4 Design and function
26 / 99
CD40131 Version 1
3DPIXA
The pattern matching algorithm and the subsequent processing of the 3D data can be
optimized with the following parameters:
Window size
Maximum and minimum disparity (corresponding to the expected height range)
Minimum correlation factor
Additional system parameters
For information on the parameter adjustment refer to the documentation of CS-3D-API and
the CS-3D-Viewer.
For pattern matching both images should be almost identical. To achieve this, evenly and
homogeneously illuminated as well as perfectly focused images are very important.
4.5
Black offset and shading correction
The 3DPIXA camera supports black offset and shading (flat field) correction with the following
properties:
Both correction methods are based on pixel-by-pixel calculation, and the effects on the
behavior of single pixels like PRNU (Photo Response Non Uniformity) and DSNU (Dark
Signal Non Uniformity) will be eliminated.
Both correction methods are carried out separately for every line (red, green, blue).
The 3DPIXA camera offers 4 data sets for black offset correction and 4 data sets for
shading correction. Therefore the user is able to utilize up to 4 different lighting systems
by selecting the necessary data sets without transferring or generating new shading data.
To measure height data with 3DPIXA dual, a valid offset and shading correction are both
needed for the master and the slave camera.
The 3DPIXA camera offers the possibility to generate black offset and shading correction
internally or offline.
For black offset correction the offline or internal calculation can be used without
limitations. For a detailed description of the black offset correction refer to chapter 6.13,
p. 76.
For shading correction the offline calculation is recommended. For the offline calculation a
scanned image that includes a moving homogeneous white object is needed. Scan the
reference object and select the scanned image as reference image in the shading wizard.
By averaging over a higher number of lines of a moving target, distortion by, e.g., dust will
be eliminated. For a detailed description of the shading procedure refer to chapter 6.14,
p. 77.
Shading data calculated internally with a static white reference in front of a camera lead to
vertical lines in the image, caused by spots of dust on the white reference or by the
structure of the white reference material itself. These vertical lines in the image lead to
incorrect or missing 3D height data which renders this method unusable.
If no offline calculation of the shading correction is possible and shading correction has to
be calculated internally, the target has to be slightly moved during the internal shading
process. In this way, distortions by, e.g., dust may also be eliminated. In this case, it is
very important to check the image quality while enabling the shading correction that no
vertical lines are apparent.
