Near infrared sensing for RRIM clone series classification utilizing optical sensing technique / Nina Korlina Madzhi, Faridatul Aima Ismail and Noor Ezan Abdullah

This paper focuses on the classification of the selected RRIM 2000 and 3000 clone series based on optical sensing technique. The Near-IR was used as sensing element in order to measure the latex from the top surface that read by photodiode which received the reflected light from the sensor via reflectance index in term of voltage. The raw obtained data was used as input parameter for ANN tool which supervised by scaled gradient backpropagation and the performance was optimized at 25 neurons with 74.4% accuracy.