Image to attribute model for trees (ITAM-T): individual tree detection and classification in Alberta boreal forest for wildland fire fuel characterization

ABSTRACT

Regional and municipal decision makers rely on fuel (vegetation) maps to inform decisions on tree stand management related to wildfire management and response. Remote sensing of trees is used in commercial applications but has limited uptake in the fire management community. A two-stage detection and identification convolutional network for high-resolution RGB drone imagery is developed to address this limitation. The detection routine is based on DeepForest, an existing convolutional neural network implementation designed to recognize trees in aerial imagery. Retraining the model and implementing an adaptive window-size workflow improves tree detection, with F₁ scores reaching 85% and averaging 72% for k-fold cross-validation in boreal forest. For classification, a VGG19 network with added data augmentation and dropout layers is trained. When this network is implemented, manually annotated trees are recognized as coniferous with an average F₁ of 97% and deciduous with an 87% F₁. Overall, the developed image-to-attribute model for trees reaches a maximum F₁ score of 85% considering classification after identification, with averages of 72% for coniferous trees and 57% for deciduous trees over six sites. Tree height, size, and stem density are extracted from the tree location output and geometric data. The calculated density is compared to the density of manual annotations, with an average R² of 0.90. A remote preliminary proximity-based hazard assessment is performed on a rural property in Alberta, demonstrating the model’s ability to detect and classify trees near values-at-risk. The results indicate a potential extension to low-cost decision support in enterprise and fire-related applications.

KEYWORDS:

• fire management
• convolutional neural network
• cnn
• tree attributes
• tree detection
• tree classification
• UAV imagery
• fire fuel
• wildfire
• deep network

Acknowledgements

We acknowledge the support of Dr. Jen Beverly of the Department of Renewable Resources at the University of Alberta for consulting on wildland fuel considerations, and Greg Sather and Bev Wilson at Alberta Agriculture and Forestry for facilitating orthophoto interpretation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data may be shared upon reasonable request. Surveys conducted within the bounds of provincial test plots may not be available due to standing agreements with the province.

Additional information

Funding

This research was funded in part by Alberta Agriculture and Forestry through the Canadian Partnership for Wildland Fire Science, grant agreement number 18GRWMB06