Loading
Hey! This thing is still a Work in Progress. Files, instructions, and other stuff might change!
3Dmaker4U

Scanoman - Full-body portable 3D scanner

by 3Dmaker4U Jul 13, 2014
Download All Files

Thing Apps Enabled

Please Login to Comment

Sir, here turn table reducer gear files are missing...so will you please upload those reducer designs?...thank you

Hi!
I'll try to do it tonight EET :)
Hopefully in one day the longest ;)
All The Best,
Daniel

Hi could you please point me in the correct direction for the kinect 360 drivers. I have reconstruct me also upgraded the graphics card but after a week of searching and not getting the camera to work is a bit disheartening.
Any help would be appreciated and a BOM please
Many Danks
Bill

Hi-
Great project.
Just found your videos on Youtube
https://www.youtube.com/watch?v=fv2Qphml9EY
Looks to take about 2 minutes to scan a person with your rig?
3 rotations of the wheel?
Can you describe the motion of the camera? What do the gears to tilt it up and down?

I like a lot of what you are doing here but IMO the turntable is too small. You really want to be able to offer scans of multiple people. You also may want to consider scan rates for subjects with high geometry outfits or couples. This is why I have stayed with a manual scan process as I can tailor the movement to the subject, resulting in a more accurate scan that is often in a much shorter period of time (40-60 seconds. http://www.thingiverse.com/thing:127968

I have a new turntable design I will be posting soon that improves upon my existing turntable, replacing the belt with a chain drive. http://www.thingiverse.com/thing:28454

The Scan-O-Tron 3000
by fredini
Heavy Duty Lazy Susan for Full Body Scanning (Legacy Design)
by fredini

Hi Fredini!

Nice to hear abou you. I have to admit that I got some inspiration from your project when I started :)
Currently I use 2 modes of operation. The first is to obtain a bust and takes 30 seconds in one rotation. The second is a full body scan in 3 turns and less than 2 minutes. For the latter, the sensor starts at mid height, then goes to face level, and finally to the knee level. Vertical movement is done without any rotation. Tilting is not actually used, and most probably the mechanism for this will be eliminated in the next version to reduce weight. Initially, I thought it will be useful to look to the top of the hear from above, just to close the mesh, but proved that data processing can do this fine. It might be still useful for some hidden areas, but I don't think this fits the automation.
You're right that the turntable is for one person only, but keep in mind that one of the key objectives was portability. The next version will be quite ambitious in this sense, aiming to make Scanoman a cabin luggage for (almost) any flight ;) Still using low-end reprap 3D printer parts to control the process.

All The Best,
Daniel

Comments deleted.

Dear Mr Daniel Constantin

First I would like to introduce my self, my name is Irfan PP, I live in indonesia. graduated bachelor degree of Computer Science (Information of Technology). now, Im magister student of Mechanical Engineering and Industry - Faculty of Engineering – Gajah Mada University (UGM) - Yogyakarta, Indonesia.
Currently my magister research topic is related to 3D scanner. why 3D scanner?, because these technology recently growing fast and also are helpfull in the field of many research and especially in design. one of them is 3D human body scanning technology, 3D human measurement methode and application would be one of the most demanding and interesting research these days.
More spesific about my research is to analyze the parameter (internal and external ) of 3D scanner devices for better (high Accuracy and presicion) output model.the scope object is full human body. But What most thing to be considered are cost, accuracy, ease of use and speed of acquisition.
so I begun doing research for this topic by searching many literatures/ journals / articles over the internet and found some good journal to read that related to my research. it is interesting because there is a device that seem has a potential to be use as 3D scanning devices based on the earlier research. this device created and meant to use for gaming purposes so the price is cheap. and also its ability is to scan 3D object (fast enough) in nearly real time and easy of use. then finally it came to my mind that my research topic would be :
"LOW Cost 3D Scanner - Parameter optimization for better Accuracy 3D Body Scanning Using Single Depth Sensor - The Micorsosft XBox 360 Kinect"
on the earlier research, some have already used this (kinect technology) 2/3/4 devices placed at different position for testing the quality of 3D body scanning using their methodes to optimized the output (accuracy and precision) model. but what they do were only involve of analyzing internal parameter (the function of the system devices). my draft idea is not only involving internal parameter but also the external parameter and how to design it. it simple by using only one kinect, turntable (object placed -> body), and tripod (where the kinect placed).
My goal is to find equation / formula ( objective function of mathematics model) using combination of the parameter (external & internal) to get high precision and accuracy model.example : external parameter to test such as lighting /background, distances of the object, sizes/ height of the object (human), speed of turntable, angle of sensor or the object, methode of scanning process ect.

When I red this Scannoman project seems to be really matched to my research. That is the reason why Im really interested in your project sir. it will be my honour to be able to meet with you, learning more further and sharing knowledge about the future of this project. I just can't wait to see for the release full manual / guideline documentation of this project which will be a really big help to accomplish my research.

sorry for my bad english grammar.

Hi Irfan!

Sent you an email a week ago (or so). Don't know if it was the right address, anyway. That's why I just replicate the actual message here, as well.

Your idea might be very valuable in improving the performance of such a 3D scanning arrangement.
And you are right, the scan results depend on both intrinsic and extrinsic parameters, as well as on various other factors that were less studied and are obvious application dependent.

I have to admit that, though I'm still very interested in further improving Scanoman, I'm facing some difficulties in allocating enough time for this. However, since many people asked for documentation, I scheduled a few webmastering sessions and hopefully the documentation will be available by the end of this month.

Based on the experience so far, I also want to produce a second version, lighter and much more... portable (well, 12 kg is not a pretty handy weight :)

The main drawback of the Kinect solution is the very poor RGB resolution. While it does not matter it terms of actual 3D model (if used for monochrome 3D printing), it is a petty for web presentation. You know perhaps the sensor made for iPad/iPhone, which has basically the same depth sensor, but uses the great back camera of those devices, hence producing very sharp colored models.

If you want to experiment AND have programming skills (or have such friends :) then I would recommend you the ReconstructMe SDK from Profactor. I just use their standalone application, which has not been yet updated to the latest developments, but works reasonably.

Meeting me in flesh is not a trivial action :) However, we can meet virtually in a Hangout. I would suggest next week (this one is already very busy).

Looking forward to hear about you.

All The Best,
Daniel

Awesome project. Can you also share BOM and details about how to programm the automatic run?
Really great project!!!

Thanks for the kind words!

Good news is that there are some pages under preparation to describing the device in detail.
Bad news is that this has started about six months ago and are yet not finished.
I'm also confronted with the dilemma of whether to document this initial version or to concentrate on the first release version.
The dilemma is standing, as the same good and bad news apply :)
However, I recently saw an increasing interest, so I'm inclining to finish the documentation of the initial version and publish it.
Hopefully this will be available in a month or so.
As for the automation, what I do now is really rudimentary, but straight and efficient: open ReconstructMe and Pronterface, set whatever is needed and then just click on a Windows macro than will call in turn each application. The actual "sources" will be part of the documentation :)

All The Best,
Daniel