Introduction
Live Video Analytics (LVA) on IoT Edge is ideal for IoT developers looking to build solutions with live video analytics capabilities such as capturing, processing, and analyzing video from existing cameras to extract business insights.
Suppose you want to build a real time monitoring solution for safety in a factory. The solution uses live feeds from cameras and detects if a person is standing too close to a machine.
Your manufacturing plant has raised safety concerns for personnel working on safety-critical systems on the factory floor. Specifically anytime someone works in the proximity of a power source. There is a high potential for causing harm to themselves. Safety incidents could also cause lost work time, higher insurance, and workers' compensation claims, all of which affect a business's operations.
Hence, maintaining employee safety is critical for your customer. To achieve higher safety standards, you want to use video analytics with machine learning algorithms on the factory floor. Your solution will capture the live video stream from a camera. Images from a video stream will be analyzed and processed in real time at the edge. Object detection will be performed in real-time, and an alert will be raised if the solution detects that a person is standing too close to a machine. Hence, the safety issue will be identified earlier and prevented.
Prerequisites
- An Azure subscription
- Ability to use Azure Cloud Shell
- Basic knowledge of Azure IoT Edge
- Basic knowledge of Custom Vision
- Basic knowledge of Live Video Analytics
- Basic knowledge of containers
- Ability to use Docker
Learning objectives
In this module, you will:
- Use Live Video Analytics on IoT Edge module to build a video analytics solution
- Deploy a set of modules to an IoT Edge virtual machine using the installer
- Set up an application that uses a virtual device for rapid inference at the edge
- Bring an AI model of your choice into the video analytics solution
- Test a solution that will detect a person at the edge from a web application