KLASIFIKASI JENIS BERAS BERDASARKAN CITRA MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN)
Ziki Azhima, 2023. Klasifikasi Jenis Beras Berdasarkan Citra Menggunakan Metode Convolutional Neural Network (CNN), Pembimbing I Sulistyowati, S.Kom., M.Cs., Pembimbing II Rudini, M.Pd. The technology of machine learning is currently experiencing rapid development and has been applied in various fields, including agriculture. One application of machine learning in agriculture is the classification of rice. Automatic rice classification can help accurately and quickly determine the type of rice. This study utilized the convolutional neural network VGG16 architecture to classify rice. The dataset consisted of photographs of five different types of rice, which were used to train and test the model. The training process utilized transfer learning techniques by utilizing a previously trained model on a large dataset of images. The dataset was pre-processed by resizing images and splitting them into folders. Hyperparameters such as batch size, epoch, and optimizer were modified to obtain the optimal model for rice classification. The test results show that the rice classification model using the convolutional neural network VGG16 architecture is capable of producing high accuracy. The evaluation results showed that the classification accuracy reached 99.86%, with an F1-score of 0.98. With the rice classification system using the CNN VGG16 method, it is hoped that it will simplify the distinction between different types of rice. An effective and accurate rice sorting process can help produce high-quality rice that is safe for consumption. As a result, consumers can choose the type of rice that suits their individual needs and preferences.