基于语义分割的页岩孔隙结构智能表征方法

刘茜; 任义丽; 汪文洁; 黄睿琦; 苏乾潇

doi:10.13700/j.bh.1001-5965.2024.0018

基于语义分割的页岩孔隙结构智能表征方法

doi: 10.13700/j.bh.1001-5965.2024.0018

中国石油勘探开发研究院，北京 100083

基金项目:

国家自然科学基金(42372175)

详细信息

通讯作者:
E-mail：renyili@petrochina.com.cn

中图分类号: P583；O235
计量
- 文章访问数: 524
- HTML全文浏览量: 174
- PDF下载量: 15
- 被引次数: 0
出版历程
- 收稿日期: 2024-01-11
- 录用日期: 2024-02-23
- 网络出版日期: 2024-04-10
- 整期出版日期: 2026-04-30

Intelligent representation method of shale pore structure based on semantic segmentation

Research Institute of Petroleum Exploration & Development，Beijing 100083，China

Funds:

National Natural Science Foundation of China (42372175)

More Information

Corresponding author: E-mail：renyili@petrochina.com.cn

摘要

摘要:
页岩孔隙类型和结构参数影响储层流体的富集和运移，是页岩储层评价的重要内容。现有评价方法存在主观性强、效率低及定量化程度低等问题，难以满足快速、精准分析页岩样品的迫切需求。基于此，提出一种基于语义分割的页岩孔隙结构智能表征方法。利用扫描电子显微镜(SEM)和多尺度采集与处理软件(MAPS)获取页岩二维灰度图像；由岩矿鉴定专家进行数据标注，分为有机孔、无机孔、裂缝和有机质；提出一种面向页岩孔隙结构分析任务的组合网络ShaleSeger及其训练范式，建立基于深度学习的页岩孔隙智能识别模型，形成基于多视域拼接的超大尺寸图像孔隙识别方案，从而实现基于MAPS图像的孔隙提取；利用图像处理技术实现孔隙结构特征智能表征。目前已实现古龙页岩孔隙结构定量分析，能够在孔隙边缘提取和类型识别的基础上自动计算各类型孔隙的面积占比，统计诸如孔径、视孔隙比表面、形状因子等孔隙结构参数。所提方法可以拓展应用于CO₂复合压裂效果评估，对CO₂复合压裂前后的微观结构特征改变进行定量比较。
- 致密页岩油 /
- 微纳米 /
- 孔隙结构 /
- 扫描电镜 /
- 古龙页岩
Abstract:
The enrichment and migration of reservoir fluids are affected by the types and structural parameters of shale pores, which are significant components of shale reservoir evaluation. Due to issues with current assessment methods, including high subjectivity, low efficiency, and low degree of quantification, it is challenging to address the urgent needs of quick and accurate examination of shale samples. Based on this, an intelligent characterization method for shale pore structure based on semantic segmentation is proposed. Firstly, the two-dimensional gray images of shale are obtained by scanning electron microscopy (SEM) and multi-scale acquisition and processing software (MAPS). Secondly, these images are annotated by the rock mineral identification experts, and divided into organic pores, inorganic pores, fractures and organic matter. Then, a combination network Shale Seger and its training paradigm for shale pore structure analysis tasks is innovatively proposed, an intelligent recognition model of shale pores based on deep learning is constructed, as well as, a pore recognition scheme of super large image based on multi-view mosaicisis established to extract pore from MAPS images. Lastly, intelligent characterisation of pore structural features is achieved by applying image processing techniques. As of right now, this study has produced a quantitative analysis of the Gulong shale's pore structure that can statistically compute pore structure parameters like pore diameter, apparent pore ratio surface, shape factor, and so forth, as well as automatically determine the area proportion of each type of pore based on pore edge extraction and type recognition. In addition, the technique described in this article can also be extended to the evaluation of CO₂ composite fracturing, through which the change of microstructure characteristics before and after CO₂ composite fracturing can be quantitatively compared.
- tight shale oil /
- micro nano /
- pore structure /
- scanning electron microscopy /
- Gulong shale

