Build a DIY Desktop 3d Scanner With Infinite Resolution.

354,660

1,900

227

About: I make custom copper signs, metalwork and prototypes. I am one of the hosts of the "3d Printing Today Podcast", available on iTunes.

Intro: Build a DIY Desktop 3d Scanner With Infinite Resolution.

If you already have a camera then this project will cost less than $50 to complete. If you like to shop on AliExpress you can probably build it for around $30. If you like to support your local middleman, more power too you, but it will cost a little more. The giant screw shown above was made from a regular drywall screw captured using this scanner and a Nikon DSLR.

Desktop 3d scanning has made great leaps in recent years but it still has great limitations. Scanner hardware is built around a specific scan volume and resolution. You can get decent results, but only if your object fits that sweet spot. If your object is too small, or too detailed or your scanner is just having a bad day your scan will look like a potato. Luckily there is another approach. Photogrammetry uses a set of regular 2d photographs taken from all angles around an object. If a point on an object can be seen in at least three pictures then its location can be triangulated and measured in three dimensions. By identifying and calculating the location of thousands, or even millions of points the software can build up an extremely accurate reproduction. Unlike a hardware based scanner there are no size or resolution limitations to this process. If you can photograph an object you can 3d scan it. It works from molecules to galaxies, or it would if they would ever approve my Hubble time.

The limiting factor with photogrammetry is the quality of the photographs and thus the skill of the photographer. Photos must be well exposed and in razor sharp focus. They must also be spaced around the object so they capture every part to be scanned and they overlap enough that the software can figure out where each shot belongs. With large objects this can be done manually with some practice, but it is virtually impossible to do it well with a small object. This scanner automates the process.

Step 1: How It Works

A high quality photogrammetry scan requires high quality photos of the subject from all angles. The easiest approach for scanning small things is to rotate the subject while photographing it. This scanner uses a stepper motor controlled by an arduino board. The stepper turns the object by a fixed amount and then an infra red LED fires off fiendishly clever series of blinks which mimics the camera's wireless remote. The camera being rather gullible and wishing to please takes the picture.

An lcd display shield with a set of buttons allows the user to control the arduino. Using the buttons the user can select the number of pictures to be taken per revolution. The scanner can run in automatic mode where it takes a picture, advances the stepper and repeats until it has completed a whole revolution. There is also a manual mode where each push of the button takes a picture, advances the stepper and waits. This is useful for scans where each picture needs to be framed and focused manually.

Step 2: Some Tips on Shooting for Photogrammetry

When photogrammetry software detects a feature in a photograph it tries to find that feature in other photos and records its location in all the shots it appears. If the feature is part of the object which is rotating we get good data. If the feature detected is on the background and is not moving while the rest of the object is it can cause a implosion of the time space continuum, at least as far as your software is concerned. There are two solutions. One is to move the camera around the object so the background stays in sync with the movement. This works well for large objects but it is much more difficult to automate. The easier solution is to keep the background featureless. This is easy to do for small objects. https://www.thingiverse.com/thing:548006 holds a sheet of regular paper like a photo studio sweep which presents a seamless white background. Combine this with proper lighting and you are well on your way to featureless backgrounds. Another tip is to over expose your images by a stop or two. This allows you the capture more detail in the shadows of an object while simultaneously burning out the background so any remaining background features vanish in the blazing white.

Step 3: What You Will Need

An arduino. Any arduino will work. I used a mega because it has pins which aren't covered by the LCD shield which makes it easier to wire.

A SainSmart 1602 LCD Shield which provides a display and a bunch of buttons to control the scanner.

A stepper motor driver (I am using an Easy Driver) This allows the Arduino to drive the stepper motor. Any stepper driver should work because we are not asking much of it except to drive the stepper which is what they do for a living after all.

A NEMA 17 stepper motor which turns the scan subject. With a larger stepper motor (with an appropriate driver and power supply) this scanner could be could scaled up scan parts of that giant robot you found in the junk yard.

A 950nm IR led which fires the camera. You could probably salvage one from an old tv remote if you are the kind of person who finds that sort of thing entertaining.

