LBP2 User Guide
Document No: 50306-001
Rev G 3/12/2020
Page 62
Important:
Loading a data file that is larger than the frame buffer size will cause the data
to wrap and overwrite the frame buffer with the last frames loaded remaining in
the buffer. LBP2 has a fixed 16 frame buffer size.
4.4
HDF5 format
The HDF5 format is compatible with many third party applications, one of which is
MATLAB. See the following web link to learn more about the HDF5 format.
The .lbp2Data file format contains the processed image output from LBP2. The HDF5
file contains a 2D array of the frame image that can be loaded into MATLAB so that
third party computations can be performed on the data. The image and pixel data
formatting is directly readable and follows standard HDF5 formatting rules.
4.4.1
Image Data Access
Pixel data is stored in the HDF5 file at folder node 1,DATA. All pixel data is in an
S32 bit signed binary fixed point format that is described below. Data is stored in
packed 32 bit words loaded left to right, top to bottom, as it appears in the 2D
display window. To determine the width and height of the date frame access
node 1,RAWFRAME, WIDTH and 1,RAWFRAME,HEIGHT.
4.4.2
Image Data Description
The input camera native source may be 8, 10, 12, 14, or 16 bits per pixel. LBP2
employs a normalized (signed 32-bit) fixed point format for storing pixel values in
HDF5 data files. The acquired and processed camera pixel data is converted to a
32-bit signed value and stored. The most significant bit of the camera’s native
data is shifted to the bit position just behind the sign bit (assuming bit positions 0
[lsb] -31 [msb] this would be position 30). The empty bits below the native
format are then available for additional precision and will be utilized via frame
averaging, background subtraction, or other image processing activities.
