Classification of bulk raisin is one of the main challenges to producers and buyers of raisin in the world. In this research, the quality of bulk raisin was investigated by using the image processing technique. For this purpose, 750 images of bulk raisin containing a different mixture of good and bad raisin with wood were used (50 images × 15 class). Different texture feature algorithms combined with different modeling methods were used to evaluate the system performance. The study results showed that the Support Vector Machine (SVM) classifier using Gray Level Run Length Matrix (GLRM) features yielded more accurate classification results. The classification accuracy of modes I (6 classes of good and bad raisin) and II (15 classes) was obtained 85.55% and 69.78%, respectively. The results also indicated that the machine vision system could be used to classify good and bad raisin successfully, but in the case of mode II, it yielded weaker results.
