The evaluation of the usefulness of textures from cross-section images obtained using a digital camera and a flatbed scanner for cultivar discrimination of quince (Cydonia oblonga Mill.)

• Published in: Food control Vol. 131; p. 108447
• Main Authors: Ropelewska, Ewa, Mieszczakowska-Frąc, Monika, Kruczyńska, Dorota
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2022
• Subjects: Discriminant analysis; Fruit cross-section; Image analysis; Quince cultivars; Texture parameters; Image analysis; Discriminant analysis; Fruit cross-section; Texture parameters; Quince cultivars
• ISSN: 0956-7135; 1873-7129
• DOI: 10.1016/j.foodcont.2021.108447

The aim of this study was to evaluate the usefulness of a digital camera and a flatbed scanner for acquiring the cross-section images intended to calculate the texture parameters for cultivar discrimination of quince. The quince belonging to cultivars ‘Bereczki’, ‘Kaszczenko’, ‘Leskovac’, ‘Marija’ and ‘Uspiech’ were used in the research. The textures selected from color channels R, G, B, L, a, b, X, Y, Z, U, V, individual color spaces Lab, RGB, XYZ, YUV and a set of textures selected from all channels (R, G, B, L, a, b, X, Y, Z, U, V) were applied for building the discriminative models. The discrimination accuracies determined in the case of models built based on selected textures from images obtained using a digital camera were higher than using a flatbed scanner. For the textures selected from all color channels, the total accuracy reached 99% in the case of images from a digital camera and 94% for images from a flatbed scanner. The models built based on the texture parameters selected from individual color spaces provided the total accuracies of up to 99% for Lab and YUV color spaces for images from a digital camera and up to 90% for Lab and YUV color spaces for images obtained using a flatbed scanner. Among the individual color channels, the color channel V ensured the highest total accuracies equal to 90% for images acquired using a digital camera and 77% in the case of images from a flatbed scanner. •A digital camera and a flatbed scanner were useful for cultivar discrimination of quince.•The discrimination accuracies for images obtained using a digital camera were higher.•For images acquired using a digital camera, the accuracy reached 99%.•In the case of images from a flatbed scanner, the correctness of up to 94% was obtained.