Intelligent Identification and Parameter Extraction of Key Joints in Rock （Mass） Based on AttentionR2U-net

doi:10.12068/j.issn.1005-3026.2024.01.013

Abstract

Abstract:

Aiming at the problem of intelligent identification and parameter extraction of key joints in complex joint networks on surfaces of rock （mass）， a method based on AttentionR2U-net network and joint geometric feature model is proposed. The attention gate was introduced on the basis of R2U-net network to improve the network， and then the accuracy test and generalization ability study of the recognition results of slope joint images and fracture images in concrete， cracked soil and common brittle rocks were carried out by qualitative and quantitative methods. Finally， the AttentionR2U-net network was used to identify the key joints by coupling the geometric features of the joints. The geometric parameters of the original and key joints are extracted， and then the trace length， area and inclination angle are analyzed. The results show that for joint identification of rock （mass）， the Dice similarity coefficient of the proposed algorithm is increased from 0.965 to 0.990 in U-net network， outperforming the traditional algorithm significantly. Therefore， the proposed algorithm has stronger reliability and superiority in joint identification of rock （mass）. For crack identification in concrete， cracked soil and brittle rocks such as marble， granite and sandstone， the similarity coefficientDice of the proposed algorithm is above 0.953， so the proposed algorithm has strong generalization ability. Compared with the original joint network， the dominant trace length of the key joint network increased significantly from 0.732 m to 1.835 m， the distribution pattern of joint dip angle and the dominant dip angle group remained unchanged， and the proportion of joints with dominant trace length and dip angle increased significantly.

Key words: rock （mass）, key joint, AttentionR2U-net network, intelligent recognition, parameter extraction

CLC Number:

TU 45

Hao SUN, Zong-sheng DAI, Ai-bing JIN, Yan CHEN. Intelligent Identification and Parameter Extraction of Key Joints in Rock （Mass） Based on AttentionR2U-net[J]. Journal of Northeastern University(Natural Science), 2024, 45(1): 101-110.

Figures/Tables 15

Fig. 1 Production process of rock slope joint dataset

Fig. 2 AttentionR2U‐net network model［23］

Table 1 Computing parameters of AttentionR2U‐net network

名称	参数	名称	参数
CPU	Intel i5-11400H	GPU	NVIDIA Geforce RTX 3050
系统	Windows 10	语言	Python 3.6
RAM	16 GB	框架	Tensorflow和Keras
Cuda	11.1

Table 2 Model superiority evaluation index

指标	含义及作用	公式
二分类交叉损失函数（Loss）	标签图像和预测图像相似程度；用于评估二分类问题中网络分割结果与给定标准结果的相似程度，也可运用反向传播算法对网络的参数进行优化.	$L o s s = - y · l g y ∧ + 1 - y · l g 1 - y ∧$
准确率（Accuracy，Acc）	正确预测出节理和背景的像素占总像素的百分比；反映分类器的分类效果.	$A c c = T P + T N T P + T N + F P + F N$
查准率（Precision，P）	预测像素中真实的节理像素占所有预测为节理像素的百分比；反映被预测为节理的结果中正确的概率.	$P = T P T P + F P$
查全率（Recall，R）	预测像素中真实的节理像素占所有真实节理像素百分比；反映节理被预测出的概率.	$R = T P T P + F N$
灵敏度（Sen）	数值上等于R值，值越大，假负例越少；体现模型对节理的敏感程度.	$S e n = T P T P + F N$
特异性（Spe）	预测出的背景像素占真实背景像素的百分比；体现模型判别背景的能力.	$S p e = T N T N + F P$
Dice相似系数（Dice）	标签图像和预测图像的相似度程度；更直观地反映模型的优劣程度.	$D i c e = 2 × T P (T P + F N) + (T P + F P)$

Table 2 Model superiority evaluation index

指标	含义及作用	公式
二分类交叉损失函数（Loss）	标签图像和预测图像相似程度；用于评估二分类问题中网络分割结果与给定标准结果的相似程度，也可运用反向传播算法对网络的参数进行优化.	$L o s s = - y · l g y ∧ + 1 - y · l g 1 - y ∧$
准确率（Accuracy，Acc）	正确预测出节理和背景的像素占总像素的百分比；反映分类器的分类效果.	$A c c = T P + T N T P + T N + F P + F N$
查准率（Precision，P）	预测像素中真实的节理像素占所有预测为节理像素的百分比；反映被预测为节理的结果中正确的概率.	$P = T P T P + F P$
查全率（Recall，R）	预测像素中真实的节理像素占所有真实节理像素百分比；反映节理被预测出的概率.	$R = T P T P + F N$
灵敏度（Sen）	数值上等于R值，值越大，假负例越少；体现模型对节理的敏感程度.	$S e n = T P T P + F N$
特异性（Spe）	预测出的背景像素占真实背景像素的百分比；体现模型判别背景的能力.	$S p e = T N T N + F P$
Dice相似系数（Dice）	标签图像和预测图像的相似度程度；更直观地反映模型的优劣程度.	$D i c e = 2 × T P (T P + F N) + (T P + F P)$

