Pit wall jointing mapping using neural network capacities

Capacities of convolutional neural networks in automatic discrimination of joint sets in rock mass are described as a case-study of aerial photo interpretation data on the Pereval marble pit in the Irkutsk Region. The research used the software products Pix4DMapper, Cloud 3D and Geomix, as well as the proprietary codes written in Python using the libraries Pillow, Ultralytics, Scipy and Numpy. The methods of the computer vision in jointing mapping, as compared with the manual identification of joints in a cloud of points, allow more accurate and complete imaging of pit wall jointing. The issues connected with the preliminary treatment of images before neural network-based detection, including adjustment of gridding, are discussed. The procedure is developed for the coordination of joints discriminated by neural networks in vertical projections of orthophotos and for fixing them in a three-dimensional space through a dense cloud of points. The difficulties faced by the authors when plotting vertical projections of orthophotos are discussed, and the potential areas of the further research are described. The studies prove applicability of convolutional neural networks in automatic mapping of joint sets, which promotes higher safety and optimization of mining operations.

Skorobogatko M. R., Batugin A. S., Belyaev E. N. Pit wall jointing mapping using neural network capacities. MIAB. Mining Inf. Anal. Bull. 2024;(11):75-87. [In Russ]. DOI: 10.25018/0236_1493_2024_11_0_75.

M.R. Skorobogatko¹, Researcher, e-mail: mikhailskorobogatko@gmail.com, ORCID ID: 0009-0001-6517-6424,
A.S. Batugin, Dr. Sci. (Eng.), Professor, NUST MISIS, 119049, Moscow, Russia, e-mail: batugin.as@misis.ru, ORCID ID: 0000-0002-9227-1160,
E.N. Belyaev¹, Senior Lecturer, e-mail: belyaeven@ex.istu.edu,
¹ Irkutsk National Research Technical University, Center for Surveying and Geodetic Innovation, 664074, Irkutsk, Russia.

M.R. Skorobogatko, e-mail: mikhailskorobogatko@gmail.com.

1. Sashurin A. D., Panzhin A. A. The role of geomechanics in ensuring safe and efficient subsoil use. Sovremennye problemy mekhaniki. 2018, no. 33(3), pp. 92—101. [In Russ].

2. Chen Na, Wu Xiaochan, Xiao Henglin, Yao Chi, Cheng Yuke Semi-automatic recognition of rock mass discontinuity based on 3D point clouds. Discover Applied Sciences. 2024, vol. 6. DOI: 10.1007/s42452-024-05876-4.

3. Jalalian Mohammad Hossein, Bagherpour Raheb, Khoshouei Esfahani Mehrbod Environmentally sustainable mining in quarries to reduce waste production and loss of resources using the developed optimization algorithm. Scientific Reports. 2023, vol. 13. DOI: 10.1038/s41598-023-49633-w.

4. Boos Yu., Kosolapov A. I., Yunakov Yu. L. [et al.] Structural mapping of slopes using multicopters and comparative analysis of modern methods for studying fracturing. Mine Surveying Bulletin. 2021, no. 5-6(144-145), pp. 105—111. [In Russ].

5. Boos Yu., Rudenko E. A., Razin A. I. [et al.] Determination of crack occurrence elements using UAVs. Moscow economic journal. 2020, no. 9, pp. 15. [In Russ]. DOI: 10.24411/2413-046X-202010640.

6. Cheng Z., Gong W., Tang H., Juang CH., Deng Q., Chen J., Ye X. UAV photogrammetry-based remote sensing and preliminary assessment of the behavior of a landslide in Guizhou China. Engineering Geology. 2021, vol. 289, article 106172. DOI: 10.1016/j. enggeo.2021.106172.

7. Kalugina A. M., Kirikov D. A., Skorobogatko M. R. Application of photogrammetry to study the structural features of the massif. Problemy razrabotki mestorozhdeniy uglevodorodnykh i rudnykh poleznykh iskopaemykh. 2023, vol. 2, pp. 52—56.

