for Robot Artificial Inteligence

25. Voxel Hashing

|

Volumetric Representation

  • Volumetric representation using a hash lookup
  • Data structures and operations(공간 효율적으로 맵핑)
    • Voxel
    • Voxel Block hash
    • Hash table and hashing function
    • Hash table Operations
  • Voxel
    • A value on a regular grid in 3D space, the extended concept of 2D pixel
    • Widely used for Realistic rendering 3D object in computer graphics.
    • Data
      • Truncated signed distance function (TSDF) value
      • TSDF Weight
      • Color value
      • Color Weight
  • Truncated signed distance function (TSDF)
    • Predefined 3D volume is subdivided uniformly into a 3D grid of voxels
    • These values are positive in-front of the surface and negative behind
    • Zero-crossing point means the surface of the object

  • Voxel block array
    • Majority of data stored in the regular voxel grid is marked either as
      • Free space
      • Unobserved space
    • Only store the surface data by efficient hashing scheme(표면 정보만 저장)
    • Grouped voxels in blocks of predefined size (ex. 8x8x8)
    • Data
      • Positions of the corner of the 8x8x8 voxel block
      • Offset in the excess list
      • Pointer to the voxel block array

  • Unordered entries는 행렬의 overide를 예방하기 위해

  • Hash table

    • To quickly and efficiently find the position of a certain voxel block in the voxel block array

  • Hash table operations
    1. Given a target 3D voxel location in world coordinates
    2. Compute its corresponding voxel block location by dividing the voxel location by the size of the voxel block array
    3. Call the hashing function to compute the index of the bucket from the ordered part of the hash table

    4. Return
      • Returns the voxel stored at the target location within the block addressed by the hash entry
    5. Insertion
      • Reserves a block inside the voxel block array

Hashing Operation

Reference

SLAM KR

Comments