top of page

The lane detection pipeline works as follows:

  1. Camera Calibration and Distortion Correction

  2. Color/Gradient Threshold

  3. Perspective transform​

  4. Detect lane line and determine its curvature

​

Step 1. The camera matrix and distortion coefficients are calculated using chess board images

Step 2. To minimize the impact of illumination on the lane detection, the image is transformed from RGB to HLS color space. Gradient value along x, y axis and gradient direction are also calculated.

Step 3.  Transform the image to a bird's eye view 

Step 4. Combine the gradient and color filters to detect lane lines. Then fit line points to a polynomial curve.

bottom of page