Shoot a photo from a Raspberry Pi controlled camera using the timelapse.py script on a configurable frequency. Send the files up to Google Cloud bucket. Rinse, wash, repeat.
Check out the setup markdown for more specific instructions. This project requires:
- a Raspberry Pi
- a camera which supports gphoto2
- power supplies, USB cables
- a Google Cloud account with storage (Google Cloud Storage, aka GCS)
- we use a service account for the upload to GCS
- Github
The STL files to 3d print my version of the enclosure are in the subdirectory "3d_print_files". That subdirectory also includes a PDF with additional information, including a visualization and complete parts list.
The script generates JPG images in the format YYYY-MM-DD_HH-MM-ss.jpg. I use ffmpeg to put those all together into an mpeg video. First I download all JPGs from the bucket onto a Google Compute instance within my cloud project (probably not the best) with this command:
gsutil cp -r gs://<bucket_name>/ .
Then cd into that directory and run the ffmepg command, to create "output.mp4":
ffmpeg -framerate 30 -pattern_type glob -i 2023-04-*.jpg -c:v libx265 output.mp4
In no particular order:
- Create an array of times for each day that finds "interesting" times, such as sunrise, sunset, or other times of day that might be visually interesting.
- Take pictures only during the interesting times (ex. - sunrise, sunset)
- Take less frequent pictures at night (maybe?)