Skip to content

Lane Detection

Hot to implementation

차선과 같은 라인을 찾는 방법.

  1. Convert original image to grayscale.
  2. Darkened the grayscale image (this help in reducing contrast from discoloured regions of road)
  3. Convert original image to HLS colour space.
  4. Isolate yellow from HLS to get yellow mask. ( for yellow lane markings)
  5. Isolate white from HLS to get white mask. (for white lane markings)
  6. Bit-wise OR yellow and white masks to get common mask.
  7. Bit-wise AND mask with darkened image .
  8. Apply slight Gaussian Blur.
  9. Apply canny Edge Detector (adjust the thresholds — trial and error) to get edges.
  10. Define Region of Interest. This helps in weeding out unwanted edges detected by canny edge detector.
  11. Retrieve Hough lines.
  12. Consolidate and extrapolate the Hough lines and draw them on original image.

Projects

Topo2D
https://github.com/homothetic/Topo2D
Enhancing 3D Lane Detection and Topology Reasoning with 2D Lane Priors
관련 프로젝트 - MMDetection3D, MapTR, PETR, OpenLane-v2, TopoNet

See also

Favorite site