Nutmeg grading system using computer vision techniques

Nutmeg is a popular spice and important commodity that has been recognized by world market. One of the problems during post-harvest handling of nutmeg is quality of the product, particularly: non-uniform quality, damage and contaminated by aflatoxin. This is caused by poor postharvest handling and the products were sorted traditionally. This study performed a computer vision system to sort or to classify quality of the nutmeg based on Indonesian national standard (SNI). In this paper, discriminant analysis and multilayer perceptron (MLP) are implemented to design an automatic classifier of the nutmeg quality. A number of 150 nutmegs are captured by using charge couple device camera. The models are then trained via the data of 124 nutmegs, and their accuracy is tested through the data of 26 nutmegs. Some visual features (texture, shape, and color) of each nutmeg are extracted by using image processing techniques. The results point out that, our proposed grading system could classify nutmeg based on the SNI accurately.