A Python-based system that uses a HeliosDAC digital-to-analog interface, a Lasercube scanning laser, and a USB camera (via PyTorch) to track and highlight 2D objects in real-time. Designed to run on an Nvidia Jetson but adaptable to other devices, the system can point a laser at detected objects, change laser colors based on detection confidence or object type, and handle multiple object tracking.
Demonstration Video: https://youtu.be/aSx_q8-XqPE
Real-Time Object Tracking: Detect and follow multiple 2D objects.
Dynamic Laser Control: Adjust laser color based on detection confidence or object type.
Flexible Hardware Compatibility: Optimized for Nvidia Jetson but usable with other devices.
Configurable Detection Settings: Easily adjust detection parameters and laser behaviors.
HeliosDAC: Digital-to-analog interface for laser control. (https://bitlasers.com/helios-laser-dac/)
Lasercube: Scanning laser for object highlighting.
USB Camera: For capturing live video feed.
Optional: Nvidia Jetson for optimized performance.
Python 3.x
PyTorch (for object detection)
HeliosDAC Python library (https://github.com/Grix/helios_dac)
git clone cd
cd examples python detectnet_centerLaser.py --camera_id 0 --laser_port /dev/ttyUSB0
Example Configuration
laser: color_mode: confidence # Options: confidence, object_type default_color: [255, 0, 0] # RGB for red
detection: confidence_threshold: 0.5 max_objects: 5
Helios_dac Team: For their open-source Python library.
PyTorch: For the object detection backbone.
License
MIT License
Contributing
Contributions are welcome! Please open an issue or submit a pull request.