煤岩裂隙图像识别方法研究

郝天轩; 徐新革; 赵立桢

doi:10.13272/j.issn.1671-251x.2022120081

煤岩裂隙图像识别方法研究

doi: 10.13272/j.issn.1671-251x.2022120081

郝天轩^{1, 2, 3,},
徐新革^1, ,,
赵立桢¹

1.
河南理工大学安全科学与工程学院，河南焦作　454000
2.
河南省瓦斯地质与瓦斯治理重点实验室—省部共建国家重点实验室培育基地，河南焦作　454000
3.
煤炭安全生产河南省协同创新中心，河南焦作　454000

基金项目: 河南省重点研发与推广专项（科技攻关）资助项目（222102320172）。

详细信息

作者简介:
郝天轩（1976—），男，河南孟州人，教授，博士，主要从事矿山安全科学、矿山信息化等方面的教学、科研及管理工作，E-mail：htx@hpu.edu.cn

通讯作者:
徐新革（1996—），女，山东肥城人，硕士研究生，研究方向为安全系统工程与安全信息，E-mail：xxg1826387@163.com。

中图分类号: TD76
计量
- 文章访问数: 108
- HTML全文浏览量: 66
- PDF下载量: 36
- 被引次数: 0
出版历程
- 收稿日期: 2022-12-27
- 修回日期: 2023-09-10
- 网络出版日期: 2023-10-25

Research on image recognition methods for coal rock fractures

HAO Tianxuan^{1, 2, 3
,},
XU Xinge^{1
, ,},
ZHAO Lizhen¹

1.
School of Safety Science and Engineering, Henan Polytechnic University, Jiaozuo 454000, China
2.
Henan Province Key Laboratory of Gas Geology and Gas Governance-the National Key Laboratory Cultivation Base Jointly Established by the Province and the Ministry, Jiaozuo 454000, China
3.
Henan Collaborative Innovation Center for Coal Safety Production, Jiaozuo 454000, China

摘要

摘要: 煤岩裂隙与瓦斯运移密切相关，且影响煤岩体稳定性，研究煤岩体中复杂的裂隙系统对于巷道支护和瓦斯抽采有重要意义。目前煤岩裂隙图像识别方法未能综合考虑煤岩图像裂隙数量、位置、形态和类别等特点，难以获取有效信息。以鹤壁煤电股份有限公司第八煤矿掘进工作面煤岩图像为研究对象，提出了一种基于U−Net网络对图像中裂隙及类别实现像素级智能识别的方法。采用直方图均衡化、高斯双边滤波和拉普拉斯算子对煤岩图像进行预处理，以提高图像质量，更有效地提取裂隙特征信息。通过观测记录煤岩裂隙特征并分为7类，对筛选出的煤岩裂隙图像进行扩增，采用Labelme软件对图像进行像素级标注，建立煤岩裂隙数据集。采用U−Net网络构建煤岩裂隙识别模型，经调试确定网络批量大小和学习率参数，实验表明当迭代次数达到300以上时，该模型的识别精确率均值为87%，召回率均值为92%，平均交并比大于85%，类别平均像素准确率大于80%。采集井下煤岩采动裂隙和实验室张性外生裂隙对煤岩裂隙识别模型进行验证，结果表明该模型可有效提取目标特征信息并与背景特征信息区分，能够较准确地定位、识别单一裂隙。
- 煤岩裂隙识别 /
- 数字图像处理 /
- U−Net /
- 特征提取 /
- 图像预处理
Abstract: Coal rock fractures are closely related to gas migration and affect the stability of coal rock. Studying the complex fracture system in coal rock is of great significance for roadway support and gas extraction. At present, the recognition methods for coal rock fracture images fail to comprehensively consider the features of the number, position, morphology, and category of fracture in coal rock images, making it difficult to obtain effective information. Taking the coal rock images of excavation face in the No.8 Coal Mine of Hebi Coal and Electricity Co., Ltd. as the research object, a pixel level intelligent recognition method based on U-Net network for coal rock fractures and categories is proposed. The histogram equalization, Gaussian bilateral filtering, and Laplace operator are used to preprocess coal rock images to improve image quality and extract fracture feature information more effectively. The features of coal rock fractures are recorded by observing and divided into 7 categories, the selected coal rock fracture images are amplified, and the images are annotated at the pixel level using Labelme software to establish a coal rock fracture dataset. The U-Net network is used to construct a coal rock fracture recognition model. After debugging, the network batch size and learning rate parameters are determined. The experiment shows that when the number of iterations reaches 300 or more, the average recognition accuracy of the model is 87%, the average recall rate is 92%, the average intersection to parallel ratio is greater than 85%, and the average pixel accuracy of the category is greater than 80%. The coal rock fracture recognition model is validated by collecting underground coal rock mining fractures and laboratory tensile exogenous fractures. The results show that the model can effectively extract target feature information and distinguish it from background feature information, and can accurately locate and recognize a single fracture.
- recognition of coal rock fractures /
- digital image processing /
- U-Net /
- feature extraction /
- image preprocessing

