shape from shade, so i do not take over Bjorn's Options

[JohnVV]

So that i do not take over his nice thread i figured i would start a new one
I am just figuring it out so bear with me
Rhea
This is a example - very early example
http://www.shatters.net/forum/viewtopic.php?f=5&t=15847

one that i am going to redo .
the program i am using "Cyclops" has a few bugs

Enceladus
this is a good example a bit better than the above



the close ups are a 8k level 3 VT with and without a texture over it
then a 4k level2 vt

unfortunately i still have to figure out a bug .I need to run a highpass on it ( the 32 bit isis dem )
Cyclops exports to a 3d *.ply this is easy to export to a isis cub
and i get this ( this happens when i combine 16 smaller images into one big one )
and this example is at 1/2 size


Just a bit of a bug !!!

[JohnVV]

why does the 2 sides of the crater rim SFS dem look higher then everything else

this is how the program works
the areas that are tangent do not pick up the height data very well

this causes the " valley" effect in the direction of the lighting

this is one of the drawbacks of using SFS

now using gmic to remove most of the low frequency data this is somewhat improved ,but not eliminated

there is still no really GREAT solution to this issue


SFS is still a problem to be solved

[JRehling]

there is still no really GREAT solution to this issue

SFS is still a problem to be solved

Indeed. And for a surface with unknown reflectance properties (e.g., an outer solar system body with very few images) not to mention albedo variation that is not due to illumination conditions, it's in principle an unsolvable problem.

A 2006 work, for reference:

https://ieeexplore.ieee.org/document/1640974

[TrappistPlanets]

a friend of mine processed this fantastic dem from one of the juno ganymede images

it captured all the little cracks and stuff very well

here is a map verson i made from that dem (fixed inaccurate elevation, and issues in some craters)

[JRehling]

a friend of mine processed this fantastic dem from one of the juno ganymede images
[...]
it captured all the little cracks and stuff very well

This isn't a DEM, and didn't capture Ganymede very well. Ganymede has tremendous variation in albedo and creating a bump map from that (which is a one-minute project in Photoshop) produces something which is largely fiction. Crater rays aren't cracks nor are they elevations, nor do the transitions from darker terrain to grooved terrain translate to regional depressions or elevation.

Moreover, the sun angle obviously varies profoundly across the image, with shadows nearly perpendicular at the terminator and almost overhead at the center of the limb. This effort would only be meaningful if the sun angle were approximately constant when it's about as different from constant as imaginable. Projecting this onto a cylindrical map makes it look like Ganymede has one rough area and another smooth area and that is total fiction.

There is plenty of data from previous missions that would enable a more complex process that would produce what this aspired to produce, but this isn't it.

There are multiple skilled image processing experts who post here and produce amazing, publishable work, but it's not easy, and there's a lot to learn from the care they put into their work.

[TrappistPlanets]

This isn't a DEM, and didn't capture Ganymede very well. Ganymede has tremendous variation in albedo and creating a bump map from that (which is a one-minute project in Photoshop) produces something which is largely fiction. Crater rays aren't cracks nor are they elevations, nor do the transitions from darker terrain to grooved terrain translate to regional depressions or elevation.

Moreover, the sun angle obviously varies profoundly across the image, with shadows nearly perpendicular at the terminator and almost overhead at the center of the limb. This effort would only be meaningful if the sun angle were approximately constant when it's about as different from constant as imaginable. Projecting this onto a cylindrical map makes it look like Ganymede has one rough area and another smooth area and that is total fiction.

There is plenty of data from previous missions that would enable a more complex process that would produce what this aspired to produce, but this isn't it.

There are multiple skilled image processing experts who post here and produce amazing, publishable work, but it's not easy, and there's a lot to learn from the care they put into their work.

but than why haven't we seen bigger DEM chunks than really tiny hunks for Ganymede (and Callisto( even fewer for Callisto)) in papers, but for Triton we have most of the cantaloupe terrain (for the approach hemisphere) height mapped, and we have a good chunk of the southern regions height mapped to (all though unreliable), according to this paper, there is plenty of usable PC DEM coverage and stereo DEM coverage (although its not going to be in the best resolution)

for callisto there is this blurry DEM from MexLab geoportal
(someone turned it into a 3d model and i processed it back into a DEM)


but it could be better though

does anyone know of any papers with significant stereo or PC DEM coverage that i haven't seen/found yet, because i am clearly missing something.