Some type of power source for the arduino and stepper motor. A 9v wall wart works fine, a 9v battery works in a pinch but the battery needs to be fresh. The stepper won't work when powered solely by a USB port.

I have attached 3d printable files for the case I made. It is extremely elegant and sure to make the rounds of Pinterest due to its beauty and innovative style. Although it would be foolhardy to attempt to equal a design this classic please feel free to come up with your own, lesser, homage. and let me know if you do because I would love a nicer looking case.

Step 4: Hooking It Up

Hooking the pieces up is pretty straightforward provided you have the eyes and fingers of a 10 year old. If you don't perhaps you can find one to help you, or at least lend you their eyes and fingers. The plan shows the pins I used in the firmware, but this is very simple and you can switch them around to whatever works best for you. The EasyDriver board is mounted on the lid of the case so it gets better ventilation. I had planned to put a heat sink on it, but it doesn't seem to need it. We aren't pushing the stepper very hard. Astute observers may notice that we are powering the EasyDriver off of the unregulated power terminal of the Arduino. Probably not the best practice, but unlikely to cause the time-space continuum to implode because we aren't drawing much power. Similarly I included a ballast resistor on the LED just because everyone who knows more than me says I should. Since the LED only sends out brief pulses it doesn't have much time to cook itself under the whopping 5v the board is putting out, but follow your own conscience here. A 9v wall wart is the best power supply, but a fresh 9v battery works too, but it won't last long. If everything else is working except the stepper check the battery because the arduino will work on a partially drained battery, the EasyDriver will not.

You can download the fritzing schematic from http://fritzing.org/projects/shapespeare-scanner

Step 5: Loading the Firmware

Download the firmware sketch and open it in the Arduino IDE. The firmware is configured by default to work with Nikon cameras. If you are using a different brand you will need to change the configuration on line 47. The scanner implements Sebastian Setz's Multi Camera IR Control library so it can work with almost any brand of camera which works with an IR remote. If you are using a Leica or a Hassablad you can probably afford to hire Mr. Setz to write the driver for you.

The controls are as follows:

SELECT: Start/Stop automatic scan.

LEFT: Manual mode. Trigger exposure and advance one unit clockwise.

RIGHT: Step one unit clockwise without triggering camera.

DOWN: Step one unit counter clockwise without triggering camera.

UP: Cycle through options for number of exposures per revolution.

Don't ask me why I laid it out like this. It made sense at the time. The sketch is pretty simple so it should be easy for you to customize if you need to.

If you get a compiler error from the DFR_Key library try installing version 1.1. It has the DFR_Key library in the download package.

Step 6: Making Your Scan

Lighting is critical for photogrammetry. I use 3-4 JANSJÖ desk lamps from Ikea to light the subject. These lamps are very easy to position, and put out a lot of light. They have small heads which makes it easy to light small subjects. They are LEDs so they put out very little heat. You can improvise diffusers with regular paper and tape and not worry about setting the secret laboratory on fire. Plus they are $10 each! I also sometimes use a cheap LED ring light on the camera lens which provides nice diffuse light, but since it is attached to the filter ring of the lens it tends to get in the way with macro shots where the subject is almost touching the lens.

I use a $25 macro rail from Amazon to aid in positioning the camera. It is about a rigid as an overcooked raman (surely that is the proper singular form) Provided the macro rail doesn't actually collapse under the weight of your terribly expensive and impressive camera (and don't blame me if it does) and assuming your aren't holding your photo session in the middle of a hurricane the noodle like flexibility doesn't really matter much because you aren't actually touching the camera when you make the exposure.

Frame your subject so it fills the picture as much as possible. Make sure as much of the image is in focus as possible. After you finish a revolution reposition the camera vertically. Shoot as many revolutions as you think you need to capture the object.

Step 7: Process Your Scan

You currently have many choices for Photogrammetry software, but there are a few options which clearly lead the pack:

Agisoft Photoscan is the best commercial package which is remotely affordable. The $180 standard version is adequate for this type of scanning. Be warned that photogrammetry is very computationally intensive so it (and you) will be happiest running on the most powerful machine you can muster. If you are running your grandma's 486 and are pissed that Microsoft no longer supports XP this will end poorly for you.

