融合亲和传播聚类与图卷积神经网络的成岩相标注方法

doi:10.19678/j.issn.1000-3428.0252025

计算机工程 ›› 2026, Vol. 52 ›› Issue (9): 457-468. doi: 10.19678/j.issn.1000-3428.0252025

• 交叉融合与工程应用 • 上一篇

融合亲和传播聚类与图卷积神经网络的成岩相标注方法

张可佳¹^,², 王晓凡¹, 刘涛¹^,*(), 刘宗堡³, 张岩¹^,², 王晨雨⁴, 王浩然³

1. 东北石油大学计算机与信息技术学院, 黑龙江大庆 163318
2. 黑龙江省石油大数据与智能分析重点实验室, 黑龙江大庆 163318
3. 东北石油大学地球科学学院, 黑龙江大庆 163318
4. 大庆油田有限责任公司数智化项目经理部, 黑龙江大庆 163453

收稿日期:2025-01-09 修回日期:2025-03-11 出版日期:2026-09-15 发布日期:2025-04-25
通讯作者: 刘涛
作者简介:
张可佳, 男, 副教授、博士, 主研方向为深度学习、时序数据挖掘
王晓凡，硕士研究生
刘涛(通信作者)，博士研究生
刘宗堡, 教授
张岩, 硕士研究生
王晨雨, 硕士研究生
王浩然, 博士研究生
基金资助:
国家自然科学基金面上项目(42172161); 黑龙江省自然科学基金联合基金重点项目(ZL2024D003); 黑龙江省自然科学基金(LH2022D011); 黑龙江省教育厅基本科研业务费专项资金(2024YSKYFX-01); 大庆市指导性科技计划(zd-2024-05)

Labeling Method for Diagenetic Facies Integrating Affinity Propagation Clustering and Graph Convolutional Neural Networks

ZHANG Kejia¹^,², WANG Xiaofan¹, LIU Tao¹^,*(), LIU Zongbao³, ZHANG Yan¹^,², WANG Chenyu⁴, WANG Haoran³

1. School of Computer and Information Technology, Northeast Petroleum University, Daqing 163318, Heilongjiang, China
2. Heilongjiang Key Laboratory of Petroleum Big Data and Intelligent Analysis, Daqing 163318, Heilongjiang, China
3. School of Earth Sciences, Northeast Petroleum University, Daqing 163318, Heilongjiang, China
4. Digital Intelligence Project Department, Daqing Oilfield Co., Ltd., Daqing 163453, Heilongjiang, China

Received:2025-01-09 Revised:2025-03-11 Online:2026-09-15 Published:2025-04-25
Contact: LIU Tao

摘要/Abstract

摘要：

成岩相样本标注是保证成岩相智能识别准确的关键环节。针对成岩相样本自动标注技术存在样本需求量大和准确率低的问题, 提出一种融合亲和传播(AP)聚类和图卷积神经网络(GCN)相结合的自动标注方法AP-GCN。该方法充分融合了AP聚类捕捉复杂关联关系的优势和GCN挖掘空间分布特征的能力, 选取松辽盆地三肇凹陷州六区块扶余油层为靶区, 实现成岩相样本的自动标注。首先, 归纳成岩相类型并预处理测井曲线数据, 进行少量标签标注, 由此构建自动标注数据集, 为后续自动标注过程奠定了基础; 其次, 利用AP聚类构造图结构, 建立测井曲线深度节点之间的关联关系; 然后, 通过图卷积层聚合节点特征, 实现成岩相的快速准确标注。最后, 设计对比实验验证所提方法的有效性。实验结果表明, AP-GCN方法对各类成岩相标注精确率在86%以上, 召回率在90%以上, F1值在88%以上, 成岩相样本自动标注准确率为90.6%, 证明了该方法的有效性和实用性, 为成岩相样本自动标注提供了新的解决思路。

关键词: 成岩相, 成岩相识别, 自动标注, 图卷积神经网络, 亲和传播聚类

Abstract:

