Use Amazon Lookout for Vision on Greengrass core devices - AWS IoT Greengrass

Use Amazon Lookout for Vision on Greengrass core devices

Note

AWS IoT Greengrass doesn't currently support this feature on Windows core devices.

Amazon Lookout for Vision is an AWS service that you can use to find visual defects in industrial products. It uses computer vision to identify missing components in an industrial product, damage to vehicles or structures, irregularities in production lines, missing capacitors on printed circuit boards, and defects in silicon wafers or any other physical item where quality is important. For more information, see What is Amazon Lookout for Vision? in the Amazon Lookout for Vision Developer Guide.

You can create Greengrass applications that use Lookout for Vision inference to find visual defects on Greengrass core devices. After you deploy a Lookout for Vision workflow to a Greengrass core device, you can perform computer vision without a connection to the Lookout for Vision service in the AWS Cloud. To create a Greengrass application that uses Lookout for Vision, you set up and deploy the following Greengrass components:

  • Lookout for Vision model components – Contains Lookout for Vision machine learning models as Greengrass artifacts. You can use the Lookout for Vision console and API to generate model components that package your pre-trained machine learning models. These components are private Greengrass components in your AWS account. For more information, see Creating a Lookout for Vision model and Packaging a Lookout for Vision model in the Amazon Lookout for Vision Developer Guide.

  • Lookout for Vision Edge Agent component – Provides a local Lookout for Vision runtime server that uses computer vision to detect anomalies using machine learning models that you provide. This component is an AWS-provided component. For more information, see the Lookout for Vision Edge Agent component.

  • Lookout for Vision client application component – Interacts with the Lookout for Vision Edge Agent component to process images for anomalies. You can develop custom client application components that send images and video streams to the local Lookout for Vision Edge Agent and reports any anomalies that the machine learning models detect. For more information, see Writing a client application component and Lookout for Vision Edge Agent API reference in the Amazon Lookout for Vision Developer Guide.

For more information about how to create, configure, and use these components, see Using a Lookout for Vision model on an edge device in the Amazon Lookout for Vision Developer Guide.