Iranian Research Organization for Science and Technology (IROST) Innovative Food Technologies 2783-350X 10 3 2023 04 21 Detection of freezing of Thomson variety orange fruit using Fourier transform-infrared spectroscopy and hyperspectral imaging methods Detection of freezing of Thomson variety orange fruit using Fourier transform-infrared spectroscopy and hyperspectral imaging methods 203 214 1262 10.22104/ift.2023.6059.2134 FA Karim Gerami Member of the academic staff of the Agricultural Engineering and Technical Research Department - Agricultural and Natural Resources Research and Education Center of West Azarbaijan - Urmia - Iran Hossein Behfar Assistant Professor of Department of Biosystem Engineering- Faculty of Agriculture- University of Tabriz- Tabriz-Iran Bahareh Jamshidi Associate Professor, Agricultural Engineering Research Institute, Agricultural Research Education and Extension Organization (AREEO), Karaj, Iran Shahin Zomorodi Associate Professor of Agricultural Engineering and Technical Research Department - Center for Research and Education of Agriculture and Natural Resources of West Azarbaijan - Urmia - Iran Journal Article 2023 02 05 Detection of oranges freezing when entering the consumer market, both from the point of view of freshness and processing, can play a very important role in the marketability of the final product.<br /><br />In this study, The freezing and non-freezing of orange fruit has been investigated using Fourier transform-infrared (FT-IR) spectroscopy and hyperspectral imaging First, the FT-IR spectrum of the skin of 20 healthy and freezing orange samples was obtained. Also, in the hyperspectral imaging method, images and spectra were obtained from 18 healthy and frozen orange samples using the system made by the company of the Physic Technologists. In order to further investigate the FT-IR and hyperspectral imaging data, linear discriminant analysis (LDA) was performed after applying different pre-processing methods to classify healthy and frozen oranges. The results of this research showed that after freezing, the peaks in the FT-IR spectrum, in the regions of 400 cm-1 to approximately 1500 cm-1, have undergone a fundamental change, and the intensity of these peaks has been greatly reduced. This shows the fundamental changes in orange peel samples due to freezing. The results showed that by applying the MF+SNV preprocessing method on spectroscopic data, it is possible to detect the freezing and non-freezing of orange fruit with high accuracy (accuracy for classification 92%). Moreover, the results for the hyperspectral imaging method showed that by applying the smoothing pre-processing methods, it is possible to detect the freezing and non-freezing oranges with good accuracy (accuracy for classification 75%).In general, the results showed that the FT-IR spectroscopic method has a higher accuracy and can detect freezing and non-freezing in orange peel samples. However, it is recommended to use the hyperspectral imaging to detect orange freezing if the non-destructive assessment of the sample is considered. Detection of oranges freezing when entering the consumer market, both from the point of view of freshness and processing, can play a very important role in the marketability of the final product.<br /><br />In this study, The freezing and non-freezing of orange fruit has been investigated using Fourier transform-infrared (FT-IR) spectroscopy and hyperspectral imaging First, the FT-IR spectrum of the skin of 20 healthy and freezing orange samples was obtained. Also, in the hyperspectral imaging method, images and spectra were obtained from 18 healthy and frozen orange samples using the system made by the company of the Physic Technologists. In order to further investigate the FT-IR and hyperspectral imaging data, linear discriminant analysis (LDA) was performed after applying different pre-processing methods to classify healthy and frozen oranges. The results of this research showed that after freezing, the peaks in the FT-IR spectrum, in the regions of 400 cm-1 to approximately 1500 cm-1, have undergone a fundamental change, and the intensity of these peaks has been greatly reduced. This shows the fundamental changes in orange peel samples due to freezing. The results showed that by applying the MF+SNV preprocessing method on spectroscopic data, it is possible to detect the freezing and non-freezing of orange fruit with high accuracy (accuracy for classification 92%). Moreover, the results for the hyperspectral imaging method showed that by applying the smoothing pre-processing methods, it is possible to detect the freezing and non-freezing oranges with good accuracy (accuracy for classification 75%).In general, the results showed that the FT-IR spectroscopic method has a higher accuracy and can detect freezing and non-freezing in orange peel samples. However, it is recommended to use the hyperspectral imaging to detect orange freezing if the non-destructive assessment of the sample is considered. https://jift.irost.ir/article_1262_0be7c784d5c0b933b226e8cb1f0ca428.pdf