HTML全文

图 1 煤岩裂隙图像预处理结果

Figure 1. Coal rock fracture image preprocessing results

下载: 全尺寸图片幻灯片

图 2 煤岩裂隙数据集扩增

Figure 2. Data set amplification of coal rock fracture

下载: 全尺寸图片幻灯片

图 3 部分可视化训练样本

Figure 3. Part of visual training samples

下载: 全尺寸图片幻灯片

图 4 煤岩裂隙识别模型网络结构

Figure 4. Network structure of coal rock fracture identification model

下载: 全尺寸图片幻灯片

图 5 煤岩裂隙识别模型混淆矩阵

Figure 5. Confusion matrix of rock fracture identification model

下载: 全尺寸图片幻灯片

图 6 煤岩裂隙识别模型的评价指标

Figure 6. Evaluation indexes of coal rock fracture identification model

下载: 全尺寸图片幻灯片

图 7 煤岩裂隙识别结果

Figure 7. Coal rock fracture identification model

下载: 全尺寸图片幻灯片

表 1 煤岩裂隙数据集

Table 1. Data set of coal rock fracture

裂隙类别	训练样本个数	测试样本个数
内生裂隙	197	49
张性外生裂隙	132	33
剪性外生裂隙	75	18
张剪性外生裂隙	43	11
压剪性外生裂隙	82	21
劈理	34	9
采动裂隙	74	18

下载: 导出CSV

表 2 不同模型的煤岩裂隙识别准确率

Table 2. Accuracy rate of coal rock fracture identification by different models

模型	准确率/%
模型	训练集	测试集
SegNet	64.63	55.38
DeepLab v3+	67.41	62.86
PspNet	68.37	53.82
U−Net	71.71	66.01

下载: 导出CSV

表 3 煤岩裂隙识别模型网络参数

Table 3. Network parameters of coal rock fracture identification model

层次结构	参数设置
卷积层_1	卷积核数：16；卷积核尺寸：5×5×3；步长：1；全零填充卷积核数：16；卷积核尺寸：5×5×3；步长：1；全零填充
池化层_1	池化核尺寸：2×2；步长：2
卷积层_2	卷积核数：32；卷积核尺寸：3×3×3；步长：1；全零填充卷积核数：32；卷积核尺寸：3×3×3；步长：1；全零填充
池化层_2	池化核尺寸：2×2；步长：2
卷积层_3	卷积核数：64；卷积核尺寸：3×3×3；步长：1；全零填充卷积核数：64；卷积核尺寸：3×3×3；步长：1；全零填充卷积核数：64；卷积核尺寸：3×3×3；步长：1；全零填充
上采样_4、拼接融合	上采样因子：2×2×3
卷积层_5	卷积核数：64；卷积核尺寸：3×3×3；步长：1；全零填充卷积核数：64；卷积核尺寸：3×3×3；步长：1；全零填充
上采样_6、拼接融合	上采样因子：2×2×3
卷积层_7	卷积核数：32；卷积核尺寸：3×3×3；步长：1；全零填充卷积核数：32；卷积核尺寸：3×3×3；步长：1；全零填充
卷积层_8	卷积核数：8；卷积核尺寸：1×1×3；步长：1；全零填充

