Deep learning implementation using convolutional neural network in mangosteen surface defect detection

Mangosteen is one of the fruits that has an enormous export potential in Indonesia. However, not all mangosteen is the defect free fruit. The quality assurance in mangosteen export is done manually by sorting expert. Therefore, this can lead unstandardized and inaccurate results. The result happens because of human error. It needs an image processing technology to help the sorting process which one is the defect and non-defect. In this study, we use one of deep learning architecture that is Convolutional Neural Network (CNN). Therefore, we use CCN as a detection of mangosteen. CNN proved to be very efficient regarding classifying images. This CNN method is implemented using 4-folds Validation Cross to validate data accuracy. In the preparation of the CNN architecture model, initializing the parameter configuration accelerates the network training process. The results of the experiments using CNN algorithm showed the performance of defect detection on the mangosteen fruit of 97%.