Annotation of diagenetic facies samples is a crucial step in ensuring the accuracy of intelligent diagenetic facies recognition. To address the challenges of high sample demand and limited accuracy in current automatic annotation technologies, this paper proposes an automatic annotation method AP-GCN, which integrates Affinity Propagation (AP) clustering and Graph Convolutional neural Network (GCN). This method effectively integrates the advantages of AP clustering in capturing complex correlations with the capability of GCN for mining spatial distribution features. The Fuyu oil layer within the Zhouliu Block of the Sanzhao Depression in the Songliao Basin is selected as the target area to implement the automatic annotation of diagenetic facies samples. First, diagenetic facies types are summarized and the logging curve data are preprocessed; a limited number of labels are annotated to construct an automatic annotation dataset, which forms the basis for the subsequent automatic annotation process. Second, a graph structure is constructed by using AP clustering to establish correlations between the depth nodes of the logging curves. Subsequently, the node features are aggregated through the graph convolutional layer to achieve rapid and accurate annotation of diagenetic facies. Finally, a comparative experiment is designed to verify the effectiveness of the proposed method. The experimental results demonstrate that the precision of the AP-GCN method for various diagenetic facies annotations exceeds 86%, with recall above 90% and F1 value above 88%. The proposed method achieves an annotation accuracy of 90.6%, confirming its effectiveness and practicality. It provides a novel solution for the automatic annotation of diagenetic facies samples.

Key words: diagenetic facies, diagenetic facies identification, automatic labeling, Graph Convolutional neural Network (GCN), Affinity Propagation (AP) clustering

张可佳, 王晓凡, 刘涛, 刘宗堡, 张岩, 王晨雨, 王浩然. 融合亲和传播聚类与图卷积神经网络的成岩相标注方法[J]. 计算机工程, 2026, 52(9): 457-468.

ZHANG Kejia, WANG Xiaofan, LIU Tao, LIU Zongbao, ZHANG Yan, WANG Chenyu, WANG Haoran. Labeling Method for Diagenetic Facies Integrating Affinity Propagation Clustering and Graph Convolutional Neural Networks[J]. Computer Engineering, 2026, 52(9): 457-468.

https://www.ecice06.com/CN/Y2026/V52/I9/457

图/表 18

图1 AP-GCN方法流程图

Fig.1 Flowchart of the AP-GCN method

图2 采用决策树评分法对目标数据集的评测结果

Fig.2 Evaluation results of the target dataset using the decision tree scoring method

图3 不同成岩相测井响应特征

Fig.3 Response characteristics of logging for different diagenetic facies

图4 三肇凹陷扶余油层部分铸体薄片图像和扫描电镜图像

Fig.4 Images of cast thin sections and scanning electron microscopy from the Fuyu oil layer in the SanZhao depression

图5 部分自动标注数据集

Fig.5 Partial automatic annotation dataset

图6 AP图结构构建流程

Fig.6 The construction process of the AP graph structure

图7 基于图卷积的样本网络结构

Fig.7 Graph convolution-based sample network architecture

图8 模型训练过程中的损失迭代

Fig.8 Loss iteration during model training process

图9 图构造方法的马修斯相关系数和准确率对比

Fig.9 Comparison of Matthews correlation coefficient and accuracy for graph construction methods

图10 4种聚类方法的测试准确率和盲井准确率对比

Fig.10 Comparison of testing accuracy and blind well accuracy for four clustering methods

