To validate the accuracy of the developed ACS in terms of proper quality assessment and appropriate settings of the whitening machine, a series of pre-tests were performed. In each test, out of the 25 qualitative grade previously defined by the experts (see the antecedent part of fuzzy rules in Table 1), 100 samples of rice kernels were prepared according to the relative MF ranges of fuzzy system inputs (Fig. 5). A certain amount of each rice sample was spread over the surface of the first conveyor belt of the singulation unit. The kernels were then separated by singulation unit and transferred into the imaging chamber. The control program was designed so that when the first part of each sample reached the imaging chamber, the conveyor belts stopped for a short moment to capture the first image. This guaranteed that no noise would be created in the captured images as a result of kernel movement. Then, conveyor belts were run again to take the first part of the samples away from the imaging chamber and subsequently, bring the next part into the imaging chamber. This process was continued to take three images from each sample. After finishing the image acquisition, the qualitative indices of rice kernels in the form of DOM and PBK were calculated by the image processing unit and the data were fed into the fuzzy inference system to make a decision about the pressure level on the discharge section of the whitening machine. It should be noted that all of the mentioned procedures were performed in the off-line mode in which the whitening machine was not working. In fact, the objective of these tests was to evaluate the decision making structure of the developed ACS in all probable working conditions. Finally, the decisions of ACS on the pressure level on the discharge section of the whitening machine were compared with those of human experts and the results were reported in the form of a confusion matrix.