Determining Natural Rubber Humidity Level using Rubber Color

Abstract

The process of drying rubber requires constant monitoring of rubber humidity. The dry rubber at the end of the drying process must have very low humidity and not be overheated. Modern drying process uses heat chamber to remove moisture from rubber. Opening the heat chamber to check the humidity of rubber often is not considered wise in terms of energy saving. However, we can install a camera inside the heat chamber to take the photo of the rubber in order to use the color of the rubber to determine the humidity of the rubber. In this research, we successfully use machine learning classification techniques to classify the color of rubber to humidity level. We found that K-nearest neighbor performs best given the data from our experiments.