Classification of Durian Fruits based on Ripening with Machine Learning Techniques

Eating fruits is good for health and an excellent source of essential vitamins and minerals. But fruits are unfit for intake if they are not fully ripened. Ripening of fruits is an important phase in Pomology based on which the fruits are categorized. In this context, various machine learning techniques were proposed to detect the classification of fruits. This paper provides a special architecture to identify the features of different classes of Durian fruits. Durian, popularly given the name, "king of fruits", is one of the rarest and sweetest fruits grown in South Asia. The model of image processing helps to divide the classification into two parts of the feature extraction part and classification part of the fruit used. In this article, the system is evaluated using the datasets of the Durian Fruit. The datasets is divided into two parts as training and testing. By applying the Edge Detection and color extraction the features of the durian are properly measured. Finally, the performance is measured using Non-Destructive Machine Learning techniques such as SVM, GNB, Random Forest. The results obtained provide the best accuracy of 89.3 % using the SVM technique and 84.3% using Random Forest.