer: Learning Convolutional Neural Networks with Interactive Visualization
journal, February 2021
- Wang, Zijie J.; Turko, Robert; Shaikh, Omar
- IEEE Transactions on Visualization and Computer Graphics, Vol. 27, Issue 2
Digital rock physics benchmarks—Part I: Imaging and segmentation
journal, January 2013
- Andrä, Heiko; Combaret, Nicolas; Dvorkin, Jack
- Computers & Geosciences, Vol. 50
Using pressure pulse decay experiments and a novel multi-physics shale transport model to study the role of Klinkenberg effect and effective stress on the apparent permeability of shales
journal, June 2020
- Li, Zihao; Ripepi, Nino; Chen, Cheng
- Journal of Petroleum Science and Engineering, Vol. 189
Cell Shape Dynamics: From Waves to Migration
journal, March 2012
- Driscoll, Meghan K.; McCann, Colin; Kopace, Rael
- PLoS Computational Biology, Vol. 8, Issue 3
Methods to measure contact angles in scCO2-brine-sandstone systems
journal, December 2018
- Dalton, Laura E.; Klise, Katherine A.; Fuchs, Samantha
- Advances in Water Resources, Vol. 122
Segmentation of X-ray computed tomography images of porous materials: A crucial step for characterization and quantitative analysis of pore structures: X-RAY CT IMAGE SEGMENTATION
journal, September 2009
- Iassonov, Pavel; Gebrenegus, Thomas; Tuller, Markus
- Water Resources Research, Vol. 45, Issue 9
ML-LBM: Predicting and Accelerating Steady State Flow Simulation in Porous Media with Convolutional Neural Networks
journal, April 2021
- Wang, Ying Da; Chung, Traiwit; Armstrong, Ryan T.
- Transport in Porous Media, Vol. 138, Issue 1
UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation
journal, June 2020
- Zhou, Zongwei; Siddiquee, Md Mahfuzur Rahman; Tajbakhsh, Nima
- IEEE Transactions on Medical Imaging, Vol. 39, Issue 6
Three-dimensional measurement of fractures in heterogeneous materials using high-resolution X-ray computed tomography
journal, October 2010
- Ketcham, Richard A.; Slottke, Donald T.; Sharp, Jack M.
- Geosphere, Vol. 6, Issue 5
Adaptive Thresholding using the Integral Image
journal, January 2007
- Bradley, Derek; Roth, Gerhard
- Journal of Graphics Tools, Vol. 12, Issue 2
Development of a Digital Rock Physics workflow for the analysis of sandstones and tight rocks
journal, July 2017
- Verri, I.; Della Torre, A.; Montenegro, G.
- Journal of Petroleum Science and Engineering, Vol. 156
Processing of rock core microtomography images: Using seven different machine learning algorithms
journal, January 2016
- Chauhan, Swarup; Rühaak, Wolfram; Khan, Faisal
- Computers & Geosciences, Vol. 86
Prediction of Porosity and Permeability Alteration Based on Machine Learning Algorithms
journal, March 2019
- Erofeev, Andrei; Orlov, Denis; Ryzhov, Alexey
- Transport in Porous Media, Vol. 128, Issue 2
Comprehensive study of the interactions between the critical dimensionless numbers associated with multiphase flow in 3D porous media
journal, September 2019
- Fan, Ming; Dalton, Laura E.; McClure, James
- Fuel, Vol. 252
Accurate Measurement of Small Features in X‐Ray CT Data Volumes, Demonstrated Using Gold Grains
journal, April 2019
- Ketcham, R. A.; Mote, A. S.
- Journal of Geophysical Research: Solid Earth, Vol. 124, Issue 4
Enhancement of oil recovery by emulsion injection: A pore scale analysis from X-ray micro-tomography measurements
journal, March 2021
- Scheffer, Kamila; Méheust, Yves; Carvalho, Marcio S.
- Journal of Petroleum Science and Engineering, Vol. 198
The Sensitivity of Estimates of Multiphase Fluid and Solid Properties of Porous Rocks to Image Processing
journal, December 2019
- Garfi, Gaetano; John, Cédric M.; Berg, Steffen
- Transport in Porous Media, Vol. 131, Issue 3
The role of the spatial heterogeneity and correlation length of surface wettability on two-phase flow in a CO2-water-rock system
journal, December 2020
- Guo, Ruichang; Dalton, Laura E.; Fan, Ming
- Advances in Water Resources, Vol. 146
Optimization of Lattice Boltzmann Simulation With Graphics-Processing-Unit Parallel Computing and the Application in Reservoir Characterization
journal, August 2016
- Chen, Cheng; Wang, Zheng; Majeti, Deepak
- SPE Journal, Vol. 21, Issue 04
A Novel Experimental Study on Density‐Driven Instability and Convective Dissolution in Porous Media
journal, November 2021
- Guo, Ruichang; Sun, Hanxing; Zhao, Qingqi
- Geophysical Research Letters, Vol. 48, Issue 23
SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation
journal, December 2017
- Badrinarayanan, Vijay; Kendall, Alex; Cipolla, Roberto
- IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 39, Issue 12
Micro-computed tomography pore-scale study of flow in porous media: Effect of voxel resolution
journal, September 2016
- Shah, S. M.; Gray, F.; Crawshaw, J. P.
