Splitting out a depth render?

[pixelplucker]

I want to take a depth map and split it into multiple images to make slices for laser cutting. Is there any easy or automatic way of doing this or do I need to fiddle in my photo editor and use threshold?

 

thanks

[John Bigboote]

You can use the 'FOG' feature as a nifty quick and intuitive depth finder... but when you say slices... I don't know- perhaps you might want to make your object and then use a BOOLEAN feature to 'slice' it, and you can move the bool deeper and deeper into your scene... also, you would want to work orthographically(non-perspective) would'nt you?

[pixelplucker]

I can render a depth image from AM, what I want to do is take that image and make multiple images out of it. Each image is one level of the depth. I would then load those into my engraver and cut each layer. Each layer becomes a pass on the machine.

Higher end lasers have the ability to read a 3d model in and carve by varying the voltage.

 

I have a small desktop unit and made a special material for it that enables me to engrave or burn off .001" at a time. It seems to be more precise than the big lasers with varying voltage because I can control the depth by layers of material.

 

That make sense?

[robcat2075]

If you know how many slices you want and want to create a series of slices as separate images, Matt's Boolean suggestion can work well. There is a thread around here somewhere where I as asking a similar need for a different purpose and someone showed how to do it.

[largento]

I wonder if you could cause the fog or depth image to band by lowering the color depth? If you could get the image to band, it would seem to me you could select layers with the wand tool in Photoshop and create individual layers.

[robcat2075]

Here is the thread that shows how to capture intersections of a plane with a solid. Note Serg's demonstration

 

http://www.hash.com/forums/index.php?showt...amp;hl=meniscus

[robcat2075]

Another reason to do the Boolean slicing method rather than the depthmap is that Boolean slicing will get you the contours that would be hidden from the single viewpoint that a depthmap represents.

[pixelplucker]

Only problem with using booleans is many of my models are created in a cad program or as voxel exports to stl and aren't native AM files.

Number of slices may vary depending on my stock thickness and the detail of the objects I want to carve. 255 would be the maximum I think I could get out of an image.

 

Basically I am trying to avoid opening 255 copies of a depth image and using threshold on each of them then saving them out.. Seems pretty painfully slow.

I don't need tool paths since laser's can work with bitmap images. There are some topics on LDI (layered depth images) which seem to be used to reconstruct 3d models from 2d images and for masking images but I can't seem to find anything on splitting the layers out to individual files.

[robcat2075]

Only problem with using booleans is many of my models are created in a cad program or as voxel exports to stl and aren't native AM files.

 

Import it into a model. It's slow but you only have to wait once.

 

 

 

 

 

Basically I am trying to avoid opening 255 copies of a depth image and using threshold on each of them then saving them out.. Seems pretty painfully slow.

 

Do you have After Effects? Put the depthmap in that and keyframe some levels/threshold operation to make your near slice and far slice and stretch that out over as many frames as you meed slices.

 

Number of slices may vary depending on my stock thickness and the detail of the objects I want to carve. 255 would be the maximum I think I could get out of an image.

 

If you use the depthmap in OpenEXR your slices can be nearly infinite because it records teh depth in floating pint format rather than 8 bit integer.

[pixelplucker]

I don't have AF, is there another program that can do that? Used to use Debabalizer until they stopped making it.

[robcat2075]

Explain to me again why importing an STL into an AM model that you can do the Boolean slicing on is not an option.

 

That is so much more what you want to do than a depth map process that I can't figure out why you are not pursuing that.

[pixelplucker]

Explain to me again why importing an STL into an AM model that you can do the Boolean slicing on is not an option.

 

That is so much more what you want to do than a depth map process that I can't figure out why you are not pursuing that.

 

How do you do a boolean with an AM model on a prop? Program crashed soon as I tried.

[robcat2075]

there's no way to add a Boolean to a prop because there's no way to add bones to a Prop.

 

When I say import it into a model, I really mean import it into a model. Not as a prop.

 

-Open a new model window.

->Plugins>Import>STL or whatever polygon format works best.

 

Yes, it will take a long time for a big model. Let it go overnight if it needs to. When it's done you have an A:M mesh you can add bones and Boolean cutters to.

[pixelplucker]

Stl's probably won't import all that well since they are triangles. Unfortunately I can't keep just quads from either Moi or 3d Coat and n-gons get funky on import.

 

Still think it is easier to take a depth scan and make multiple images, I'll see about getting AE for that.

[robcat2075]

Stl's probably won't import all that well since they are triangles.

 

It costs nothing to try, which is what I'd do before declaring it impossible..

 

For a dense mesh it wont' matter if they are triangles or not, especially since all you are going to do with it is run a boolean through it.

 

You can select the whole thing and hit P to turn them into flat triangles just as the polygon program thought they were gong to be anyway.

[pixelplucker]

True,

thanks