图11 AP-GCN方法标注结果与真实值对比

Fig.11 Comparison of AP-GCN method annotation results with true values

图12 Precision、Recall和F1值计算结果

Fig.12 Results of Precision, Recall, and F1 value calculation

图13 添加高斯噪声后mAA和mAP变化情况

Fig.13 The change in mAA and mAP after adding Gaussian noise

参考文献 34

1	ZHAO X B , YAO G Q , CHEN X J , et al. Diagenetic facies classification and characterization of a high-temperature and high-pressure tight gas sandstone reservoir: a case study in the Ledong area, Yinggehai Basin. Marine and Petroleum Geology, 2022, 140, 105665. doi: 10.1016/j.marpetgeo.2022.105665
2	LIU H , REN Y L , LI X , et al. Rock thin-section analysis and identification based on artificial intelligent technique. Petroleum Science, 2022, 19 (4): 1605- 1621. doi: 10.1016/j.petsci.2022.03.011
3	周能武, 卢双舫, 王民, 等. 中国典型陆相盆地致密油成储界限与分级评价标准. 石油勘探与开发, 2021, 48 (5): 939- 949.
	ZHOU N W , LU S F , WANG M , et al. Limits and grading evaluation criteria of tight oil reservoirs in typical continental basins of China. Petroleum Exploration and Development, 2021, 48 (5): 939- 949.
4	刘宏坤, 艾勇, 王贵文, 等. 深层、超深层致密砂岩储层成岩相测井定量评价: 以库车坳陷博孜-大北地区为例. 地质科技通报, 2023, 42 (1): 299- 310.
	LIU H K , AI Y , WANG G W , et al. Quantitative well logging evaluation of diagenetic facies of deep and ultra deep tight sandstone reservoirs: a case study of Bozi-Dabei area in Kuqa Depression. Bulletin of Geological Science and Technology, 2023, 42 (1): 299- 310.
5	魏兆胜, 覃建华, 李映艳, 等. 混积页岩油储层成岩相特征及其成储意义——以吉木萨尔凹陷芦草沟组为例. 中国石油勘探, 2024, 29 (6): 99- 115.
	WEI Z S , QIN J H , LI Y Y , et al. Diagenetic facies of diamictic shale oil reservoir and its significance on reservoir formation: a case study of Lucaogou Formation in Jimsar Sag. China Petroleum Exploration, 2024, 29 (6): 99- 115.
6	SONG Z X , YANG X L , XU Z L , et al. Graph-based semi-supervised learning: a comprehensive review. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34 (11): 8174- 8194. doi: 10.1109/TNNLS.2022.3155478
7	YANG X L , SONG Z X , KING I , et al. A survey on deep semi-supervised learning. IEEE Transactions on Knowledge and Data Engineering, 2023, 35 (9): 8934- 8954. doi: 10.1109/TKDE.2022.3220219
8	GUI J , CHEN T , ZHANG J , et al. A survey on self-supervised learning: algorithms, applications, and future trends. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46 (12): 9052- 9071. doi: 10.1109/TPAMI.2024.3415112
9	CHEN L Y , LI S B , BAI Q , et al. Review of image classification algorithms based on convolutional neural networks. Remote Sensing, 2021, 13 (22): 4712. doi: 10.3390/rs13224712
10	PATTNAIK S, CHEN S H, HELBA A, et al. Automatic carbonate rock facies identification with deep learning[C]//Proceedings of the SPE Annual Technical Conference and Exhibition. Virtual, Richardson: Society of Petroleum Engineers, 2020: SPE 201673-MS.
11	DENG T Q , XU C C , LANG X Z , et al. Diagenetic facies classification in the arbuckle formation using deep neural networks. Mathematical Geosciences, 2021, 53 (7): 1491- 1512. doi: 10.1007/s11004-021-09918-0
12	蒋瑞刚. 松辽盆地三肇凹陷致密储层地质评价. 石油天然气学报, 2021, 43 (1): 42- 50.
	JIANG R G . Geological evaluation of tight reservoirs in Sanzhao Sag, Songliao Basin. Journal of Oil and Gas Technology, 2021, 43 (1): 42- 50.
13	ANAND S K , KUMAR S . Experimental comparisons of clustering approaches for data representation. ACM Computing Surveys, 2023, 55 (3): 1- 33.
14	ABDULAH S , ATWA W , ABDELMONIEM A M . Active clustering data streams with affinity propagation. ICT Express, 2022, 8 (2): 276- 282. doi: 10.1016/j.icte.2021.08.017
15	BRUSCO M J , STEINLEY D , STEVENS J , et al. Affinity propagation: an exemplar-based tool for clustering in psychological research. British Journal of Mathematical and Statistical Psychology, 2019, 72 (1): 155- 182. doi: 10.1111/bmsp.12136
16	LU S Y , CAI C Y , ZHONG Z , et al. Ultradeep carbonate reservoir lithofacies classification based on a deep convolutional neural network-a case study in the Tarim Basin, China. Interpretation, 2023, 11 (3): 551- 566. doi: 10.1190/INT-2022-0020.1
17	WANG A T , ZHAO S Y , XIE K , et al. Attention mechanism-enhanced graph convolutional neural network for unbalanced lithology identification. Scientific Reports, 2024, 14, 17319. doi: 10.1038/s41598-024-64871-2
