What’s new in Vision AI Assistant 2023
Posted: January 9, 2024
Use images from existing general-purpose cameras and convert them into image classification-based analytics. Vision AI Assistant monitors real-time image streams and provides alerts and notifications to operators for immediate action. Integrated with AVEVA’s HMI/SCADA software and AVEVA™ Insight, the solution employs deep learning to train and deploy machine learning models from an easy-to-use web-based interface. Simply use live camera feeds and preferred-state images for the analytical model to learn from. Vision AI Assistant processes the live images to automatically generate an alert when an image deviates from the normal or expected state.
- User Defined Pipeline (UDP) model is the most sophisticated skill type available with Vision AI Assistant.
- Self-specify a series of automated steps in a chain for pre-processing and transformation of desired images. This can include image alignment, area isolation, colorization, artifact measurement and more. The input image travels through this series of defined processing blocks, transformations, or actions (referred to as a pipeline), to achieve the desire prediction output.
- Highly flexible and powerful for a range of use cases: measuring dimensions, image isolation, determining color, finding defects, and more.
- Includes a new UI for user to interact and design their pipelines, and view the results of the pre-processing, measurement, prediction, and action blocks.
This will enable license distribution, usage monitoring, and future enhancements.
Improved integration with AVEVA Insight
- Near real-time API for alerts
- Improved the property window, helps for configuring properties and validations
- Icons on each UDP skill block
- Path of assets shown in skill property
- Anomaly preview organizes good/bad images with clear positive and negative labels
- Asset dropdown in skill property page
Related blog posts
Stay in the know: Keep up to date on the latest happenings around the industry.
