Barro Colorado Island 50-ha Plot Crown Maps, CanopyRS Preprocessed Tiles: Drone RGB Tree Crowns of a Panamanian Tropical Forest (Smithsonian)
- Provider
- Vasquez et al. (Smithsonian Tropical Research Institute); preprocessed by the CanopyRS authors
- License
- CC-BY-4.0 (source)
- Size
- 55 files
- Format
- Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/CanopyRS/BCI50ha)
- Access
- Free and public. No account with the data provider is needed.
- DOI
- 10.25573/data.24784053
- Category
- Environmental Science
- Source
- https://huggingface.co/datasets/CanopyRS/BCI50ha
About this dataset
A tiled version of the Barro Colorado Island (BCI) 50-hectare forest dynamics plot crown maps of Vasquez, Cushman, Ramos, Williamson, Villareal, Gomez Correa and Muller-Landau (Smithsonian Tropical Research Institute, Panama), prepared by the CanopyRS authors as an evaluation set for the SelvaBox tropical tree crown detection benchmark. The source deposit combines drone (UAV) RGB orthomosaics of the BCI 50-ha plot with manually delineated and instance-segmented tree crown polygons, part of the Panama Forest Landscapes UAV collection. Here the imagery is cut into 513 test tiles with COCO-format crown annotations, stored as 53 Parquet shards under data/ (about 27 GB), readable or streamable with datasets.load_dataset. Useful for testing tree crown detectors and segmenters in closed tropical canopy, canopy tree demography and linking crowns to the long-term BCI census. Source dataset DOI 10.25573/data.24784053, licensed CC BY 4.0 by the Smithsonian authors; cite Vasquez et al. (2023) and the SelvaBox paper.
How to load this dataset in Python
from datasets import load_dataset
ds = load_dataset("CanopyRS/BCI50ha")
print(ds)Data structure
File types in sampled files: Parquet
Files
Example file: data/test-00000-of-00053.parquet
Analyze this dataset without downloading it
On Scigantic, Barro Colorado Island 50-ha Plot Crown Maps, CanopyRS Preprocessed Tiles: Drone RGB Tree Crowns of a Panamanian Tropical Forest (Smithsonian) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Environmental Science datasets
- WCS Camera Traps
- The Wilds camera traps: images and video from four trap stations, Ohio, summer 2025
- Izembek Lagoon Waterfowl Aerial Imagery (Alaska)
- HKH Glacier Mapping (ICIMOD, Landsat 7 and SRTM)
- BIRCH-Trees: UAV imagery benchmark for individual tree height and species estimation
- Eastern Bering Sea Summer Bottom Trawl Survey: Specimen-Level Crab Data, 1975-2018 (NOAA AFSC)
- EPA Risk-Screening Environmental Indicators
- Boston BERDO building energy and emissions disclosure, 2015 to 2024