/Lane_RSR

Primary LanguagePythonMIT LicenseMIT

Lane_RSR

Overview

1 2 3

Lane_RSR is a cascaded framework for robust lane detection.

It consists of reconstruction module, segmentation module, and restoration module.

This implementation uses ENet-SAD model as segmentation module.

framework

Dependencies

pip3 install -r requirements.txt

Datasets

Lane_RSR is trained and tested on CULane dataset.

CULane dataset is available in https://xingangpan.github.io/projects/CULane.html

Lane Reconstruction

cd UDLR

train

python3 main.py --mode = train

test

python3 main.py --mode = test

Lane Restoration

TBU