DEVELOPMENT OF MATHEMATICAL MODEL FOR ESTIMATING THE RICE MILLING DEGREE BASED ON FLUORESCENCE IMAGE

This research aims to develop a mathematical model for estimating the milling degree of milled rice based on fluorescent imaging. The materials used were the Ciherang, IR64 and Mekongga varieties which are widely grown and consumed by Indonesian people. The experiment was conducted by varying the polishing time starting from 0 seconds to 34 seconds with 1 second intervals. Six grains of polished rice were taken to record their images using a camera with 365 nm UV fluorescent ring light. The data collected in this research were milling degree obtained by gravimetric method, milling degree obtained using a milling meter and color values of fluorescent images of milled rice by image processing. The results showed that prediction of milling degree using the RGB color model has the coefficient of determination between 0.8001 – 0.8652, which is considered as potential to be used as a model for estimating the degree of milled rice based on fluorescence images. The RGB color model shows that the image red signal has the highest coefficient of determination compared to green and blue signals. For all of the three varieties in this study, the Ciherang variety has a predictive model equation for the image red signal y = 3.9027x - 429.61, the IR64 variety has a predictive model equation for the image red signal y = 3.7344x - 415.01, and the Mekongga variety has a predictive model equation for the image red signal y = 3.5627x - 388.86.