下载: 导出CSV

表 4 煤岩裂隙识别模型训练结果

Table 4. Training results of coal rock fracture identification model

批量大小	不同学习率下的识别准确率/%
	0.01		0.001		0.000 5		0.000 1		0.000 01
	训练集	测试集	训练集	测试集	训练集	测试集	训练集	测试集	训练集	测试集
2	30.19	27.99	76.31	77.26	90.36	71.21	83.05	82.93	79.13	68.96
4	48.38	46.95	78.55	69.34	90.36	71.19	89.23	83.07	80.52	68.52
6	63.52	52.68	72.43	73.88	85.74	68.01	87.69	89.43	84.13	77.53
8	42.41	41.26	87.14	74.67	89.31	68.61	82.45	87.31	88.39	72.45
10	42.87	38.57	84.91	80.50	89.72	71.70	86.03	82.60	82.60	65.42

下载: 导出CSV

参考文献(22)

[1]	刘建华,王生维,粟冬梅. 二连盆地群低煤阶煤储层裂隙地质建模与精细描述[J]. 煤炭科学技术,2022,50(5):198-207. LIU Jianhua,WANG Shengwei,SU Dongmei. Geological modeling and fine description of fractures in low coal rank coal reservoirs of Erlian Basin Group[J]. Coal Science and Technology,2022,50(5):198-207.
[2]	韩文龙,王延斌,倪小明,等. 多期构造运动对深部煤储层物性特征影响研究[J]. 煤炭科学技术,2021,49(10):208-216. doi: 10.13199/j.cnki.cst.2021.10.028 HAN Wenlong,WANG Yanbin,NI Xiaoming,et al. Study on impact of multi-period tectonic movement on deep coal reservoir physical properties[J]. Coal Science and Technology,2021,49(10):208-216. doi: 10.13199/j.cnki.cst.2021.10.028
[3]	王登科,魏强,魏建平,等. 煤的裂隙结构分形特征与分形渗流模型研究[J]. 中国矿业大学学报,2020,49(1):103-109,122. WANG Dengke,WEI Qiang,WEI Jianping,et al. Fractal characteristics of fracture structure and fractal seepage model of coal[J]. Journal of China University of Mining & Technology,2020,49(1):103-109,122.
[4]	MAJIDIFARD H,ADU-GYAMFI Y,BUTTLAR W G. Deep machine learning approach to develop a new asphalt pavement condition index [J]. Construction and Building Materials,2020,247. DOI: 10.1016/j.conbuildmat.2020.118513.
[5]	JU Huyan,LI Wei,TIGHE S,et al. Detection of sealed and unsealed cracks with complex backgrounds using deep convolutional neural network [J]. Automation in Construction,2019,107. DOI: 10.1016/j.autcon.2019.102946.
[6]	刘勇,崔洪庆. 基于裂隙形态特征的煤层图像裂隙识别研究[J]. 工矿自动化,2017,43(10):59-64. LIU Yong,CUI Hongqing. Research on coal-bed image fractures identification based on fracture shape characteristics[J]. Industry and Mine Automation,2017,43(10):59-64.
[7]	孙月龙,崔洪庆,关金锋. 基于图像识别的煤层井下宏观裂隙观测[J]. 煤田地质与勘探,2017,45(5):19-22. SUN Yuelong,CUI Hongqing,GUAN Jinfeng. Image recognition-based observation of macro fracture in coal seam in underground mine[J]. Coal Geology & Exploration,2017,45(5):19-22.
[8]	覃木广. 基于matlab图像识别王庄矿后备区裂隙发育方位[J]. 采矿技术,2021,21(1):173-176. QIN Muguang. Identification of fracture development orientation in Wangzhuang Coal Mine reserve area based on matlab image[J]. Mining Technology,2021,21(1):173-176.
[9]	谢配红,谭海英. 基于图像识别技术对出露危岩体裂隙发育规律统计分析[J]. 煤炭技术,2021,40(7):66-67. XIE Peihong,TAN Haiying. Statistical analysis of fracture development law of exposed dangerous rock mass based on image recognition technology[J]. Coal Technology,2021,40(7):66-67.
[10]	张庆贺,陈晨,袁亮,等. 基于DIC和YOLO算法的复杂裂隙岩石破坏过程动态裂隙早期智能识别[J]. 煤炭学报,2022,47(3):1208-1219. ZHANG Qinghe,CHEN Chen,YUAN Liang,et al. Early and intelligent recognition of dynamic cracks during damage of complex fractured rock masses based on DIC and YOLO algorithms[J]. Journal of China Coal Society,2022,47(3):1208-1219.
