Build your own 3d scanner optical triangulation for beginners

Advantages and disadvantages of

scanners Both time-of-flight and triangulation scanners have their own strengths and weaknesses, which determines their choice for each specific situation. The advantage of time-of-flight devices is that they are optimally suited for operation over very long distances up to several kilometers. They are ideal for scanning buildings or geographic features. At the same time, their disadvantages include measurement accuracy. After all, the speed of light is quite high, so when calculating the time it takes for the beam to overcome the distance to and from the object, some flaws (up to 1 mm) are possible. And this makes the scan results approximate.

As for triangulation rangefinders, the situation is exactly the opposite. Their range is only a few meters, but the accuracy is relatively high. Such devices can measure distance with an accuracy of tens of micrometers.

The study of the edge of an object negatively affects the accuracy of the TOF scanners. The laser pulse is sent one, and is reflected from two places at once. The coordinates are calculated based on the position of the scanner itself, and the average value of the two reflections of the laser beam is taken. This causes the point to be defined in the wrong place. When using scanners with high resolution, the chances that the laser beam hits the exact edge of the object increase, but noise will appear behind the edge, which will negatively affect the scan results. Scanners with a small beam can solve the edge scanning problem, but they have limited range, so the beam width will exceed the distance. There is also special software that allows the scanner to perceive only the first reflection of the beam, while ignoring the second.

At 10,000 dots per second, low resolution scanners can do the job within seconds. But for scanners with high resolution, you need to do several million operations, which will take minutes. It should be borne in mind that the data may be distorted if the object or the scanner moves. So, each point is fixed at a certain point in time in a certain place. If the object or scanner moves in space, then the scan results will be false. That's why it's so important to mount both the object and the scanner on a fixed platform and keep the possibility of vibration to a minimum. Therefore, scanning objects in motion is practically impossible. Recently, however, there has been active research on how to compensate for the effect of vibration on data corruption.

It is also worth considering that when scanning in one position for a long time, a slight movement of the scanner may occur due to temperature changes. If the scanner is mounted on a tripod and one side of the scanner is exposed to strong sunlight, then the tripod will expand and the scan data will gradually distort from one side to the other. However, some laser scanners have built-in compensators that counteract any movement of the scanner during operation.

Conoscopic holography

In the conoscopic system, a laser beam is projected onto the surface of an object, after which the beam is reflected along the same path, but through a conoscopic crystal, and is projected onto a CCD (charge-coupled device). The result is a diffraction pattern from which frequency analysis can be used to determine the distance to the surface of an object. The main advantage of conoscopic holography is that only one beam path is needed to measure the distance, which makes it possible to determine, for example, the depth of a small hole.

Handheld laser scanners

Handheld laser scanners create a 3D image using the triangulation principle described above. A laser beam or stripe is projected onto an object from a hand-held emitter, and a sensor (often a CCD or position-sensitive detector) measures the distance to the surface of the object. The data is collected relative to the internal coordinate system and therefore, to obtain results, if the scanner is in motion, the position of the device must be accurately determined. This can be done using basic features on the scanned surface (adhesive reflective elements or natural features) or using the external tracking method. The latter method often takes the form of a laser tracker (providing a position sensor) with a built-in camera (to determine the orientation of the scanner). You can also use photogrammetry, provided by 3 cameras, which gives the scanner six degrees of freedom (the ability to make geometric movements in three-dimensional space). Both techniques typically use infrared LEDs connected to the scanner. They are observed by cameras through filters that ensure the stability of ambient lighting (reflecting light from different surfaces).

Scan data is collected by a computer and recorded as points in 3D space, which after processing are converted into a triangulated grid. The computer-aided design system then creates a model using a non-uniform rational B-spline, NURBS (a special mathematical form for creating curves and surfaces). Handheld laser scanners can combine this data with passive visible light sensors that capture surface texture and color to create or reverse engineer a complete 3D Models .

Structured light

Structured light 3D scanners are a projection of a light grid directly onto an object, the deformation of this pattern is a model of the object being scanned. The grid is projected onto the object using a liquid crystal projector or other constant light source. A camera positioned just to the side of the projector captures the shape of the network and calculates the distance to each point in the field of view.
Structured light scanning is still an active field of research, with quite a few research papers devoted to it each year. Ideal maps are also recognized as useful as structured light patterns that can solve matching problems and allow errors to be corrected as well as detected.

