PEMANFAATAN KECERDASAN BUATAN UNTUK DETEKSI PENYAKIT DAUN BUNCIS (PHASEOLUS VULGARIS L.) MENGGUNAKAN PENDEKATAN TRANSFER LEARNING PADA MODEL YOLOV8

Agriculture is an important sector in improving the economy and becoming a source of livelihood for the community, both from food crops and horticulture. One of the plants that is widely cultivated and has high economic value is the bean plant (Phaseolus Vulgaris L.). However, in its cultivation, bean leaves are susceptible to pests and diseases. In addition, the lack of farmer knowledge in identifying bean leaf diseases is crucial and reduces bean productivity. Therefore, a technology- based solution is needed that can detect bean leaf diseases quickly and accurately. This study aims to broadcast the performance of the YOLOv8 model with a transfer learning approach to detect bean leaf diseases found at the research location, precisely in Sindang Village, Hegarmanah Village, Sukabumi Regency. The dataset used is 2,222 bean leaf images classified into three categories, namely leaf spots (leaf spots), leaf rust (leaf rust), and healthy leaves (healthy leaves). The training model uses a transfer learning approach on the YOLOv8-s architecture which produces an accuracy of 85% in the 30th training (epoch). Performance evaluation was conducted using precision, recall, and mean average precision (mAP) metrics. The results of the model were integrated into a web-based system using the flask framework, which is expected to help in early mitigation of bean leaf disease and protect agricultural ecosystems, the environment and health.
Keywords: Bean Plant, Leaf Disease, Artificial Intelligence (AI), Yolov8, Transfer learning