3. PROBLEM STATEMENT Demand for rice nowadays increasing tremendously, hence producing and sorting of rice becoming more faster than usual requirement. Based on previous research [3], for a conventional rice sorter to recognition percentage is below 90 percent if the rice flow exceeds few thousands [kg/h]. Recognition ability will be limited and hard to differentiate between good and bad quality of rice. Moreover, since huge amount of rice need to be sorted, there might be some overlooked rice during sorting process. Quality of rice will not be preserved if grading performance is not been developed wisely. Rice grading scope is very large; it can start with rice grading or mapping area [29-30], infected diseases on rice leaf detection [20, 32], nitrogen or moisture status of rice [1,4] and etc. All these scope can be covered or solve by using machine vision, however they are still some problems will occur or might affect the grading process. Below are the difficulties faced in image processing in getting good quality of rice: