VibrantVS Canopy Height Model (CHM)

Validation Sample

Published by
Vibrant Planet
Story Last updated on
March 13, 2025

About

This dataset release provides a small validation sample of canopy height model (CHM) data for select areas across the western United States. The data includes high-resolution GeoTIFF files for NAIP imagery, LiDAR-derived CHM, and VibrantVS CHM, allowing users to compare model outputs against benchmark LiDAR measurements.

Datasets Included: 

  1. NAIP (National Agriculture Imagery Program) Imagery
    • 4-band aerial imagery (red, green, blue, near-infrared)
    • Serves as the primary input for vegetation structure modeling
  2. LiDAR-Derived Canopy Height Model (CHM)
    • High-precision elevation data from 3DEP LiDAR surveys (2014–2021)
    • Provides ground-truth canopy height measurements for validation
  3. VibrantVS Canopy Height Model (CHM)
    • AI-generated CHM using a multi-task Vision Transformer (ViT) deep learning model
    • Trained on NAIP imagery and benchmarked against LiDAR data
    • Designed for broader spatial coverage and frequent updates (~3-year cycle)

The data is provided in resolution in (EPSG:) and stored as GeoTIFF tiles. Each tile is × pixels, covering about ~ × km on the ground.

This tiling grid is aligned to , making it directly interoperable with those products.

To keep downloads efficient and analysis straightforward, the dataset is organized into a fixed tiling grid. This approach prioritizes optimal file sizes and regional coverage for analytical workflows. While the grid follows standard geospatial conventions, it is not pixel-aligned to USGS ARD or other standardized grids.

How to Get Started:

  1. Explore the Map: Zoom and pan to locate your area of interest. The tile grid outlines the spatial boundaries of the dataset.
  2. Select Your Tiles: Click one or more tiles to highlight them. Each tile corresponds to a geographic area you can download.
  3. Review & Adjust: Use the checkboxes in the selection list to confirm or deselect tiles.
  4. Download Your Data: Click the "Download Data" button to bundle your selected tiles into a single ZIP file.
A Few Things to Keep in Mind:
  • Download Limitations: For best performance, please select no more than 20 tiles per download. Larger selections may slow your browser or cause issues.
  • Need more coverage? No problem—just download in smaller batches. All tiles use the same folder structure, so you can easily merge them later in your GIS software for seamless regional analysis.
  • Working at larger scales? For very large areas, contact us to discuss alternative delivery options.

The dataset available for download below corresponds to the findings presented in the publication, "VibrantVS: A High-Resolution Multi-Task Transformer for ForestCanopy Height Estimation which details the methodology, analysis, and key insights.

For a deeper dive, access the full publication HERE, or read a summary on Vibrant Planet's blog.

Selected Data Tiles: 

Download

VibrantVS Canopy Height Model (CHM)

This dataset provides a small validation sample of canopy height data for select areas across the western U.S. It includes GeoTIFF files for NAIP, LiDAR, and VibrantVS CHM.

MB

Direct Download (Full File) - HTTPS Link

Use this link to download the entire file directly to your local storage through the browser. This method is ideal for quick and easy access but requires sufficient local storage and bandwidth for the full download.

What’s inside

Direct Access

Get hands-on access to this dataset using interactive notebooks. Choose between the Google Colab notebook for quick exploration in your browser or access the hosted Jupyter Notebooks via Binder or GitHub for more advanced workflows.

Direct access to the Google Collab notebook

Open and explore instantly

Click the button to the left to launch an interactive notebook directly in your browser. This pre-configured Colab notebook provides a quick and easy way to explore, visualize, and analyze the data—no setup required.

GitHub hosted Jupyter Notebooks

Flexible access for advanced workflows

Access the full collection of Jupyter Notebooks hosted on GitHub. These notebooks can be used on your local machine or via cloud platforms like Binder or Google Colaboratory, providing flexibility for more advanced customizations.