METODE KLASIFIKASI ALGORITMA CONVOLUTIONAL NEURAL NETWORK (CNN) PADA PENYAKIT DAUN TEH

Abstract

Tea is one of the agricultural commodities that plays an important role in the Indonesian economy. The tea industry generates revenue, provides jobs, and drives local growth. Tea leaf disease is one of the factors that has caused tea production in Indonesia to decline. Algae leaf disease is a common disease in tea leaves, as well as anthracnosis and bird eye. Exposure to pathogens in the tea leaf can cause significant declines in global yields. Detecting disease on tea leaves at an early stage is essential to reduce the loss of production yields. Detection methods using visual observation may be less effective and do not help in controlling the disease well. Deep learning has successfully collected and analyzed large amounts of data, enabling the diagnosis of tea plant diseases quickly and accurately. The research aims to improve the accuracy of tea leaf disease classification by optimizing the used Convolutional Neural Network (CNN) algorithms as well as improving feature management. The research methods carried out included data collection, image data pre-processing, CNN model design, accuracy classification, testing and results. The results showed that the CNN model managed to classify tea leaf disease with satisfactory levels of accuracy, precision, and recall. The study successfully indicated that the use of the CNN algorithm in the classification of tea leaf diseases, namely Algal Leaf, Antracnose, and Bird Eye Spot, had significant accuracy with an achievement of 79.36%.