Model Prediksi Nitrogen Tanah Menggunakan Indeks Vegetasi dari Citra Sentinel 2 di Kecamatan Sumbersari

Abstract

The main form of nitrogen in the soil comes from organic materials, which is influenced by organic carbon. Nitrogen content analysis is generally done using the PUTS method and laboratory analysis. The problem with these two methods is that there are still limitations, both in terms of the completeness of PUTS results and the relatively long duration of laboratory analysis. An alternative solution that can be used is remote sensing, such as the Sentinel-2 satellite. The aim of this study is to find out the correlation between vegetation indices from satellite data and organic carbon with soil nitrogen, and to build a prediction model by combining vegetation indices with organic carbon. The research steps are collecting soil samples in the field, then analyzing them in the lab and matching the imagery data with the research locations. Image data is processed using vegetation indices such as NDVI, GNDVI, and OSAVI. The prediction model is formulated using polynomial regression. The results show that there is no relationship between vegetation indices and soil nitrogen, while organic carbon and soil nitrogen have a small correlation. The prediction model that was obtained has a relatively low determination coefficient percentage, with the lowest RMSE value found in the combination of GNDVI x OSAVI x Organic Carbon.