Codes for the paper "Individual tree detection from Unmanned Aerial Vehicle (UAV) derived point cloud data in a mixed broadleaf forest using hierarchical graph approach", which is published in European Journal of Remote Sensing.
The contributors are Seyed Ali Ahmadi, Arsalan Ghorbanian, Farshad Golparvar, Ali Mohammadzadeh, and Sadegh Jamali.
In this study, graph-based approach was developed for detecting individual trees in a broadleaf, complex forest region based on UAV-derived point cloud data. Horizontal cross-sections at different heights were applied to the Canopy Height Model (CHM) to extract initial candidates for graph nodes. The graph was processed in multiple steps, and individual treetop locations were detected based on graph nodes’ properties. The impact of various parameters, such as minimum area of connected components and minimum tree heights, on the performance of the developed method, were investigated. The evaluation step demonstrated the potential of the proposed graph-based method for individual tree detection in a complex forest region in Mazandaran, Iran. In particular, the graph-based method obtained precision, recall, and F1-score values of 0.64, 0.73, and 0.68, respectively. Furthermore, the intercomparison with the well-known and most used Local Maximum (LM) suggested the applicability of the proposed method. After point cloud generation, the proposed method was implemented entirely in Python using open-source packages, which increases its usability for other scholars and managers.
https://doi.org/10.1080/22797254.2022.2129095
@article{ahmadi2022individual,
title={Individual tree detection from unmanned aerial vehicle (UAV) derived point cloud data in a mixed broadleaf forest using hierarchical graph approach},
author={Ahmadi, Seyed Ali and Ghorbanian, Arsalan and Golparvar, Farshad and Mohammadzadeh, Ali and Jamali, Sadegh},
journal={European Journal of Remote Sensing},
volume={55},
number={1},
pages={520--539},
year={2022},
publisher={Taylor \& Francis}
}