PERANCANGAN MACHINE LEARNING UNTUK DETEKSI PENYAKIT TANAMAN KUBIS DENGAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN) ARSITEKTUR ALEXNET

Indonesia, as an agrarian country, relies heavily on the agricultural sector as the main pillar of its economy. In mountainous areas, cabbage cultivation serves as the primary source of income for local communities due to fertile soil and favorable climate conditions. However, farmers in Krisik Village, Pudak Sub-district, face major challenges, particularly the susceptibility of cabbage plants to diseases such as leaf blight and soft rot, which can significantly reduce crop yields. Manual disease detection has proven to be less accurate, time-consuming, and prone to errors. To address this issue, this research aims to develop a computerized cabbage leaf disease detection system. The system employs a Convolutional Neural Network (CNN) with the AlexNet architecture, which is well known for its effectiveness in recognizing visual patterns in digital images. CNN can extract key features from leaf images and automatically classify disease types efficiently, providing detection results along with accuracy, disease charts, and treatment recommendations. The implementation results demonstrate excellent performance. Evaluation using classification metrics such as precision, recall, and F1-score shows that the model achieved an F1-score of 0.98 for soft rot, 0.94 for blight, and 0.96 for healthy leaves. The macro-average and weighted-average scores were also high at 0.96, indicating consistent performance across all classes. The outcome of this research is a system that can be utilized as a reliable early diagnostic tool, helping farmers reduce losses and improve crop quality.