The advantage of the Structured Light 3D Scanners is their speed and accuracy. Instead of scanning one point at a time, structured scanners scan several points at the same time or the entire field of view at once. Scanning the entire field of view takes a fraction of a second, and the generated profiles are more accurate than laser triangulations. This completely solves the problem of data corruption caused by motion. In addition, some existing systems are capable of scanning even moving objects in real time. For example, the VisionMaster, a 3D scanning system, has a 5-megapixel camera, so each frame contains 5 million dots.

Real-time scanners use digital edge projection and a phase-shifting technique (one of the techniques for applying structured light) to capture, reconstruct, and create a high-density computer model of dynamically changing objects (such as facial expressions) at 40 frames per second. A new type of scanner has recently been created. Various models can be used in this system. The frame rate for capturing and processing data reaches 120 frames per second. This scanner can also process individual surfaces. For example, 2 moving hands. Using the binary defocusing method, the shooting speed can reach hundreds or even thousands of frames per second.

Modulated light

When using the 3D scanners based on modulated light, the light beam directed at the object is constantly changing. Often the change of light passes along a sinusoid. The camera captures the reflected light and determines the distance to the object, taking into account the path that the light beam has traveled. Modulated light allows the scanner to ignore light from sources other than the laser, thus avoiding interference.

Volumetric Techniques

Medical

Computed tomography (CT) is a special medical imaging technique that creates a series of three-dimensional x-ray images of an object's interior space. Magnetic resonance imaging works on a similar principle - another imaging technique in medicine, which is distinguished by a more contrast image of the soft tissues of the body than CT. Therefore, MRI is used to scan the brain, the musculoskeletal system, the cardiovascular system, and to search for oncology. These techniques produce volumetric voxel models that can be rendered, modified, and transformed into a traditional 3D surface using isosurface extraction algorithms.

Production

Although MRI, CT or microtomography are more widely used in medicine, they are also actively used in other areas to obtain a digital model of an object and its environment. This is important, for example, for non-destructive testing of materials, reverse engineering or the study of biological and paleontological samples.

Non-contact passive scanners

Passive scanners do not emit light, instead they use reflected light from the surrounding area. Most scanners of this type are designed to detect visible light, which is the most accessible form of ambient radiation. Other types of radiation, such as infrared, may also be involved. Passive scanning methods are relatively cheap, because in most cases they do not need special equipment, a conventional digital camera is enough.
Stereoscopic systems involve the use of 2 video cameras located in different places, but in the same direction. By analyzing the differences in the images of each camera, you can determine the distance to each point in the image. This method is similar in principle to human stereoscopic vision.

Photometric systems typically use a single camera that captures multiple frames in all lighting conditions. These methods attempt to transform the object model in order to reconstruct the surface for each pixel.

Silhouette techniques use contours from successive photographs of a three-dimensional object against a contrasting background. These silhouettes are extruded and transformed to get the visible skin of the object. However, this method does not allow you to scan the recesses in the object (for example, the inner cavity of the bowl).

There are other methods that are based on the fact that the user himself discovers and identifies some features and shapes of the object, based on many different images of the object, which allow you to create an approximate model of this object. Such methods can be used to quickly create a three-dimensional model of objects of simple shapes, for example, a building. You can do this using one of the software applications: D-Sculptor, iModeller, Autodesk ImageModeler or PhotoModeler.

This 3D scan is based on the principles of photogrammetry. In addition, this technique is in some ways similar to panoramic photography, except that the photographs of the object are taken in three-dimensional space. Thus, it is possible to copy the object itself, rather than taking a series of photos from one point in three-dimensional space, which would lead to the reconstruction of the object's environment.

Reconstruction

From point clouds

Point clouds generated by 3D Scanners can be directly used for measurements or visualizations in architecture and engineering.
However, most applications use non-homogeneous rational B-spline, NURBS, or editable CAD models (also known as solid models) instead of polygonal 3D models.

