National Instruments IMAQTM ユーザーズマニュアル

5.

Set the tolerances and parameters to specify how the algorithm 
operates at run time using 

CWIMAQMatchColorPatternOptions

.

6.

Test the search algorithm on test images using 

CWIMAQVision.MatchColorPattern

.

7.

Verify the results using a ranking method.

The selection of a effective template image plays a critical part in obtaining 
accurate results with the color pattern matching algorithm. Because the 
template image represents the color and the pattern that you want to find, 
make sure that all the important and unique characteristics of the pattern are 
well defined in the image.

Several factors are critical in creating a template image. These critical 
factors include color information, symmetry, feature detail, positional 
information, and background information.

A template with colors that are unique to the pattern provides better results 
than a template that contains many colors, especially colors found in the 
background or other objects in the image.

A rotationally symmetric template in the luminance plane is less sensitive 
to changes in rotation than one that is rotationally asymmetric.

A template with relatively coarse features is less sensitive to variations in 
size and rotation than a model with fine features. However, the model must 
contain enough detail to identify it.

A template with strong edges in both the x and y directions is easier to 
locate.