Optical Distortion Calculation

[glass99]

I am looking for a software implementation for the measurement of optical distortion in glass per ASTM C1652 for digital photography of grids.

My application is relatively large pieces of architectural glass, and the idea is that you take a photo of a grid without the glass then with the glass then measure the difference between them.

Are there any online open source or other libraries which might contain this? I don't mind paying for it.

 

[IRstuff]

You could possibly look through these:

http://www.mathworks.com/matlabcentral/fileexchange/index?utf8=✓&term=image++distortion

and see if anything could work for you.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529


Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:

http://www.engineering.com/AskForum/aff/32.aspx

 

[glass99]

thanks IRStuff. Seems like Matlab is a good environment for implementing this kind of thing at least. I didn't find the specific thing I was looking for in their library unfortunately.

 

[IRstuff]

If you don't have Matlab, you could try Octave, which is a free-ware similar to Matlab.

There are just a lot of things on the web, like:

http://www.vision.caltech.edu/bouguetj/calib_doc/

http://docs.opencv.org/doc/tutorials/calib3d/camera_calibration/camera_calibration.html

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529


Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:

http://www.engineering.com/AskForum/aff/32.aspx

 

[glass99]

Thanks IRStuff. There is loads of stuff for correcting pin cushion distortion of camera lenses, but having trouble finding the specific calculation of optical distortion in millidiopters of a random piece of glass. I guess there are more people with an interest in photography than in inspection of architectural glass.

 

[IRstuff]

The same general approach can still apply. While typical optical distortion can be relatively easily fitted with a polynomial, your glass will not have the same characteristics as spherical optics. Therefore, you will need to brute-force it, and simply take the delta between the correct coordinates vs. the distorted coordinates. The delta will then be the distortion, and you can still interpolate to determine the positions of points not on your original grid. Because there probably isn't any sort of systematic distortion, aside from the waviness imposed by the rollers that formed the glass, that's probably as good as it will get.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529


Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:

http://www.engineering.com/AskForum/aff/32.aspx

 

[glass99]

Ideally I would have a map of the surface of the glass showing the optical distortion measured in millidiopters for every square 1"x 1" or so. The way I imagine it is that each 1" square of glass has an angle away from where it should be, which corresponds with a tangent of a circle, which corresponds to a lense radius and therefore a lense power. The output from the analysis is that 90% of the glass has an optical distortion of less than say 51mD with a peak lense power of 85mD. The ASTM tells you to use a zebra board with stripes between 2" and 1/4" depending on whether you are doing transmission or reflection.

Tempered glass tends to have a sine wave shape with a wavelength of approx 10in and depth peak to trough of approx 0.004in. But there are other optical distortion sources too, like lamination and tool marks.