🛣️ Road Lane Detection Using Computer Vision

A computer vision pipeline for detecting lane lines on roads, built as part of the Self-Driving Car Engineer Nanodegree. Uses Canny edge detection, Hough transforms, and region-of-interest masking to identify and overlay lane markings on images and video.

✨ Features

• 🔍 Grayscale conversion and Gaussian blur for noise reduction
• 📐 Canny edge detection for identifying lane boundaries
• 🎯 Region-of-interest masking — dynamic triangular ROI based on image dimensions
• 📏 Hough line transform for detecting line segments
• 🧮 Slope-based lane separation — distinguishes left vs. right lane markings
• 📈 Line averaging & extrapolation — smooth, full-length lane overlay
• 🎬 Video processing — frame-by-frame lane detection on MP4 files using MoviePy
• 🎨 Color-based lane detection — HLS color space for yellow and white lane identification

🔄 Pipeline Overview

Input Image/Frame
  → Grayscale conversion
  → Gaussian blur (kernel=11)
  → Canny edge detection (50/150 thresholds)
  → Region-of-interest masking
  → Hough line transform
  → Slope filtering & lane averaging
  → Weighted overlay on original image

🛠️ Tech Stack

• Python 3 — core language
• OpenCV — image processing and computer vision
• NumPy — numerical operations
• Matplotlib — visualization
• MoviePy — video processing

🚀 Getting Started

Prerequisites

pip install opencv-python numpy matplotlib moviepy

Run the Notebook

jupyter notebook "Finding Lane Lines- CARND-Term-1- Submission.ipynb"

The notebook walks through the full pipeline with test images and video processing.

📁 Project Structure

Road-Lane-Detection-Using-Computer-Vision/
├── Finding Lane Lines- CARND-Term-1- Submission.ipynb   # Main notebook
├── output_4_1.png                                        # Sample output
├── output_29_*.png                                       # Test image results
└── README.md

📸 Sample Results

The pipeline processes multiple test images (solid white, solid yellow, curves) and three video clips (white right, yellow left, challenge) with lane overlay rendering.

👤 Author

Sanjay Santhanam