Step 14: Finish platform

Place the small platform (with flange) over the short dowel that is attached to the bottom plate
I have read several of your comments and given the problems listed some thought. <br>It seems to me that one could just have a stationary table. Bevel the outer edge of the table top. Bend pvc into the bevel. Use a three way connector to attach the ends together. Attach another length of pipe to the third opening in the connector (the opening sticking out and away from the table top). Insert another pipe ( smaller in diameter) (longer in length) into that one. Drill pin holes at intervals into the outside pipe and a single set into the inside pipe. Use an adapter coupling on the inside pipe to add another very short piece of pipe that is the outside pipe's same size. Put another t connector on the very short pipe so that there are two ends aligned vertically. Attach a length of pipe to the bottom (enough so that it can rest on the floor). Slide the smaller pipe into the top. Make pin holes as before. Attach camera to top of small pipe. The end result should be something like a large, side ways T attached to a hoop that is wrapped around a table. Each leg of the T is adjustable (via the pin holes) and the hoop should spin freely around the table. <br>Below is a simple hand drawn picture to give you a better idea of what I mean. I hope this helps. <br>If any one is interested in this I would be willing to make a video on how to do this or make a 3d model of it. <br>Either one may take some time since I work offshore and will only be able to work on this once I get home.
I don't understand why this is better (or even as good as) having your camera on a tripod and the model on a turntable. This seems needlessly complex to an extreme degree. Maybe I missed something?
Since 123D Catch is stitching images by looking at surface features, a turntable would confuse it in two ways:<br><br>1. The object would appear to be moving in relation to the background.<br>2. The lighting would change from shot to shot, and the shadows would move.<br><br>Other scanning techniques (structured light, laser) are less subject to these constraints, but have their own limitations in terms of cost/accuracy/ability to capture a texture.
1. The object would appear to be moving in relation to the background.<br><br>you could attach background to the turntable---<br><br>2. The lighting would change from shot to shot, and the shadows would move.<br><br>The lighting should probably come from around the camera, in either case.<br><br>
I agree. I've used 123D Catch with a turntable. I made a background for the turntable (with a unique geometric colored shape for each corner). The lighting was consistent and 123D Catch had no problems. I suppose a large object may be more suited to having a moving camera mount.
Do you have a photo somewhere that shows what you mean? I'm having trouble visualizing what the background looks like and where the unique geometric shapes at the corner fit into the whole scheme...
Here is a Sketchup mockup of what I did. The white square is the paper sheet I printed with geometric shapes. The gray shape is just a model of a rock.
<p>Wow I have to give this a try. I'm working on developing a protocol for museums to use free apps to capture 3D image of specimens but have had little success so far. </p>
Thanks a lot! We're still refining the process for our models (primate skulls), so I will give this a try.
You are absolutely right!<br>I have tried to use 123D catch with a turntable and did not work even after giving it several tries.<br><br>Then I found the trick and the turntable method worket very fine.<br>I used a colored surface with large pattern for the turn table. I clicked photos with such angles that the background other than the turntable does not get into the photo. Also focused it nicely so that, the background was totally out of focus<br><br>Your method seems to be very useful one.<br><br>but clicking photos by holding the camera by hands, doen't give bad results in 123D catch. If you click photos with progression near 10% (90% prev data, 10% new data), then you can scan the objects with very sharp details.<br><br>Well, thanks for nice instructable and the design.<br>Please post the scanning videos and scanned object pics too!
<p>I have to give your method a try! I've been trying to capture fossils for a museum project with little success using a handheld camera.</p>
<p>It was easy to make, but it's a pity the measures were only in inches and not metric.</p>
An intriguing and well-executed, if a little impractical, project. The horizontal&nbsp;'swing-arm' effect&nbsp;requires a lot of floor space (though that would be good for lots of (natural) light), and does demand a FLAT floor.<br> <br> It does, however, seem a reasonable solution to modelling larger objects than might be practical on a turntable. Four more pieces similar to the 45-degree(?) braces for the vertical pole, but pivoted by their flats, to allow for larger platforms, could connect by dowels to the sleeve on the platform-support (possibly, it would also need pivoted attachments), to support larger platforms without blocking the 'swing-arm'.<br> <br> I would like to see some length/distance markings ruled on the 'swing-arm', and height&nbsp;marks on the vertical pole (so that you can set-up regularly-used points more accurately), as well as a line along the upright's entire length, which could be matched up with a mark on the camera collar, to keep it always pointed in the same direction&nbsp;.
I think it has been said before, but 123D Catch does not work well with a turntable. I've found that lighting changes confuse it.
It <em>has</em>, indeed, been said before - by many users - and I'd hope that the programmers are working on what seems to be a problem of scale.<br> <br> I say &quot;a problem of scale&quot; because&nbsp;this <em>is</em>, essentially, a giant, stripped-down, turntable - albeit with the camera moving, rather than the 'table' and subject - and 123D Catch doesn't seem to have a problem with <em>it</em>.<br> <br> The main issue with this particular rig is the need for&nbsp;a flat, smooth, level (circlular) space, something&nbsp;over 6-ft in diameter, which doesn't seem particularly suited to 'home use' (not that I think it was/is intended to be such).
Right, I agree, a turntable version of this would not be useful with the current 123D Catch software. <p><br><br>Sure, you need space to use this rig; no question about that. I like this Instructable because in my experience 123D Catch has worked best with consistently lit, well overlapped photos. This rig can help the less experienced user get that right. A person might attach lighting to the stable central platform to provide unmoving light on the subject.</p><p><br><br>I wonder about your suggestion to enhance the program to allow for a rotating subject with stationary camera. One beautiful feature of 123D Catch is that the photos can be taken with a handheld camera in arbitrary positions. I don't want to lose that. If the software needs consistent lighting cues to be able to handle the camera motions, that's a tradeoff I'm willing to accept. </p>
You're right - it <em>would</em> be a pity to lose the flexibility - hopefully a not-too-distant update will come with either a 'mode'(?) option so you can tell it if you're working freehand or static, or it might have to be 2 different 'versions' (though I'd hope -&nbsp;and assume -&nbsp;not, as they might conflict if you tried to install both on&nbsp;the one&nbsp;machine).<br> <br> However, looking anew at previous comments, some people are claiming success with turntables, so maybe there are ways around the problem, without waiting for software fixes that may never come (even if they are possible)...
I'm interested to know if anyone has practical experience with this rig? Did it improve the quality of your captures? I'm using a tripod and remote trigger at the moment but I'm looking for something better. <br>
Hi. I'm excited to explore your rig, but I am finding myself obfuscated trying to download the spl files and/or finding them on Thingverse. Any helpful hints? Thanks!
I really need/want a 3d printer. I am jealous (in the nicest possible way) that you are to bring your 3d ideas to life. One day, I will join you (figuratively speaking) in making good ideas into useful ideas. Keep up the nice work. Keep posting Instructables. And thanks, by the way, for including the stl files.
How would you scale this up to do the side of a whole car? Then how would you use the program? This is very interesting to me. <br> <br>Thanks in advance! <br> <br>Yaughn
Great work and nice prints. I especially like the red fittings. I'm waiting for my Replicator to arrive and am experimenting with both the Autodesk imaging system and the David laser system. I'll be working with smaller objects and am wondering, have you experimented with a turntable and stationary camera? <br> <br>...Man, I really like those red fittings:)
Nice work! Now you need a people free space so people won't keep walking into your pictures and messing up the results!<br /><br />A similar design made from PVC plumbing fixtures would be interesting as a more accessible/cheaper solution.<br /><br />What models have you managed to catch nicely so far?
I'll be posting models shortly! I think the Makerbot blog has images of some scans already.<br><br>I looked at PVC initially - the problem is that it is difficult to find PVC pipe fittings (especially flanges) with a 45 degree angle, and slip joints don't exist. PVC couplers generally contain a small ring inside to keep the pipes from sliding through, and for this design I need the pipes to slide.<br><br>However, the build stage lends itself to being built from metal plumbing fixtures, a pipe and two flanges. That would make it more stable. However, it would cost more than all the printed plastic parts put together...
I would respectfully disagree although I do live in an area where PVC supplies are rather abundant. Printing 3D unfortunately is exorbitantly expensive...assuming one can actually find a place with one of these devices for public use. Buying a machine is just straight out of the question considering huge expense. Sure, there is a really big coolness factor to 3D printing but it is just plain expensive. I think I will try this great project but I will use PVC and improvise wherever needed.
I want a 3D printer. :'( It seems that instructibles is becoming more and more a showcase for these machines and I can't do any of them. This makes me sad. <br><br>Nice idea btw. :D
I concur<br>
El cheapo.<br><br>An adjustable pole, a level,a camera, 123D Catch, and a piece of string.<br><br>1) Attach the camera and the level to the top of the pole.<br>2) Loop the string loosely around the object and tie a knot.<br>3) Tie string to the top of the pole securely at a distance from the object that suits your need.<br>4) Use the level (attached to the top of the pole) to ensure the camera is level.<br>5) take picture<br>6) move around object until you have photographed all aspects.<br>7) Detach string from subject and wrap arounf pole. <br>8) Detach camera.<br>7) input pictures to 123D catch.<br>9) Done!
Great idea and Amazing tutorial man, and the files... wow! Congratulations! Excelent work!
Okay work with me here - I can see a potential application / adaptation of this that would really rock! Don't forget where you heard it...<br><br>What if you built a structure or booth that the object - or even a person! - could sit in. The rig as described in this instructible would be flipped upside-down so that the axis is directly above the object (or person's) head. You could actually do a 3-D bust of an individual and from there create a mold to cast the likeness in plaster, cement or bronze!
Great idea...123 Catch is really really handy when fed properly.
Very cool. First time I've sen something useful come out of a makerbot. Usually it's the statue of liberty or star wars figures. How long did it take to print the components?
It took about 12 hours to print all the components, at stock Thing-O-Matic settings. I've had some success with using an experimental firmware that supports acceleration, and thus higher speeds, using that I estimate all parts could be printed in 4 hours.
I've not played with 123D yet... but might I suggest adding a <br>http://www.texample.net/tikz/examples/degree-wheel/<br>and a pointer to the base and boom?<br><br>having your photos accurate to within a part of a degree can't hurt any, and might help? If 123D can compensate without issue... the small addition may also make the rig useful for something else?
on second thought, a MUCH more useful addition would be a flood light mounted on top of the camera support pole.<br><br>20-30 pictures, twice = dead AA's for MY camera anyhow.<br>Without needing a flash, it'd handle 60 image captures without even losing a bar of battery...
Actually, putting a light on the rig would really throw things off, since the shadows would be in a different position from picture to picture, and would confuse the stitching algorithm. I tested running a string of LEDs up the pole.<br><br>If this didn't matter, then it would be a simple matter of putting the target object on a turntable.<br><br>The degree wheel is an excellent idea.

About This Instructable