This project is an AI-based ADAS prototype that performs lane detection and object detection from uploaded road-driving videos. It uses OpenCV for lane detection, YOLOv8 for object detection, and Streamlit for real-time dashboard visualization.
- Upload road-driving video through Streamlit dashboard
- Frame-by-frame video processing using OpenCV
- Lane detection using ROI masking, color filtering, Canny edge detection, and Hough Transform
- ADAS lane overlay visualization
- Lane status warning: Safe, Move Left, Move Right
- Exponential smoothing for stable lane line display
- YOLOv8-based object detection
- Detection of vehicles, pedestrians, bicycles, buses, trucks, motorcycles, traffic lights, and stop signs
- Live dashboard with ADAS parameters
- Python
- OpenCV
- NumPy
- Streamlit
- YOLOv8
- Ultralytics
- PyTorch