Skip to main content
ARS Home » Plains Area » College Station, Texas » Southern Plains Agricultural Research Center » Aerial Application Technology Research » Research » Publications at this Location » Publication #425063
Research Project: Advancing Precision Aerial Application for Sustainable Crop Production and Protection

Location: Aerial Application Technology Research

Title: Effects of flight and processing parameters on UAS image-based point clouds for plant height estimation

Author
item Yang, Chenghai
item Suh, Charles
item Fritz, Bradley

Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/19/2026
Publication Date: 1/21/2026
Citation: Yang, C., Suh, C.P., Fritz, B.K. 2026. Effects of flight and processing parameters on UAS image-based point clouds for plant height estimation. Remote Sensing. https://doi.org/10.3390/rs18020360.
DOI: https://doi.org/10.3390/rs18020360

Interpretive Summary: Improving the efficiency and accuracy of drone-based crop monitoring can lead to more reliable crop management practices in precision agriculture. This study evaluated how the number and placement of ground control points (GCPs) and processing parameters affect the accuracy of drone image-based 3D models and plant height estimation. Fieldwork was conducted on an experimental and a commercial field, testing various GCP configurations and processing settings. Results showed that increasing the number of GCPs improved accuracy, but the improvements were minimal beyond five for the small field or seven for the large field. Processing settings directly influenced both processing time and model quality. These findings offer practical guidance for selecting optimal GCPs and image processing parameters to enhance plant height estimation in both small and large fields.

Technical Abstract: Point clouds and digital surface models (DSMs) generated from unmanned aircraft system (UAS) imagery are commonly used for plant height estimation in precision agriculture. This study assessed the effects of ground control point (GCP) quantity, placement, and image processing parameters on point cloud and DSM generation for plant height estimation. Fieldwork was conducted in a 2-ha experimental field with four crops (corn, cotton, sorghum, and soybean) in 2019 and 2022, and in a 32-ha commercial cotton field in 2024. In 2019, 36 panels were placed at three height levels (0, 0.75, and 1.5 m); in 2022, 25 panels were used at two levels (0 and 1.5 m); and in 2024, 13 panels were evenly distributed at ground level across the large field. UAS RGB images were captured on four dates, including two in 2022. Orthomosaics, point clouds, and DSMs were generated using Pix4Dmapper with various GCP configurations and processing parameters. Results from 2019 and 2022 showed that increasing GCPs from 1 to 5 reduced the 3D positional root mean square error (RMSE) to below 0.03 m, with minimal improvement beyond five. Panel height positions had little effect on accuracy. In the large field, 7–9 GCPs provided consistent positional accuracy of 0.12–0.13 m. Correlations between ground-measured and point cloud-extracted plant height values were significant and consistent with four or more GCPs. Processing parameter evaluation showed their impact on 3D point creation, processing time, and plant height estimation. This study provides practical guidance for selecting optimal GCPs and processing parameters for accurate 3D modeling using UAS imagery in both small and large fields.