Morphology-based correction improves plant height estimation from drone RGB SfM photogrammetry in dense salt marsh canopy

Abstract

The height of salt marsh vegetation is a key biophysical trait used to assess ecological health, monitor marsh restoration, and evaluate coastal protection from wave attenuation. This study defined a morphology-based correction that improved the accuracy of digital surface model (DSM)-derived canopy height for Spartina alterniflora . Using imagery collected at six drone altitudes (3 m to 120 m), corresponding to ground sampling distances (GSDs) of 1 mm/pixel to 34 mm/pixel, mean canopy height was estimated within 0.5 m by 0.5 m quadrats and validated against in situ RTK-GNSS Rover measurements. DSMs from drone altitudes of 60 m or lower (GSD ≤ 17 mm/pixel) yielded consistent elevation estimates with mean error < 5 cm. Increasing point cloud densification by changing image scale and point density settings did not improve DSM accuracy. Canopy height derived from DSMs was, on average, 60% of the true canopy height measured by Rover. A novel, quantitative assessment of canopy structure showed that the mean vertical position of projected horizontal canopy area was also 60% of the canopy height, suggesting that SfM reconstruction was capturing this position in the canopy, rather than the uppermost plant tips. The canopy structure methodology could be used for other species to estimate the underestimation of canopy height derived from SfM. Overall, this study provides a framework for selecting drone flight settings, processing parameters, and predicting canopy-height correction factors to improve DSM-based plant height measurements in dense canopies and heterogeneous vegetation surfaces.

Plant phenotyping relevance

ドローンRGB-SfMによる植物群落高の推定補正法を開発し、RTK-GNSSで検証しており、植物形質取得手法が中心である。

Code and data availability

公開論文であることは確認できましたが、現在の公式API・許可済み取得経路では本文を自動取得できませんでした。

No evidence-backed public reproduction asset is currently recorded.