HTML全文

图 1 页岩样本示例

Figure 1. Example of shale samples

下载: 全尺寸图片幻灯片

图 2 网络分割结果与标签的对比

Figure 2. Comparison between network segmentation results and labels

下载: 全尺寸图片幻灯片

图 3 ShaleSeger 的网络结构

Figure 3. Network Architecture of ShaleSeger

下载: 全尺寸图片幻灯片

图 4 基于MiT的主干特征提取

Figure 4. Backbone feature extraction based on MiT

下载: 全尺寸图片幻灯片

图 5 边缘特征提取模块

Figure 5. Edge feature extraction module

下载: 全尺寸图片幻灯片

图 6 All-MLP 解码器结构

Figure 6. All-MLP decoder structure

下载: 全尺寸图片幻灯片

图 7 ShaleSeger网络半监督训练方案

Figure 7. Semi-supervised training scheme for ShaleSeger

下载: 全尺寸图片幻灯片

图 8 基于多视域拼接的超大尺寸页岩孔隙识别方法

Figure 8. Pore recognition scheme of super large image based on multi-view mosaicis

下载: 全尺寸图片幻灯片

图 9 孔隙结构标注方案

Figure 9. Pore Structure Annotation Scheme

下载: 全尺寸图片幻灯片

图 10 边缘特征提取分支增加前后的实验结果对比

Figure 10. Comparison of experimental results before and after the addition of edge feature extraction branches

下载: 全尺寸图片幻灯片

图 11 孔隙分割结果

Figure 11. Pore segmentation results

下载: 全尺寸图片幻灯片

图 12 大尺度孔缝识别结果

Figure 12. Large scale hole and seam recognition results

下载: 全尺寸图片幻灯片

图 13 CO₂复合压裂实验前后孔隙度变化

Figure 13. Changes in porosity before and after CO₂ composite fracturing experiment

下载: 全尺寸图片幻灯片

图 14 CO₂复合压裂对孔隙结构改变的直观效果展示

Figure 14. Visual display of the effect of CO₂ composite fracturing on pore structure changes

下载: 全尺寸图片幻灯片

表 1 分割效果定量对比

Table 1. Quantitative comparison of segmentation effects

模型	半监督训练	像素准确率/%	平均交并比/%	平均准确率/%	平均F₁分数/%	平均精确率/%	平均召回率/%
SegFormer^[25]	否	99.25	66.70	75.57	77.82	80.64	75.57
OCRNet^[19]	否	98.99	61.59	71.94	73.48	75.47	71.94
DeepLabv3^[21]	否	98.57	52.43	59.49	64.11	71.37	59.49
PSPNet^[22]	否	98.54	53.95	62.67	65.94	70.04	62.67
SegFormer+边缘特征增强	否	99.25	66.84	75.49	77.84	81.06	75.49
ShaleSeger	是	99.50	72.28	79.67	82.06	85.03	79.67

下载: 导出CSV

表 2 各类型孔隙、有机质面积占比

Table 2. Area proportion of various types pores and organic matter

样本名称	无机孔占比/%	有机孔占比/%	裂缝占比/%	有机质占比/%	其他占比/%
样本a	0.400	0.014	1.997	0.213	97.377
样本b	2.953	0	5.93×10⁻³	0	97.041

下载: 导出CSV

表 3 每个孔隙的特征参数

Table 3. Characteristic parameters of each pore

面积/mm²	直径/$ \mathrm{nm} $	周长/$ \mathrm{nm} $	视孔隙比表面/nm⁻¹	形状因子
214.577	16.529	246.510	1.463	0.044
643.728	28.629	153.593	0.304	0.343
5257.087	81.814	367.252	0.089	0.490
1823.911	48.190	208.632	0.146	0.527
858.308	33.058	146.484	0.217	0.503
751.023	30.923	150.039	0.254	0.419
2253.051	53.560	208.632	0.118	0.650
$\vdots $	$\vdots $	$\vdots $	$\vdots $	$\vdots $
4184.242	72.990	403.657	0.123	0.323

