Milad Fathi; Seyed Mohammad Ali Razavi
Abstract
In this study, potential application of image texture analysis as a non-destructive method for automation and prediction of mechanical properties of carrot chips was investigated. Samples were fried at different processing conditions and moisture content, colour parameters (i.e. L*, a*, b* and E) ...
Read More
In this study, potential application of image texture analysis as a non-destructive method for automation and prediction of mechanical properties of carrot chips was investigated. Samples were fried at different processing conditions and moisture content, colour parameters (i.e. L*, a*, b* and E) and mechanical properties (i.e. hardness and apparent modulus) were determined. Hardness and apparent modulus increased by increasing frying temperature and time. Four image texture features namely contrast, correlation, energy and homogeneity were calculated using gray level co-occurrence matrix. The results showed contrast and energy of gray level images were well correlated with hardness of fried samples in compression and puncture tests. Correlation coefficients of 0.97 and 0.98 between four image texture features and hardness were obtained in compression and puncture tests, respectively. Results indicate that image texture analysis can be successfully applied as a non-destructive method for estimation of mechanical properties of carrot.
