Real-time object detection (YOLOv11)

TensorRT INT8 detection on dual RGBD streams with 3D position.

An OpenKinematics engineer running robot demos from a live control console — Real-time object detection (YOLOv11) runs against the same observable edge stack used for testing and deployment.

This entry has an implementation in openbrain-ros. Review its README, tests, and hardware requirements before use.

About this demo

TensorRT-optimized YOLOv11 detector running on both camera streams in parallel. Bounding boxes are projected back into the dashboard, and 3D positions are estimated from depth. Custom model upload supported.

Highlights

→TensorRT INT8 quantization
→Dual-stream parallel inference
→3D position from depth
→Custom model upload

Supported robots

Any robot with RealSense

Related demos

View all →

Perception

Edge NeRF capture

Neural Radiance Fields for environment capture, optimized for Jetson.

Open demo →Teleop

CockPit

Predefined teleop dashboard with dual cameras, 3D model, map, and controls.

Open demo →Dashboard

My UI

Drag-and-drop dashboard — every CockPit widget, your layout.

Open demo →