Transformation in Sand

[Michael Gerald-Yamasaki]

Greetings. As I mentioned in the Introduce Yourself Forum, I became interested in photography after working many years in scientific visualization and my wife gave me a point and shoot and a copy of Photoshop. She had noticed a similarity between my work and what PS does.

The following is an example of my sci vis work...



This image was created with a technique I developed called vector field marquetry. It is created almost completely from transformations on the small image on the right of wind blown sand. Elevation data is processed to provide the "recipe" for shading, transforms, masking and coloring to create the visualization image.

If you look closely you will see the wind blown sand pattern replicated throughout the image. From the elevation data, a vector field of the gradient of the elevation is calculated (Think of at each point there is an arrow representing the direction of uphill and how steep is, the magnitude of the vector).

The gradient vector is used to select a transformed sand image so the sand lines are oriented as though the wind was flowing in the direction of the gradient (the sand hillocks are formed perpendicular to the wind flow direction). In essence, a layer is created with the direction rotated as determined by the gradient and masked for the areas of the same gradient direction).

A similar process is used to select a lightness value... The model is for side light coming from the right. So, right facing hillsides will be lighter, while left facing hillsides will be in shade. Finally, a hue for each pixel is selected on a linear hue scale corresponding to relative elevation from violet (low elevation) to red (high elevation).

You may notice some additional lines which are perpendicular to the wind blown sand lines. This is one added element consistent with the hand drawn lines in topo maps indicating regions of high gradient (very steep regions). These are created by random seeded line integrals of the gradient field in regions of high gradient magnitude. ;)

Ah, I see my attempt at a simple explanation has failed :D

Happy to answer questions, else thanks for looking.

Cheers,

-Michael-

[Arthur Morris]

I do not see the pattern and I have no clue what you are talking about but I do love the image. It certainly is OOTB and I would love to see more like it. It does look like an aerial topographic image of a really lush place.

[Cheryl Slechta]

Michael, it's beautiful and I'm sure I will never understand what you wrote:o It does look like aerial topography and the colors are lovely. I hope you post more of them.

[denise ippolito]

Michael, Really beautiful image and I can see this as a large print. It does look like a topo map in there somewhere.:)

[Michael Gerald-Yamasaki]

Artie,

Maybe this one will help you see the idea of the transforms (?). I used the small image of Spanish roof tiles and much simpler data for this one. The tiles are aligned upon the gradient. And only side light shading was used (no mapping of hue to height).

In any event you might notice the tiles pattern is present throughout but oriented and shaded to express the underlying "shape" of the data.



Artie, Cheryl, Denise, Thanks for the nice comments. I haven't looked at this for some years and hadn't looked at it through any experience as a photographer before... I think I might re-engage with the processing with art and photography as the main purpose (as opposed to science) and see where it leads.

Cheers,

-Michael-

[Arthur Morris]

Thanks Michael, I can see much more of a connection here but the understanding of the process is not there but that is not needed to enjoy the wonderful shapes and patterns.

[Judy Howle]

It's certainly an interesting technique and very effective. Stunning image!