Prediction of Isoflavone Content in Soybeans with Sentinel-2 Optical Sensor Data by Means of Regressive Analysis

Abstract

Prediction of isoflavone content in soybeans with Sentinel-2 of optical sensor data by means of regressive analysis is conducted. There is strong demand on estimation of isoflavone content for soybeans famers because the isoflavone is good for health. Soy isoflavones are a component that is abundant in the germ of soybeans. Soy isoflavones have a structure similar to female hormone (estrogen) and are known to exhibit female hormone-like action. Namely, isoflavone supports joint movement; promotes bone metabolism; maintains walking ability. There is a possibility to estimate isoflavone content from space. That is to estimate isoflavone content through regressive analysis with NDVI derived from remote sensing satellite based optical sensor and truth data of isoflavone obtained by chemical measurements. If the isoflavone content can be estimated before harvest, then soy farmers may control fertilizer and water supply. Through experiments of isoflavone content estimation, it is found that the proposed method allows prediction of isoflavone content at three weeks before harvest.KeywordsAgricultural product quality predictionIsoflavone contentSoy plantationTime series analysisNDVI: Normalized Deviation Vegetation IndexSentinel-2 of optical sensor