Evaluation Crown Height Model Extracted from the UAV in Individual Tree Detection in Sisangan Forest Park
2022
Mohammad reza kargar (Master of Remote Sensing, General Department of Natural Resources and Watershed Management of Fars Province, Shiraz, I.R. Iran) | farhad khabazi (PhD student in Forest Science and Engineering, Tarbiat Modares University Faculty of Natural Resources and Marine Sciences, Noor, Mazandaran, I.R. Iran) | aref hesabi (PhD student in Forest Science and Engineering, Tarbiat Modares University Faculty of Natural Resources and Marine Sciences, Noor, Mazandaran, I.R. Iran)
Extended AbstractIntroduction 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.
显示更多 [+] 显示较少 [-]