Ultrafast Lane Detection Inference Pytorch

Example Jupyter-Notebook file for the detection of lanes using the ultra fast lane detection model in Pytorch.

Source: https://www.flickr.com/photos/32413914@N00/1475776461/

Requirements

• OpenCV, Scikit-learn and pytorch.

Conda and Carla Requirements

# conda install pytorch torchvision torchaudio cudatoolkit=10.1 -c pytorch
# conda create --name pytorch-gpu-cuda10.1   **-c pytorch**  pytorch torchvision  cudatoolkit=10.1 jupyterlab scikit-learn opencv


cd workspace/github/Ultrafast-Lane-Detection-Inference-Pytorch-
conda info --envs
conda activate pytorch-gpu-cuda10.1
jupyter-lab

cd Downloads\CARLA_0.9.12\WindowsNoEditor
CarlaUE4 /Game/Carla/Maps/Town03 -windowed -ResX=256 -ResY=128 -carla-port=3000 -benchmark -fps=30   -quality-level=Epic

Installation

pip install carla

Pytorch: Check the Pytorch website to find the best method to install Pytorch in your computer.

Pretrained model

Download the pretrained model from the original repository and save it into the models folder.

Ultra fast lane detection - TuSimple(link)

• Input: RGB image of size 1280 x 720 pixels.
• Output: Keypoints for a maximum of 4 lanes (left-most lane, left lane, right lane, and right-most lane).

Examples

• Carla RGB Cam inference:

carlaLaneDetection.ipynb 

Inference video Example

Original video: https://youtu.be/2CIxM7x-Clc (by Yunfei Guo)