In this post I will introduce you to 3D scanning. What some people certainly see as a means to an end in the production of colorful figures is in itself a separate – very beautiful – discipline in our universe. First, I'll give you an overview of different 3D scanners. After that, you'll learn about the different methods of 3D scanning and hopefully, by the end of the post, you'll know a lot more than you do now.
What is 3D scanning?
- portabler 3D-Handscanner für Intel iCore5 Prozessoren (oder höher) USB 3.0
- Auflösung x/y 0.9mm - Tiefenauflösung 1mm - Scan Volume min. 0.2 m x 0.2 m x 0.2 m Scan...
* Stand: 2023-07-05 / Bilder: Amazon API
What are the different 3D scanning methods?
It sounds almost too simple to be true: Instead of tediously and lengthily familiarizing yourself with 3D graphics software, simply take a standard digital camera, take a few photos and then have the software create a complete, three-dimensional image of the photographed object – which is suitable for 3D printing on any standard 3D printer. What initially reads like science fiction actually works. And even quite well! As an alternative to the digital camera, line lasers from the hardware store are also suitable – or the Kinect camera from Microsoft, which is part of the X-Box peripherals.
The current common methods in the field of 3D scanning include photogrammetry (also called photometric), the light-sectioning method (triangulation), working with coded light, and the use of the Kinect in the context of the structured light method.
How does photogrammetry work?
For the process of photogrammetry, you basically need exactly two things: a digital camera and an Internet connection. With the camera, you circle the object, which will be printed later, and shoot photos from all perspectives. The resulting photos are then uploaded to specialized web services such as my3Dscanner.com or Autodesk's 123D Catch. After the upload, the services start analyzing the images and create a three-dimensional image of the object. For this process to work, the photos must contain distinctive points – for example, sharp lines or light-dark patterns that repeat in all photos. This is the only way that the service provider's software can recognize where exactly the photographed object is located in the room. During the creation of the three-dimensional object, the software first generates a “point cloud” from the data in the image files, then refines this cloud and finally generates a digital surface mesh between the points. Once the calculation of the 3D object is complete – which can take up to a day – the user receives the finished file conveniently by mail.
How does the light section process work?
A digital camera is also required for the light sectioning process. However, the convenient support by a web service is omitted here, additional hardware is required instead. This hardware can be found in any hardware store in the form of a line laser. The line laser is used to mark the object to be digitized slice by slice so that it can then be recorded with the camera placed to the side. In this way, angles between the laser, the object and the camera can be detected – which, in combination with the distortions of the image recordings, allow the triangulation of the depth for each point on the laser line.
In order to generate a complete image of the object, the laser line must be completely swept across the object once – and from all sides. The resulting individual images then “only” have to be assembled using appropriate software. We have deliberately put the “only” in brackets, because in fact this part is the most difficult task. This is because overlapping points on the individual images have to be found manually and then communicated to the program. Satisfactory results can only be achieved here with sufficient practice!
How does working with coded light work?
Working with coded light is also one of the light-sectioning techniques, but in an optimized form. Instead of a line laser, you need a beamer in addition to the digital camera – which naturally drives up the price of the equipment. When working with coded light, each point on the surface of an object corresponds to a coordinate. The projector now casts different light-dark patterns in stripes onto the object, while the camera simultaneously registers whether the currently observed point is in light or dark. Binary code is used for processing: Light corresponds to a 1, Dark to a 0. For the temporal sequence of the pattern, the number of each strip of the projector is encoded in Grey Code, where the number corresponds to a coordinate in the projector. Once the coordinates of the image and the projector are generated, the depth of each point on the object can be calculated. In order to create a complete, three-dimensional image of the object, it must be scanned from several sides, as with the other processes, and the scans must then be merged using software.
How does the 3D scan work with Structured Light and the Kinect?
