discerning expiration status of edible vegetable oils based on color changes during oxidation process: using digital image and linear discriminant analysis in both primary and secondary oxidations

Discerning the expiration status (nonrejected and rejected) of edible vegetable oils is very significant because of the hazardous primary and secondary oxidation products. therefore, it is of outmost importance to monitor the quality and safety of these oils. based on previous literature, reports and experimental observation, the oil color changes during oxidation. thus, the present study investigates the use of image processing and linear discriminant analysis (lda) for the classification of nonrejected and rejected edible vegetable oils during oxidation process at 85°c, with respect to the induced period in both primary and secondary oxidation of four oil type (olive, sunflower, palm and soybean). the purpose of this study was to find less costly and quicker methods with environmental protection, by using the color spaces (rgb, hsi, l*a*b* with grayscale) instead of chemical analyses to determine the expiration status of edible vegetable oils. results of this study indicated that the best classification for expiration status of known oils according to induced period of peroxide value in  each color space, was achieved with lda model were for palm with 100% (hsi and grayscale), olive with 84.61% (l*a*b* and rgb), soybean with 95% (grayscale) and sunflower with 100% (rgb and hsi), also in induced period of carbonyl value test, the best classification performance was achieved in palm with 100% (l*a*b*), olive with 100% (l*a*b*), soybean with 89.47% and sunflower with 95% (hsi).

Discerning expiration status of edible vegetable oils based on color changes during oxidation process: Using digital image and linear discriminant analysis in both primary and secondary oxidations

Discerning the expiration status (nonrejected and rejected) of edible vegetable oils is very significant because of the hazardous primary and secondary oxidation products. Therefore, it is of outmost importance to monitor the quality and safety of these oils. Based on previous literature, reports and experimental observation, the oil color changes during oxidation. Thus, the present study investigates the use of image processing and linear discriminant analysis (LDA) for the classification of nonrejected and rejected edible vegetable oils during oxidation process at 85°C, with respect to the induced period in both primary and secondary oxidation of four oil type (Olive, Sunflower, Palm and Soybean). The purpose of this study was to find less costly and quicker methods with environmental protection, by using the color spaces (RGB, HSI, L*a*b* with Grayscale) instead of chemical analyses to determine the expiration status of edible vegetable oils. Results of this study indicated that the best classification for expiration status of known oils according to induced period of peroxide value in  each color space, was achieved with LDA model were for palm with 100% (HSI and Grayscale), olive with 84.61% (L*a*b* and RGB), soybean with 95% (Grayscale) and sunflower with 100% (RGB and HSI), also in induced period of carbonyl value test, the best classification performance was achieved in palm with 100% (L*a*b*), olive with 100% (L*a*b*), soybean with 89.47% and sunflower with 95% (HSI).