Neural Network-based Classification of Germinated Hang Rice Using Image Processing

Germinated Hang rice is produced using traditional folklore wisdom. It has drawn a lot of attention by researchers due to its high nutritional value to the human body. Conventionally, the quality of germinated Hang rice grains has been assessed manually into good/bad. However, this method is very time consuming and relies primarily on human skills and experience. Thus, the purpose of this research was to develop an algorithm capable of automatically determining the quality of germinated Hang rice by dividing it into six groups comprised of good, broken, discoloured, un-husked paddy, deformed and withered grains. The algorithm is based on image processing techniques and extracts the shape, colour and texture features, after which they are fed into a neural network classifier with PCA feature selection. The experimental results showed that the overall classification accuracy achieved was 94.0%.