Skip to main content
. 2023 Feb 13;23(4):2103. doi: 10.3390/s23042103
Algorithm 1. Pseudo-code for LiDAR to BEV conversion
start:
  PointCloud2BEV function (points)
input variables:
  Resolution = 0.1
  Side range = (−20, 20) # left side 20 m, right side 20 m view
  Forward range = (0, 50) # we do not consider points that are behind the vehicle
  Height range = (−1.8, 1.25) # assuming the position of LiDAR to be on 1.25 m
processing:
  Creates a 2D birds eye view representation of the point cloud data
  Extract the points for each axis
  Three filters for the front-to-back, side-to-side, and height ranges
  Convert to pixel position values based on resolution
  Shift pixels to have minimum be at the (0, 0)
  Assign height values between a range (−1.8, 1.25 m)
  Rescale the height values between 0 and 255
  Fill the pixel values in an image array
output:
  Return BEV image