Document Type : Research article
Authors
1
Ph.D. Student of Forestry, Faculty of Natural Resources, Lorestan University, Khorramabad, Iran
2
Corresponding author, Associate Prof., Department of Forestry, Faculty of Natural Resources, Lorestan University, Khorramabad, Iran
3
Associate Prof., Department of Photogrammetry and Remote Sensing, Faculty of Geodesy and Geomatics Engineering, K.N. Toosi University of Technology, Tehran, Iran
4
Associate Prof., Department of Forestry, Faculty of Natural Resources, Lorestan University, Khorramabad, Iran
Abstract
Background and objectives: Over the last two decades, the use of 3D remote sensing technologies such as mobile phone LiDAR and photogrammetric methods to extract point cloud information in forestry has grown significantly. Mobile phone LiDAR, due to its laser scanning capability, allows accurate 3D measurements of objects in a short time for single trees. This study measured crown base height and diameter at breast height under similar atmospheric and lighting conditions using close-range photogrammetry and mobile phone LiDAR approaches. The measurement process time from data collection to dense point cloud generation was also considered. The aim was to evaluate and compare the performance of these two methods in terms of accuracy for breast diameter and stem height estimation and acquisition time for 3D tree stem modeling.
Methodology: Sixteen individual oak trees were selected from four central Zagros sites in western Iran. The close-range photogrammetry method involved capturing images with a 360-degree rotation around the tree under manual focus and adequate lighting. For mobile LiDAR, the stem was scanned by walking around the tree. After image acquisition, processing steps included interior orientation, camera location determination, and input of control and check lengths previously marked on the stems, to produce 3D visualizations and dense point clouds. The iPhone LiDAR data were processed using Scaniverse software. Control and check lengths were measured directly with calipers before data collection. In Metashape, distances between check and control points were optimized by marking. Statistical parameters including RMSE, RMSE%, MAE, and MAE% were calculated by comparing intervals from photogrammetry and LiDAR point clouds to direct caliper measurements.
Results: The photogrammetry workflow and processing steps averaged across the 16 trees are tabulated. Obtaining a dense point cloud via photogrammetry requires completing ten processing steps. Processing time depends on the system used. In terms of timing, close-range photogrammetry requires ten steps after data acquisition to produce dense point clouds, whereas mobile phone LiDAR only involves two stages: scanning and direct processing. Results showed that the time needed to generate dense point clouds using photogrammetry was roughly 21 times longer than with mobile LiDAR for each tree. The photogrammetry approach yielded better accuracy for crown base height (RMSE = 6.63%), while mobile phone LiDAR performed better for breast height diameter estimation (RMSE = 1.86%).
Conclusion: Zagros oak trees, being monoecious and seed-propagated, provide a suitable test case for applying these technologies in forest cover studies. Tree characteristics such as thin crowns, sufficient light penetration to the stem, low altitude, and low crown branching facilitate measuring stem height. Comparing photogrammetric and LiDAR methods remains challenging due to different quality and accuracy criteria used since these technologies emerged. Ultimately, the choice between these technologies depends on whether the focus is single-tree or mass tree measurements.
Keywords
Main Subjects
)