You can deploy this model and start earning money today!

forestinvlab/
chm-builder

Views

No views yet

🌲 CHM Model Builder

Generate Canopy Height Models (CHM) directly from LiDAR point clouds (LAS/LAZ).

What it does

Reads any LAS/LAZ point cloud file
Classifies ground points (via LAS classification codes, percentile thresholds, or lowest-returns)
Rasterizes a Digital Surface Model (DSM = max elevation per pixel)
Rasterizes a Digital Terrain Model (DTM = ground elevation per pixel)
Produces CHM = DSM − DTM as a GeoTIFF

Usage

Python API

python

from chm_builder import CHMBuilder, CHMConfig

builder = CHMBuilder(CHMConfig(resolution=1.0, ground_method="las_class"))
meta = builder.build("input.laz", "output_chm.tif")
print(meta)

Command Line

python chm_builder.py input.laz output_chm.tif --resolution 0.5 --ground-method las_class

Dependencies

pip install laspy[lazrs] rasterio scipy numpy

Methods

Ground Method	Description
`las_class`	Uses LAS classification code 2 (standard ground)
`percentile`	Points below a global elevation percentile
`lowest`	Lowest return in each raster cell
`csf`	Cloth Simulation Filter (requires `cloth-simulation-filter`)

Output

GeoTIFF (GTiff) with CHM heights
Metadata JSON with bounds, resolution, and statistics (min, max, mean, median, std)

Gradio Demo

Try the live demo: https://huggingface.co/spaces/forestinvlab/chm-model-builder

License

MIT

forestinvlab/chm-builder