18	VENKATESHWARAN B , RAMKUMAR M , SIDDIQUI N A , et al. A graph convolutional network approach to qualitative classification of hydrocarbon zones using petrophysical properties in well logs. Natural Resources Research, 2024, 33 (2): 637- 664. doi: 10.1007/s11053-024-10311-x
19	徐朝晖, 刘钰铭, 周新茂, 等. 基于卷积神经网络算法的自动地层对比实验. 石油科学通报, 2019, 4 (1): 1- 10.
	XU Z H , LIU Y M , ZHOU X M , et al. An experiment in automatic stratigraphic correlation using convolutional neural networks. Petroleum Science Bulletin, 2019, 4 (1): 1- 10.
20	ALSWAIDAN Z , ALFARRAJ M , LUQMAN H . Geology-constrained dynamic graph convolutional networks for seismic facies classification. Computers & Geosciences, 2024, 184, 105516.
21	SNELL J, SWERSKY K, ZEMEL R S. Prototypical networks for few-shot learning[C]//Proceedings of Advances in Neural Information Processing Systems. Long Beach, USA: [s. n], 2017: 30.
22	朱国文, 王小军, 白雪峰, 等. 松辽盆地北部致密油勘探新领域及资源潜力. 石油学报, 2025, 46 (1): 33- 47.
	ZHU G W , WANG X J , BAI X F , et al. New exploration fields and resource of tight oil in northern Songliao Basin. Acta Petrolei Sinica, 2025, 46 (1): 33- 47.
23	杨嘉明, 王璐奕, 杜欣蓉, 等. 大庆-三肇凹陷扶余油层致密储层成岩作用类型及序列. 化工管理, 2023 (15): 70- 73.
	YANG J M , WANG L Y , DU X R , et al. Diagenesis types and sequences of tight reservoirs in Fuyu oil layer in Daqing Sanzhao depression. Chemical Engineering Management, 2023 (15): 70- 73.
24	邓庆杰, 胡明毅, 胡忠贵, 等. 浅水三角洲分流河道砂体沉积特征——以松辽盆地三肇凹陷扶Ⅱ-Ⅰ组为例. 石油与天然气地质, 2015, 36 (1): 118- 127.
	DENG Q J , HU M Y , HU Z G , et al. Sedimentary characteristics of shallow-water delta distributary channel sand bodies: a case from Ⅱ-Ⅰ Formation of Fuyu oil layer in the Sanzhao Depression, Songliao Basin. Oil & Gas Geology, 2015, 36 (1): 118- 127.
25	张响响, 陶士振, 朱如凯, 等. 成岩作用对致密油储层分布的影响——以松南让字井斜坡带扶余油层为例. 沉积学报, 2023, 41 (2): 559- 568.
	ZHANG X X , TAO T Z , ZHU R K , et al. Influence of diagenesis on the distribution of tight oil reservoirs: a case study of the fuyu reservoir, Rangzijing slope zone, southern Songliao Basin. Acta Sedimentologica Sinica, 2023, 41 (2): 559- 568.
26	贾承造, 邹才能, 李建忠, 等. 中国致密油评价标准、主要类型、基本特征及资源前景. 石油学报, 2012, 33 (3): 343- 350.
	JIA C Z , ZOU C N , LI J Z , et al. Assessment criteria, main types, basic features and resource prospects of the tight oil in China. Acta Petrolei Sinica, 2012, 33 (3): 343- 350.
27	唐振兴, 赵家宏, 王天煦. 松辽盆地南部致密油"甜点区(段)"评价与关键技术应用. 天然气地球科学, 2019, 30 (8): 1114- 1124.
	TANG Z X , ZHAO J H , WANG T X . Evaluation and key technology application of"sweet area"of tight oil in southern Songliao Basin. Natural Gas Geoscience, 2019, 30 (8): 1114- 1124.
28	LIU T , LIU Z B , ZHANG K J , et al. Research on the generation and annotation method of thin section images of tight oil reservoir based on deep learning. Scientific Reports, 2024, 14, 12805. doi: 10.1038/s41598-024-63430-z
29	李盛谦, 曾溅辉, 刘亚洲, 等. 东海盆地西湖凹陷孔雀亭地区古近系平湖组储层成岩作用及孔隙演化. 岩性油气藏, 2023, 35 (5): 49- 61.
	LI S Q , ZENG J H , LIU Y Z , et al. Reservoir diagenesis and pore evolution of Paleogene Pinghu Formation in Kongqueting area of Xihu Sag, East China Sea Basin. Lithologic Reservoirs, 2023, 35 (5): 49- 61.
30	张强, 李家金, 王毛毛, 等. 基于改进主成分分析法的测井曲线岩性分层技术. 吉林大学学报(地球科学版), 2022, 52 (4): 1369- 1376.
	ZHANG Q , LI J J , WANG M M , et al. Logging curve rock layering technology based on improved principal component analysis. Journal of Jilin University (Earth Science Edition), 2022, 52 (4): 1369- 1376.
31	ZHANG J Y , AMBROSE W , XIE W . Applying convolutional neural networks to identify lithofacies of large-n cores from the Permian Basin and Gulf of Mexico: the importance of the quantity and quality of training data. Marine and Petroleum Geology, 2021, 133, 105307. doi: 10.1016/j.marpetgeo.2021.105307
32	刘畅, 孙雨, 杨远礼, 等. 改进YOLOv10n算法在番茄叶片病害检测上的应用. 光电工程, 2025, 52 (8): 250036.
	LIU C , SUN Y , YANG Y L , et al. Application of improved YOLOv10n algorithm in tomato leaf disease detection. Opto-Electronic Engineering, 2025, 52 (8): 250036.
33	ANTARIKSA G , MUAMMAR R , LEE J . Performance evaluation of machine learning-based classification with rock-physics analysis of geological lithofacies in Tarakan Basin, Indonesia. Journal of Petroleum Science and Engineering, 2022, 208, 109250. doi: 10.1016/j.petrol.2021.109250
34	PEREIRA N. PereiraASLNet: ASL letter recognition with YOLOX taking mean average precision and inference time considerations[C]//Proceedings of the 2nd International Conference on Artificial Intelligence and Signal Processing (AISP). Washington D.C., USA: IEEE Press, 2022: 1-6.

