Garry Oak Tree Detection in Victoria and Saanich Using LiDAR and Orthophotography

A study examined Garry oak crown detection in urban forests using high-density airborne laser scanning and fused high-resolution orthophotography. The fusion model achieved an overall accuracy of 86.2%, increasing producer's accuracy for Garry oak to 74.3%. The dataset was created by Jiaju Wu and last updated on 2026-05-02.

Use Cases

• Comparing LiDAR-only versus fused remote sensing models for tree species classification based on crown-level structural and spectral features.
• Training random forest classifiers for urban tree species detection based on canopy height models and spectral variability.
• Mapping the spatial distribution of a specific tree species in a fragmented urban ecosystem based on inventoried locations.
• Analyzing the variable importance of near-infrared information versus LiDAR metrics in species classification models.

Strengths

• The study reports specific model performance metrics, including an overall accuracy increase from 77.5% to 86.2%.
• The fusion approach is described as reducing false positives and creating a more realistic spatial distribution aligned with inventoried locations.
• Data sources are specified as high-density (45 points/m²) airborne laser scanning and high-resolution (10cm) four-band orthophotography.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale modeling tasks.

Provenance

Source

Borealis Harvested Dataverse

Collection Method

Airborne laser scanning and orthophotography, with individual tree crown segmentation and feature extraction for classification.

Time Range

null

Freshness

Last updated 2026-05-02 04:11:01; freshness should be verified.

Geography

Urban forests of Victoria and Saanich, Greater Victoria, British Columbia, Canada.