Estimating maize plant height using a crop surface model constructed from UAV RGB images

Authors: NIU, YAXIAO - Northwest A&f University; HAN, WENTING - Northwest A&f University; Zhang, Huihui; ZHANG, LIYUAN - Northwest A&f University; CHEN, HAIPING - Northwest A&f University

Submitted to: Biosystems Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/2/2024
Publication Date: 4/6/2024
Citation: Niu, Y., Han, W., Zhang, H., Zhang, L., Chen, H. 2024. Estimating maize plant height using a crop surface model constructed from UAV RGB images. Biosystems Engineering. 241:56-67. https://doi.org/10.1016/j.biosystemseng.2024.04.003.
DOI: https://doi.org/10.1016/j.biosystemseng.2024.04.003

Interpretive Summary: Plant height is an important agronomic trait for assessing crop growth. The study aims to develop a method to extract maize plant height from UAV RGB images. We demonstrated a process to construct an accurate digital terrain model (DTM) and choose the optimal plant height feature from the crop surface model (CSM). Moreover, we investigated the influence of fractional vegetation cover (FVC) on the construction accuracy of DTM for the first time, and the influences of view angle (oblique and nadir) and spatial resolution on the accuracy of plant height estimation. Results show that the accuracy of DTM constructed based on the selected soil scatters and the inverse distance weighted algorithm was significantly influenced by FVC conditions. Compared to the DTM constructed using UAV images over bare soil shortly after sowing, FVC less than 0.4 was demonstrated as the necessary condition for the accurate construction of DTM. This study provided an instructive guidance for the accurate DTM construction for PH estimation using UAV-based high-resolution RGB imagery.

Technical Abstract: Plant height (PH) is an important agronomic trait and can be used for assisting crop breeding pipelines, assessing crop productivity, and making crop management decisions. Two key steps for PH estimation are 1) accurate construction of the digital terrain model (DTM), and 2) selection of optimal PH feature from the crop surface model (CSM) obtained by using structure from motion technique on unmanned aerial vehicle (UAV) images. The influence of fractional vegetation cover (FVC) on DTM construction accuracy was investigated for the first time, and the influences of view angle (oblique and nadir) and spatial resolution on the accuracy of maize PH estimation were explored. Results show that the accuracy of DTM constructed based on selected soil scatters and inverse distance weighted (IDW) algorithm was significantly influenced by FVC conditions. Compared to the DTM constructed using UAV images over bare soil shortly after sowing, FVC less than 0.4 was necessary for accurate construction of DTM, with averaged estimation errors of 15 cm in 2018 and 9 cm in 2019. The optimal PH feature, with the smallest errors compared to ground-truth PH, was the 99th percentile in 2018 (root mean square error (RMSE) of 22.69 cm, mean absolute error (MAE) of 18.64 cm) and 100th percentile in 2019 (RMSE of 18.73 cm, MAE of 14.53 cm); the differences were mainly caused by different UAV flight missions. Compared to the grid mission with the nadir view, the double grid mission with the oblique view resulted in a more accurate 3-D reconstruction. When the original spatial resolution of 15 mm was upscaled to 20, 30, 60, and 120 mm by the bilinear interpolation algorithm, a decreasing trend of PH estimation accuracy was observed, with RMSE increasing from 34.70 to 39.98 cm and MAE increasing from 28.96 to 36.15 cm. Overall, this study provides an instructive guidance for the accurate DTM construction for PH estimation using UAV-based high-resolution RGB imagery.