Results of primary interferogram processing for machine learning model construction
DOI:
https://doi.org/10.33216/1998-7927-2023-279-3-11-15Keywords:
data preparation, interferogram,, image processingAbstract
The work is dedicated to processing interferometric videos and images for disciplines that require precise and dynamic measurement of physical parameters. The application of machine learning models enhances the analysis of interferometric data, making their utilization more efficient and accurate. Interferometry [6-10] is employed for various measurements based on interference phenomena to determine statistical and dynamic object parameters. Measurement of static parameters using interferometry may include surface height, surface deformation, material layer thickness, optical properties such as light transmission or reflection coefficients, material stress and deformation, object positioning, and angular measurements. Interferometers can measure dynamic parameters such as the speed and direction of movement of objects in transportation systems and biological cells. The integration of machine learning methods into interferogram analysis can significantly improve the efficiency and accuracy of results, especially in conditions involving large datasets and complex patterns. Key tasks for applying machine learning methods include noise filtration, object segmentation, change prediction, artifact correction, and data processing optimization. To build a machine learning model, it is necessary to investigate real interferometric patterns, determine the model's key parameters, and implement automatic image processing methods. Thus, the study explores real interferometric patterns, provides their description, and automates the process of determining their quality, offering recommendations for application. In this work, we analyze a video experiment to obtain an interferometric image using the Mach-Zehnder interferometer. Real interferometric patterns obtained from the video recording of the Mach-Zehnder interferometer operation are investigated. The behavior of interferometric fringes along the X and Y axes is analyzed, identifying specific areas and characteristics of the real signal in the interferogram cross-section. Based on the results obtained, image segmentation is performed, determining the characteristic behavior of light over time in these segments. The study identifies which segments are most and least suitable for analysis and within what time intervals.
References
1. Andrew Wolf, Machine Learning Simplified: A gentleintroduction to supervised learning, 2022, p. 199URL:https://www.goodreads.com/book/show/60297347-machine-learning-simplified
2. Maheshwari A. Business intelligence and data mining. –Business Expert Press, 2014. URL:https://www.amazon.com/Business-Intelligence-Data-Mining-Analytics/dp/1631571206
3. Yang X. S. Introduction to Algorithms for Data Mining andMachine Learning. – Academic Press, 2019. Doi:https://doi.org/10.1016/C2018-0-02034-4
4. Fernandes M. Statistics for business and economics. –Bookboon, 2008. URL: https://bookboon.com/en/statisticsfor-business-and-economics-ebook?mediaType=ebook
5. Ryazantsev O., Khoroshun G., Riazantsev A., Ivanov V.,Baturin A. Statistical Optical Image Analysis forInformation System // Proceedings of 2019 7thInternational Conference on Future Internet of Things andCloud Workshops (FiCloudW), Istanbul, Turkey, IEEE2019, pp. 130-134, DOI: 10.1109/FiCloudW.2019.00036
6. Senthilkumaran, P., Masajada, J., & Sato, S. (2012). Interferometry with vortices. International Journal of Optics, 2012.
7. Bekshaev, A. Y., Basistiy, I. V., Slyusar, V. V., Soskin, M.S., & Vasnetsov, M. V. (2002). Observation of therotational Doppler effect with an optical-vortex one-beaminterferometer. Укр. фіз. журн., 47(11), 1035-1040.
8. Khoroshun, Anna. "Shearing Interferometers with aSingular Source of Light." Interferometers: Fundamentals,Methods and Applications (2015): 253-267.
9. Angelsky, O. V., Mokhun, I. I., Mokhun, A. I., & Soskin,M. S. (2002). Interferometric methods in diagnostics ofpolarization singularities. Physical Review E, 65(3),036602.
10. Bekshaev, A. Ya, et al. "Spectral characteristics of theFabry-Perot interferometer transmission upon illuminationby an arbitrary light beam." arXiv preprintarXiv:1812.11687 (2018).