The Kinect camera from Microsoft was originally developed for the motion control of the Xbox 360. However, the basic technical structure of the Kinect also makes the device interesting for applications beyond computer games – including 3D scanning. Kinect works with an infrared laser that “draws” small, bright and dark dots (speckles) in space. Several points are always combined to form a pattern. By comparing stored reference patterns with the recorded pattern, the Kinect then calculates a depth image. Since the Kinect can capture the complete spatial depth of all points in an image with a single shot, the process is exceedingly fast.
* Stand: 2023-07-05 / Bilder: Amazon API
The scan setup
Each of the above four methods requires its own typical setup of camera and peripherals. We present the basic structure in short form.
Scan setup for photogrammetry
Camera in hand and get started – with photogrammetry, a thorough setup is not necessary. More important here is the way an object is photographed. Best results are achieved when the object is circled laterally and continuously photographed. One round trip should generate 30 images or more. If possible, the object should be photographed once from the side, once with a slight overhead view and once from the frog's perspective. It is also possible to record the object as a video and then export it as a series of images.
Scan setup for the light section method
The scan setup for the light section method can be considerably more complex. For panning the laser over the surface, manual guidance of the laser and the use of free, self-printed calibration aids are sufficient – but if you want to deal with the process more professionally and want correspondingly high-quality results, you need a camera with at least 2 megapixels, a base plate and calibration surfaces in addition to a line laser. The scan becomes even more professional and, above all, faster if the panning process is automated by means of a motor and control electronics.
Scan setup when working with coded light
In the coded light method, a line projector (beamer) and a camera are set up at a predefined distance, the triangulation base, away from the object to be recorded. This distance must remain stable, even if the object is photographed from several sides.
Scan setup during 3D scan with Kinect
Since the Kinect completely captures the spatial depth of an object immediately with a single shot from the infrared camera, there is nothing special to consider during setup except for a minimum distance between the Kinect and the object.
The advantages and disadvantages of each method
Each of the 3D scanning methods mentioned here has its own advantages – and of course also disadvantages. We will therefore deal with both sides of the coin in bullet points.
Photogrammetry – Advantages
- Comparatively cheap, especially for first attempts
- No CAD software or corresponding knowledge required.
- No complicated setup necessary: a standard digital camera is all that is needed.
- Works well when shooting outdoors under cloudy skies.
Photogrammetry – Disadvantages
- Does not work in direct sunlight or backlight.
- Does not work with glossy or reflective object surfaces.
- Service providers need up to 24 hours to provide the 3D data.
- Apart from the actual object, there must be no other motifs on the images, as this would overtax the algorithm of the analysis software.
- For really good results, a high-quality camera and an elaborate scan setup are required.
Light sectioning – Advantages
- Works without web service
- Initial tests can be carried out very inexpensively with material from the hardware store.
- Necessary software available as freeware.
- Starter kits for semi-professional use of the technology are also available at reasonable prices.
Light section method – disadvantages
- System must be calibrated before each scan, which increases the effort.
- The individual scans must be manually combined in the software to form an overall image.
- Unsuitable when there is too much ambient light.
- Works poorly when scanning very dark objects.
- Accuracy of the process always depends on the thickness of the laser beam.
- Capture of colors not possible.
Coded light – advantages
- Easier handling than light sectioning.
- Scan speed is significantly higher.
- Very precise results.
Coded light – disadvantages
- Not suitable for 3D scanning of small objects.
- Unsuitable for extraneous light.
- Very high entry price, even for starter kits.
Kinect – Advantages
- Very fast
- One image is enough to capture the spatial depth of an object.
- With a purchase price of only around 100 euros, it is very affordable.
- Free software available for home users.
Kinect – Disadvantages
- Commercial applications only allowed through Kinect for Windows. This camera can do the same as the Xbox version, but costs significantly more.
- Calculation of 3D data immensely memory hungry.
- Allergic to reflections and transparent object details.
- Mains voltage is required for operation, which makes outdoor use difficult.
Which method is best suited for what?
