DEFECT SORTING OF DRY DATES BY IMAGE-ANALYSIS

Spectral characteristics were obtained for good quality and defective fruit of two dry date varieties, 'Majhul' and 'Zahidi'. An image proce ssing algorithm was developed to sort good and defective fruit using s imple thresholding operations. To eliminate the effects of varying sys tem characteristics, defective date spectral characteristics were norm alised based on an average 'good date' reference spectrum. Thresholdin g parameters were determined by integrating the normalized reflectance curves over the wavelength range in which defect/good date contrast w as maximised. The reflectance data were also used to select optical fi lters to enhance the features of interest in the images. Preliminary r esults indicate that the method proposed offers a promising means of r apidly detecting defective fruit. Further research is needed to estima te sampling errors associated with the technique and to investigate re al-time sorting speed.