Edge AI Video Analytics on Existing Cameras: How It Works

How does edge AI video analytics work on existing cameras?

Edge AI video analytics reads the camera streams a site already runs, over RTSP or ONVIF, on a small edge computer at that site. The computer decides on each stream locally and flags each event with the model's confidence so an operator knows what to check first. Raw footage stays on site, and nothing gets ripped out or replaced.

Which cameras does it work with?

DHI supports RTSP and ONVIF-compatible cameras. Because DHI reads RTSP directly, it runs alongside the VMS you already have, including Genetec, Milestone and Avigilon environments, without replacing it.

What hardware does the edge computer need?

DHI runs on any GPU edge node, from a small 8 GB edge module to an industrial x86 server. Size the node to the number of cameras, not to a fixed product.

What leaves the site?

Inference happens at the edge. Only structured safety events or policy-approved clips are transmitted off-site when your workflow allows it. Full-resolution video and historical archives stay on the premises unless explicitly exported.

How do events reach the people who respond?

DHI provides REST APIs, webhooks, VMS integrations, and event-routing patterns for common incident management workflows. Routing DHI events into a specific VMS alarm workflow is scoped integration work rather than a shipped connector.

What does the analytics cover?

DHI covers 31 safety and security use cases, from people on rail tracks to camera tampering, on the camera streams a site already has.

What is the first step?

A Focused Pilot runs for 30 days in one defined zone, with one workflow, using your current cameras, video management system and response workflow. You can also take a click-through product tour of the DHI console with demo data, with no form, no email and no call. Clips and screens on this site are test footage and demo data, not customer sites.