Use computer vision at the edge to improve operations with AWS Panorama
summary
AWS Panorama brings real time computer vision to on‑prem cameras, claiming subsecond inference by compiling SageMaker models to run on an AWS Panorama appliance or Panorama‑enabled cameras using ONVIF and RTSP streams. It runs multiple models across streams on hardware like Nvidia Xavier AGX and Ambarella CV2X, pairs predictions with AWS Lambda business logic, and ships a developer kit that registers, deploys, and displays results to HDMI in the demo. Customers (Fender, Cargill, BPX) use it to count people, track parts, and route trucks while addressing bandwidth and privacy limits and the talk cites 98% of enterprise video as unanalyzed.
AI-written summary · about
chapters (11)
- 0:00Introduction to AWS Panorama
- 1:26What is computer vision
- 3:39Why run vision at the Edge
- 5:25Introducing AWS Panorama
- 9:41How Panorama works
- 10:47Panorama Edge devices
- 12:14Building Panorama applications
- 14:25Customer use cases
- 18:02Live demo walkthrough
- 22:58Getting started with Panorama
- 24:47How to access the preview
related talks
29mApplied Custom Vision
30mCustomer centricity in manufacturing with operational improvements
48mAccelerate science, research, and innovation with AWS
40mDemocratise public agency user access to AI/ML on VAS
27mComputer vision for automated quality inspection
14mComputer vision on edge devices
28mVision, Big Data and AI
23mComputer Vision Toolbox - Connected Computer Vision & AI Quality Inspection at Scale!