Grading of Yemeni raisin (Razigi Cv.) using non-contact machine vision techniques [electronic resource].

Includes references.

This research aimed at establishing quality standards for noncontact grading of Yemeni raisin (Vitis vinifera) of Razigi variety, by developing a digital image processing algorithm based on color and shape information. This was accomplished through developing preprocessing procedures to segment transparent interior areas in raisin color images as the regions of interest and highlight their morphology for extracting shape features. A distinct signature for each raisin grade was generated by calculating number of matches of a set of twenty selected morphological features. A minimum distance classifier was developed, trained and tested in grading raisins by sorting them into three grades, namely, A, B, and C. The classifier was successful in sorting two raisin grades, namely, A and C with 100% CCR and 0% MCR for each. Its performance was not good enough in sorting raisins of grade B as the CCR dropped to 50%. Some difficulties were encountered by the classifier in sorting half of the raisins of grade B due to their great similarity with grade A raisins

Summary in Arabic.

1