SISTEM KLASIFIKASI HAMA JAMUR PADA CITRA DAUN KENTANG BERBASIS KECERDASAN BUATAN

Abstrak

Potatoes are one of the important food crops that are susceptible to fungal pest attacks. Classification of  fungal pests on potato leaves is very susceptible to human error if done manually. This research aims to efficiently classify fungal pests on potato leaves Research begins with data collection, data preprocessing, and model design. The model was then trained for 20 epochs using training data against validation data. Despite fluctuations, the model shows significant improvement during training, and ultimately achieves good accuracy. The results of the research evaluation show that the classification of fungal pests is greatly influenced by variations in images and the amount of data. The noise level in the image also has a significant impact. Accuracy results through model testing on test data show an accuracy level of 96.67 percent. The confusion matrix for each data also shows fairly even results, indicating that the system can classify fungal pests on potato leaves quite well. Thus, this system can help farmers to take more efficient preventive measures. Suggestions for future research are to expand the dataset to increase the diversity of potato leaf images and improve system reliability.   Keywords: Artificial intelligence, Potato Leaf Fungal Pests, Image Processing, Classification