Proposals:LiDAR Data Support: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
(New page: As LiDAR sensors become cheaper and more readily available, it seems important for ITK to support data acquired by these sensors. New functionality can include *PTX IO - Leica scanners ou...)
 
No edit summary
Line 1: Line 1:
__TOC__
= Background =
As LiDAR sensors become cheaper and more readily available, it seems important for ITK to support data acquired by these sensors.
As LiDAR sensors become cheaper and more readily available, it seems important for ITK to support data acquired by these sensors.


New functionality can include
= File formats =
*PTX IO - Leica scanners output data in this format.
*PTX (Leica scanners)
*Other scanner file formats (Velodyne, etc.)
*Other scanner file formats (Velodyne, etc.)
= Data structures =
*itkGridMesh - a class that combines the neighborhood functionality of itkImage (getUpperNeighbor, getLowerNeighbor, getRightNeighbor, getLeftNeighbor) with the 3D mesh type operations (FindPointsInRadius(center, radius), etc)
*itkGridMesh - a class that combines the neighborhood functionality of itkImage (getUpperNeighbor, getLowerNeighbor, getRightNeighbor, getLeftNeighbor) with the 3D mesh type operations (FindPointsInRadius(center, radius), etc)
**Ray/Triangle (Ray/Mesh) intersections
**Ray/Triangle (Ray/Mesh) intersections
**Normal estimation
**Normal estimation
= Algorithms =
*LiDAR segmentation
*LiDAR segmentation
**Region growing (possibly on the proposed itkGridMesh structure? Or on a standard itk Vector Image?)
**Region growing (possibly on the proposed itkGridMesh structure? Or on a standard itk Vector Image?)

Revision as of 14:12, 10 August 2009

Background

As LiDAR sensors become cheaper and more readily available, it seems important for ITK to support data acquired by these sensors.

File formats

  • PTX (Leica scanners)
  • Other scanner file formats (Velodyne, etc.)

Data structures

  • itkGridMesh - a class that combines the neighborhood functionality of itkImage (getUpperNeighbor, getLowerNeighbor, getRightNeighbor, getLeftNeighbor) with the 3D mesh type operations (FindPointsInRadius(center, radius), etc)
    • Ray/Triangle (Ray/Mesh) intersections
    • Normal estimation

Algorithms

  • LiDAR segmentation
    • Region growing (possibly on the proposed itkGridMesh structure? Or on a standard itk Vector Image?)
  • LiDAR/3D Model registration
    • Spin Images
    • ICP (already some ICP methods in ITK)