<![CDATA[This study aims to design and develop a leaf moler disease detection system on shallots (Allium cepa L.) based on digital image processing in Enrekang Regency, South Sulawesi. Moler disease caused by the fungus Fusarium oxysporum f. sp. cepae is one of the main factors that reduce the quality and productivity of shallots. So far, disease identification is still done manually through direct observation by farmers, which is subjective and time-consuming. To overcome this problem, this study applies the Convolutional Neural Network (CNN) algorithm to automatically classify shallot leaf images into two categories, namely healthy and infected with moler disease. The number of datasets used is 502 images, consisting of 251 healthy images and 251 infected images, with data division of 70% for training, 15% for validation, and 15% for testing. The CNN architecture used consists of convolution, pooling, flatten, and fully connected layers with ReLU and sigmoid activation functions in the output layer. The training process used the Adam optimizer with a learning rate of 0.001 and a binary cross-entropy loss function. Test results showed a training accuracy of 97.14%, a validation accuracy of 94.73%, and a testing accuracy of 97.37%, indicating the model has a good level of precision and generalization ability without overfitting. This system is implemented as a Flask-based web application that allows users to upload leaf images and obtain detection results instantly. This system is expected to help farmers detect diseases more quickly and increase shallot productivity in Enrekang Regency.]]>
Copyrights © 2026
