Overview
A compact starting point for real-time computer vision. The script loads the YOLO11 medium weights and runs detection on a webcam stream or a video file, drawing boxes, class labels and confidence scores on every frame.
Detection work like this later moved onboard robots, in RoverX and in the synthetic data pipeline.
How it works
Load
The YOLO11 medium model is loaded from its weights file.
Capture
OpenCV reads frames from a webcam or a video file.
Detect
Each frame runs through the model and comes back with boxes, classes and confidence scores.
Display
The results are drawn onto the frame and shown in real time.
What's next
- GPU acceleration
- Training on custom datasets
- Edge deployment on a Raspberry Pi or Jetson Nano