Use of an unmanned aerial vehicle for monitoring and prediction of oilseed rape crop performance

Abstract

The flowering stage of oilseed rape (Brassica napus L.) is of vital interest in precision agriculture. It has been shown that data describing the flower production of oilseed rape (OSR), at stage 3, in spring can be used to predict seed yield at harvest. Traditional field-based techniques for assessing OSR flowers are based on a visual assessment which is subjective and time consuming. However, a high throughput phenotyping technique, using an unmanned aerial vehicle (UAV) with multispectral image (MSI) camera, was used to investigate the growth stages of OSR (in terms of crop height) and to quantify its flower production. A simplified approach using a normalised difference yellowness index (NDYI) was coupled with an iso-cluster classification method to quantify the number of OSR flower pixels and incorporate the data into an OSR seed yield estimation. The estimated OSR seed yield showed strong correlation with the actual OSR seed yield (R2 = 0.86), as determined using in-situ sensors mounted on the combine harvester. Also, using our approach allowed the variation in crop height to be assessed across all growing stages; the maximum crop height of 1.35 m OSR was observed at the flowering stage. This methodology is proposed for effectively predicting seed yield 3 months prior to harvesting.

Plant phenotyping relevance

UAVマルチスペクトル画像と画像分類により、ナタネの草高・花数を定量化し収量予測へ利用する高スループット表現型計測手法が中心である。

Code and data availability

The paper reports UAV multispectral phenotyping of oilseed rape (canopy height models, NDYI iso-cluster classification, yield prediction), but no public phenotype dataset, imagery, code, or trained model is deposited. The Data Availability Statement states all relevant data are within the paper, and the analysis used商业

No evidence-backed public reproduction asset is currently recorded.