8. Sergunin M. P., Eremenko V. A. Learning of neural network to predict overlying rock mass displacement parameters by the data on jointing in terms of the Zapolyarny Mine. MIAB. Mining Inf. Anal. Bull. 2019, no. 10, pp. 106—116. [In Russ]. DOI: 10.25018/0236-1493-2019-10-0-106-116.

9. Sergunin M. P., Eremenko V. A. Application of artificial intelligence algorithm in processing structural geological models. MIAB. Mining Inf. Anal. Bull. 2023, no. 9, pp. 2—13. [In Russ]. DOI: 10.25018/0236_1493_2023_9_0_2.

10. Semukhin M. V., Yadryshnikova O. A. Serkin M. F., Chertina K. N. Development of a fracture recognition system to read off experimental data from petrographic thin section images. Oil Industry. 2017, no. 5, pp. 27—31. DOI: 10.24887/0028-2448-2017-5-27-31.

11. Chen Jiayao, Huang H., Zhou Mingliang, Chaiyasarn Krisada Towards semi-automatic discontinuity characterization in rock tunnel faces using 3D point clouds. Engineering Geology. 2021, vol. 291, article 106232. DOI: 10.1016/j.enggeo.2021.106232.

12. Denisenko A. A. Solving the problem of binary classification using convolutional neural networks using the Tensorflow framework. Aktual'nye voprosy tekhnicheskikh nauk: Materialy V Mezhdunarodnoy nauchnoy konferentsii [Current issues of technical sciences: Proceedings of the V International Scientific Conference], Saint-Petersburg, 2019. [In Russ].

13. Nikolaev E. I., Zakharov V. V., Zakharova N. I. Technology for training deep models on a synthesized data set for recognition and detection of gestures. Vestnik Severo-Kavkazskogo federal'nogo universiteta. 2017, no. 4(31), pp. 237—240. [In Russ].

14. Lee Chung-Hong, Yang Hsin-Chang, Wang Shih-Hao An image annotation approach based on bag-of-keypoints for geospatial location search. 25th International Conference on Computers and Their Applications, 2010, CATA. 2010, pp. 37—42.

15. Liu Xiangquan, Huang Xiaoming Weakly supervised salient object detection via boundingbox annotation and SAM model. Electronic Research Archive. 2024, vol. 32, pp. 1624—1645. DOI: 10.3934/era.2024074.

16. Van den Broeck Wouter, Goedemé Toon Combining deep semantic edge and object segmentation for large-scale roof-part polygon extraction from ultrahigh-resolution aerial imagery. Remote Sensing. 2022, vol. 14, article 4722. DOI: 10.3390/rs14194722.

17. Pchelintsev S. Y., Liashkov M. A., Kovaleva O. A. Method for creating synthetic data sets for training neural network models for object recognition. Informatsionno-upravliaiushchie sistemy. 2022, no. 3, pp. 9—19. [In Russ]. DOI: 10.31799/16848853-2022-3-9-19.

18. Molin A. E., Blekanov I. S., Mitrofanov E. P., Mitrofanova O. A. Methods for generating synthetic data for training neural networks in the task of segmenting the level of nitrogen regime of plants in images of unmanned aerial vehicles on an agricultural field. Vestnik Sankt-Peterburgskogo Universiteta, Prikladnaya Matematika, Informatika, Protsessy Upravleniya. 2024, vol. 20, no. 1, pp. 20—33. [In Russ]. DOI: 10.21638/11701/spbu10.2024.103.

19. Kurakin A. V., Abramov N. S., Karpesh S. V., Khachumov V. M. Method for synthesizing an improved image from a series of Earth remote sensing images. Modern high technologies. 2019, no. 102, pp. 268—272. [In Russ]. DOI: 10.17513/snt.37735.

20. Panishchev V. S. Algorithm for restoring image pixels based on a neural network. Proceedings of the South-West state university. 2018, vol. 22, no. 6(81), pp. 183—188. [In Russ]. DOI: 10.21869/22231560-2018-22-6-183-188.