3D scanning has gained immense popularity over recent years as it continues to revolutionize various industries. Many different scanning technologies are now available, each with their pros and cons, which can be daunting to new users just getting to know the field.
In this article, we aim to provide you with a basic overview of the main types of scanning technologies, how they work and which one fits best depending on your needs.
The basic question always is, “What will the 3D model be used for?” For example, do you need a perfect geometry for industrial engineering or the most natural look possible for your e-commerce website? Scanning people, on the other hand, places different demands on technology than scanning objects, buildings or even landscapes.
So let’s have a closer look at different approaches to 3D scanning:
If your main goal is a perfect shape with accurate detail and you don’t care about the actual appearance in terms of color or texture, you might prefer one of the following technologies:
• Computed Tomography (CT)
• Laser Scanning
• Structured Light
• Contact Scanning
If you are aiming for the most natural look of your 3D model, be it because you want a product presentation in e-commerce or you are in need of a true digital twin, then photogrammetry, or a combination of photogrammetry and other approaches, will probably work best.
• Photography
• Combining different technologies
So let’s take a deeper look at the different technologies:
Computed Tomography, also known as Computerized Tomography, CT or CAT scan, takes multiple X-ray scans of a subject and superimposes them to create a 3D model. This means that at first a series of X-ray images are taken in several layers – as if you were cutting an object into wafer-thin slices and then reassembling the individual slices to form an overall image.
CT scans not only provide very detailed geometries, but in contrast to all other methods which can only capture the external shape of an object, they can also display the interior of an object. This is particularly advantageous for machine components or complex structures such as clocks. The only disadvantage is that the resulting 3D models are completely untextured, i.e. colorless.
Texture and color are completely irrelevant for use in industrial design or construction. However, if you wanted to optimize a model for a sales presentation, for example, you may combine it with another scanning method, such as photogrammetry, which provides a perfect texture to create a more realistic look.
Laser scanning technologies involve the use of a laser beam and a receiver. There are two main types of laser scanning techniques: Time of Flight and Triangulation.
Time of Flight scanners, also known as Laser Pulse technology, work by emitting a laser beam towards the subject and measuring the time taken for the reflected light to be received. The distance of the subject from the scanner can then be calculated by knowing the speed of light. This process must be repeated from different angles in order to obtain enough data to create a 3D representation of the subject. Often the scanner head is capable of rotating and contains mirrors to change the angle of the beams, making it possible to scan a full 360° view of an environment.
The main advantage of this technique is the huge ranges it is capable of scanning, by the order of kilometers. This allows very large objects, buildings or terrain to be scanned using just one device. However, in comparison to other scanning technologies, Time of Flight scanners are not as accurate, and struggle to scan smaller, intricate objects. They can also be relatively slow, sometimes taking minutes to scan something, meaning they cannot scan any subject in motion.
While very similar to the Time of Flight technique, triangulation measures the level of distortion or angle of the reflected light in order to calculate the distance of the subject from the scanner. It gets its name from the triangle formed between the laser emitter, the laser beam and the receiver. The distance and size of the subject being scanned can be calculated by knowing the distance between the laser emitter and receiver and the angle between the laser beam and the scanner.
Just like CT scanning, Triangulation has a very high level of accuracy, capable of picking up extremely fine details of a subject’s geometry. However, unlike CT scanning, it can only capture the exterior form of an object. Triangulation also struggles to scan highly reflective or shiny objects, as the surface can distort the reflection of the laser beams, resulting in inaccurate data being collected.
Scanning with structured light involves projecting a grid or pattern onto the subject and analyzing the deformation of the pattern to obtain the 3D data. Imagine, for example, rows of straight parallel lines of light (or even a grid pattern)being projected onto a balloon. The lines (or grid) will follow the shape of the balloon and show a distortion depending on the amount of curvature. Analyzing the images taken by one (or more) cameras, the software then detects the projected pattern and uses the various curvature factors to calculate the shape of the object.
This technique is very fast and able to scan entire objects at once instead of different points, so one can scan subjects in motion. It also comes with a very high resolution in terms of geometry. Apart from the lack of texture, this technology has its limits not only on shiny or reflective surfaces, but also on dark surfaces.
Contact scanning machines scan objects using a highly sensitive probe that touches the object and records its X, Y and Z coordinates in 3D space. This process takes very long, as the probe must cover the entire object in order to obtain enough information to create a 3D model, however it is very accurate.
The advantages of this technology are its high precision and the fact that it is not affected by shiny, reflective or transparent surfaces, as it scans physically, not optically. This is also the direct disadvantage, as this type of scanner requires physical contact with the object and can therefore damage sensitive surfaces. Contact scanners quickly reach their limits even on very soft or porous surfaces. In addition, a scan takes much longer compared to faster optical methods.
Finally, photogrammetry is the 3D scanning method of choice if, in addition to accurate geometry, you also desire the lifelike appearance of scanned objects. As the name suggests, photogrammetry works on the basis of photos. Multiple images of an object are captured from different angles and analyzed for common reference points by specialized software. The software then uses these points to create a “mesh” of the scanned subject, which is the basis of a 3D model. The rule of thumb is: the higher volume and quality of images, the better the scan. Investing in good cameras and lenses quickly pays off.
This technique is very versatile as, unlike other scanning technologies, the only physical equipment needed is one or more cameras. Photogrammetry can be used both at close range for smaller subjects or long range using wide-lens cameras or aerial images to model a landscape or building. As long as high resolution cameras are used in the process, this technology is extremely accurate, creating very precise digital models.
Photogrammetry is a “non-contact” scanning process, meaning the subject being scanned does not come into physical contact with the technology and is rather scanned from a distance. There is no risk of damage or danger to the subject, which makes this technique perfect for scanning fragile or soft items or even human beings.
There are three main areas of application for photogrammetry:
1. Landscape and Architecture (where the entire discipline has its origins)
2. 3D Scanning of Objects
3. 3D Scanning of People
While photogrammetry in landscape and architecture is a field of its own, the main difference in scanning objects and people is the general setup in terms of numbers of cameras.
Objects are easier to scan because they typically don’t move. In this respect, theoretically, you can position them on a turntable and take pictures with only a single camera, rotating the turntable and varying the height of the camera position after each complete rotation until you cover all possible perspectives, from a worm’s eye view to a bird’s eye view.
For high-quality scans, up to 300 individual images are required. While a single high-quality camera and lens is sufficient for private use, you save valuable time and rework for professional use by using multiple cameras. Depending on customer requirements, botspot’s 3D object scanners work with between 3 and 7 cameras, with fully automated processes.
3D full-body scans of people are much more demanding, as even micro-movements such as eyes blinking, breathing in and out or maintaining balance distort the scan result. Therefore, the recordings for a 3D whole-body scan cannot be taken successively, but rather simultaneously, which creates the need for a significantly higher number of cameras.
Depending on the intended application and quality requirements of the 3D models, we achieve the best results for our customers by equipping our BOTSCAN NEO with between 70 cameras (e.g. for pure body measurements) and 160 cameras (e.g. as preparation for the fashion industry or CGI / VFX applications in the film and gaming industry).
We meet the typical challenges of photogrammetry, such as very light or very dark skin tones and fabrics and very shiny or reflective materials (like black patent leather shoes), with the additional use of powerful laser projections. These produce a “digital spray”, which allows the software to recognize more reference points and so creates a much more geometrically accurate model.
Please feel free to contact us for further information! Our Innovation Lab is always happy to help and find solutions for even the most demanding tasks.