Preprocessing Library¶
TUM Autonomous Motorsport package for preprocessing
Implemented Preprocessing Formats¶
-
General Preprocessing Functionalities:
-
Crop Filter: Apply spatial cropping based on the distance of a point from the origin given a minimum and maximum range
-
Crop Footprint: Apply spatial cropping based on the x- and y-coordinate of a point given the longitudinal and lateral size of the vehicle footprint
-
-
LiDAR Preprocessing:
- the default preprocessing for dense LiDAR scans
- applies only the general functionalities above (range crop, then vehicle-footprint crop) and requires no sensor-specific point attributes
- available as a CPU implementation and a CUDA variant (
CUDA_LIDAR) that runs the same crops on the GPU