USelfDrivingSimulator
USelfDrivingSimulator copied to clipboard
How to get PointCloud from RawImage (Lidar v2)?
I hope, I could get distances for every pixel from LIDAR image, calculate corresponding angles and than apply spherical to Cartesian conversion and get point cloud.
But there are several problems with it, so I want to ask, if somebody has experience of using LidarV2 from this repo for robotics application. Problems, I found, are:
- frag_sphere_mapping return logisticDistance. For the purpose of point cloud calculation (not visualization only) linearDistance should be returned, I think.
- It's unclear, how to get actual distance from Lidar Image pixel value - what scale should be applied. There is MeasurementRange value, but it's never used.
return linearDistance + Z_far scaling seems to work.
@frontw Did you realize point cloud in your repo? I'd love to see it!