下载: 导出CSV

表 4 基于MAPS图像的各类型孔隙、有机质面积占比

Table 4. Area proportion of various types pores and organic matter based on MAPS

无机孔占比/%	有机孔占比/%	裂缝占比/%	有机质占比/%	面孔率/%
0.529	0.016	0.305	0.353	0.849

下载: 导出CSV

表 5 CO₂复合压裂前后孔隙结构变化

Table 5. Changes in pore structure before and after CO₂ composite fracturing

状态	无机孔占比/%	有机孔占比/%	裂缝占比/%	其他占比/%
复合压裂前	1.107	0.541	2.751	95.601
复合压裂后	1.019	0.561	3.211	95.209

下载: 导出CSV

参考文献(32)

[1]	朱林奇, 张冲, 周雪晴, 等. 融合深度置信网络与与核极限学习机算法的核磁共振测井储层渗透率预测方法[J]. 计算机应用, 2017, 37(10): 3034-3038. ZHU L Q, ZHANG C, ZHOU X Q, et al. Nuclear magnetic resonance logging reservoir permeability prediction method based on deep belief network and kernel extreme learning machine algorithm[J]. Journal of Computer Applications, 2017, 37(10): 3034-3038(in Chinese).
[2]	张益明, 张繁昌, 丁继才, 等. 基于混合深度学习网络的致密砂岩甜点预测[J]. 石油物探, 2021, 60(6): 995-1002. ZHANG Y M, ZHANG F C, DING J C, et al. Sweet spot prediction in tight sand reservoirs by a hybrid deep-learning network[J]. Geophysical Prospecting for Petroleum, 2021, 60(6): 995-1002(in Chinese).
[3]	ZHOU S X, SHENG W, WEI X, et al. Fast image analysis on pore structure of concrete based on deep learning[J]. Journal of the Chinese Ceramic Society, 2019, 47(5): 653-663.
[4]	ZHOU S X, SHENG W, WANG Z P, et al. Quick image analysis of concrete pore structure based on deep learning[J]. Construction and Building Materials, 2019, 208: 144-157.
[5]	CARPENTER C. Machine-learning techniques characterize source-rock images at the pore scale[J]. Journal of Petroleum Technology, 2022, 74(1): 92-95.
[6]	LIU M L, MUKERJI T. Multiscale fusion of digital rock images based on deep generative adversarial networks[J]. Geophysical Research Letters, 2022, 49(9): e2022GL098342.
[7]	WANG H, GUO R, DALTON L E, et al. Comparative assessment of U-net-based deep learning models for segmenting microfractures and pore spaces in digital rocks[J]. SPE Journal, 2024, 29(11): 5779-5791.
[8]	廖广志, 李远征, 肖立志, 等. 利用卷积神经网络模型预测致密储层微观孔隙结构[J]. 石油科学通报, 2020, 5(1): 26-38. LIAO G Z, LI Y Z, XIAO L Z, et al. Prediction of microscopic pore structure of tight reservoirs using convolutional neural network model[J]. Petroleum Science Bulletin, 2020, 5(1): 26-38(in Chinese).
[9]	陈雁, 李祉呈, 程超, 等. FLU-Net: 用于表征页岩储层微观孔隙的深度全卷积网络[J]. 海洋地质前沿, 2021, 37(8): 34-43. CHEN Y, LI Z C, CHENG C, et al. Flu-net: a deep fully convolutional neural network for shale reservoir micro-pore characterization[J]. Marine Geology Frontiers, 2021, 37(8): 34-43(in Chinese).
[10]	YU Q Y, XIONG Z W, DU C, et al. Identification of rock pore structures and permeabilities using electron microscopy experiments and deep learning interpretations[J]. Fuel, 2020, 268: 117416.
[11]	ZHANG H, ZHANG R, SUN D Q, et al. Analyzing the pore structure of pervious concrete based on the deep learning framework of Mask R-CNN[J]. Construction and Building Materials, 2022, 318: 125987.
[12]	陈宗铭, 唐玄, 梁国栋, 等. 基于深度学习的页岩扫描电镜图像有机质孔隙识别与比较[J]. 地学前缘, 2023(3): 208-220. CHEN Z M, TANG X, LIANG G D, et al. Identification and comparison of organic matter pores in shale scanning electron microscopy images based on deep learning[J]. Geological Frontiers, 2023(3): 208-220(in Chinese).
[13]	毕飞宇, 肖占山, 张学忠, 等. 基于深度学习的页岩孔隙类型自动识别方法[J]. 测井技术, 2022, 46(4): 439-445. BI F Y, XIAO Z S, ZHANG X Z, et al. Automated identification method of shale pore types based on deep learning[J]. Well Logging Technology, 2022, 46(4): 439-445(in Chinese).
