Skip to content

About

Water-aware photogrammetry and Gaussian splatting for shoreline structures

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

1 star

Watchers

0 watching

Forks

Latest commit

 

History

105 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

shoreline-recon icon: textured mesh of the Reposalmi pier

shoreline-recon

Water-aware photogrammetry and Gaussian splatting for shoreline structures

CI Containers Rust License GitHub stars

Vartiosaari pier from one camera position: Gaussian splat on the left, textured mesh with its synthetic water plane on the right

Water, sky and transient objects are masked out of every reconstruction stage, and a synthetic water plane replaces the water surface afterwards. One drone or phone clip of a pier, shore or waterfront goes in; a textured mesh with its water plane and a Gaussian splat come out.

Quick start

brew tap laurigates/tap
brew trust laurigates/tap      # Homebrew loads third-party taps only once trusted
brew install shoreline-recon   # brings ffmpeg, COLMAP and Brush
shoreline-recon pier.mp4

Docker must be running: the segmentation model and OpenMVS run in public images, pulled on first use. Everything runs on the local machine, on the CPU apart from Brush; a clip that keeps 117 frames took about 90 minutes on Apple Silicon.

The deliverables land in ./pier-recon/:

File Contents
model.glb textured mesh with the water plane (glTF binary)
splat.ply Gaussian splat, floaters over the water culled
water_plane.json the fitted water plane and the splat-to-mesh transform
report.json every stage's metrics, fingerprint and tools, and the resolved config

Other options: shoreline-recon --help and docs/usage.md.

A remote CUDA host runs the heavy stages faster and also renders the showcase and processing-step videos (--videos). How each tool runs, locally, in a container or on that host, is machine configuration: see docs/configuration.md.

Other installs: the shell installer and archives for macOS (Apple Silicon) and Linux (x86_64) on the releases page, or from source (docs/development.md). Homebrew installs the shell completions; for the others, see shell completions.

Documentation

Page For
Usage the command line, dataset configs, outputs, batches, demo video, viewer, bundles
Configuration config and tool layers, user files, runtimes, the remote GPU host
Pipeline what each stage does, backends, technology
Development building, testing, project layout, releases, contributing
Kickoff design and milestones
Decisions architecture decision records

License

MIT

About

Water-aware photogrammetry and Gaussian splatting for shoreline structures

Topics

Resources

Code of conduct

Contributing

Security policy

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages