creating a single 3D image from 2D jpgs

Hello, I have several hundred 2D OCT scans in jpg format and i've been told I could use AutoCAD to create a single 3D image of the analysed probe starting from those photos.

Sadly, I am a complete beginner at this software so any indication or hint would be highly apreciated.

Thank you for your time.

I've never heard of that technique. Who told you this?

What are you trying to construct a 3D drawing of? A building? An object (such as a machine part)?

No. I have a few hundred pictures of an OCT analysis, taken as the laser walked over a probe. They kind of look like this:

My task is to create a single image of the probe in 3 dimensions from them, and I've been told to use either AutoCAD or 3DS MAX. As I said, I'm absolutely clueless about how something like this could be done, so I hope someone here isn't . Thank you.

Edit:the guys that analysed the probe did this with a software of their own making, to which i sadly do not have access, so I have to try with something else.

Just looking at the image makes my eyes hurt. I can't think of a way to do it but maybe someone else can. I guess I need to read up on OCT analysis and running a lazer over a proble. I'm clueless. Sorry.

remark, oct is Optical Coherence Tomography. Its basically an xray of a part for material science purposes(and other applications). The image he is showing is basically a slice of some object i think.

Miaz, im sorry but i know of no way you can take jpeg images and create a 3d object. You can import jpeg images as raster image files and "trace" over your white lines creating a reference line that can be used. im assuming the jpegs you have are at different "heights" and you need to stack these images and create a solid of some sort? This will be very imprecisise and impractical.

What is the probe of? do you have a picture or an actual part? What are you going to get by recreating this? The OCT stuff should be used to find cracks or intersticial spaces in the material. Are they wanting you to create the imperfections?

Thank you shift. I had yet to crack open my Encyclopedia Brittanica!

Maybe Miaz is thinking of stereolithography? I know of some copmpanies that take stereo photographs, of piping for example, and from the resultant images can produce a 3D drawing for use in AutoCAD. But this lazer image of a probe is quite a different application.

The probe was a piece of mandible with two screws bolted in it. I do not possess an actual picture of it. In the end I should obtain just that, the 3D image of the "x-rayed" probe.

Shift, that stacking thing might be helpful, I will look into it. I don't know how to put it in technical terms, but placing the photos "one over the other" in a way or another should do the trick.

is it possible for you to upload a few jpegs. This may help out if we can work with/manipulate them.

can you get any other file types from the scan company? or are the jpegs all you have to work from?

The problem is jpegs really dont carry any information with them like a dwg, dxf, stl, iges etc....

what do you need to do with the model you end up with?

I only have the jpegs. A friend told me something about inserting a jpg, placing dots, raising the ucs (what's that?) inserting another photo and ading dots and so on. Didn't understand much really .

I'll place the photos here with image tags if you don't mind:

For the first part i only have to get it done. I really don't know what's in store for me afterwards.

Also, notice that loop in the upper left corner, it's a strange aberation but i can neglect it for the time being.

You want to insert jpg, place dots, raise UCS all to create a 3D drawing of this probe? And you don't know much about AutoCAD? How big is this probe? Can't you take it somewhere to be digitized in 3D?

Yes, I need to create a 3D drawing of this probe. There's no certain technique i must use, I simply have to so it one way or another. I've also asked the 3dsmax users for help, i'm waiting for answers.

How big is the probe? Well, as big as a piece of mandible with two screws bolted in it . Oh yes, I've forgot to mention. The photos measure 3 mm on the OY axis (vertical in the convetion we use for such scans) at 1024x512 resolution. This means that a pixel is equal to 5,8 microns.

It doesn't have to be of too much finesse for starters, an inaccurate rudimentary image is quite enough for the beginning, improvments will be made in time.

It's a tad painful and frustrating for me, I've never been a fan of computer stuff (I'm the old fashioned pen & paper guy) but I guess I will have to see things through.

So, once again, I guess that for starters it is really enough to manage to somehow place the photos one atop another and then be able to move around the 3D object thus obtained. Is something like this possible in autoCAD?

I still have no idea how big this probe is. Is is bigger than a car or smaller than a toaster?

You want to place all these "points" in 3D space then there would have to be some kind of "reference line" to work off of. Maybe you could take individual jpgs and "digitize" the points then rotate everything into the proper axis. Know anyone with a digitizing board?

I did this real fast for one of the images for you. This is not a precise method. You will never get a precise part off of a jpeg!

I used insert, raster image and put the jpeg in my drawing. then i used the spline command and turned Osnap off(or else you will snap to the edges of your drawing) and traced the while lines as best as i could. I did this in multiple line segments. each time i ended a segment and started a new one i turned osnap back on so it would snap to the end of the last line, then turn it back off and continue tracing. After i was done i drew vertical lines representing my height between photos. At this point i would start a new layer and turn this layer off moving my ucs to the end of one of my vertical lines. You will always default to drawing in the XY plane. For this i just copied my first line up to show you cause i dont have the time to draw a second one. then i used edgesurf to create a surface. you need 4 lines to use edge surf. two vertical and two of my splines. You can see this will be a very long inaccurate process.

oh and remark, the drawing is huge. my vertical lines are a few inches tall. I drew a line 1" tall and i couldnt find it:)

Thank you very much shift. Will this procedure also conserve the information regarding the depth the laser penetrated the probe? (basically the "intensity" of white over the surface).

Remark, as I said, the probe consits of a piece of mandible (so several inches) with two screws bolted in it. The screws were quite big also, but all in all, I guess it could fit in your hand, provided you have a big hand.

Then it is of a size that it could be digitized using a 3D digitizing arm or a handheld digitizer.

when i insert the jpeg into my cad model to scale it comes in 800 units(inches in my case) long. Even if this were a mm number it would be 31.5inches. This is an issue when using a jpeg image. While it may retain its lxw ratio it does not keep a scale.

This does not preserve any information. You are just "tracing" over the white line manually in the sketch plane. Everything is on the same plane. Without some other file format that includes some information there is no way to get this back. You may want to look into grey scale imaging. basically the difference in color from black grey white holds an intensity value. Cad is not a program that will handle this(as far as i know).

This type of imagining will only handle one picture at a time and i dont know how you would layer them.

The procedure i showed you above will only be as good as taking the drawing, laying it under trace paper and manually tracing by hand. Doing this in cad gives you the ability to create a surface between curves but without some sort of point cloud data from the scane(which doesnt happen with this type) i dont see an easy way to recreate what you want.

i just had a friend of mine look at it who does photo editing and we still arent really sure what we are looking at but we both came to the consensus that this is a very tough task.

thinking about it, how would you align each later to another? are all the scans centered the same on the page?

Yes, the photos are hard to follow for me too, not being a real expert and all.

Basically you have a surface seen from an eagle eye (the laser's) point of view. The laser shoots across it and what you basically see is how deep it goes. Then the probe (or the laser) is moved by a certain distance and the process repeats. The photos were coordinated with the laser.

The lab-boys that made the scan had a program built specialy for this and they got a great result. Sadly, we do not have access to that program so I've got to find a way to deal with something else.

I guess the best thing that could happen would be this:

It really doesn't have to be that fabulous, a stationary 3D model retaining the information of the tomography would do.

are these 4jpegs each side of a model? I understand what the picture is saying but im having trouble deciphering what the object actually is.