LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Image correlation

Hi,

 

I try to write a program which makes use of the image correlation VI to obtain the correlation coefficient over the two attached images. The file 200 is the template, and 201 is the sub-image with a narrow mark by pen on it. However when I swept over the sub-image and compare it with the template, the output 2D correlation coefficient does not change even when the sub-image enters into the pen mark zone. Therefore it does not make sense to me. Any advice on this? Thanks.

 

Wallace Lai

Download All
0 Kudos
Message 1 of 6
(3,668 Views)

Hi Wallace

 

Unfortunately the code you linked contained subVI's that did not open. From looking at this code you will need to specify a rectangle smaller than the image you are correlating over. As a problem solving strategy try placing probes on the input and output of your correlation. Use this to confirm that the output is what you intend it to be before transforming the data into an array. Also please see chapter 6 of the Vision Concepts Manual for more information on image comparison and the algorithms used.


Thank You
Eric Reid
National Instruments
Motion R&D
0 Kudos
Message 2 of 6
(3,635 Views)

Sorry I forget to attach the subVI. Here they are.

 

Wallace

Download All
0 Kudos
Message 3 of 6
(3,618 Views)
The maximum value of the correlation seems to change only by one.  I don't think you have enough of a change in the two images to detect the change.  The images and the correlation are 8-bit and that may not be enough resolution to detect the change in maximum correlation value.
Randall Pursley
0 Kudos
Message 4 of 6
(3,607 Views)

In other words, the program itself is fine?

 

Wallace

0 Kudos
Message 5 of 6
(3,590 Views)
Seems to be to me.
Randall Pursley
0 Kudos
Message 6 of 6
(3,561 Views)