0

Is there any software that will assist me in the distortion described in 2 below. Another way of doing this is numerical, which is likely to be labourious, error prone, and exceptionally time consuming.

I am trying to produce large scale prints of some of my landscape format photographs which are going to be placed on complex 3-D surfaces. The approach I am taking is:

  1. Create a 2-D pattern from the 3-D surface to create a shape that will cover the 3-D surface. This picture shows the type of shape that might arise.

  1. Distort the picture in such a way that it maps the rectangular shape of the picture to the irregular pattern created in stage one. The second picture shows all the features that might appear on a pattern, though no pattern will have all of these features.

These are:

A Obtuse angles at corner

B Acute angles at corner   

C Convex curved sides   

D Concave curved sides   

E Edges that are straight and curved   

F Right angled corners    

G Straight edges   

H 'Darts' to fit the 2-D pattern to the 3-D surface

The features shown in this picture are gross exaggerations of reality - in fact most of the features are likely to be subtle, though not necessarily small scale. When distorting or resizing, I can obtain as many measurements as necessary as shown by 1 and 2 on this picture.

I cannot obtain any mathematical description of the surface to be covered. The intention is to ensure that the picture looks normal once it is in place on the 3-D surface.

  1. Resize the distorted image to the intended size.- approximately 1600 mm x 600 mm. The initial size of the images is 4288 x 2848 pixels. The likely viewing distance of the finished image is likely to be 1-4 metres. The answers to this question may provide assistance in the process of actual resizing and printing.

  2. Have the image printed - this would be of the distorted, resized image, within the minimum bounding rectangle of the image so the print would be rectangular with spare white space round the image. The print will them be laser cut to the pattern.

closed as off-topic by MikeW, NickM, inkista, John Cavan Sep 13 '15 at 21:13

  • This question does not appear to be about photography within the scope defined in the help center.
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    I flagged this question as off-topic. This is not about photography but about image projection. You should look at software recommendation and graphics SE. – Olivier Sep 7 '15 at 19:28
  • I think it passes the generally-broad "photographer's perspective" test for topicality. That is, it's a photographer asking for help in showing his photographs in a novel way. If this were about putting abstract patterns on shapes in the same way, that'd be different. (Or probably even photographs obtained from someone else.) – mattdm Sep 7 '15 at 20:36
  • 3
    Not about photography but about image projection. – MikeW Sep 8 '15 at 2:57
2

There's your problem

I cannot obtain any mathematical description of the surface to be covered.

A software capable of doing that distortion needs that mathematical description.

It's unclear why you claim that you are unable obtain such description. There are two solutions I can think of:

  1. Take the high tech approach and find a 3D scanning service near you.
  2. Go low tech and measure it out by hand: create a grid (in a plane, not curved) above the surfaces and take a height (or rather depth) measurements at the intersection points of the grid. (adjust the spacing of the grid depending on how precise you want to be) This produces a height or z map, which is pretty much the same thing as a plain old map of a landscape, that tells you the elevation at every point

This suggestion remains a partial answer as I can only speculate that Photoshop is probably able to do the rest of the job, but can not give a guarantee. I'd also look at 3D modelling software like Blender, which should also be helpful. Again, no software can help you if you cannot find some way to describe the 3D surface. Maybe even fiddling around fitting some 3D surfaces by hand in blender might be a valid option.

And now for something completely different

The intention is to ensure that the picture looks normal once it is in place on the 3-D surface.

Another word for such mapping would be a projection. Now what tool is good at making projections?

A projector.

Wait! Give it a thought while I change from my captain obvious costume to my regular clothing. Your problem boils down to "I want to see the 2D image when I look in that direction, no matter what freaky 3D surface might be there" And that's exactly what a projector does: it throws a 2D image at whatever you point it at.

Sometimes the easiest way to do math and physics is to not do them at all.

Using a projector to develop prints was common practice a few years ago. The usual process is to project onto a flat surface. Now yours is curved, so what? Increase the depth of field of the projector and project for a longer time to get better sharpness across the 3D surface, but in the end, it's not too different from a regular analog film workflow.

The problem with this approach might be to get the paper onto the 3D surface. In your question it looks like you are capable of doing the transformation with a piece of paper. This makes me think that you can do this with a piece of photographic paper as well.

Covering the unknown unmeasurable 3D surface with light sensitive photographic paper and then just throwing the 2D image at it with a projector might be easier than doing it in software.

  • The need to take depth measurements had escaped me entirely. Am I correct in inferring that such 3-D measurements would fulfil the need for a mathematical description. I anticipate that with such measurements I would have to derive an equation for the surface, and it was this that I said was not able to derive. I will have to go the low tech route as this is an amateur, personal project, and I have already had a rough quote for just the printing and cutting of £160 per panel. Other major expenses such as 3D scanning service are likely to make the entire project infeasible. – Chris Walton Sep 9 '15 at 16:25
  • +1 for "Sometimes the easiest way to do math and physics is to not do them at all". Your suggestion as Captain Obvious is very useful - I had not considered projection as a mechanism. I cannot take your suggestion on completely, as the panels on which I want to mount my photographs are fixed, outdoors, and not in a position where it will be easy to set up darkroom type facilities to utilise light sensitive paper. However use of projection may aid me in describing the image manipulation required. – Chris Walton Sep 9 '15 at 16:38
  • @ChrisWalton Most things 3D that you see in movies for example are made from discrete points. This is a lot easier to handle than analytic equations for the surfaces. But even if you need such equation it can be derived from the points you measured. There are mathematical ways to fit curves described by equations (more or less exactly) to discrete points. (regression is one example) The problem I see is that if the surfaces are fixed, you cannot project anything, but you cannot measure them either. You'd have to get the off. – null Sep 10 '15 at 21:56
1

The intention is to ensure that the picture looks normal once it is in place on the 3-D surface.

The bad news. On a 3D object you can only have one, and only one, point of view where your image looks normal. From there, every other point of view will start having distortions.

I can not obtain any mathematical description of the surface to be covered.

You do not need a mathematical description, you need the unwraping of the 3D object. The people making the 3D real life object could provide you the dimensions.

I can not

You don't, but a computer can.

1) Make a 3D model.

2) Select the point of view.

3) Project the image point of view to the uvmaping coordinates of the 3D object.

4) Use your image as a texture.

5) Unwrap a High resolution mapping using software like this: http://www.tamasoft.co.jp/pepakura-en/

Ask help from a designer familiar with a 3D sofwate to do this.

  • I don't understand what you said here. Typos; unfamiliar (to me) technical language; and my lack of familiarity with large scale, complex, printing; are to blame. My difficulties are: What is unwraping? Is this the same as unwrapping; and what do you mean by the term? Nobody is making a 3D real life object; I have some surfaces onto which I wish to print my photographs. These exist, but with no measurements available, except those I make What is meant by uvmaping - is this un-mapping; mapping to some sort of auxiliary coordinate system; or something else? What does puse mean? – Chris Walton Sep 9 '15 at 16:37
  • +1 for the bad news. My original statement of requirement was too strong - the photographs will be viewed from comparatively restricted positions; and the requirement is, rather than normal, is not to look too distorted from these restricted positions. – Chris Walton Sep 9 '15 at 16:38

Not the answer you're looking for? Browse other questions tagged or ask your own question.