Dataset: Census of individual trees of Tucunare Vichada (2020)

Ruiz Hurtado, A.F.; Perez Bolanos, J.; Arrechea Castillo, D.A.; Matiz Rubio, N.; Costa Junior, C.; Arango Mejia, J.; Cardoso Arango, J.A. (2024) Dataset: Census of individual trees of Tucunare Vichada (2020). https://doi.org/10.7910/DVN/RTO8MM

This dataset contains information of individual trees in Tucunare, Vichada-Colombia extracted from remote sensing imagery processed using AI models. Information from the High Resolution 1-meter Global Canopy Heights map from Meta-AI. The dataset includes a CSV file containing attributes related to tree dimensions and geolocation and a shapefile with polygons representing each tree

Metodology:Tree information was obtained by leveraging the High-Resolution 1 m Global Canopy Height map generated by Meta's AI model, trained on a large dataset of satellite images and LIDAR data, to predict canopy heights with a mean absolute error of 2.8 meters. The map tiles were downloaded from Amazon Web Services (AWS) S3 Bucket and processed using additional steps: merging and clipping tiles to cover the region of interest of Tucunare, vectorizing, filtering, and georeferencing individual trees to derive attributes like area, perimeter, equivalent diameter, and circularity representing the visual crown of each tree, and finally applying zonal statistics to assign tree height values. The resulting vector layer contains detailed tree metrics, supporting advanced canopy analysis.