Rock solid image transfer

[Machriam]

Problems

• The current implementation does not detect an ir-image outage correctly
• It does not transport all images sometimes - despite the fact, that all images are available on the device

Proposal

• The device should not send the images to the server after the image connection was closed
• The device should copy the desired images to another folder and serve those images to the server if requested
• There should be 4 endpoints defined:
  1. availablephototourimages --> all available images to transfer
  2. deletephototourimages --> delete available images, when the gateway server received everything
  3. getphototourimages --> fetch available images
  4. phototourinfo --> A summary of the last 100 phototrips. Should be stored on the device and gives information about how many images were captured for ir and vis for those phototrips
• This has the advantage, that the device holds images, even if only a low bandwidth connection exists between device and gateway server
• The device should furthermore offer an endpoint to serve recent phototrip information to the server. This endpoint should show how many proper images were stored for ir and vis
• Based on this endpoint, the server should power cycle or reboot the device

[Machriam]

During a week long test, it seems like the internet connection is not a problem at all. It may only happen, that no images are taken from the videostream - for example the ir-camera needs a restart. To resolve this issue, the initial test should just start the videostream and then an endpoint is called to retrieve the current image count in the corresponding streaming folders.

The new endpoint should be called countofstreamedimages and it should take the camera sensor type as an argument and returns the count of images in the corresponding folder.

The initial check can then be as simple as: If the count of both streams is greater than 0, then both cameras are functional. Else a restart might be required.

Restarting behaviour should be tiered. First a reboot, then an actual power cycle