• Polygon mesh models: In polygonal representation shapes curved surfaces consist of many small flat surfaces with edges (a striking example is a ball in discotheques). Polygonal models are very in demand for visualization in the field of CAM - an automated system for technological preparation of production (for example, mechanical processing). At the same time, such models are quite « heavy" (accommodate a large amount of data) and are quite difficult to edit in this format. Reconstruction into a polygonal model involves searching and combining neighboring points with straight lines until a continuous surface is formed. For this, you can use a number of paid and free programs (MeshLab, Kubit PointCloud for Au toCAD, 3D JRC Reconstructor, ImageModel, PolyWorks, Rapidform, Geomagic, Imageware, Rhino 3D, etc.).
• Surface models: This method represents the next level of sophistication in the field of modeling. It applies a set of curved surfaces that give your object its shape. It can be NURBS, T-Spline or other curved objects from the topology. Using NURBS converts, for example, a sphere to its mathematical equivalent. Some applications require manual processing of the model, but more advanced programs also offer automatic mode. This option is not only easier to use, but also provides the ability to modify the model when exporting to a computer-aided design system (CAD). Surface models are editable, but only in a sculptural way. Organic and artistic forms lend themselves well to modeling. Surface modeling is available in Rapidform, Geomagic, Rhino 3D, Maya, T Splines.
• 3D CAD Models: From an engineering and manufacturing perspective, this type of modeling is a full digitized form of a parametric CAD model. After all, CAD is the industry's common "language" for describing, editing, and preserving the shape of an enterprise's assets. For example, in CAD, a sphere can be described by parametric functions that are easy to edit by changing their value (say, radius or center point).

These CAD models don't just describe the shell or shape of an object, but they also enable design intent (ie, critical features and their relationship to other features). An example of design intent that is not expressed in form would be the ribbed bolts of a brake drum, which should be concentric with the hole in the center of the drum. This nuance determines the sequence and method of creating a CAD model, so the engineer, taking into account these features, will develop bolts tied not to the outer diameter, but, on the contrary, to the center. Thus, to create such a CAD model, you need to correlate the shape of the object with the design intent.

There are several approaches to get a parametric CAD model. Some involve only exporting a NURBS surface, leaving the CAD engineer to complete the modeling (Geomagic, Imageware, Rhino 3D). Others use the scan data to create an editable and verifiable function model that can be fully imported into CAD with an intact fully functional tree, providing a complete fusion of shape and design intent of the CAD model (Geomagic, Rapidform). However, other CAD applications are powerful enough to manipulate a limited number of points or polygonal models in a CAD environment (CATIA, AutoCAD, Revit).

From a set of 2D slices

3D reconstruction of the brain or eyeballs based on CT results is performed using DICOM images. Their peculiarity is that the areas on which air is displayed, or bones with a high density are made transparent, and the sections are superimposed in a free alignment interval. The outer ring of biomaterial surrounding the brain is made up of the soft tissues of the skin and muscles on the outside of the skull. All sections are made on a black background. Since they are simple 2D images, when added one-to-one when viewed, the borders of each slice disappear due to their zero thickness. Each DICOM image is a slice about 5 mm thick.

CT, industrial CT, MRI or microCT scanners do not create a point cloud, but 2D slices (referred to as a “tomogram”) that are superimposed on each other, resulting in a kind of 3D model. There are several ways to do this scan, depending on the desired result:

• Volume rendering: Different parts of an object usually have different thresholds and grayscale densities. Based on this, a three-dimensional model can be freely designed and displayed on the screen. Several models can be made from different thresholds, allowing different colors to represent a specific part of an object. Volumetric rendering is most often used to render a scanned object.
• Image segmentation: When different structures have similar threshold or midtone values, it may not be possible to separate them simply by changing volume rendering parameters. The solution to the problem will be segmentation - a manual or automatic procedure that will remove unnecessary structures from the image. Special programs that support image segmentation allow you to export segmented structures to CAD or STL format, which will allow you to continue working with them.
• Meshing based on image analysis: When 3D image data (CFD and FEA) is used for computer analysis, simple data segmentation and meshing from a CAD file can be quite time consuming. In addition, some typical image data may not be inherently suitable for a complex topology. The solution lies in image analysis meshing, which is an automated process for generating an accurate and realistic geometric description of the scanned data.

Application

Material Handling and Manufacturing

3D Laser Scanning describes a general way to measure or scan a surface using laser technology. It is used in several areas at once, differing mainly in the power of the lasers that are used and the results of the scan itself. Low laser power is needed when the scanned surface should not be influenced, for example, if it only needs to be digitized. Confocal or 3D laser scanning are methods that provide information about the scanned surface. Another low power application involves a projection system that uses structured light. It is applied to solar panel plane metrology involving voltage calculation with a throughput of more than 2,000 plates per hour.

The laser power used for laser scanning of industrial equipment is 1W. The power level is typically 200mW or less.

