Discovered by @engelsjk on Twitter.
The contents of this repository are a work in progress but in general aim to automate (to whatever degree possible) the processing of related data to understand the actual motion of the spacecraft.
The files starting with a number can be run in order to process the core data and some visualizations.
Some directories are deleted automatically on-rerun and some are not. Trying to be friendly to people altering the data between phases for R&D purposes.
Everything after "phase 4" (4-warp.py) is not really functional.
Install GDAL, ImageMagick, ffmpeg, and the Python dependencies (in a virtual environment below. On a Mac with Homebrew this might look like:
$ brew instal python gdal imagemagick ffmpeg
$ python -m venv venv
$ source venv/bin/activate
(venv) $ pip install affine rasterio requestsExtract only the squiggle from the full EarthExplorer dataset.
Scrape low-res thumbnails from the USGS site.
Add georeferencing information to the thumbnails.
Reproject thumbnails to ortho-space. This produces one set of images that are naively reprojected to a target resolution and size that GDAL determines to be appropriate, and another set of images that are reprojected to a fixed grid. The latter are assembled into video frames.
Generate individual video frames. Unclear how necessary this is and seems to depend on if ffmpeg or ImageMagick is used to generate a video. The latter has a method for compositing new data into the previously rendered frame, which avoids some more bespoke compositing options.
Convert composited frames to files that either ffmpeg or ImageMagick can interpret. Again, this isn't quite done
Render a video with either ffmpeg or ImageMagic,. Unscoped.
Applies all possible versions of a thumbnail with GCPs applied. The idea is that a human should just brute-force this process since the list is fairly small. Probably needs a reference layer to help determine if a given GCP order is correct.
Country borders provide some context:
$ wget https://www.naturalearthdata.com/download/10m/cultural/ne_10m_admin_0_countries.zip