Fig. 3 Evolution curves of model evaluation index

Fig. 4 Comparison of joint recognition results of different algorithms

Table 3 Quantitative results of predicting different joint images based on AttentionR2U‐net

指标	样品1	样品2	样品3	样品4
TP	4022	10921	3923	4570
TN	58404	51487	58501	57857
FN	19	69	11	34
FP	55	23	65	38
Sen	0.995	0.994	0.997	0.992
Spe	0.999	0.999	0.998	0.999
P	0.986	0.998	0.984	0.992
Dice	0.991	0.996	0.990	0.992

Table 4 Calculation results of each index in the test set

指标	Acc	Sen	Spe	P	Dice
平均	0.997	0.991	0.998	0.990	0.990
最高	0.999	0.998	0.999	0.998	0.997

Fig. 5 Indicators of the results of different algorithms

Fig. 6 Comparison of fracture identification in different fields

Table 5 Diceresults for fracture identification in other fields

算法	混凝土	大理岩	花岗岩	砂岩	龟裂土
本文算法	0.985	0.953	0.974	0.983	0.963
大津法	0.879	0.867	0.375	0.258	0.785
边缘检测法	0.263	0.341	0.185	0.157	0.438
区域生长法	0.846	0.367	0.735	0.786	0.579
血管增强法	0.564	0.459	0.708	0.723	0.738

Fig. 7 Principle of enclosing minimum rectangle

Table 6 Comparison of results of two coupling methods for critical joint identification

方法	关键节理识别率	关键节理精度	单条节理精度
原模型	0.568	0.842	0.990
方法1	0.738	0.916	0.953
方法2	1.000	0.990	0.990
方法3	1.000	0.953	0.953

Fig. 8 Intelligent identification process of key joints

Fig. 9 Histogram of geometrical parameters andprobability density curves of joints

