Automatic Particle Recognition Based on Digital lmage Processing

The purpose of the research is to develop and compare various methods and algorithms for effective particle analysis based on their visual characteristics. Тhe purpose of this article is to develop and compare various methods and algorithms for effective particle analysis based on their visual characteristics. Тhe paper considers two fundamentally different approaches: the analysis of grayscale gradients and the machine learning method.
Methods.Тhe research methodology includes the analysis of particle images obtained by precipitation from colloidal solutions after laser ablation and images of powder particles for selective laser melting. Тhe materials were obtained using a Quanta 200 3D electron microscope (FЕ/). For the analysis, threshold brightness binarization, contour recognition methods by the Kenny operator and the Hough algorithm are used to combine boundary points into connected contours. For comparison, the U-Net neural network solution was used, and a dataset generator was created to train the neural network. Hand-cut images of aluminum alloy powder particles and micro and nanoparticles of various metals are used as data for generation.
Results.Тhe results of the study show that the Hough method provides recognition of the number of particles at the level of 80%, and the machine learning method achieves 95% accuracy in recognizing the shape of particles. Both methods can be used to analyze microand nanoparticles, including irregularly shaped particles.
Conclusion.Тhe findings of the work confirm that neural networks are the optimal solution for automatic particle recognition in digital images. However, in order to create a dataset of sufficient volume, it is necessary to develop a generator of labeled images, which requires a detailed study of the subject area.

1. Gulyaev P. V., Shelkovnikov E. Yu., Ermolin K. S. Obrabotka i raspoznavanie izobrazhenij repernyh otmetok dlya lokalizacii zonda v skaniruyushchem tunnel'nom mikroskope [Image processing and recognition of reference marks for probe localization in a scanning tunneling microscope]. Himicheskaya fizika i mezoskopiya = Chemical Physics and Mesoscopy, 2018, vol. 20, no. 3, pp. 437–445.

2. Myasnyankin M. B., Kuzmin A. A., Serebrovskiy V. V., Aldokhin E. A. Mnogopotochnaya arhitektura programmnogo obespecheniya obrabotki mnogokanal'nyh medicinskih signalov [Multithreaded software architecture for processing multicameral medical signals]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta. Seriya: Upravlenie, vychislitel'naya tekhnika, informatika. Medicinskoe priborostroenie = Proceedings of the Southwest State University. Series: Control, Computer Engineering, Information Science. Medical Instruments Engineering, 2022, vol. 12, no. 2, pp. 76–97.

3. Allaberganov A. A. Metod raspoznavaniya tekstovoj informacii iz cifro-voj formy (izobrazhenie), opredelenie ocherednosti naneseniya rekvizitov [The method of recognizing textual information from a digital form (image), determining the order of application of details]. Zakon i pravo = Law and Law, 2020, no. 4, pp. 187–193.

4. Gundin A. A., Gundina M. A., Cheshkin A. N. Obrabotka cifrovyh izobrazhe-nij pri defektoskopii poverhnostej promyshlennyh ob"ektov [Processing of digital images during flaw detection of surfaces of industrial objects]. Nauka i tekhnika = Science and Technology, 2016, vol. 15, no. 3, pp. 225–232.

5. Lapko A. V., Lapko V. A. Neparametricheskie algoritmy ocenivaniya sostoyanij prirodnyh ob"ektov [Nonparametric algorithms for determining the position of primary objects]. Avtometriya = Autometry, 2018, vol. 54, no. 5, pp. 33–38. https://doi.org/10.15372/AUT20180504

6. Arlazarov V. V., Bulatov K. B., Uskov A. B. Model' sistemy raspoznavaniya ob"ektov v videopotoke mobil'nogo ustrojstva [Model of the object recognition system in the video stream of mobile device]. Trudy Instituta sistemnogo analiza Rossijskoj akademii nauk = Proceedings of the Institute of System Analysis of the Russian Academy of Sciences, 2018, special is., pp. 73–82.

7. Borzov S. M., Potaturkin O. I. Issledovanie effektivnosti spektral'no-prostranstvennoj klassifikacii dannyh giperspektral'nyh nablyudenij [Investigation of the effectiveness of spectral-spatial classification of hyperspectral observation data]. Avtometriya = Autotometry, 2017, vol. 53, no. 1, pp. 32–42.