Just take a few photos on the way and a short time later hold the photographed object in your hands as a 3D print: Photogrammetry makes it possible. Without a doubt, the process is a beginner-friendly way to turn objects of all kinds into printable data. For those who do not want to deal with technical details, are not interested in CAD programs and simply want to receive 3D prints, there is currently no way around photogrammetry.
Light section method
To deal more deeply with the matter, not to spend a lot of money on basic technology and to achieve a useful result in the end – the light-sectioning method is much less beginner-friendly than photogrammetry, but on the other hand you are not dependent on external service providers. Especially the compositing of the single images in the software can be quite annoying, here you definitely need patience and a lot of practice to get usable results.
A high entry-level price for the necessary hardware should not be a deterrent if you want to work with the coded light method. On the other hand, this form of 3D scanning is fast, reliable and accurate. Those who mainly want to digitize larger objects and are not dependent on the model surfaces being reproduced in true color can definitely delve deeper into the technology. However – and this is something this process has in common with the light sectioning technique – it does not rely on a promising 3D scanning process. For this, the effort is simply too high compared to photogrammetry or the Kinect.
For just 100 Euros, the Kinect is a powerful 3D scanner that convinces with its speed and flexibility. An active open source community ensures that suitable software for processing and preparing the scanned data is made available free of charge and, above all, is constantly being further developed.
But beware: If the 3D scan data is to be used commercially, the Windows version of the Kinect must be purchased, which is significantly more expensive than the Xbox version. A certain affinity and experience in CAD, 3D printing and image processing is also necessary for the Kinect.
What to look for when buying 3D scanners?
The most important thing in a 3D scanner is, of course, the camera. The more detailed the images, the easier it is to prepare the data in the software. The good news is that even simple, modern digital cameras serve their purpose here. If you do not want to rely on external service providers to process the data, but want to work with the appropriate programs yourself, you will need peripheral hardware in addition to the camera.
However, lasers, software and computing power are quite expensive, especially if you rely on semi-professional or even professional equipment. A good, affordable and above all extremely flexible way to go is to rely on Microsoft's Kinect. Those who like to play on the console probably already have the 3D scanner from Microsoft in their house anyway. Extensive open source software helps keep costs down.
Special case Touch Probe Scanning
The four methods mentioned so far have one thing in common: they operate without contact by means of optical devices. 3D Touch Probe Scanning takes a different approach. Here, a probe is used to scan an object, making it available as a digital model. The basis for the procedure is, in addition to an appropriate probe (for example, from the Renishaw company), a CNC-controlled machine with at least three controlled axes. The object to be scanned is clamped on the table of the machine and the following parameters are adjusted in the CNC control:
- Diameter of the measuring ball on the probe
- Two defined points in the coordinate system of the machine
- Direction of the scan (X or Y)
- Zero point
Once the entries have been stored in the controller, the scan process begins. The measuring ball is brought up to the object. By adjusting the Z-axis (vertical axis), the measuring ball is repeatedly moved to the surface of the object, the coordinates are stored, then the X-axis or Y-axis is adjusted, a new point is measured, until at the end a digital image of the object is available in the form of coordinates. The corresponding three-dimensional image is then created from these coordinates.
In the professional environment, touch probe scanning is inexpensive and economical to implement – especially if the corresponding technology is already available. However, home users can also assemble appropriate machines through kits. Suitable probes of sufficient quality are available from around 300 euros.
Shoot a few photos on vacation and find the photographed sights as a small model in your mailbox when you return – what was fiction not so long ago is now feasible for everyone. 3D scanning is easier than ever and, with an understanding of technology and software, can be done at home. Whereby “simple” must be clearly put into relation here, because the 3D scan is not yet really “simple”, i.e. to be carried out without great knowledge and corresponding hardware and software. It will still take a year or two before it can be used on a mass scale. Currently, 3D scanning is clearly still a matter for specialists and enthusiasts. Each of the available methods has its own advantages and disadvantages. But the fact is that the future belongs to 3D scanning and 3D printing of all kinds of objects. Because what is more beautiful than a souvenir photo from your vacation? That's right – a souvenir object to touch.