Using the solution

At this point, your ML model is deployed at the edge server. You can use the solution in different ways.

The following sections show different scenarios of use:

• Local inference with images from a camera
• Streaming inference results to BigQuery
• Streaming inference results to local MQTT for local actions
• Launching physical actions based on inference results
• Triggering inspection remotely with an MQTT command
• Using multiple cameras, with dedicated ML models, triggered simultaneously
• Batch processing inference against a set of image files

---

Table of contents

• Local inference
• Inference to BigQuery
• Inference to MQTT
• Physical triggers
• Remote MQTT trigger
• Multiple Cameras
• Batch processing