search

coloradocube / balloonsat

COTS balloonsat mission to test the RPi 4 with a battery UPS, a quad camarray, a GPS module, a RockBLOCK module, and a small number of Qwiic sensors
0 stars 5 forks source link

The camera code should be in a separate process #80

Open ivogeorg opened 2 years ago

ivogeorg commented 2 years ago

Description

Recording video and still images does not belong in the main telemetry gathering loop.

  1. The camarray recording should be in a separate process.
  2. The camarray should record videos of no more than 30 seconds at once unless it has a mechanism to stop upon being sent a shutdown notification (not necessarily a Linux signal). If it does, than it can record for longer.
  3. The videos should be interspersed with stills.
  4. Catching the bursting of the balloon requires continuous operation once at altitude close to expected bursting altitude. This requires monitoring the altitude to start continuous video recording early enough. Also, it requires scheduling an end of the recording after the bursting and the payloads drop out, possibly timed to be after the deployment of the parachute. Can the camera/camarray be programmed to operate this way?
  5. Similarly, if the landing is to be recorded, the altitude should also be monitored to start a video. In this case, stopping the video should be timed to happen after landing, and therefore the state of flight-vs-ground should be monitored.
  6. The camarray is recording very large frames. The recording process should monitor the filling of memory and throttle, if necessary.