Parameter configuration to maximize accuracy in point clouds acquired with lidar-based terrestrial mobile laser scanners in fruit tree orchards
The aim of this paper is to determine the optimal parameter configuration to minimize LiDAR point cloud errors and maximize accuracy for agronomic applications in field conditions in the framework of Precision Agriculture. The structure and geometry of trees are closely linked to vegetative paramete...
| Autores: | , , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Universitat de Lleida (UdL) |
| Repositorio: | Repositori Obert UdL |
| OAI Identifier: | oai:repositori.udl.cat:10459.1/469214 |
| Acceso en línea: | https://doi.org/10.1016/j.atech.2025.101573 https://hdl.handle.net/10459.1/469214 |
| Access Level: | acceso abierto |
| Palabra clave: | Configuration of parameters PPK and TC SLAM Point cloud error LiDAR |
| Sumario: | The aim of this paper is to determine the optimal parameter configuration to minimize LiDAR point cloud errors and maximize accuracy for agronomic applications in field conditions in the framework of Precision Agriculture. The structure and geometry of trees are closely linked to vegetative parameters and productivity in fruit orchards. Therefore, achieving precise and accurate geometric characterization of tree canopies is fundamental for developing site-specific management strategies for resource optimization while reducing costs and environmental impacts associated with agricultural practices. Among the various sensing technologies available, LiDAR-based sensors have emerged as one of the preferred options for obtaining accurate geometric data in orchards. For this reason, the paper focuses on analyzing how different parameter configurations used during point cloud production contribute to reconstruction and location errors. Specifically, it examines the influence of type of SLAM algorithm (triple resolution or SLAM3 and feature or SLAMF), the GNSS constellations in use, orbit and clock products, trajectory correction (PPK or TC) and navigation constraint resolution, and the distance to the GNSS base station. Regarding the reconstruction error, since the maximum difference between the best configuration (featuring SLAM3 and PPK) and the worst (featuring SLAMF and TC) is just 0.006 m, all parameter configurations are capable of efficiently reconstructing point clouds. Location error analysis showed that GPS-based solutions consistently exceeded 0.1 m, while GNSS-based solutions - involving any combination of two or more constellations among GPS, GLONASS, Galileo, and BeiDou - remained below 0.07 m. The best configurations include SLAM3 paired with PPK, regardless of the navigation constraint resolution, achieving an error of 0.032 m in the most favorable case. Although PPK trajectories performed better in GNSS-based solutions, they were less effective than TC when only GPS-based solutions were available. Additionally, using PPK trajectories ensures better time efficiency, as they are a prerequisite for generating TC trajectories. SLAM3 outperformed SLAMF under the tested conditions. This study represents a significant advancement in the optimization of LiDAR-derived point clouds generation. By identifying and fine-tuning key parameter configurations, it effectively minimizes both reconstruction and location errors, while simultaneously enhancing the temporal efficiency of the point cloud production process. |
|---|