Terminology

The following terms are important for understanding how to design and implement a Visual Inspection AI Edge deployment:

  • Edge server. A server that you deploy at the edge of your environment which is running the Visual Inspection AI Camera applications to capture photos, invoke machine learning models, fetch inference results and send results to the backend.

  • Cloud Resources. Services that run in the Google Cloud platform to receive inference results, storing results, developing machine learning models and deploy the containerized models to the edge server.

  • Visual Inspection AI edge applications. Applications that are running on the edge server, this includes the camera application that connects to the camera, captures photos and gets machine learning model inference results from a machine learning model, and the Visual Inspection AI machine learning models which are packed as a container image and running on the edge server.

  • Visual Inspection AI Service. A Google Cloud service hosted, managed and operated by Google Cloud. You use the service to train your Visual Inspection AI models.

  • Target edge environment. The physical location where your edge server will be deployed.

  • Setup workstation A Linux or macOS desktop or laptop, with access to the internet and USB ports, that will be used to setup the Google Cloud assets and prepare the Edge server configuration and operating system.

During the deployment process, you will be asked to perform some actions in the setup workstation or in the edge server. To help you identify more clearly which one to use, you will find labels before the group of steps:

Run on Setup Workstation

Run on Edge Server