Skip to content

Latest commit

 

History

History
14 lines (10 loc) · 934 Bytes

readme.md

File metadata and controls

14 lines (10 loc) · 934 Bytes

SuperDrive

A live-processing capable, clean(-ish) implementation of lane & path detection based on comma.ai's SuperCombo neural network model

Tested with Python 3.8.2, OpenCV 4.2.0, and TensorFlow 2.2.0 on Ubuntu 20.04

SuperDrive Demo Video

Running

To get a list of all the options, run python3 drive.py -h To run the code with visualizations, run python3 drive.py --input <input> --show-opencv-window To run this code with an actual camera, use something in /dev/v4l/by-id/<camera> for your <input> :)

Note

I don't really know what I'm doing here, so this code will probably contain bugs. This is not production-quality by any means. While the model itself has proven to be quite successful at driving millions of miles elsewhere, that doesn't mean it'll translate to this quick implementation! Use with caution.