Autodesk Memento is a close runner up. It sometimes does better than Agisoft if you have low quality photos, or not many of them. It has better tools for processing after the scan is made, but less control of the scan itself. Memento is cloud based which is nice if you have an old slow computer, but can be annoying if you get the feeling you could process your scan faster on your hampster powered Babbage engine. It is currently free, but may not always be so.

123d Catch is Autodesk's consumer level photogrammetry package. It has the advantage that you can manually align photos which can help salvage a bad scan with not enough overlap between the shots. But the number of photos you can upload is limited so that makes it difficult to scan something very detailed. And just because you can manually align photos does not mean that it is an amusing way to spend your spare time.

Step 8: Where to Take It From Here

So make yourself a scanner. Take out a second mortgage on the secret lair and buy a really expensive and impressive camera, or at least a decent point and shoot with a macro mode. Find a innocent looking dead bug, scan it and use it as the basis for your army of giant robotic monsters, scheduled to take over the world any day now.

If you want to learn more about photogrammetry and 3d printing check out the "3dPrintingToday Podcast", available on iTunes and wherever fine podcasts are given away. We talked about the creation of this scanner in show #111. We talk about awesome 3d printing tips and tricks and stuff in all 125 (and counting) episodes.

11 People Made This Project!

Recommendations

  • Plastics Contest

    Plastics Contest
  • Audio Contest 2018

    Audio Contest 2018
  • Electronics Tips & Tricks Challenge

    Electronics Tips & Tricks Challenge

227 Discussions

0
None
shapespeare

2 years ago

I uploaded version 1.1 on the same download page as before. It has DFR_Key included in the download package. It compiles and installs on my system using Arduino 1.6.7 (of course the old version did too). Hopefully this will fix the issue some of you have been having

3 replies
0
None
NikosN14shapespeare

Reply 3 months ago

Hi there and thank you for you amazing work!

I had a hard time making some clone board and screen work and looking deeper I found that the ADC key filter in DFR_Key, and in general the DFR lib, did not work for me. Digging into the sample LCD code on DFR's wiki

https://www.dfrobot.com/wiki/index.php/LCD_KeyPad_...

I found a different ADC routine that solved my problem. Take a look, it might help you simplify things a bit and provide a more robust ADC method.

Again, awesome work!

0
None
mberlandsNikosN14

Reply 25 days ago

Hello, NikosN14! It could be nice if you can explain what you have changed to make it work! I also have some 3rd party boards and no success. Can I ask for your code that works?

0
None
brianalipashapespeare

Reply 2 years ago

thanks i'll try that and let you know how i get on --- I also looked at the arduino forums and it seem there has been a change in layout of the Dir structure, the libraries now go in two places one for included libraries and one for user libraries this may also be part of the problem I'll post my results when i get to them later to-day

thanks again for the great support

Brian

0
None
BraidenP

15 days ago on Introduction

What changes to line 47 regarding diffeeent cameras? I was thinking of either Sony or to use the HP 3D scan cameras. Also would you agree that lower megapixel and higher resolution is ideal for scanning?

0
None
shapespeareMickH34

Answer 4 weeks ago

LIDARs are getting cheaper all the time. One of the units they sell for robots could work. Let me know if you do it.

0
None
gaudia

Question 6 weeks ago on Introduction

Hello Shapespeare, almost working but the canon mark ii takes the picture only when

the table is moving. Is there any way to adjust when the picture is taken? thanks, alan

1 more answer
0
None
shapespearegaudia

Answer 4 weeks ago

Its been a long time since I looked at the code, but I recall a timer which delays the shot after the turntable stops moving to allow things to be absolutely still.

0
None
Kyler EllisonF

Question 4 weeks ago on Step 5

I cannot get the multi CameraIR library to work. I've uninstalled and reinstalled it to IDE about 20 times, restarted my computer, double checked that it is in the right folder location and it is - no matter what I do it always says it cannot find the library - that there is no such file or library. Hitting the banging head on table point; any help would be much appreciated.

