Linear fit not working as expected?

[joptimus]

Hi,

I'm using the NI_AALPro.lvlib:Linear Fit.vi on an array of vertically stacked points. However, the result makes absolutely no sense, the fitted line doesn't represent the input data.

But if I rotate the points by 1 degree, the fit seems to work...

Why is that and is there a function or a way to properly fit the points regarding their orientation?

Thanks in advance!

I have attached:

• The VI itself
• A screenshot of the VI
• The test image (no rotation)
• The test image (rotated 1 degree)
• A screenshot of the fitting result (no rotation)
• A screenshot of the fitting result (rotated 1 degree)

[joptimus]

Pictures part 2

[GerdW]

Hi optimus,

can you attach your VI with a typical set of "blob coordinates" made default in the indicator?

What's the point of the FOR loop? Why not use simple math:

[joptimus]

Thanks for your quick reply! I adjusted the VI, it seems I thought a bit too complicated.

I included the blob coordinates (in the 1600x1200 px image) as a constant.

[GerdW]

Unfortunately you did not attach the VI according to your image.

Please attach your VI with that "blob coordinates" constant!

[joptimus]

Sorry, my mistake. I'll upload it shortly

I adjusted the file paths, you might need to update them with the location of the test picture I uploaded if you want to use that.

Thanks again!

[joptimus]

Sorry, my mistake, here it is:

[GerdW]

Hi Optimus,

using a XY graph the interpolation looks quite ok. See attachment…

[joptimus]

Hm, yes it does. But why is it so weird in case of my blobs?

[GerdW]

Hi optimus,

I thought that "blob coordinates" constant were real blob coordinates?

Why should it display different on your IMAQ indicator than on my XY graph?