References 30

1	黄润秋.20世纪以来中国的大型滑坡及其发生机制［J］.岩石力学与工程学报，2007，26（3）：433-454.
	Huang Run‐qiu.Large‐scale landslides and their sliding mechanisms in China since the 20th century［J］.Chinese Journal of Rock Mechanics and Engineering，2007，26（3）：433-454.
2	Chen S W， Walske M L， Davies I J.Rapid mapping and analysing rock mass discontinuities with 3D terrestrial laser scanning in the underground excavation［J］.International Journal of Rock Mechanics and Mining Sciences，2018，110：28-35.
3	Ismail A， Ahmad‐Safuan A R， Sa'ari R，et al.Application of combined terrestrial laser scanning and unmanned aerial vehicle digital photogrammetry method in high rock slope stability analysis：a case study［J］.Measurement，2022，195：111161.
4	Agar Ozbek A S， Weerheijm J， van Breugel K.High speed photography technique for measuring impact strength of porous concrete［J］.Construction and Building Materials，2018，186：1092-1104.
5	Sturzenegger M， Stead D.Close‐range terrestrial digital photogrammetry and terrestrial laser scanning for discontinuity characterization on rock cuts［J］.Engineering Geology，2009，106（3/4）：163-182.
6	金爱兵，陈帅军，赵安宇，等.基于无人机摄影测量的露天矿边坡数值模拟［J］.岩土力学，2021，42（1）：255-264.
	Jin Ai‐bing， Chen Shuai‐jun， Zhao An‐yu，et al.Numerical simulation of open‐pit mine slope based on unmanned aerial vehicle photogrammetry［J］.Rock and Soil Mechanics，2021，42（1）：255-264.
7	贾曙光，金爱兵，赵怡晴.无人机摄影测量在高陡边坡地质调查中的应用［J］.岩土力学，2018，39（3）：1130-1136.
	Jia Shu‐guang， Jin Ai‐bing， Zhao Yi‐qing.Application of UAV oblique photogrammetry in the field of geology survey at the high and steep slope［J］.Rock and Soil Mechanics，2018，39（3）：1130-1136.
8	Talab A M A， Huang Z C， Xi F，et al.Detection crack in image using Otsu method and multiple filtering in image processing techniques［J］.Optik，2016，127（3）：1030-1033.
9	Kheradmandi N， Mehranfar V.A critical review and comparative study on image segmentation‐based techniques for pavement crack detection［J］.Construction and Building Materials，2022，321：126162.
10	Nhat‐Duc H， Nguyen Q L， Tran V D.Automatic recognition of asphalt pavement cracks using metaheuristic optimized edge detection algorithms and convolution neural network［J］.Automation in Construction，2018，94：203-213.
11	Wang L T， Gu X Y， Liu Z，et al.Automatic detection of asphalt pavement thickness：a method combining GPR images and improved Canny algorithm［J］.Measurement，2022，196：111248.
12	Kang C C， Wang W J， Kang C H.Image segmentation with complicated background by using seeded region growing［J］.AEU-International Journal of Electronics and Communications，2012，66（9）：767-771.
13	Tang Y D， He L， Lu W，et al.A novel approach for fracture skeleton extraction from rock surface images［J］.International Journal of Rock Mechanics and Mining Sciences，2021，142：104732.
14	Zhu J S， Song J B.Weakly supervised network based intelligent identification of cracks in asphalt concrete bridge deck［J］.Alexandria Engineering Journal，2020，59（3）：1307-1317.
15	Cui X N， Wang Q C， Dai J P，et al.Pixel‐level intelligent recognition of concrete cracks based on DRACNN［J］.Materials Letters，2022，306：130867.
16	柳厚祥，李汪石，查焕奕，等.基于深度学习技术的公路隧道围岩分级方法［J］.岩土工程学报，2018，40（10）：1809-1817.
	Liu Hou‐xiang， Li Wang‐shi， Zha Huan‐yi，et al.Method for surrounding rock mass classification of highway tunnels based on deep learning technology［J］.Chinese Journal of Geotechnical Engineering，2018，40（10）：1809-1817.
17	梁世豪.基于深度学习的野外露头区岩石裂缝识别方法研究［D］.大庆：东北石油大学，2020.
	Liang Shi‐hao.Research on identification method of rock fractures in outcrop area based on deep learning［D］.Daqing：Northeast Petroleum University，2020.
18	Chaiyasarn K， Buatik A， Mohamad H，et al.Integrated pixel‐level CNN‐FCN crack detection via photogrammetric 3D texture mapping of concrete structures［J］.Automation in Construction，2022，140：104388.
19	Chen J Y， Zhou M L， Huang H W，et al.Automated extraction and evaluation of fracture trace maps from rock tunnel face images via deep learning［J］.International Journal of Rock Mechanics and Mining Sciences，2021，142：104745.
20	张紫杉，王述红，王鹏宇，等.岩坡坡面裂隙网络智能识别与参数提取［J］.岩土工程学报，2021，43（12）：2240-2248.
	Zhang Zi‐shan， Wang Shu‐hong， Wang Peng‐yu，et al.Intelligent identification and extraction of geometric parameters for surface fracture networks of rocky slopes［J］.Chinese Journal of Geotechnical Engineering，2021，43（12）：2240-2248.
21	Xie Z Y， Asari V K， Haritashya U K.Evaluating deep‐learning models for debris‐covered glacier mapping［J］.Applied Computing and Geosciences，2021，12：100071.
22	Obeso A M， Benois‐Pineau J， García Vázquez M S，et al.Visual vs internal attention mechanisms in deep neural networks for image classification and object detection［J］.Pattern Recognition，2022，123：108411.
23	于文玲，刘波，刘华，等.基于Attention Gates和R2U‐net的遥感影像建筑物提取方法［J］.地理与地理信息科学，2022，38（3）：31-36，42.
	Yu Wen‐ling， Liu Bo， Liu Hua，et al.Building extraction from remote sensing images based on the R2U‐net model and attention gates［J］.Geography and Geo-Information Science，2022，38（3）：31-36，42.
24	Turkan Y， Hong J， Laflamme S，et al.Adaptive wavelet neural network for terrestrial laser scanner‐based crack detection［J］.Automation in Construction，2018，94：191-202.
25	Zheng M J， Lei Z J， Zhang K.Intelligent detection of building cracks based on deep learning［J］.Image and Vision Computing，2020，103：103987.
26	Zhao Y X， Sun B， Liu S M，et al.Identification of mining induced ground fissures using UAV and infrared thermal imager：temperature variation and fissure evolution［J］.ISPRS Journal of Photogrammetry and Remote Sensing，2021，180：45-64.
27	Tang C A， Webb A A G， Moore W B，et al.Breaking earth’s shell into a global plate network［J］.Nature Communications，2020，11：3621.
28	Shi Y， Cui L M， Qi Z Q，et al.Automatic road crack detection using random structured forests［J］.IEEE Transactions on Intelligent Transportation Systems，2016，17（12）：3434-3445.
29	金爱兵，陆通，王本鑫，等.基于力学等效的岩体关键节理迹长阈值研究［J］.岩石力学与工程学报，2022，41（5）：904-915.
	Jin Ai‐bing， Lu Tong， Wang Ben‐xin，et al.Study on the threshold of key joint trace length in rock mass based on mechanical equivalence［J］.Chinese Journal of Rock Mechanics and Engineering，2022，41（5）：904-915.
30	Domínguez C， Heras J， Pascual V.IJ‐OpenCV：combining ImageJ and OpenCV for processing images in biomedicine［J］.Computers in Biology and Medicine，2017，84：189-194.

[1]	GUO Jun¹, YIN Wan-zhong¹, YANG Bin², ZHU Zhang-lei³. Effect of Acid Etching on the Floatability of Dolomite and Its Mechanism of Action [J]. Journal of Northeastern University(Natural Science), 2024, 45(1): 93-100.
[2]	JIA Peng, MAO Song-ze, QIAN Yi-jin, LU Jia-liang. Dynamic Splitting Characteristics of Freeze-Thawed Sandstone at Different Loading Rates [J]. Journal of Northeastern University(Natural Science), 2024, 45(1): 111-119.