[14]	蔡宇恒, 滕奇志, 涂秉宇. 基于深度学习的岩石铸体薄片图像孔隙自动提取[J]. 科学技术与工程, 2020, 20(28): 11685-11692. CAI Y H, TENG Q Z, TU B Y. Automatic extraction of pores in thin slice images of rock castings based on deep learning[J]. Science Technology and Engineering, 2020, 20(28): 11685-11692(in Chinese).
[15]	王庆, 曾齐红, 张友焱, 等. 基于多尺度区域卷积神经网络的露头孔洞自动提取[J]. 现代地质, 2021, 35(4): 1147-1154. WANG Q, ZENG Q H, ZHANG Y Y, et al. Automatic extraction of outcrop cavity based on multi-scale regional convolution neural network[J]. Geoscience, 2021, 35(4): 1147-1154(in Chinese).
[16]	LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2015: 3431-3440.
[17]	BADRINARAYANAN V, KENDALL A, CIPOLLA R. SegNet: a deep convolutional encoder-decoder architecture for image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(12): 2481-2495.
[18]	RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[M]. Berlin: Springer, 2015: 234-241.
[19]	YUAN Y H, CHEN X L, WANG J D. Object-contextual representations for semantic segmentation[C]//Proceedings of the Computer Vision-ECCV. Berlin: Springer, 2020: 173-190.
[20]	CHEN L C, PAPANDREOU G, KOKKINOS I, et al. DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(4): 834-848.
[21]	CHEN L C, PAPANDREOU G, SCHROFF F, et al. Rethinking atrous convolution for semantic image segmentation[EB/OL]. (2017-12-05)[2023-07-01]. https://arxiv.org/abs/1706.05587.
[22]	ZHAO H S, SHI J P, QI X J, et al. Pyramid scene parsing network. [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2017: 6230-6239.
[23]	DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16×16 words: transformers for image recognition at scale[EB/OL]. (2021-06-03)[2023-07-01]. https://arxiv.org/abs/2010.11929.
[24]	ZHENG S X, LU J C, ZHAO H S, et al. Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2021: 6877-6886.
[25]	XIE E, WANG W, YU Z, et al. SegFormer: simple and efficient design for semantic segmentation with transformers[EB/OL]. (2021-10-283)[2023-07-01]. https://arxiv.org/abs/2105.15203.
[26]	OUALI Y, HUDELOT C, TAMI M. Semi-supervised semantic segmentation with cross-consistency training[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2020: 12671-12681.
[27]	LIU Y Y, TIAN Y, CHEN Y H, et al. Perturbed and strict mean teachers for semi-supervised semantic segmentation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2022: 4248-4257.
[28]	LAI X, TIAN Z T, JIANG L, et al. Semi-supervised semantic segmentation with directional context-aware consistency[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2021: 1205-1214.
[29]	HUNG W C, TSAI Y H, LIOU Y T, et al. Adversarial learning for semi-supervised semantic segmentation[EB/OL]. (2018-07-24) [2023-07-01]. https://arxiv.org/abs/1802.07934.
[30]	WANG Y C, WANG H C, SHEN Y J, et al. Semi-supervised semantic segmentation using unreliable pseudo-labels[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2022: 4238-4247.
[31]	TAKIKAWA T, ACUNA D, JAMPANI V, et al. Gated-SCNN: gated shape CNNs for semantic segmentation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE Press, 2019: 5228-5237.
[32]	杨运龙, 梁路, 滕少华. 一种双路网络语义分割模型[J]. 广东工业大学学报, 2022, 39(1): 63-70. YANG Y L, LIANG L, TENG S H. A two-way network model for semantic segmentation[J]. Journal of Guangdong University of Technology, 2022, 39(1): 63-70(in Chinese).