Figure 4 - uploaded by Onur Can Bayrak
Content may be subject to copyright.
Confusion matrix of YOLOv8-l prediction results (Green: True Positives, Red: False Positives and False Negatives)
Source publication
Forest monitoring and tree species categorization has a vital importance in terms of biodiversity conservation, ecosystem health assessment, climate change mitigation, and sustainable resource management. Due to large-scale coverage of forest areas, remote sensing technology plays a crucial role in the monitoring of forest areas by timely and regul...
Context in source publication
Context 1
... illustrated in Figure 4, number of true positives for Abies, Alnus, Betula, Cleare, Fagus, Fraxinus, Larix, Picea, Pinus, Pseudotsuga and Quercus classes are relatively higher than underrepresented classes Populus, Prunus, and Tilia. It is thought that the small number of training and test data causes this situation. ...
Similar publications
![UAV based fire detection and management system [16]](profile/Divneet-Kapoor/publication/374476715/figure/fig1/AS:11431281234021391@1712196379445/UAV-based-fire-detection-and-management-system-16_Q320.jpg)
![Ground-based fire detection systems overview [27]](profile/Divneet-Kapoor/publication/374476715/figure/fig2/AS:11431281234010932@1712196379819/Ground-based-fire-detection-systems-overview-27_Q320.jpg)
According to the most current Global Forest Resources Assessment (GFRA) report, from 2003 and 2012, there were about 67 MHa of forest fires each year, with 98 MHa of the total occurring in 2015. IoT applications like real-time forest monitoring systems can assist to mitigate the growing environmental effect of forest fires. The research paper prese...
Citations
... During this application, YOLOv8 achieved a precision of 0.923, a recall of 0.951, and an F1 score of 0.937 [51]. Authors evaluated the performance of YOLOv8 variants through the classification of tree species from aerial imagery, during which YOLOv8-l outperformed other YOLOv8 variants, achieving weighted and micro-average scores of 71.55% and 72.70%, respectively [52]. Studies have also been carried out for the detection of weeds in lawns using YOLO object detectors. ...
The objective of this project was to automate the detection and counting process of stacked eucalypt (hybrid Eucalyptus urophylla x Eucalyptus grandis) timber in the forestry industry using the YOLOv8 model. The dataset consists of 230 diverse images of eucalypt roundwood, including images of roundwood separated on a rail and stacked timber. The annotations were made using LabelImg, ensuring accurate delineation of target objects on the log surfaces. The YOLOv8 model is customized with a CSPDarknet53 backbone, C2f module, and SPPF layer for efficient computation. The model was trained using an AdamW optimizer and implemented using Ultralytics YOLOv8.0.137, Python-3.10.12, and torch-2.0.1 + cu118 with CUDA support on NVIDIA T1000 (4096MiB). For model evaluation, the precision, recall, and mean Average Precision at a 50% confidence threshold (mAP50) were calculated. The best results were achieved at epoch 261, with a precision of 0.814, recall of 0.812, and mAP50 of 0.844 on the training set and a precision of 0.778, recall of 0.798, and mAP50 of 0.839 on the validation set. The model’s generalization was tested on separate images, demonstrating robust detection and accurate counting. The model effectively identified roundwood that was widely spaced, scattered, and overlapping. However, when applied to stacked timber, the automatic counting was not very accurate, especially when using images. In contrast, when using video, the relative percentage error for automatic counting significantly decreased to −12.442%. In conclusion, video proved to be more effective than images for counting stacked timber, while photographs should be reserved for the counting of individual roundwood pieces.
