Stochastic 3D rock reconstruction using GANs

被引:0
作者
Valsecchi, Andrea [1 ,2 ]
Damas, Sergio [1 ,3 ]
机构
[1] Univ Granada, Andalusian Res Inst Data Sci & Computat Iniellige, Granada, Spain
[2] Panacea Cooperat Res, Ponferrada, Spain
[3] Univ Granada, Dept Software Engn, Granada, Spain
来源
2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR) | 2021年
关键词
stochastic 3D image reconstruction; generative adversarial neural networks; porous media;
D O I
10.1109/ICPR48806.2021.9412345
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The study of the physical properties of porous media is crucial for petrophysics laboratories. Even though micro computed tomography (CT) could he useful, the appropriate evaluation of flow properties would involve the acquisition of a large number of representative images. That is often unfeasible. Stochastic reconstruction methods aim to generate novel, realistic rock images from a small sample, thus avoiding a large acquisition process. In this contribution, we improve a previous method for 3D-to-3D reconstruction of the structure of porous media by applying generative adversarial neural networks ( GANs). We compare several measures of pore morphology between simulated and acquired images. Experiments include Beadpack, Berea sandstone, and Kelton limestone images. Results show that our GANs-based method can reconstruct three-dimensional images of porous media at different scales that are representative of the morphology of the original images. Furthermore, the generation of multiple images is much faster than classical image reconstruction methods.
引用
收藏
页码:7969 / 7976
页数:8
相关论文
共 27 条
[1]  
[Anonymous], 2016, PHYS REV E
[2]  
[Anonymous], 2018, ARXIV180301045
[3]   Industrial applications of digital rock technology [J].
Berg, Carl Fredrik ;
Lopez, Olivier ;
Berland, Havard .
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING, 2017, 157 :131-147
[4]  
Blunt M. J, 2017, MULTIPHASE FLOW PERM, DOI [10.1017/9781316145098, DOI 10.1017/9781316145098]
[5]   Pore-scale imaging and modelling [J].
Blunt, Martin J. ;
Bijeljic, Branko ;
Dong, Hu ;
Gharbi, Oussama ;
Iglauer, Stefan ;
Mostaghimi, Peyman ;
Paluszny, Adriana ;
Pentland, Christopher .
ADVANCES IN WATER RESOURCES, 2013, 51 :197-216
[6]   Pros and cons of GAN evaluation measures [J].
Borji, Ali .
COMPUTER VISION AND IMAGE UNDERSTANDING, 2019, 179 :41-65
[7]  
Brock A., 2018, Large scale gan training for high fidelity natural image synthesis, P1
[8]   THE ORIGIN OF LATE SPAR CEMENTS IN THE LINCOLNSHIRE LIMESTONE, JURASSIC OF CENTRAL ENGLAND [J].
EMERY, D ;
HUDSON, JD ;
MARSHALL, JD ;
DICKSON, JAD .
JOURNAL OF THE GEOLOGICAL SOCIETY, 1988, 145 :621-633
[9]  
FINNEY JL, 1970, PROC R SOC LON SER-A, V319, P479, DOI [10.1098/rspa.1970.0189, 10.1098/rspa.1970.0190]
[10]  
Goodfellow IJ, 2014, ADV NEUR IN, V27, P2672