Technical Specifications

VisionAI Observ is deployed on a Kubernetes architecture, supporting both edge and cloud computing. After establishing a connection to cameras via a browser, AI detection tasks can be initiated.

Architectural Diagram

Camera Details

For best practice, deployed cameras must:

  • Use RTSP streaming protocol with support for TCP or UDP transport.

  • Require static physical IP address for stable connection.

  • Have an upload speed of at least 10 mbps per camera for consistent and high-quality streaming.

  • Have a resolution of 480p or 720p at 10 fps. Excessive resolution may degrade AI inference speed and accuracy.

Software

Software requirements for deployment include:

  • the latest version of Google Chrome.

  • 400MB of available disk space.

  • A screen resolution of 1024 x 768 pixels or higher.

  • At least 10 mbps download speed for smooth real-time playback.

Hardware

Different detection scenarios and the number of cameras will require different setup specifications, which can be adjusted according to the user's needs. Contact the Linker Vision team for more details.

Last updated