Classification of rice plant fertilizer needs based on leaf color chart using radial basis function neural network

Abstract

Management of nitrogen fertilizers in rice plants plays an important role in the productivity of rice plants. Excessive use of fertilizers may result in rice leaves becoming pale, yellowish that can lead to its death. Fertilizer should be given based on the content of nitrogen in rice leaves which is indicated by the color of rice leaf. The color of rice leaf can be checked through tools called Leaf Color Chart (LCC). However, farmers are still likely to experience differences in color perception while classifying the color of the leaf so that farmers sometimes give an inaccurate dose of nitrogen fertilizer. Therefore, this study has the objective to organize the needs of rice nutrient based on leaf color chart and to provide information on the fertilizer dose which is recommended to be given to reduce errors in the use of nitrogen fertilizer. This study used the classification process by looking at the greenish color of the leaf based on leaf color chart using the Radial Basis Function (RBF). The classification process was divided into three classes, namely Green1, Green2, and Green3. This study used 60 images as training data and 30 images as testing data with 90% of accuracy rate.