Volume 10, Issue 20 (11-2022)
ifej 2022, 10(20): 193-203 |
Back to browse issues page
Master of Remote Sensing, General Department of Natural Resources and Watershed Management of Fars Province, Shiraz, I.R. Iran
Abstract: (1705 Views)
Extended Abstract
Introduction and Objective: Advances in the field of UAVs and sensors, has provided access to high-resolution images and 3D data of them can be used to monitor of forests and evaluate the characteristics of trees. In this study, low-cost UAV sensors and algorithms SFM images have been used to extract the crown height model. then Local Maxima algorithm was used to detect trees.
Materials and Methods: Current research was conducted in the Sisangan Forest Park that located in 30 km east of Nowshahr. Based on the Orthomosaic of drone images, six sample plots with 30 to 30 m dimensional were designed.
Results: A total of 209 trees were recorded as ground reference, which allowed the algorithm to accurately detect 139 of them with the F-Score of 0.63. Applied algorithm detected 88 points by mistake, and lost 72 trees. Our r findings showed that the less variety of species and the crown of the trees are symmetrical and in a crown diameter range, the performance of the detection algorithm is better. Also, by using LiDAR sensors, better results may be achieved. generally, the results of this investigation indicated that the in open- canopy forests with the tangier and Also, the trees are symmetrical trees, the acceptable outcome might be access.
Conclusion: Overall, the local Maxima algorithm did not detect single-trees. These results therefore need to be interpreted with caution in broadleaves forests.
Introduction and Objective: Advances in the field of UAVs and sensors, has provided access to high-resolution images and 3D data of them can be used to monitor of forests and evaluate the characteristics of trees. In this study, low-cost UAV sensors and algorithms SFM images have been used to extract the crown height model. then Local Maxima algorithm was used to detect trees.
Materials and Methods: Current research was conducted in the Sisangan Forest Park that located in 30 km east of Nowshahr. Based on the Orthomosaic of drone images, six sample plots with 30 to 30 m dimensional were designed.
Results: A total of 209 trees were recorded as ground reference, which allowed the algorithm to accurately detect 139 of them with the F-Score of 0.63. Applied algorithm detected 88 points by mistake, and lost 72 trees. Our r findings showed that the less variety of species and the crown of the trees are symmetrical and in a crown diameter range, the performance of the detection algorithm is better. Also, by using LiDAR sensors, better results may be achieved. generally, the results of this investigation indicated that the in open- canopy forests with the tangier and Also, the trees are symmetrical trees, the acceptable outcome might be access.
Conclusion: Overall, the local Maxima algorithm did not detect single-trees. These results therefore need to be interpreted with caution in broadleaves forests.
Type of Study: Research |
Subject:
سنجش از دور
Received: 2021/11/3 | Accepted: 2022/05/1 | Published: 2022/10/23
Received: 2021/11/3 | Accepted: 2022/05/1 | Published: 2022/10/23
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |