Edge Application for Scan Tunnels

We have an application in mind that would be for scan tunnel monitoring. It would look roughly like this:

  1. The 16 cameras on the scan tunnel send images and messages to an edge device each scan

  2. The edge device has onboard storage that holds the images and the messages

  3. Certain analytics are done at the edge and then made available to the onsite app and/or are sent offsite to central storage with MQTT or Kafka

  4. Using a UUID in the messages, the edge device would also act as a web server wherein HTTP calls could be used to retrieve desired images from remote clients

  5. Whatever device the edge app runs on could also host an AI app that could make direct use of the images without the need to send them over the network. Produced classifications could also be stored locally on the edge device with certain key metrics/classifications also made available to the local app and/or sent offsite to central storage

For the message portion, I already have an image in my head of how it might be. However, we don't often work with images in Ignition, we tend to handle more primitive data.

Is anyone working with camera image capture and storage on the edge and is it being done with something involving Ignition (for example a custom module) or something else?

Thanks,

Nick

My Image Streamer module was originally created for a sponsor's use with an older-model Keyence Vision System, to save reject images from a 100% inspection application. It expanded to include OpenCV to improve performance, which enabled the RTSP streaming option. All of the java bindings for OpenCV are exposed to jython, so it can be scripted to do fairly complex image processing.