Intelligent paddy rice color recognition suitable for harvesting / Athirah A.Rahim

This paper presents an automatic recognition of paddy rice color using RGB color extraction. In this work, five sets of paddy rice images from paddy field at Kampung Tua, Semanggol Perak are digitally captured at ICS (Image Capturing Studio) room. The identified regions of interest (ROI) of these paddy's images are processed to quantify the reflectance indices in RGB color model. Paddy rice images are then processed to produce the dominant RGB pixel indices in the primary color model. These reflectance indices gained under standard and controlled environment are then used to design a ANN diagnosis model for paddy rice using MATLAB software. The optimized model is evaluated and validated through analysis of the performance indicators regularly applied in classification models. From the findings, this work has shown that the best model has produced percentage accuracy of 88.75%, 92% specificity and 85.5% sensitivity when measured at 0.1 threshold with a balanced percentage rate of training dataset