Designing an Automatic System for Grading Raisins and Determining its Tailing Percentage Using Image Processing Techniques

Raisins are one of the most important agricultural products obtained from drying grapes. At present, grading raisins and determining the percentage of different types of raisins in a sample, as well as identifying or not having a tail in it is done manually, and therefore requires a lot of time. In this study, the aim is to provide effective and powerful algorithms using image processing techniques in the field of machine vision for grading raisins as well as detecting and determining the percentage of tailed and tailless raisins. In order to analyze the proposed algorithm, the performance of the algorithm is evaluated by preparing photos of different samples of raisins and implementing the proposed algorithms on these samples in MATLAB software and comparing the results obtained with manual methods. To evaluate the performance of the proposed methods, the criteria of total accuracy, sensitivity and accuracy of positive output were calculated that the findings obtained from the evaluation of the proposed method for rating raisins total accuracy 98/65%, sensitivity 98/47% and positive output accuracy 97/83% as well as the findings of the proposed method for Determination of raisin tailing percentage showed total accuracy of 98%, sensitivity of 92/32% and positive output accuracy of 98/69%, indicating the optimal and reliable performance of these methods with low cost (low software calculations) compared to traditional methods.