Skip to content
RamaiProductions
RamaiProductionsArchviz · Websites · IT

Labs · Prototype

Multi-camera simulation and AI object detection

Simulated stadium with four synchronized cameras, tracking players with object detection boxes. Built as a proof of concept for automated sports camera systems.

Prototype, in developmentUnreal Engine 5simulationcomputer visionobject detection

Four camera feeds of the same simulated stadium, played side by side with the players tracked in each view. The goal: automatic camera switching based on where the action is.

33-second preview at 1280 × 720, muted, no autoplay. Unreal Engine 5, 30 fps.

What it does

One scene, four angles, the same players in every frame.

This is a working prototype, not a finished product. Here is exactly what runs today.

Four feeds from one scene

Camera 0 to 3 render the same stadium scene at the same time, each labelled, so there is always a usable angle on the action.

Every player tracked in every view

Red boxes follow each player across all four camera feeds at once — the input a switching system would need.

Built for automatic switching

The long-term goal is a system that picks the feed with the action by itself, instead of a director cutting by hand.

Where it stands: the simulation, the four synchronized feeds and the on-screen tracking boxes are working. Automatic selection between the feeds is the next stage. We publish this as a prototype and will update the page as the switching logic is built.

Building something in simulation or broadcast?

Tell us the scene and the camera problem. We will reply with what is possible and what it would take — no obligation.

WhatsApp usCall