pylynceus 0.1.0.dev0

Open-source photogrammetry and LiDAR processing.

pyLynceus is named for the lookout of the Argonauts, whose sight was so sharp he could see through the earth itself — which is more or less what a good LiDAR pipeline does through canopy. The project aims to be a practical, production-minded toolkit for people who process aerial mapping data for a living: point cloud I/O and filtering, georeferencing and datum work, terrain products, and (on the roadmap) structure-from-motion.

Why another geospatial library?

Existing open tools are excellent at what they do (PDAL for pipelines, OpenDroneMap for end-to-end SfM, laspy for I/O). pyLynceus sits in the gap between them: a Python-native, importable library for the workflows a working photogrammetrist actually runs — classification filtering, spatial deduplication of obstacle datasets, control point residual reporting, CRS/datum transforms, and quick DEM gridding — without leaving a script or notebook.

Install

pip install pylynceus            # core: numpy, laspy, pyproj
pip install "pylynceus[all]"     # + rasterio, opencv, scipy, matplotlib

Quickstart

from pylynceus.io.las import read_las, write_las
from pylynceus.cloud.filters import by_class, dedup_xy, GROUND
from pylynceus.georef.transform import reproject
from pylynceus.terrain.grid import grid_dem

# Read a LAZ tile
cloud = read_las("tile.laz")
print(cloud)                     # PointCloud(4,213,880 points, crs=2240, ...)

# Ground points only, deduplicated within 1 cm in XY
ground = dedup_xy(by_class(cloud, GROUND), tolerance=0.01)

# Reproject and grid a 1-unit DEM
ground = reproject(ground, dst_epsg=6447)
dem, origin = grid_dem(ground, cell_size=1.0)

write_las(ground, "ground.laz")

Or from the command line:

lynceus info tile.laz

Module map

Roadmap

• Vertical datum transforms (ellipsoid ↔ orthometric via geoid grids)
• TIN and IDW DEM interpolation; contour extraction
• Tiled/streaming processing for datasets larger than memory
• Feature extraction & matching (SfM pipeline, sfm extra)
• Obstacle analysis helpers (surface penetration against OIS-style surfaces)
• QGIS plugin wrapper

Development

git clone https://github.com/pyLynceus/pylynceus
cd pylynceus
pip install -e ".[dev]"
pytest
ruff check src tests

Contributions welcome — this project is young and the API is still settling. Open an issue before large PRs.

License

MIT. See LICENSE.