[11]	靳阳阳,韩现伟,周书宁,等. 图像增强算法综述[J]. 计算机系统应用,2021,30(6):18-27. JIN Yangyang,HAN Xianwei,ZHOU Shuning,et al. Review on image enhancement algorithms[J]. Computer Systems & Applications,2021,30(6):18-27.
[12]	LAURENT C,JEAN-PHILIPPE T,PIERRE C. The guided bilateral filter:when the joint/cross bilateral filter becomes robust[J]. IEEE Transactions on Image Processing,2015,24(4):1199-1208.
[13]	陈春谋. 基于直方图均衡化与拉普拉斯的铅条图像增强算法[J]. 国外电子测量技术,2019,38(7):131-135. CHEN Chunmou. Lead line image enhancement algorithm based on histogram equalization and Laplacian[J]. Foreign Electronic Measurement Technology,2019,38(7):131-135.
[14]	王兆会,孙文超,水艳婷,等. 预制裂隙类岩石试件表面变形场演化与裂隙扩展机理研究[J/OL]. 煤炭科学技术:1-11 [2022-12-11]. https://doi.org/10.13199/j.cnki.cst.2022-1447. WANG Zhaohui,SUN Wenchao,SHUI Yanting,et al. Surface deformation field and fracture propagation mechanism of rock-like samples with pre-existing fracture[J/OL]. Coal Science and Technology:1-11 [2022-12-11]. https://doi.org/10.13199/j.cnki.cst.2022-1447.
[15]	曾勇,屈永华,宋金宝. 煤层裂隙系统及其对煤层气产出的影响[J]. 江苏地质,2000(2):91-94. ZENG Yong,QU Yonghua,SONG Jinbao. The coal seam system of fissures and their influence on the occurrence of coal seams[J]. Jiangsu Geology,2000(2):91-94.
[16]	SUN C,SHRIVASTAVA A,SINGH S,et al. Revisiting unreasonable effectiveness of data in deep learning era[C]. IEEE International Conference on Computer Vision,Venice,2017. DOI: 10.1109/ICCV.2017.97.
[17]	HESTNESS J, NARANG S, ARDALANI N, et al. Deep learning scaling is predictable, empirically[EB/OL]. [2022-12-02]. https://doi.org/10.48550/arXiv.1712.00409. HESTNESS J,NARANG S,ARDALANI N,et al. Deep learning scaling is predictable,empirically[EB/OL]. [2022-12-02]. https://doi.org/10.48550/arXiv.1712.00409.
[18]	JABRI A,JOULIN A,MAATEN L,et al. Learning visual features from large weakly supervised data[C]. European Conference on Computer Vision, Amsterdam,2016. DOI: 10.1007/978-3-319-46478-7_5.
[19]	薛珊,张振,吕琼莹,等. 基于卷积神经网络的反无人机系统图像识别方法[J]. 红外与激光工程,2020,49(7):250-257. XUE Shan,ZHANG Zhen,LYU Qiongying,et al. Image recognition method of anti UAV system based on convolutional neural network[J]. Infrared and Laser Engineering,2020,49(7):250-257.
[20]	周飞燕,金林鹏,董军. 卷积神经网络研究综述[J]. 计算机学报,2017,40(6):1229-1251. ZHOU Feiyan,JIN Linpeng,DONG Jun. Review of convolutional neural network[J]. Chinese Journal of Computers,2017,40(6):1229-1251.
[21]	林丽,刘新,朱俊臻,等. 基于CNN的金属疲劳裂纹超声红外热像检测与识别方法研究[J]. 红外与激光工程,2022,51(3):475-483. LIN Li,LIU Xin,ZHU Junzhen,et al. Research on vibrothermography detection and recognition method of metal fatigue cracks based on CNN[J]. Infrared and Laser Engineering,2022,51(3):475-483.
[22]	杨凯,罗帅,王勇,等. 基于U−Net的列车轮对激光曲线提取[J]. 无损检测,2021,43(1):19-23. YANG Kai,LUO Shuai,WANG Yong,et al. Laser curve extraction of train wheelset based on U-Net[J]. Nondestructive Testing,2021,43(1):19-23.