Configure to stream to harddrive

[KamiCreed]

Both Jetson's should save the files to the harddrive attached to the raspi. Could use SSHFS and see how that goes.

• [x] Test hot swapping and what happens during the time where there's no disk
  • [x] Test whether the USB port matters for the auto mounting
  • [x] Test hotswapping on the deployable raspis
• [x] Test what happens when the Starlink network goes down (Possibly deliberatly)
• [x] Test the Starlink sleeping feature
• [x] Test if static IPs can be set for each device
  • Set 192.168.1.* static IP with 192.168.1.1 gateway (Essentially, the same as a regular Starlink)
  • If it works through the network switch, it should work when connected to the Starlink router
• [x] Test parallelized processing and streams on Jetsons
• [x] Finish setting up static IPs for all devices (Except temp gateway)
• [x] Test each Jetson streaming to harddrive
• [x] Setup Jetsons to point to respective camera IPs
• [x] Set raspi 192.168.1.1 (gateway) back to their respective IP address (192.168.1.5) to not block the Starlink router
• [x] Set the two Jetsons to point to correct raspi (192.168.1.5)
• [x] Do a final harddrive streaming test

[KamiCreed]

The BARLUS cameras have an HTTP API, so I can technically change any of the camera settings by executing the right API from any device. It does not require a Windows or Mac machine to interface with the camera to change the settings.

[KamiCreed]

I found how to save the lighting settings. It's a little hidden in the PTZ settings where you needed to input a value of 250 and press the "Call" button. It's outlined in the manual.

[KamiCreed]

Seems when the raspi is offline, the sshfs folder is reverted back to normal, which unfortunately allows the recording to continue and not reconnect to the rapsi when it comes back online.

A solution may be to set the permissions of the folder to deny writing unless it's created with sshfs.

[KamiCreed]

Fixed an issue where artifacts were appearing due to the delay between processing and grabbing a new frame from the camera. I instead parallelized the frame grabbing and processing to hopefully prevent this issue.