Learn how to design and deploy reliable AI-powered perception systems for the edge-where conditions are unpredictable and performance matters. In this session, experts from Au-Zone Technologies and Toradex will explore the tools, architecture, and best practices for building spatial detection systems that run efficiently in production.

You’ll get a detailed look at the full development workflow including remote data collection, dataset creation, model training and deployment, real-time video analysis, and handling intermittent connectivity. The session features a real-world example from The Ocean Cleanup’s ADIS (Autonomous Debris Identification System), which uses 4K video and AI to identify and track marine debris-illustrating how these tools work together in demanding field environments.

• Accelerating edge AI development and deployment with EdgeFirst Studio plus Torizon OTA
• Using EdgeFirst Perception and Torizon to build maintainable, production-ready vision systems
• Managing real-time video, object tracking, localization, and metadata at the edge
• Designing for remote environments with unreliable connectivity
• Lessons from deploying edge AI in real-world applications like The Ocean Cleanup