- Advances in Water Resources, Vol. 95
Digital Rock Segmentation for Petrophysical Analysis With Reduced User Bias Using Convolutional Neural Networks
journal, February 2020
- Niu, Yufu; Mostaghimi, Peyman; Shabaninejad, Mehdi
- Water Resources Research, Vol. 56, Issue 2
Linking Morphology of Porous Media to Their Macroscopic Permeability by Deep Learning
journal, October 2019
- Kamrava, Serveh; Tahmasebi, Pejman; Sahimi, Muhammad
- Transport in Porous Media, Vol. 131, Issue 2
Deep neural networks for improving physical accuracy of 2D and 3D multi-mineral segmentation of rock micro-CT images
journal, June 2021
- Wang, Ying Da; Shabaninejad, Mehdi; Armstrong, Ryan T.
- Applied Soft Computing, Vol. 104
Pore-scale analysis of permeability reduction resulting from colloid deposition: COLLOID DEPOSITION REDUCES PERMEABILITY
journal, April 2008
- Chen, Cheng; Packman, Aaron I.; Gaillard, Jean-François
- Geophysical Research Letters, Vol. 35, Issue 7
Computational methods for quantitative analysis of three-dimensional features in geological specimens
journal, January 2005
- Ketcham, Richard A.
- Geosphere, Vol. 1, Issue 1
Entropy-assisted image segmentation for nano- and micro-sized networks: ENTROPY-ASSISTED IMAGE SEGMENTATION
journal, December 2015
- Kim, D.; Choi, J.; Nam, J.
- Journal of Microscopy, Vol. 262, Issue 3
Contact Angle Measurements Using Sessile Drop and Micro-CT Data from Six Sandstones
journal, May 2020
- Dalton, Laura E.; Tapriyal, Deepak; Crandall, Dustin
- Transport in Porous Media, Vol. 133, Issue 1
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning
journal, May 2016
- Shin, Hoo-Chang; Roth, Holger R.; Gao, Mingchen
- IEEE Transactions on Medical Imaging, Vol. 35, Issue 5
Application of high resolution X-ray computed tomography to mineral deposit origin, evaluation, and processing
journal, March 2015
- Kyle, J. Richard; Ketcham, Richard A.
- Ore Geology Reviews, Vol. 65
PoreFlow-Net: A 3D convolutional neural network to predict fluid flow through porous media
journal, April 2020
- Santos, Javier E.; Xu, Duo; Jo, Honggeun
- Advances in Water Resources, Vol. 138
3D-visualization and analysis of macro- and meso-porosity of the upper horizons of a sodic, texture-contrast soil
journal, June 2007
- Jassogne, L.; McNeill, A.; Chittleborough, D.
- European Journal of Soil Science, Vol. 58, Issue 3
Fuzzy c-means clustering with spatial information for image segmentation
journal, January 2006
- Chuang, Keh-Shih; Tzeng, Hong-Long; Chen, Sharon
- Computerized Medical Imaging and Graphics, Vol. 30, Issue 1
Acquisition, optimization and interpretation of X-ray computed tomographic imagery: applications to the geosciences
journal, May 2001
- Ketcham, Richard A.; Carlson, William D.
- Computers & Geosciences, Vol. 27, Issue 4
Pore-scale characteristics of multiphase flow in porous media: A comparison of air–water and oil–water experiments
journal, February 2006
- Culligan, K. A.; Wildenschild, D.; Christensen, B. S. B.
- Advances in Water Resources, Vol. 29, Issue 2
NIH Image to ImageJ: 25 years of image analysis
journal, June 2012
- Schneider, Caroline A.; Rasband, Wayne S.; Eliceiri, Kevin W.
- Nature Methods, Vol. 9, Issue 7
Classification and quantification of pore shapes in sandstone reservoir rocks with 3-D X-ray micro-computed tomography
journal, January 2016
- Schmitt, Mayka; Halisch, Matthias; Müller, Cornelia
- Solid Earth, Vol. 7, Issue 1
Deep semantic segmentation of natural and medical images: a review
journal, June 2020
- Asgari Taghanaki, Saeid; Abhishek, Kumar; Cohen, Joseph Paul
- Artificial Intelligence Review, Vol. 54, Issue 1
Deep residual U-net convolution neural networks with autoregressive strategy for fluid flow predictions in large-scale geosystems
journal, April 2021
- Jiang, Zhihao; Tahmasebi, Pejman; Mao, Zhiqiang
- Advances in Water Resources, Vol. 150