Automatic Identification of Flour Types by Absorption Spectra

Introduction: The need for a comprehensive flour quality assessment system, including the presence of impurities, additives, and improvers, involves the use of machine vision and machine learning, an advanced field of artificial intelligence. An important area of obtaining data for analysis is the use of optical spectral methods.

Purpose: To study the optical photoluminescent properties of various types of flour in order to develop methods for automatic identification of the composition of mixtures during storage and production of bakery products.

Materials and Methods: For spectral measurements, flour samples from wheat, rye, oats, rice, peas, buckwheat and chickpeas were used. Optical spectral measurements of the obtained flour samples were carried out in the extended spectral range of 200–500 nm on a CM 2203 diffraction spectrofluorimeter. 

Results: All spectral characteristics contain peaks at the following wavelengths: 290nm, 272 nm, 286 nm, 362 nm, as well as a weak maximum of 424 nm. Rice flour has the strongest excitement, and buckwheat flour has the least. For chickpea flour, strong absorption occurs in the short-wavelength (260–290 nm) and long-wavelength (420 nm or more) regions. The other types of flour studied have approximately similar characteristics. The integral parameters H in the entire studied range of 220–500 nm are determined with an error of up to 10.9 % (for oat flour) and differ by 4.3 times for all the studied types of flour. However, for the design of machine vision systems, it is advisable to identify flour types by integral parameters in the narrower areas of λ1-λ2 and by their ratios. Based on the selected features, classification models were built that showed accuracy of up to 88.9 % during testing, while problematic pairs of classes such as pea and buckwheat flour were identified.

Conclusion: The integral and statistical parameters of the spectra have a high separation ability. The ratio H220-500/H470-500 (85.3 %), kurtosis (84.8 %) and mathematical expectation (84.6 %) are the most complex estimates, which are recommended to be used as the basis for constructing classification algorithms. Practical testing on machine learning models has confirmed the possibility of automatic identification of flour types with accuracy that meets the requirements of industrial control.

1. Berezina, N. A., Khmeleva, E. V., Artemova, E. N., Osipova, G. A., & Shvedova, M. N. (2019). Modeling of a polycomposite mixture for bakery products. Polzunovsky Bulletin. 1. 11-16. (In Russ.).

2. Vershinina, O. L., Gonchar, V. V., Roslyakov, Yu. F., & Tychina, A.V. (2019). The use of gluten-free buckwheat flour in the production of bread from a mixture of rye and wheat flour. News of higher educational institutions. Food Technology. 5-6(371-372). 35-38. (In Russ.) https://doi.org/10.26297/0579-3009.2019.5-6.8.

3. Donskaya, G. A., Bychkova, T. S., & Yurova, E. A. (2024). Investigation of the radioprotective potential of flour from oilseeds. Polzunovsky Bulletin. 4. 86-92. (In Russ.) https://doi.org/10.25712/ASTU.2072-8921.2024.04.013

4. Zhazykbaeva, G. M. (2022). A modern method of flour quality control using INFRANEO. Innovative Engineering and Technology. 9(4). 17-21. (In Russ.).

5. Kladov, E. A., Yesin, S. B., & Egorova, E. Y. (2022). Computer image processing and their interpretation in the analysis of the granulometric composition of oilseed flour. Polzunovsky Bulletin. 4-1. 48-56. (In Russ.) https://doi.org/10.25712/ASTU.2072-8921.2022.04.006.

6. Kradetskaya, O. O., Chilimova, I. V., Dashkevich, S. M., Utebaev, M.O., & Kairzhanov, E.K. (2022). The relationship of the color differences of flour and grain with the ash content and strength of flour of various varieties of soft wheat. Bulletin of Science of the Kazakh Agrotechnical University named after S. Seifullin. 2-1(113). 24-34. (In Russ.) https://doi.org/10.51452/kazatu.2022.2(113).951.

7. Kunasheva, J. M., & Kodzokova, M. H. (2018). The use of crushed rice flour in food technology. New Technologies. 4. 62-67. (In Russ.).

8. Lebedev, D. V., Rozhkov, E. A., Leonov, V. A., & Malnev, I. D. (2019). The use of electrotechnological optoelectronic methods in bakery production to determine the quality of wheat flour and the concentration of flour dust in the air. AgroEcoInfo. 4(38). 35. (In Russ.).

9. Miroshina, T. A., & Reznichenko, I. Y. (2025). The potential of buckwheat flour in the production of gluten-free bakery, pasta and confectionery products. Bulletin of KrasGAU. 8(221). 204-216. (In Russ.) https://doi.org/10.36718/1819-4036-2025-8-204-216 .

10. Nikonorova, Yu. Yu., Volkova, A.V., & Makushin, A. N. (2021). Investigation of rheological properties of dough and bread made from a mixture of premium wheat flour and sorghum flour. Bulletin of KrasGAU. 4(169). 155-160. (In Russ.) https://doi.org/10.36718/1819-4036-2021-4-155-160.

11. Pochitskaya, I. M. (2022). Scientific and practical foundations of the development of a comprehensive food quality assessment system: specialty 05.18.15 "Technology and commodity science of food products and functional and specialized purposes and catering": dissertation for the degree of Doctor of Technical Sciences, 514. (In Russ.).