[JRehling]

but than why haven't we seen bigger DEM chunks than really tiny hunks for Ganymede

Possibly nobody finds DEM chunks a really valuable thing to produce. Someone interested in the science of Ganymede's geology can do work like this without trying to publish a map.

https://www.sciencedirect.com/science/artic...019103510003945

for callisto there is this blurry DEM from MexLab geoportal

Again, this is not a DEM. You seem to think that every bump map generated from albedo is a map of elevation. The bright spots on Callisto are craters with bright subsurface material revealed. Using that brightness to produce a map of elevation is pure fiction.

A bump map might happen to be very close to a DEM if a lot of conditions were being met which are not being met in the cases you're playing around with. Eg, fixed albedo, constant illumination geometry, a Lambertian surface or one with minimal specular properties under the given illumination conditions. When those conditions aren't being met, you get a work of fiction.

[TrappistPlanets]

Possibly nobody finds DEM chunks a really valuable thing to produce. Someone interested in the science of Ganymede's geology can do work like this without trying to publish a map.

https://www.sciencedirect.com/science/artic...019103510003945



Again, this is not a DEM. You seem to think that every bump map generated from albedo is a map of elevation. The bright spots on Callisto are craters with bright subsurface material revealed. Using that brightness to produce a map of elevation is pure fiction.

A bump map might happen to be very close to a DEM if a lot of conditions were being met which are not being met in the cases you're playing around with. Eg, fixed albedo, constant illumination geometry, a Lambertian surface or one with minimal specular properties under the given illumination conditions. When those conditions aren't being met, you get a work of fiction.

http://cartsrv.mexlab.ru/geoportal/

thats where the original callisto "dem" came from (you haft to do some playing around to get to it, there is also a blurry europa, and Io "dem") before someone turned it into this model https://www.thingiverse.com/thing:3686246 witch i processed a "dem" from

how is DEM data not a viable thing, we need to know topography elevation, like if we wanted to land a probe on an object, we need to know its terrain and as much of its climate as possible?

[JRehling]

First of all, the standard of citing research is to provide an author name of a paper, not that there is a blurry image on a website for people to go looking for with no information about who created it (not "someone"), how their results may be duplicated, what data and method was used to derive it.

how is DEM data not a viable thing, we need to know topography elevation, like if we wanted to land a probe on an object, we need to know its terrain and as much of its climate as possible?

And does a blurry image on a website that has none of the above information, hasn't been reviewed, and doesn't indicate what grayscale corresponds to what elevation help mission planners? Not a bit.

As far as I know, there are two different consumers of such products:

People who provide space art can use a DEM to generate a view of the world in viewing situations that have never actually occurred. We can easily do this with global DEMs of Mars, the Moon, etc. This is fun stuff… not particularly essential… but also depends on a pretty complete global map. With little fragments that cover a few percent, this doesn't work, unless you only want to show a little fraction of the world, which is also fun stuff, but not really a burning need.

People who do real science and want to know, as in the paper I linked to, how the geology of a world is evidenced in its topography, and those people want to know how tall a crater rim is vs. the floor. Not a blurry DEM that someone posted without any of the specific information that I listed above.

What you posted is a fiction. It shows that Ganymede is smooth in an area where it isn't smooth… it just had different illumination, in one image, on the limb than on the terminator. Who benefits from that fiction? It also shows changes in vertical elevation where bright crater rays exist. Who needs that fiction? Can someone planning a mission look at your image determine what grayscale corresponds to what elevation? Will they bet a billion dollar mission on that? No.

People aren't generating those because they're not valuable for any purpose. They aren't even real DEMs. You could just as easily apply the same Photoshop filter to the Mona Lisa and that won't generate a 3D version of a person's face, either.

Now an actual DEM where you began to take into account variation in albedo and the variation in illumination across a curved surface might be of some use. That's going to be vastly harder work to come up with than applying a Photoshop filter to an image. For starters, you have to understand that "shape from shade" only works when in the real world, shade is from shape; e.g., that variations in albedo and illumination geometry do not exist. Ganymede and Callisto are utterly disqualified from that category.

There's a lot of paying attention to detail that has to occur before products of value emerge. Then a lot of hard work. Those aren't steps to skip.