Lane Assist System

A lane detection system built for deployment in an RV, assisting with real-time lane tracking and departure alerts. It supports both traditional computer vision and deep learning approaches, depending on available compute.

Note: This repository contains the lane detection module I developed as part of a larger RV automation project. The full system integrated multiple subsystems (sensors, control logic, UI), but this repo highlights my contributions in computer vision and deep learning for real-time lane tracking.

• Traditional CV Mode
  Uses OpenCV techniques like Canny edge detection, perspective transforms, and Hough lines for efficient lane detection.

  

• Deep Learning Mode
  Uses U-Net or DeepLab models to generate segmentation masks for more robust lane detection if sufficient power and compute is available.

  

Tools & Technologies

• Languages: Python
• Libraries & Frameworks: OpenCV, PyTorch, NumPy, Matplotlib
• Models: U-Net, DeepLab
• Techniques: Canny edge detection, perspective transforms, Hough line detection, semantic segmentation

Project Structure

• cv_detection.py – OpenCV-based lane detection
• dl_detection.py – Deep learning-based segmentation
• cv_utils/ – OpenCV-based image processsing
• ml_utils/ – Deep learning inference and training helpers
• test_videos/ – Sample driving clips
• output/ – Output visualizations

Requirements

• Python 3.8+
• OpenCV
• PyTorch
• NumPy
• Matplotlib

Install all dependencies:

pip install -r requirements.txt

Running

This system is tuned for a specific RV camera setup. Other vehicles or cameras may require retraining or calibration for best results.

Run detection on a video:

For traditional computer vision mode:

python cv_detection.py --video path/to/video

For deep learning mode:

python dl_detection.py --video path/to/video