Application of Sentinel Images 2-A in Estimation of Potassium in Potatoes

Abstract

Potato production in Indonesia decreased by 2.43% from 1,314,657 in 2019 and 1,282,768 tons in 2020. One of the causes of the decline in potato production is a lack of potassium. Potassium nutrient deficiency can be caused by fertilization that is not yet precise and is still done conventionally. The purpose of this study was to estimate the nutrient content of potassium using Sentinel 2-A. This study observed 50 points that were determined through the free grid method. Sentinel 2-A was transformed into GLI, GNDVI, NDVI which is the vegetation index and NDSI, and SAVI which is the soil index. The results showed that plant K correlated with GLI CS index (r = -0,46), NDVI CS (r = -0,48) and NDSI CS (r = -0,46). NDVI CS (R2 =2 3%) is the most accurate index in estimating the nutrient content of Potassium than GLI CS (R2 = 21%) and NDSI CS (R2 = 21%). Based on the results of the plant K regression test and NDVI CS, the regression equation y = 1,8003 + (-0,5716 NDVI CS) was obtained. The results of the validation test showed that the t table (-3.18) > t count (2.15) so that there is a significant difference in the estimation results of potassium with the results of potassium obtained in the field. Based on the results of the validation test which were significantly different, the productivity estimation model could not be used to estimate the potassium nutrient in potatoes.