Steps to build the output:
- Record videos, preferrably in 60fps, save them in a
videos
directory and name them according to thestations.csv
spreadsheet - Run the
extract_frames.sh
script which dumps all the video frames to PNG and colour-corrects a couple of stations (note, this will result in approximately 36,000 images, for a total of around 4 gigabytes. - Go through each station and find the first frame number that shows movement, and the last frame number that shows the platform, and record those in
stations.csv
- run
render.py
to build the final output