Innovative Flood Impact Monitoring and Harvest Analysis in Oil Palm Plantations Utilizing Geographic Information Systems and Deep Learning

Floods, as a form of disaster, significantly affect individuals and farmers in impacted areas, particularly through crop damage and the inability to harvest due to prolonged and extensive flooding. Among the most severely affected agricultural sectors are oil palm plantations, which regularly experience such disruptions annually. Current methods of assistance and relief during flooding rely on field surveys conducted manually by personnel, a process constrained by its time-intensive nature. Moreover, existing applications or platforms do not support the classification and inspection of oil palm plantations affected by floods during harvesting. This research aims to develop a method and application for inspecting oil palm plantations impacted by floods during harvesting. The approach utilizes deep learning and geographic information systems (GIS) to classify and analyze flood-affected areas and determine the ripeness of oil palm bunches on trees, enabling accurate and rapid identification of flood-affected areas. The study results demonstrate that the proposed method achieves a flood classification accuracy ranging from 96.80% to 98.29% and ripeness classification accuracy for oil palm bunches on trees ranging from 97.60% to 99.75%. These findings indicate that the proposed model effectively and efficiently monitors flood-affected areas. Additionally, the developed application serves as a valuable tool for flood management, facilitating timely assistance and relief for farmers impacted by flooding.