Topic: Deconvolution for pinhole photographs?

Yes, I know this is a bit exotic - but/and perhaps one of you can see through whether what I have in mind is possible :-)

A very, very old photographic technique is based on making the exposure, not through a lens, but simply through a pinhole, typically a fraction of a mm in diameter. Somehow, it would be neat if we could combine this antique technique with a digital SLR + appropriate image analysis, wouldn't it to create nice and special photographs?

The bad news is that pictures you take with a digital SLR through a pinhole (in eg. a lens cap, without a lens) get very blurred by diffraction. The potential good news is that it seems that I would not be difficult to make an exposure of a point light source - which I assume should be a useful thing to have when attempting a deconvolution.

How could one proceed with this - ideas, anyone?

Kind regards
Karsten

Re: Deconvolution for pinhole photographs?

It might be possible to determine the point spread function corresponding to the blur, and then correct it. Unfortunately deconvolution is quite noise sensitive, so it will only work if there is very little noise in the images.

Another thing is that the blur must be the same over the whole image. If you e.g. take a picture of a piece of paper with small dots in the center and near the edges of the image, does the dots look similar or are they blurred differently? I guess that if the problem is diffraction then there might be some pattern?

Michael Vinther

> software developer <

Re: Deconvolution for pinhole photographs?

Thanks for the reply! With regard to noise I think I may be in pretty good shape, as the camera I have in mind has very little electronic noise. I will follow the suggestion of taking a picture with similar dots, and let you know how that works out.

Re: Deconvolution for pinhole photographs?

You never really answered his question (one which I have too) and that is do you plan to add deconvolution (such as Richardson Lucy) to this program? I like to do amateur astrophotography, and stars tend to need PSF correction. But the only software that can do that so far is really expensive. If you could add deconvolution to Image Analyzer, that would make a nice free way to do this type of processing.

Re: Deconvolution for pinhole photographs?

Not understood. IA already has a deconvolution feature where you can specify the PSF.

Michael Vinther

> software developer <

Re: Deconvolution for pinhole photographs?

admin wrote:

Not understood. IA already has a deconvolution feature where you can specify the PSF.

Actually from what I saw, it has convolution (kernel filter, useful for making blurs, edge-detect, emboss,etc). But it doesn't have DE-convolution. Deconvolution is needed to REMOVE blurs from point sources in images, and I don't think that Image Analyzer is currently able to do that. If it is, then please tell me how to find it in the menu.

Re: Deconvolution for pinhole photographs?

Operations | Filters | Restoration by deconvolution

Michael Vinther

> software developer <