University of Tehran Press Iranian Journal of Biosystem Engineering 2008-4803 55 2 2024 06 21 Application of fuzzy clustering algorithm and hyperspectral images for rice authentication Application of fuzzy clustering algorithm and hyperspectral images for rice authentication 79 92 100478 10.22059/ijbse.2025.384493.665571 FA Mahsa Edris Ph.D. student, Mechanical Engineering of Biosystems Department, Faculty of Agriculture, Shahrekord University, Shahrekord,, Iran 0009-0006-9701-5699 Mahdi Ghasemi-Varnamkhasti Associate Professor, Mechanical Engineering of Biosystems Department, Faculty of Agriculture, Shahrekord University, Shahrekord, Iran 0000-0001-6339-2062 Sajad Kiani Assistant Professor, Biosystems Engineering Department, Faculty of Agriculture, Sari University of Agricultural Sciences and Natural Resources, Sari, Iran 0000-0003-4152-2991 Hassan Yazdanpanah Associate Professor, Toxicology and Pharmacology Dept., School of Pharmacy, Shahid Beheshti University of Medical Sciences, Tehran, IR Iran 0000-0002-2509-5013 Zahra Izadi Associate Professor, Mechanical Engineering of Biosystems Department, Faculty of Agriculture, Shahrekord University, Shahrekord, Iran 0000-0002-1730-9851 Journal Article 2024 10 28 Rice is a vital and strategic product that is used as a major food source. The high demand for purchasing and consuming rice leads to the adulteration of this product globally. Hence, a non-destructive and rapid method is needed to verify the authenticity of rice. Hashemi rice, a high-priced and high-quality rice in the market, is combined with rice such as Neda and Shiroudi, which are very similar in shape but lower in quality and price than Hashemi rice. This study used hyperspectral imaging (HSI) coupled with a fuzzy clustering algorithm to assess adulteration in Hashemi rice samples. First, to reduce the data's dimensionality, the principal component analysis method was applied to the preprocessed data using the multiplicative dispersion correction and Savitzky-Golay methods. Then, the fuzzy unsupervised clustering algorithm was applied using the whole spectrum wavelength (400-1000 nm). It was able to separate the original sample from the adulterated samples well. Also, the fuzzy membership diagram separated the original and self-adulterated samples, mixing 5% to 50%, confirming the correctness and capability of the fuzzy method. Therefore, the HSI system with fuzzy unsupervised algorithms can be used as a reliable and out-of-laboratory method for rapid rice authenticity evaluation. Rice is a vital and strategic product that is used as a major food source. The high demand for purchasing and consuming rice leads to the adulteration of this product globally. Hence, a non-destructive and rapid method is needed to verify the authenticity of rice. Hashemi rice, a high-priced and high-quality rice in the market, is combined with rice such as Neda and Shiroudi, which are very similar in shape but lower in quality and price than Hashemi rice. This study used hyperspectral imaging (HSI) coupled with a fuzzy clustering algorithm to assess adulteration in Hashemi rice samples. First, to reduce the data's dimensionality, the principal component analysis method was applied to the preprocessed data using the multiplicative dispersion correction and Savitzky-Golay methods. Then, the fuzzy unsupervised clustering algorithm was applied using the whole spectrum wavelength (400-1000 nm). It was able to separate the original sample from the adulterated samples well. Also, the fuzzy membership diagram separated the original and self-adulterated samples, mixing 5% to 50%, confirming the correctness and capability of the fuzzy method. Therefore, the HSI system with fuzzy unsupervised algorithms can be used as a reliable and out-of-laboratory method for rapid rice authenticity evaluation. https://ijbse.ut.ac.ir/article_100478_e81f4f4f1faebda4a0b12a7b8fc4a1e4.pdf