8. Segedi S., Ioffe S., Vanhoek V., eds. Inception-v4, Inception-Resnet and the influence of residual connections on learning. Proceedings of the AAAI conference on Artificial intelligence. San Francisco, California, USA, AAAI Press Publ., 2017, pp. 4278–4284.

9. Jianjun Ni, Zubair Khan, Shihao Wang, Kang Wang and Syed Kamran Haider. Automatic detection and counting of overlapping circular objects using circular Hough transform and contour detection. 12th World Congress on Intelligent Control and Automation (WCICA). Guilin, China, IEEE Publ., 2016. pp. 2902–2906.

10. Konanykhin A. Yu., Konanykhina T. N., Panishchev V. S. Metody uluchsheniya vydelennoj oblasti izobrazheniya pri bystrodejstvuyushchej obrabotke simvol'noj informacii [Methods of improving the divided image area with high-speed processing of symbolic information]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta Seriya: Upravlenie, vychislitel'naya tekhnika, informatika. Medicinskoe priborostroenie = Proceedings of the Southwest State University. Series: Control, Computer Engineering, Information Science. Medical Instruments Engineering, 2021, vol. 11, no. 4, pp. 106–116.

11. Mestetsky L. Nepreryvnaya morfologiya binarnyh izobrazhenij. Figury, skelety, cirkulyary [Continuous morphology of binary images. Figures, skeletons, circulars]. Moscow, Fizmatlit Publ., 2009. 286 p.

12. Panishchev V. S., Trufanov M. I., Dobroserdov O. G., Khomyakov O. O. Raspoznavanie simvol'noj informacii dlya avtomatizacii proizvodstvennyh processov [Recognition of symbolic information for automation of production processes]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University, 2021, no. 25(1), pp. 122–137. https://doi.org/10.21869/2223-1560-2021-25-1-122-137

13. Divya Meena S., Viramachaneni Gayatri Shiva Samiraja, Nagineni Sai Lasya, Mada Satvika, Veluru Harshita, J. Shila. Architecture of a hybrid neural network for recognizing objects with multiple labels using feature fusion. Procedia Computer Science, 2022, no. 215, pp. 78–90.

14. Manasvi N. K. Deep learning using Python applications. New York, USA, Springer Science – Business Media Publ., 2018. 219 p.

15. Bougourzi F., Distante C., Dornaika F., Taleb-Ahmed A. PDAtt-Unet: Pyramid DualDecoder Attention Unet for COVID-19 infection segmentation from CT-scans. Medical Image Analysis, 2024, no. 86(17), p. 102797.

16. Law Y. N., Jian H., Lo N. W. S., eds. Low-cost automated whole smear microscopy screening system for detection of acid fast bacilli. PLoS One, 2018, no. 1 (13), pp. 1–8.

17. Polkovnikova N. A. Nejrosetevye sistemy avtomaticheskogo raspoznavaniya morskih ob"ektov [Neural network systems for automated recognition of marine objects]. Ekspluataciya morskogo transporta = Encyclopedia of Marine Transport, 2019, no. 4, pp. 207–219. https://doi.org/10.34046/aumsuomt94/29

18. Abdulkadirov R. I., Lyakhov P. A. Novyj podhod k obucheniyu nejronnyh setej s pomoshch'yu natural'nogo gradientnogo spuska s impul'som na osnove raspredelenij dirihle [A new approach to training neural networks using natural gradient descent with an impulse based on Dirichlet distributions]. Komp'yuternaya optika = Computer Optics, 2023, no. 1, pp. 160–170.

19. Popov G. A., Popova T. A. Sravnenie algoritmov raspoznavaniya ob"ektov [Comparison of object recognition algorithms]. Innovaciya v informatike = Innovation in Computer Science, 2020, vol. 14, no. 2, pp. 21–25.

20. Ronneberger O., Fischer P., Brox T. U-Net: Convolutional networks for segmentation of biomedical images. Medical image processing and computer intervention – MICCAI 2015. Cham, Springer Publ., 2015, pp. 234–241. https://doi.org/10.1007/978-3-319-24574-4_28