Advanced Lane Finding

This project was forked from Udacity, it is part of the Self-Driving Car NanoDegree course projects.

Overview

In this project, I used computer vision techniques like edge detection, perspective transform, camera calibration ... to detect highway lane informations (e.g. position, curvature ...) using pure camera data.

How To

1. Setup environment using Udacity Starter Kit
2. start docker by running docker run -it --rm -p 8888:8888 -v pwd:/src udacity/carnd-term1-starter-kit bash
3. run python image_pipeline.py

The Project

• Project report in writeup_report.md
• Video pipeline source code in P4.ipynb
• Saved camera calibration matrix at distort_mtx.p
• Result at results.mp4

The result looks like this

More detail

For more informations please reference writeup_report.md. I also wrote a medium post about it (Mandarin Chinese)