Plant height prediction of maize varieties with varying maturity based on temperature

Abstract

Abstract The use of unmanned aerial vehicles for remote sensing is an effective method for monitoring crop growth, particularly for tall crops such as maize. High‐resolution imagery obtained from unmanned aerial vehicles enables the measurement of plant height, which is a critical indicator of crop growth. However, a reference plant height is required to assess growth. This study aimed to develop a model to predict the reference height for growth assessment using temperature data. Furthermore, a methodology was proposed to estimate model parameters from the relative maturity, thus enabling adaptation to a range of maize varieties. In 2022 and 2023, maize plant height was measured using an unmanned aerial vehicle at two flying altitudes (40 and 100 m) several times for 12 varieties with varying relative maturity. Moreover, a regression model was developed to predict the silking stage and identify the optimal sensing time 1 week before the silking stage. The results showed that the growth rate was not statistically different among the varieties, indicating that maximum plant height was determined by the duration of the growth period. A growth model was developed based on these results. The root mean square error (RMSE) for the model was 0.16 and 0.15 m for data sets from 40‐ and 100‐m altitudes, respectively. In estimating plant height, this growth model performed marginally better than the logistic curves used in existing studies. Additionally, a linear relationship was observed between relative maturity and the parameters of the developed growth model. Consequently, the newly developed growth model can predict the plant height for new varieties because the parameters of the model can be inferred from the relative maturity.

Plant phenotyping relevance

UAV画像によるトウモロコシ草高推定のための成長モデルと成熟度に基づく適応手法を開発・評価しており、植物形質取得が中心である。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.