1 more answer
0
None
shapespeareKyler EllisonF

Answer 4 weeks ago

I'm sorry, that's beyond my area of expertise. I would check the paths again, but I'm sure you've done that.

0
None
Dwal001

Question 4 weeks ago on Introduction

I’m a novice so these questions might sound stupid. Will the 3D scanner work using old cell phone cameras? Is it possible to get better scans if more than one camera is used? Last question for now, I have a old Anet A8 motherboard and a 4 button display, are They usable for this project?

Thanks,
Dave

1 more answer
0
None
shapespeareDwal001

Answer 4 weeks ago

The firmware is very simple. You could probably adapt it to run on almost anything if you knew what you were doing. Phone cameras would work, but would limit what you could scan to things they could photograph well. Multiple cameras speed things up but don't improve the scan quality unless you are scanning something which moves. Then you need a hundred cameras and fire them all off at once.

0
None
MattiV

3 months ago

interesting,but how i connect my go pro clone cam it thats ? and 2 laser. poor men version if only rotation table and no more.and need add future cheap projector too who show raster image parts top.

0
None
offtherails2010

9 months ago

howdy shapespeare

Um, im at a point where i need to have a cry lol

.

i dont think i will ever be able to afford such a camera such as any DSLR types, especially the Nikon DSLR :(

.

all i have is a spare iphone 4S and a current iphone 5S with a very aged Kodak C713 camera (7MP) but none have this special ability to take photo's via IR/RF/Anything except manual operation :(

is there any hope for me with this kind of device ? lol

thank you loads in advance for any and all help, merry Christmas too :)

4 replies
0
None
kqb64n4offtherails2010

Reply 9 months ago

The cheapest method I can think of to use a manual-button-only camera would involve making something close to the simplest robot possible by using a small servo on a bracket attached to the camera, then replace the Arduino code for the IR with a servo PWM code to run the servo arm down and back up and have some type of small arm on the servo (they usually come with the servo) press and release the shutter button. A 9g servo should work fine and you can find them all over the web for well under ten bucks, or go for free and tear one out of a cheap RC toy that quit working. As for the bracket, anything from popsicle sticks and duct tape to diamond coated, laser machined titanium should all work about equally well.

0
None
offtherails2010kqb64n4

Reply 9 months ago

howdy kqb64n4

Sorry for pestering you, but i have a quick question;

Suppose i was going to use a point & shoot camera specifically for this project and that will be the only use for this camera - so knowing this, if i take the camera apart and find the button/momentary switch that takes a photo - how can i connect this switch directly to the Arduino Pins ??

Would i need additional circuitry ? im guessing that it would be 5v (or 3v) on the switch but the camera is powered by two AA Batteries in series, so thats 3v total, roughly...

Would i be able to get away with simply wiring up the switch of the camera directly to the Arduino Mega Pins ?

If not, what extra Devies/IC's/Components would i need ???

i already have some servo's but these are for a different project and i don't want to use a servo on this camera when i can wire up the switch directly

- i think i can do it but i don't know if i would need anything else to go along with this hack to make it work.

Naturally the code would be needing to be tweaked too.

Any help appreciated greatly :)

0
None
offtherails2010kqb64n4

Reply 9 months ago

thanks so much for the reply kqb64n4:)

thats helps a hell of alot, thanks for the inspiration ! hope restored that i might actually be able to make one of these lol

Parts collection started lol !

0
None
florianschwander

1 year ago

For those of you who experience the problem when the stepper is moving too early (post-shutter wait not working): The problem is probably caused by not updating the "currentTime" variable before the next "if"statement.

Solution:

Just move the "if (waitFlag == 2)"-block under the "currentTime = millis())-line, like seen below. Leave the rest untouched.

-------------------------------------------------------
currentTime = millis();

if (waitFlag == 2) // wait after triggering shutter before moving motor
{
if (currentTime - startWait >= postWait) //wait time has expired
{
waitFlag = 0; //done waiting
}
}

if (waitFlag == 0) // advance stepper and start wait timer
{

And so on....

---------------------------------------------------------

I'm still waiting for some parts, but now it seems to work just fine!

THX to shapespeare!!!