09:30 〜 09:45
[AOS10-03] Vegetation indices using unmanned aerial vehicle to monitor plant-level tomato plant height under till and no-till farming conditions
キーワード：分光反射率、作物体スケール、精密農業
Precise plant height (PH) estimation is consequential for precision farming. Most vegetation indices derived from remote sensing measurements are used for monitoring and predicting PH. However, few studies have estimated tomato (Solanum lycopersicum L.) PH at the plant level using multi-spectral reflectance data obtained from unmanned aerial vehicle (UAV). In this study, 14 vegetation indices were compared for estimating tomato PH at the plant level under till (with mulching) and no-till (without weed clipping treatments) conditions. The experiment was carried out at the Field Museum Fuchu, Tokyo University of Agriculture and Technology during the growing season from 13 May to 30 July 2020. Results showed good logistic growth relationships obtained between measured tomato PH and some vegetation indices. In particular, good performance for predicting PH was observed for: Infrared Percentage Vegetation Index and Transformed Vegetation Index under till treatment (R2 of 0.688 and 0.717 and RMSE of 0.139 and 0.136 m, respectively); and Green Normalized Difference Vegetation Index (GNDVI) and Weighted Difference Vegetation Index (WDVI) under no-till treatment (R2 of 0.571 and 0.561 and RMSE of 0.166 and 0.147 m, respectively). However, the variability of each vegetation index increased with increasing PH, and the prediction accuracy under fully expanded leaf cover conditions was lower than that under the early growth stage. In addition, all vegetation indices except for the Enhanced Vegetation Index, GNDVI, and WDVI under no-till treatment had lower performance for estimating PH as compared with those under till treatment due to soil background effects and presence of weeds. Overall, results indicate that certain vegetation indices provided an effective method to a certain degree for monitoring tomato PH at plant scale.