Read an rtsp video feed, extract image for use in making timelapse video and website snapshot and write to the appropriate directories.
I built this to work with a TP-link Tapo C320WS camera, but should work with any camera providing an RTSP feed.
It's not overly complicated, but assumes you're a bit competent with python, opencv and yaml. I'm no guru, so have tried to limit anything overly elegant.
- Copy example_config.yaml to config.yaml, and modify to suit your situation.
- create the any directories needed for your configuration, both local and on remote servers
- Change the cameraImage.py file to be executable.
- Install any necessary python libraries.
I call it once per minute with a cron job.
* * * * * cd /home/fred/bin; /home/fred/bin/cameraImage.py
Thanks to https://www.briandorey.com/post/tplink-tapo-tc65-camera-capture-opencv-webcam for the starting point.
segmentAssemble.py merges the timelapse images and any motion videos together into one integrated video.
- create directories ./motion and ./assemble under the directory holding your timelapse images download_motions.py syncs the motion detection videos on the camera onto your server
- I use a cronjab every 5 minutes to do this
**
5,15,25,35,45,55 * * * * /home/fred/bin/download_motions.py
Each day after 0600h, run python segmentAssemble.py, then go into the assemble directory and run bash assemble.sh. On my poor old PC, it typically takes most of an hour to process. My GPU cannot handle 2K video size. Update: limiting motion videos to their first ten seconds helps.