Detection Storage Time of Mild Bruise’s Loquats Using Hyperspectral Imaging

Bruise may cause spoilage, reduce commodity economic value, and give rise to food quality safety concerns. Therefore, it is crucial detect whether a loquat bruised when save storage transportation costs. At present, the bruise of loquats mainly discriminated by operator’s naked eye, which affected personal habits, light intensity, subjective psychological factors. The detection method time-consuming, inaccurate, inefficient, difficult identify bruise’s time loquats. Due fact that color features can be used perform conditions darkened brownish regions in loquats, combined spectral information proposed accurately mild this study. In order losses, different methods are deal with at corresponding time. Loquats four types time, including 6, 12, 24, 36 h, studied. Models characteristics, information, RGB HSI mixed (mixed-spectral), established based on linear discriminant analysis (LDA), support vector machine (SVM), least-squares (LS-SVM). investigated 400 independent samples utilized assess classification ability methods. results indicate Mixed-RBF-LS-SVM model has lowest errors, accuracies h 100%, 92%, respectively. overall accuracy LS-SVM mixed-spectral 96%, demonstrates Finally, optimized UVE, SPA, CARS, GA, respectively; found UVE-LS-SVM best, 92%. It also lays foundation for future studies about detecting fruits high-precision, rapid, nondestructive measurement.