Skip to content

Final Project of Computer Vision course in the spring semester at NYCU.

License

Notifications You must be signed in to change notification settings

joycenerd/CV-Lane_Detection

Repository files navigation

Lane Detection

Final project codes for Computer vision in the spring semester at NYCU. This is a basic lane detection using TuSimple dataset.

109-2 CV Final

Technologies

Prerequisite:

conda environment
1080Ti or 2080Ti GPU device
pytorch
torchvision
visdom
cudatoolkit=10.1
opencv
sklearn
ujson
matplotlib

Method Description

Deep Learning Method

2 Hourglass Network + aggregation of SGPN loss

Traditional Method

  • Method 1: Edge detection + Hough transform
  • Method 2: Perspective transform

Results

License

License: MIT

About

Final Project of Computer Vision course in the spring semester at NYCU.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published