Ground-Aware Interpolation for UAV-Based Sugarcane Height Estimation: Achieving High Accuracy without GCPs or RTK
<p dir="ltr">Accurate and frequent monitoring of plant height is critical for precision management in sugarcane cultivation. However, conventional UAV photogrammetry workflows are often hampered by the operational inefficiency of using Ground Control Points (GCPs) or the high cost of...
Saved in:
| Main Author: | |
|---|---|
| Published: |
2025
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <p dir="ltr">Accurate and frequent monitoring of plant height is critical for precision management in sugarcane cultivation. However, conventional UAV photogrammetry workflows are often hampered by the operational inefficiency of using Ground Control Points (GCPs) or the high cost of Real-Time Kinematic (RTK) systems. A fundamental challenge for low-cost approaches is the generation of an accurate Digital Terrain Model (DTM) once the ground is obscured by a dense crop canopy. This study introduces a novel method for sugarcane height estimation that operates without GCPs or RTK by intelligently deriving a DTM from on-season RGB imagery. The proposed approach, Ground-Aware Point Interpolation (GAPI), identifies sparse, true-ground elevation points from inter-row spaces within the photogrammetric point cloud and uses them to construct a robust DTM. A naive attempt to filter a DTM from the dense canopy point cloud without this method resulted in a high Root Mean Square Error (RMSE) of 2.81 m, highlighting the severity of this challenge. In contrast, the GAPI method, validated across five distinct sugarcane field types, achieved an average RMSE of 0.39 m and a Mean Absolute Percentage Error (MAPE) of 10% when compared to manual field measurements. By eliminating the need for GCP setup, the proposed approach reduces field operation time for a 10-hectare survey by 80%, from approximately 100 minutes to 20 minutes. This provides a practical, scalable, and cost-effective solution for high-frequency crop monitoring, enabling more timely and data-driven agricultural management.</p> |
|---|