Low-Latency Multimodal Perception for Robotic Assembly and Real-Time Inspection of Automotive Wiring Systems using Frame-Based and Event-Based Vision
A masters research project for deformable linear object (wire) detection using two complementary vision sensors:
- Basler frame camera (grayscale, high-resolution frames)
- iniVation DVXplorer event camera (asynchronous events, high dynamic range)
The system combines AprilTag-based ROI motion detection, homography calibration, and real-time wire tracking with spline fitting.
- Dual-camera acquisition (Basler + DVX)
- AprilTag ROI & motion detection inside a defined region
- Homography calibration (DVX → Basler alignment using ORB features)
- Real-time overlay of event data on frame-based images
- Spline-based wire detection from masked regions
- Screenshot saving & ROI cropping with keyboard shortcuts
- Interactive UI with visualization windows
Requirements
- Python 3.8+
- Basler Pylon SDK
- DVXplorer SDK
Core dependencies:
- opencv-python
- numpy
- scipy
- pypylon (Basler)
- dv-processing (DVXplorer)
- pupil-apriltags
Keyboard Shortcuts (wire detection window)
- ESC → exit
- p → save overlay screenshot
- m → save cropped ROI
- q/w → adjust low threshold
- a/s → adjust high threshold
- r/e → decrease/increase ROI width
- f/t → decrease/increase ROI height
- i/j/k/l → move ROI rectangle (up/left/down/right)
License
This project is licensed under the MIT License. See LICENSE for details.
Acknowledgements
- Basler Pylon SDK
- iniVation DVXplorer SDK
- pupil-apriltags -OpenCV community
- FAU-FAPS
Citation / Bibtex If you use this project in your research, please cite:
@projekt{chaudhry2025lowlatency,
title = {Low-Latency Multimodal Perception for Robotic Assembly and Real-Time Inspection of Automotive Wiring Systems using Frame-Based and Event-Based Vision},
author = {Chaudhry, Affyief},
supervisor = {Hartmann, Annalena},
institution = {FAU-FAPS},
year = {2025},
note = {Project Arbeit},
url = {https://github.com/Affyief/Project-arbeit---Event-based-vision}
}