[1]	王忠美, 聂芃轩, 刘建华, 吴海波, 郑良. 基于多约束模态不变图卷积融合网络的轴承故障诊断[J]. 计算机工程, 2026, 52(8): 422-430.
[2]	吴迪, 张乐, 陈尧. 跨行为特征正则化和交互对比学习多行为推荐[J]. 计算机工程, 2026, 52(7): 143-154.
[3]	胡静丹, 李波, 杨静. 基于增强数值表示的数学应用题求解模型研究[J]. 计算机工程, 2026, 52(5): 95-102.
[4]	付碧超, 盛捷, 王雷. 基于多步态特征自适应融合的情绪识别[J]. 计算机工程, 2026, 52(4): 82-89.
[5]	程俊军, 王明文. 基于双分支深度图卷积网络的指静脉识别研究[J]. 计算机工程, 2026, 52(3): 152-160.
[6]	胡书林, 张华军, 邓小涛, 王征华. 结合依存图卷积的中文文本相似度计算研究[J]. 计算机工程, 2025, 51(3): 76-85.
[7]	张洋, 刘畅, 李少青. 基于可控制性度量的图神经网络门级硬件木马检测方法[J]. 计算机工程, 2024, 50(7): 164-173.
[8]	李政学, 李枝名, 彭德中, 陈杰. 基于特征对比学习和图卷积的社交网络用户分类[J]. 计算机工程, 2024, 50(4): 258-266.
[9]	白宇, 王珺, 冉红雷, 安胜彪. 半导体器件内部缺陷标注与检测方法研究[J]. 计算机工程, 2024, 50(12): 245-253.
[10]	代祖华, 刘园园, 狄世龙. 语义增强的图神经网络方面级文本情感分析[J]. 计算机工程, 2023, 49(6): 71-80.
[11]	衡红军, 苗菁. 语义与句法信息加强的二元标记实体关系联合抽取[J]. 计算机工程, 2023, 49(4): 77-84.
[12]	徐康, 李霏, 姬东鸿. 结合依存图卷积与文本片段搜索的方面情感三元组抽取[J]. 计算机工程, 2023, 49(4): 61-67.
[13]	陈治旭, 靳雁霞, 芦烨, 杨晶, 刘亚变, 史志儒. 基于子图卷积神经网络的多精度服装建模方法[J]. 计算机工程, 2023, 49(4): 174-181.
[14]	翟社平, 张宇航, 柏晓夏. 融合实体邻域信息的知识图谱嵌入负采样方法[J]. 计算机工程, 2023, 49(3): 95-104.
[15]	HAYTHEM Alhag, 杨军. 基于隐式描述符的三维模型对应关系计算[J]. 计算机工程, 2022, 48(5): 229-234.

选择文件类型/文献管理软件名称

选择包含的内容

融合亲和传播聚类与图卷积神经网络的成岩相标注方法

Labeling Method for Diagenetic Facies Integrating Affinity Propagation Clustering and Graph Convolutional Neural Networks

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 34

相关文章 15

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

融合亲和传播聚类与图卷积神经网络的成岩相标注方法

Labeling Method for Diagenetic Facies Integrating Affinity Propagation Clustering and Graph Convolutional Neural Networks

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 34

相关文章 15

编辑推荐

Metrics

本文评价