Reference Manual
Chapter
4
IVC-2D
©
SICK AG • Advanced Industrial Sensors • www.sick.com • All rights reserved
125
Matching
Optimizing Robustness and Speed
The parameter settings have a large impact on the tool’s processing time.
The concept of image pyramids (down-sampling) is used to speed up the initial, coarse
search for the pattern. The number of pyramid levels (max 8) is chosen automatically by
the algorithm, based on a combination of the Robustness vs. Speed parameter and the
pattern ROI size. The refined, final search is typically performed on the original image.
Rules-of-thumb:
•
Small patterns (relative to search rectangle) make the tool very slow in the robust
mode. Try to reduce the search rectangle size and use the normal mode
•
Large patterns (relative to search rectangle) with solid features are well-suited for
the fast search mode
•
Patterns that consist of fine details or lines, i.e. lack solid features, require the ro-
bust mode, otherwise the characteristic details may get lost in the down-sampling
•
To optimize speed, only use the rotation tolerance interval that the application re-
quires. If the pattern is expected at different fixed angles, for example around 0
• 5 and 180 • 5 , it is faster and more robust to use the tool twice (once for
each interval) than to search • 180
The Robustness vs. Speed parameter has some effect on the accuracy of the pattern
pose, where the robust and normal modes are twice as accurate as the fast mode.
Input Parameters
01 = Source Bank
The image bank where to search for the pattern.
02 = Data Block
The data block number in the working memory, where the
reference pattern is stored.
03 = Shape Locator
Step
Reference to a Shape Locator step, to enable automatic
pose adjustment of the search rectangle. The default value
-2 means that no Shape Locator pose is used.
04 = Shape Index
A number that specifies which of the Shape Locator’s
found shapes to refer to. The default value 0 means the
first shape found by the Shape Locator.
05 = ROI X Coordinate
The x coordinate of the search rectangle’s left edge.
06 = ROI Y Coordinate
The y coordinate of the search rectangle’s top edge.
07 = ROI Width
The width of the search rectangle.
08 = ROI Height
The height of the search rectangle.
09 = Rotation Tolerance
(degrees)
The greatest angle the search for a match should be
performed for. The value is symmetric, so that for example
the value 180 means
±
180 degrees, i.e. full rotation.
10 = Rotation
Offset (degrees)
Expected angle offset of the pattern in the live image
compared to the reference pattern. For example, if the
offset is 45 and the tolerance is 5, the pattern will be
searched for in the interval from 40 to 50 degrees.
11 = Min Score
The minimum score in % that will result in a match. If the
best match is below the threshold, the pose and score will
still be returned, but the status flag and visualization will
indicate that no valid match was found.
12 = Robustness vs.
Speed
The trade-off between robustness and speed: Robust,
Normal and Fast.
13 = Create Pattern ROI
If enabled (true), the found pattern will generate a ROI in
