An Intelligence System for Identifying Copra Maturity Level Using Convolutional Neural Network
The North Sulawesi Province of Indonesia (SULUT) is known as the Coconut Waving Province because of the many coconut trees that grow. One of the processed products of coconut meat is copra which is the main commodity of SULUT. This study aims to propose a model for identifying copra maturity levels using the Convolutional Neural Network (CNN) based on its digital image with 3 classification classes of raw, half-ripe, and ripe. 10-fold cross-validation techniques were used to evaluate models. Results show a good performance of the models, even for the worst model provided a good accuracy of 87,78% in the training and validation processes while the best model provided a perfect accuracy of 100%. Besides, the testing process using new data got a good accuracy of 83,34% for the worst model whereas the best model provided a perfect accuracy of 100% then the proposed model can be used to identify copra maturity with a satisfactory performance. In the future, an automatic machine can be developed based on these findings to sort the quality of copra massively which is useful for farmers and industries side.