Deep Learning Classification Model for Oil Palm Tree Health Assessment

The oil palm tree is a highly valued plant. It is a tropical palm tree known as Elaeis guineensis. It is valued for its fruit which produces two varieties of oil: palm oil and palm kernel oil. This study explored the application of satellite and Unmanned Aerial Vehicle (UAV) imagery in assessing vegetation health through the calculation of the Visible Atmospherically Resistant Index (VARI) and Normalized Difference Vegetation Index (NDVI). Utilizing the capabilities of ArcGIS software, the imagery data are processed to generate VARI and NDVI raster layers, providing insights into oil palm tree health conditions. The workflow involves data collection, preprocessing, VARI and NDVI calculations, analysis, and visualization. The grayscale VARI and NDVI images reveal spatial patterns of vegetation health. This result contributes to land management, environmental monitoring, and agricultural decision-making processes. This study demonstrates the comparison of VARI and NDVI in terms of accuracy for oil palm tree health assessment. A new novelty from this study is that VARI has higher accuracy than NDVI.