Predicting rice grain yield using normalized difference vegetation index from UAV and GreenSeeker

Abstract

A precise, simple, and rapid growth diagnosis method using normalized difference vegetation index (NDVI) obtained by unmanned aerial vehicle (UAV), which will help determine nitrogen (N) application rate to increase grain yield in numerous farmers' fields, is necessary for the development of a robust production system for rice (Oryza sativa L.). In the present study, we examined the relationship between UAV-NDVI and NDVI measured with the GreenSeeker handheld crop sensor (GS-NDVI), and between grain yield and UAV-NDVI or GS-NDVI at the reproductive stage in the plant communities at 4–1 ​week (wk) before heading in 2018 and 2019 and in 2020 and 2021, respectively. In the data of each measurement day in 2018 and 2019, the relationship between UAV-NDVI and GS-NDVI was strongly positive. However, in the pooled data of different measurement days, the relationship between UAV-NDVI and GS-NDVI was weakly positive. This was because GS-NDVI was more constant under various climatic conditions and across various time of day than UAV-NDVI at the reproductive stage. Furthermore, in the pooled data of different years in 2020 and 2021, GS-NDVI correlated more strongly with grain yield than UAV-NDVI between 3 and 1 ​wk before heading. To increase the efficiency of growth diagnosis and yield prediction in the numerous farmers’ fields, UAV-NDVI could be used with correction by a few measurements of GS-NDVI determined on the same day.