NOTE ON UPDATES: we are working on improvements and extensions to this weekend hackathon project in a private repository for now. This code will likely be open-sourced at a later date. Planned features include data stucture/format conversions, facies boundary modeling, visualization utilities, and GPU acceleration of raw point cloud processing.
3D digital outcrop models are cool. Gee Whiz is it fun to spin them around! Now let's do something interesting with them.
Base:
numpy
matplotlib
sklearn
To do shapefile line projections:
To use mesh
data instead of [XYZ,RGBA]
point cloud:
- PyMesh (Fair Warning: install requires building a lot of
C++
dependencies.)
data/
: Input data, and slice frames directory (the latter not uploaded).
media/
: Presentation slides, figures, movies, and simple ffmpeg
movie generation script.
notebooks/
: Jupyter notebooks. See README
inside for more info.
models/
: pickle
d sklearn.ensemble.AdaBoost
model for (HSV color) --> (sand/shale/heterolithic) classification
mesh_process.py
: Data loading + point cloud slicing + some convenience functions. Should add some things, rename, and have PyMesh
functionality in a seperate module.