|
The type of point content defines whether a point provides information about various cues, namely orientation or color. The tree main sources of information are available per point: 1 The euclidean coordinates of the point in millimeters. These are represented by a float 32 bit value per coordinate. Since the order of the coordinates and other elements is fundamental (when integrating ArmarX with other frameworks) the notation "Pfx" means, Point-float-x coordinate. Thus, the sequence PfxPfyPfz means Point XYZ 96 bits with values in millimeters. 2 The notation "Cbr mean, Color-byte-read channel 3 The notation "Nfx" means Normal-float-x coordinate. Finally, The combination(s) clearly represent the content of each point. Hence, from this definition it is also possible to infer the whole size in bytes of a point.
Color value RGBA \in N⁴ := 0,..,255^4 represented with 4 bytes = 32 bits, the value A (Alpha channel) is actually used for padding when the point cloud is directly obtained form the active camera. The further usage may be containing alpha (transparency) values or other labels. Euclidean normal vector NXNYNZ \in R³ represented with 3 floats = 12 bytes = 96 bits, no padding added. Euclidean point vector XYZ \in R³ represented with 3 floats = 12 bytes = 96 bits, no padding added. Total point size: 28 bytes = 224 bits
Color value RGBA \in N⁴:= 0,..,255^4 represented with 4 bytes = 32 bits, the value A (Alpha channel) is actually used for padding when the point cloud is directly obtained form the active camera. The further usage may be, containing alpha (transparency) values, or other application specific labels. Euclidean point vector XYZ \in R³ represented with 3 floats = 12 bytes = 96 bits, no padding added. Total point size: 16 bytes = 128 bits
Euclidean normal vector NXNYNZ \in R³ represented with 3 floats = 12 bytes = 96 bits, no padding added. Euclidean point vector XYZ \in R³ represented with 3 floats = 12 bytes = 96 bits, no padding added. Total point size: 24 bytes = 192 bits
Euclidean point XYZ \in R³ represented with 3 floats = 12 bytes = 96 bits, no padding added. Total point size: 12 bytes = 96 bits