Volume 14, Issue 1 (7-2026)
Ecol Iran For 2026, 14(1): 121-133 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
hasanvand H, akbari H, shataee Jouibary S, lotfalian M, hosseinpour A. (2026). Investigating the Elevation Accuracy of a Digital Surface Model Obtained from UAV Images in Forest Areas (Case Study: Darabkola Forests of Sari). Ecol Iran For. 14(1), 121-133. doi:10.61882/ifej.2026.561
URL: http://ifej.sanru.ac.ir/article-1-561-en.html
URL: http://ifej.sanru.ac.ir/article-1-561-en.html
1- Department of Forestry, Faculty of Natural Resources, Sari Agricultural Sciences and Natural Resources University, Sari, Iran
2- Department of Forestry, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran
3- Employee of Natural Resources of Gologah, Faculty of Forestry, Sari Agricultural Sciences and Natural Resources University, Sari, Iran
2- Department of Forestry, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran
3- Employee of Natural Resources of Gologah, Faculty of Forestry, Sari Agricultural Sciences and Natural Resources University, Sari, Iran
Abstract: (842 Views)
Extended Abstract
Background: The construction of forest roads serves as a vital lifeline for the management and planning of the Hyrcanian forests. It is essential for utilizing various forest services, implementing conservation and restoration operations, and transporting timber and non-timber forest products, thereby playing a significant role in the national income. However, road construction can disrupt natural watercourses, create impervious surfaces, and induce runoff through side ditches, cross-drains, and cut-and-fill slopes. If forest roads are properly designed and distributed, they cause minimal damage to the biological mass and habitats, allowing the forest to maintain its optimal conditions for efficient management. Given the necessity of obtaining up-to-date information regarding forest micro-topography and the surface condition of forest roads to assess surface degradation, there is a need for a precise study of modern remote sensing methods, such as UAV imagery, and their required processing requirements. One effective quality control method for roads is a management and planning system based on UAV technology, which offers comprehensiveness, high precision, cost-effectiveness, and speed compared to traditional methods involving manual field inspections by technical experts. Furthermore, the Digital Surface Model (DSM) is a key product of remote sensing that serves as the foundation for many practical projects; photogrammetry enables the generation of accurate 3D models from 2D images. The objective of this study was to investigate the vertical accuracy of DSMs derived from UAV imagery, comparing models generated with varying numbers and distributions of accurate Ground Control Points (GCPs) against those generated without GCPs, using both high and medium-quality processing methods in the Type-1 Darabkola forests of Sari.
Methods: UAV images in suitable weather conditions with a flight height of100 meters above the ground, overlapping length of 85% and width of 75%, and with a speed of 5 meters per second by Phantom 3 Pro quadcopter UAV images in days 3, 4 February 2023 in part of Darabkla forests with a length of about 2 km, 200 meters wide road around the road and in three separate flight projects were received during different days. The position of the points of ground reality observations was done using differential DGPS. Several ground control points were made for accurate ground referencing of the images and the preparation of a precise and integrated orthophoto mosaic of the harvested area using differential DGPS. Digital elevation models of the area prepared with ground control points and without control points were processed and compared using Agiosoft Photoscan software in high and medium-quality modes.
1.01 m for the digital height model prepared with high-quality control points and 1.21 m for the medium quality method. However, it was 11.05 meters high and 10.72 meters with average quality for the digital height model prepared without high-quality control points. Moreover, the root mean square absolute error (RMSE) of the height values was 2.33 meters for the height digital model prepared with control points with high-quality method and 2.39 meters with medium quality . For the digital model, the height prepared without control points was 13.67 meters with high quality and 13.42 meters with medium quality. The root means square relative error (RMSE%) values were was 0.87% and 0.90% for the height digital model prepared with high-quality control points and with medium quality, respectively. However, values of 5.14% and 5.04% were obtained for the height digital model prepared without high-quality control points and average quality, respectively. In addition, the results of the paired t-test at 5 and 1% probability levels of error showed that no significant difference between the height of the ground collection points and the digital height model with high and medium quality control points. However, there was a significant difference with high and medium quality methods between the height of the ground collection points and the digital height model of the preparation without control points.
Conclusion: The use of accurate Ground Control Points (GCPs) for georeferencing and orthomosaic production significantly improves the accuracy of DSM generation. Furthermore, comparing the two processing levels showed that high-quality processing yields lower RMSE values; therefore, it is recommended to use high-quality settings in Agisoft Metashape Professional 1.8. The results demonstrate that integrating accurate GCPs with UAV imagery ensures the production of highly accurate DSMs, and that high-quality processing enhances both vertical and horizontal precision. Notably, due to the sensitivity of the parameters, image acquisition was conducted during the deciduous season when the trees were leafless, ensuring a clear view of the forest floor (absence of canopy cover). To minimize errors in DSM production, the simultaneous use of GCPs.
Background: The construction of forest roads serves as a vital lifeline for the management and planning of the Hyrcanian forests. It is essential for utilizing various forest services, implementing conservation and restoration operations, and transporting timber and non-timber forest products, thereby playing a significant role in the national income. However, road construction can disrupt natural watercourses, create impervious surfaces, and induce runoff through side ditches, cross-drains, and cut-and-fill slopes. If forest roads are properly designed and distributed, they cause minimal damage to the biological mass and habitats, allowing the forest to maintain its optimal conditions for efficient management. Given the necessity of obtaining up-to-date information regarding forest micro-topography and the surface condition of forest roads to assess surface degradation, there is a need for a precise study of modern remote sensing methods, such as UAV imagery, and their required processing requirements. One effective quality control method for roads is a management and planning system based on UAV technology, which offers comprehensiveness, high precision, cost-effectiveness, and speed compared to traditional methods involving manual field inspections by technical experts. Furthermore, the Digital Surface Model (DSM) is a key product of remote sensing that serves as the foundation for many practical projects; photogrammetry enables the generation of accurate 3D models from 2D images. The objective of this study was to investigate the vertical accuracy of DSMs derived from UAV imagery, comparing models generated with varying numbers and distributions of accurate Ground Control Points (GCPs) against those generated without GCPs, using both high and medium-quality processing methods in the Type-1 Darabkola forests of Sari.
Methods: UAV images in suitable weather conditions with a flight height of
Conclusion: The use of accurate Ground Control Points (GCPs) for georeferencing and orthomosaic production significantly improves the accuracy of DSM generation. Furthermore, comparing the two processing levels showed that high-quality processing yields lower RMSE values; therefore, it is recommended to use high-quality settings in Agisoft Metashape Professional 1.8. The results demonstrate that integrating accurate GCPs with UAV imagery ensures the production of highly accurate DSMs, and that high-quality processing enhances both vertical and horizontal precision. Notably, due to the sensitivity of the parameters, image acquisition was conducted during the deciduous season when the trees were leafless, ensuring a clear view of the forest floor (absence of canopy cover). To minimize errors in DSM production, the simultaneous use of GCPs.
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
