Application of methods of system analysis and machine learning to automate the reading of drawings


Аuthors

Devyatov D. A.1, Koroleva M. R.2*, Lyalin M. S.1, Mishchenkova O. V.1, Chernova A. A.1**

1. Kalashnikov Izhevsk State Technical University, 7, Studencheskaya str., Izhevsk, 426069, Russia
2. Udmurt Federal Research Center Ural Branch of the Russian Academy of Sciences, 34 Tatiana Baramzina St., Izhevsk, 426067

*e-mail: koroleva@udman.ru
**e-mail: alicaaa@gmail.com

Abstract

The work is devoted to the issues of building an effective and universal approach to automating the recognition of engineering drawings. The aim of the work is the construction and development of effective algorithms for recognising graphical information from drawings for the subsequent formation of initial information used in the automation of technological preparation of production. In the proposed approach the recognition of information from engineering drawings is implemented through the joint application of methods of system analysis of image objects in drawings, an adapted method of rules and mathematical apparatus of machine learning - a specialised neural network. The systematisation and classification of elements of drawings has been carried out, which allowed to identify common characteristics and features necessary for their structured description. An algorithm for identification of technological elements implemented using a neural network is proposed. The model is adapted for searching and analysing key structural components in drawings, demonstrating its effectiveness under different data types. The formulated approach and methods are tested on the task of finding and describing holes in drawings. The validation revealed that the training model requires further training on a more massive sample. In the current version of the model makes some errors in pattern recognition, the accuracy of classification of the drawing element provided by the neural network, when retested on the validation sample, reaches 80%, which is acceptable for the prototype.
Thus, the proposed approach allows to automate the process of reading drawings and allows to form the initial information for automation of technological preparation of production. Thus, a new approach to the automation of engineering drawings analysis is formed and proposed in the work. The testing of the approach has shown that the developed system achieves the minimum accuracy required for the prototype when recognising and analysing drawing elements, which confirms its effectiveness and practical applicability. The obtained results lay the foundation for further development of systems for automation of drawing analysis, providing integration with modern engineering processes.

Keywords:

drawing, automation, algorithm, machine learning, neural networks

References

  1. Devyatov D.A., Chernova A.A. Assessment of the possibility of automating the formation of technological processes in small-scale production. Vestnik IzhGTU imeni M.T. Kalashnikova. 2023. V. 26, No 3. P. 67-74. (In Russ.). DOI: 10.22213/2413-1172-2023-3-67-74
  2. Balashov A.I., Koshevoi V.V., Savchenko E.A. Upravlenie zhiznennym tsiklom produktsii (Product Lifecycle Management). Moscow: Mashinostroenie Publ., 2003. 
  3. Polyakov E.Yu., Kondusov D.V., Kondusova V.B. Unified model of the process of automating the coordination of documentation for technological preparation of production. Avtomatizatsiya v promyshlennosti. 2023. No. 6. P. 56–58. (In Russ.). DOI: 10.25728/ avtprom.2023.06.13 
  4. Popov V.N., Batyreva A.B., Kovtun L.I., Boyarov I.P. Automation of accumulating experience in optimizing technological processes and minimizing risks of distributed virtual production during the operation of a technological preparation service. Ekonomika. Pravo. Innovatsii. 2019. No. 1. P. 42–48. (In Russ.)
  5. Khrustaleva I.N., Lyubomudrov S.A., Romanov P.I. Automation of technological preparation for single and small-scale production. Nauchno-tekhnicheskie vedomosti SPBPU. Estestvennye i inzhenernye nauki. 2018. V. 24, No. 1. P. 113–121. (In Russ.). DOI: 10.18721/JEST.240111
  6. Kordyukov A.V., Savchenkov M.I., Chernova A.A. Adaptation of the algorithm for the operation of a computer-aided process planning system for technological preparation in small-scale and single production. Vestnik IzhGTU imeni M.T. Kalashnikova. 2023. V. 26, No. 3. P. 26-32. (In Russ.)
  7. Aleksandrov I.A. Principles of automating technological preparation of production through neural network modeling. Inzhenernyi vestnik Dona. 2019. No. 5 (56). 12 p. (In Russ.)
  8. Chigrinets E.G., Verchenko A.V. CAD/CAM/CAE systems, OMV technologies and neural network based data analysis algorithms at the aviation industry enterprises. Trudy MAI. 2019. No. 104. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=102420
  9. Yasutomo Sugisawa, Keigo Takasugi, Naoki Asakawa. Machining sequence learning via inverse reinforcement learning. Precision Engineering journal. 2022. V. 73, P. 477-487. DOI:10.1016/J.PRECISIONENG.2021.09.017
  10. Kasimov D.R., Chernova A.A., Koroleva M.R. Automatic identification and description of surfaces in 2D drawings. Avtomatizatsiya v promyshlennosti. 2024. No.1. P. 55-58. (In Russ.). DOI: 10.25728/avtprom.2024.01.14 
  11. Babic B., Nesic N., Miljkovic Z. A review of automated feature recogition with rule-based pattern recognition. Computers Industry. 2008. No. 59 (4). P. 321–337. DOI: 10.1016/j.compind.2007.09.001
  12. Mazin Al-wswasi, Atanas Ivanov, Harris Makatsoris. A survey on smart automated computer-aided process planning (ACAPP) techniques. The International Journal of Advanced Manufacturing Technology. 2018. V. 97 (1), P. 809-832. DOI: 10.1007/s00170-018-1966-1
  13. Sakurai H., Dave P. Volume decomposition and feature recognition. Part II: Curved objects. Computer-Aided Desing. 1996. V. 28, P. 519–537.
  14. Ning F., Shi Y., Cai M., Xu W., Zhang X. Manufacturing cost estimation based on the machining process and deep-learning method. Journal of Manufacturing Systems. 2020. V. 56, P. 11–22. URL: https://doi.org/10.1016/j.jmsy.2020.04.011
  15. Khomyakov O.O., Panishchev V.S., Titov V.S., Vatutin E.I. Mathematical model and a parallel algorithm for processing images containing symbolic information about products.Trudy MAI. 2024. No. 137. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=181884
  16. Malygin I.V., Bel'kov S.A., Tarasov A.D., Usvyatsov M.R. Machine learning methods in classification of radio signals. Trudy MAI. 2017. No. 96. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=85797
  17. Gusev P.Yu., Tavolzhanskii A.V., Potsebneva I.V. Modeling the selection of time series haracteristics for machine learning methods based on the correlation matrix. Trudy MAI. 2024. No. 138. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=182676
  18. Terven J., Cordova-Esparza D.-M., Romero-González J.-A. A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS. Machine Learning and Knowledge Extraction. 2023. No. 5 (4). P. 1680-1716. DOI: 10.3390/make5040083
  19. Ren Sh., He K., Girshick R., Sun J. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2015. No. 39 (6). P. 1137-1149. DOI: 10.1109/TPAMI.2016.2577031
  20. Hosang J.H., Benenson R., Schiele B. Learning Non-maximum Suppression. IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2017. P. 6469-6477. URL: https://doi.org/10.48550/arXiv.1705.02950


Download

mai.ru — informational site MAI

Copyright © 2000-2025 by MAI

 Вход