One of the main causes of citrus postharvest losses is fungal diseases. Some fungal diseases cause surface rot, while Alternaria alternata fungus causes internal rot, which appears apparent in advanced stages. Spectroscopy as an internal quality measurement method was used for early detection of citrus black rot disease caused by Alternaria alternata. The spectra of inoculated Jaffa oranges were collected in the range of 200 to 1100 nm in three weeks after inoculation. Soft independent modeling of class analogy (SIMCA) was implemented to select the important wavelengths. The most effective wavelengths in the detection of healthy and infected samples of the first week were 507, 933, 937 and 950 nm and for the second week were 522 and 787 nm. Classification of healthy and infected samples was done by support vector machine and Neural Network using optimal wavelengths. The overall classification accuracies of 100, 82 and 92 percent were achieved by the optimal selected wavelengths for samples of all weeks in training, cross validation and test stages, respectively. Keywords: Fast detection, Fungi infection, Postharvest decay, Spectroscopy, Spectral analysis

Keywords