Справочник Пользователя для National Instruments IMAQ Vision for LabWindows TM /CVI

© National Instruments Corporation

The pattern matching algorithms in IMAQ Vision measure the similarity 
between an idealized representation of a feature, called a template, and the 
feature that may be present in an image. A feature is defined as a specific 
pattern of pixels in an image. Pattern matching returns the location of the 
center of the template and the template orientation. Complete the following 
generalized steps to find features in an image using pattern matching.

1.

Define a template image in the form of a reference or fiducial pattern.

2.

Use the reference pattern to train the pattern matching algorithm with 

imaqLearnPattern2()

.

3.

Define an image or an area of an image as the search area. A small 
search area reduces the time to find the features.

4.

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

imaqMatchPattern2()

.

5.

Test the search algorithm on test images using 

imaqMatchPattern2()

.

6.

Verify the results using a ranking method.

The selection of a good template image plays a critical part in obtaining 
good results. Because the template image represents 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. 

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