12. Rukshan, L. V., & Novozhilova, E. S. (2023). Investigation of the technological properties of flour obtained from lupine seeds processed in various ways. Food Industry, 6, 30-32. (In Russ.) https://doi.org/10.52653/PPI.2023.6.6.009.

13. Serebrenikova, E. S., & Anisimova, L. V. (2022). The quality of sorghum flour and the rheological properties of a dough made from a mixture of wheat and sorghum flour. Polzunovsky Bulletin. 3. 71-80. (In Russ.) https://doi.org/10.25712/ASTU.2072-8921.2022.03.010.

14. Chernykh, V. Ya., Sarbashev, K. A., Shulenin, A.V., & Zhirnova, E. V. (2017). Determination of the color characteristics of wheat flour in the production of bakery and pasta products. Bread products. 2. 44-47. (In Russ.).

15. Shabolkina, E. N., Shevchenko, S. N., & Anisimkina, N. V. (2021). The results of studying the biochemical properties of bare oat grain and the physical properties of dough mixtures with wheat flour. Grain Industry of Russia. 4(76). 61-65. (In Russ.). https://doi.org/10.31367/2079-8725-2021-76-4-61-65.

16. Steinberg, T. S., & Morozova, O. V. (2021). Evaluation of the metrological characteristics of a scanning flour analyzer in determining the optical properties of components by digital image analysis. Bread products. 7. 36-38. (In Russ.) https://doi.org/10.32462/0235-2508-2021-30-7-36-38 .

17. Akın, P.A., Sezer, B., Bean, S.R., Peiris, K., Tilley, M., Apaydın, H. & Boyacı, İ.H. (2021). Analysis of corn and sorghum flour mixtures using laser-induced breakdown spectroscopy. J Sci Food Agric, 101, 1076-1084. https://doi.org/10.1002/jsfa.10717

18. Brabec, D., Grothe, S., Perez-Fajardo, M., Pordesimo, L., & Yeater, K. (2024). Potential of a flatbed scanner for evaluation of flour samples for dark specks and flour color. Cereal Chemistry, 101, 508–517. https://doi.org/10.1002/cche.10758

19. Guo, X., Meng, X., Li, L., Qiu, X., Li, C., Guo, G., Sun, X., Gong, T., & Tian, Y. (2024). A moisture detection sensor for flour based on the gas in scattering media absorption spectroscopy. Microw Opt Technol Lett, 66, e34011. doi:10.1002/mop.34011

20. He, P., Wu, Y., Wang, J., Ren, Y., Ahmad, W., Liu, R., Ouyang, Q., Jiang, H., & Chen, Q. (2020). Detection of mites Tyrophagus putrescentiae and Cheyletus eruditus in flour using hyperspectral imaging system coupled with chemometrics. J Food Process Eng. 43, e13386. https://doi.org/10.1111/jfpe.13386

21. Hernández-Aguilar, C., Domínguez-Pacheco, A., Cruz-Orea, A., & Ivanov, R. (2019). Photoacoustic Spectroscopy in the Optical Characterization of Foodstuff: A Review. Journal of Spectroscopy, 5920948. https://doi.org/10.1155/2019/5920948

22. Jenfa, M. D., Adelusi, O. A., Aderinoye, A., Coker, O. J., Martins, I. E., & Obadina, O. A. (2024). Physicochemical compositions, nutritional and functional properties, and color qualities of sorghum–orange-fleshed sweet potato composite flour. Food Science & Nutrition, 12, 2364–2378. https://doi.org/10.1002/fsn3.3922

23. Nagel-Held, J., Kaiser, L., Longin, C. F. H., & Hitzmann, B. (2022). Prediction of wheat quality parameters combining Raman, fluorescence, and near-infrared spectroscopy (NIRS). Cereal Chemistry, 99, 830–842. https://doi.org/10.1002/cche.10540

24. Pompeu, D.R., Souza, J.N.S. & Pena, R.S. (2021), Fast and non-destructive determination of simultaneous physicochemical parameters of Manihot esculenta flour using FT-NIR spectroscopy and multivariate analysis. Int. J. Food Sci. Technol., 56, 3826-3833. https://doi.org/10.1111/ijfs.14998

25. Rai, D., Rai, A.K., Rai, A.K., Singh D. B., & Yadav A. K. (2025). LIBS–a promising technique for control of food quality. Journal of Optics. https://doi.org/10.1007/s12596-024-02436-2

26. Shao, Y., Xuan, G., Hu, Z., & Wang, Y. (2019). Detection of adulterants and authenticity discrimination for coarse grain flours using NIR hyperspectral imaging. J Food Process Eng. 42, e13265. https://doi.org/10.1111/jfpe.13265

27. Yue, M., Zhang, S., Zhang, J., Wang, S., Yu, X., Li, H., Yin, X. & Ma, C. (2023). Potato flour content determination of potato–wheat flour mixture based on hyperspectral imaging. Int J Food Sci Technol, 58, 4780-4789. https://doi.org/10.1111/ijfs.16587