Construction industry

• Robot control: Laser scanner acts as the eye of the robot
• Executive drawings of bridges, industrial plants, monuments
• Documenting Historic Sites
• Site modeling and layout
• Quality control
• Measurement of works
• Reconstruction of highways
• Marking an existing shape/state to identify structural changes after extreme events - earthquake, ship or truck impact, fire.
• Creation of GIS (Geographic Information System), maps and geomatics
• Scanning of subsurface in mines and karst voids
• Court records

Benefits of 3D scanning

Creating a 3D model by scanning has the following benefits:

• Increases efficiency when working with complex parts and shapes
• Encourages product design when needed to add a part created by someone else.
• If CAD models become outdated, 3D scanning will provide an updated version
• Replaces missing or missing parts of

Entertainment

3D scanners are widely used in the entertainment industry to create 3D digital models in film and video games. If the model being created has a counterpart in the real world, then scanning will allow you to create a three-dimensional model much faster than developing the same model through 3D modeling. Quite often, artists first sculpt a physical model, which is then scanned to get a digital equivalent, instead of creating such a model on a computer.

Reverse engineering

Reverse engineering of mechanical components requires a very precise digital model of the objects to be recreated. This is a good alternative to converting many points of a digital model to a polygon mesh, using a set of NURBS flat and curved surfaces, or, ideally for mechanical components, creating a 3D CAD model. A 3D scanner can be used to digitize objects that freely change shape. As well as the prismatic configuration, for which a coordinate measuring machine is usually used. This will allow you to determine the simple dimensions of the prismatic model. This data is further processed by special programs for reverse engineering.

3D printing

3D scanners are also actively used in the field of 3D printing, as they allow you to create fairly accurate 3D models of various objects and surfaces in a short time, suitable for further refinement and printing. In this area, both contact and non-contact scanning methods are used, both methods have certain advantages.

Cultural heritage

An example of copying a real object through 3D scanning and 3D printing. There are many research projects that have been carried out using the scanning of historical sites and artifacts to document and analyze them. The combined use of 3D scanning and 3D printing makes it possible to replicate real objects without the use of a traditional plaster cast, which in many cases can damage a valuable or delicate cultural heritage artifact. The sculpture of the figure on the left was digitized using a 3D scanner, and the resulting data was converted in the MeshLab program. The resulting digital 3D model was printed using a rapid prototyping machine that allows you to create a real copy of the original object.

Michelangelo

There are many research projects that have been carried out using scanning of historical sites and artifacts to document and analyze them.

In 1999, 2 different research groups started scanning Michelangelo's statues. Stanford University, along with a team led by Mark Levoy, used a conventional laser triangulation scanner built by Cyberware specifically to scan Michelangelo's statues in Florence. In particular, the famous David, "Slaves" and 4 more statues from the Medici chapel. Scanning is performed with a dot density of 0.25 mm, sufficient to see the traces of Michelangelo's chisel. Such a detailed scan involves obtaining a huge amount of data (about 32 gigabytes). It took about 5 months to process them.

Around the same time, a research group from IBM was working, led by H. Raschmeyer and F. Bernardini. They were tasked with scanning the Florentine Pieta sculpture to obtain both geometric data and color information. The digital model obtained from a Stanford University scan was fully used in 2004 to further restore the statue.

Medical applications CAD/CAM

3D scanners are widely used in orthopedics and dentistry to create a 3D patient shape. Gradually, they replace the outdated gypsum technology. CAD/CAM software is used to create prostheses and implants.
Many dentistry uses CAD/CAM as well as 3D scanners to capture the 3D surface of a dentifrice (in vivo or in vitro) in order to create a digital model using CAD or CAM techniques (e.g. , for a CNC milling machine (computer numerical control), as well as a 3D printer). Such systems are designed to facilitate the process of 3D scanning of the drug in vivo with its further modeling (for example, for a crown, filling or inlay).

Quality assurance and industrial metrology

The digitization of real world objects is of great importance in various fields of application. 3D scanning is very actively used in industry to ensure product quality, for example, to measure geometric accuracy. Predominantly all industrial processes such as assembly are quite complex, they are also highly automated and are usually based on CAD (computer-aided design data). The problem is that the same degree of automation is required for quality assurance. A striking example is the automated assembly of modern cars, because they consist of many parts that must match exactly with each other.
Optimum performance levels are guaranteed by quality assurance systems. Geometrical metal parts need special checking, because they must be of the correct size, fit together to ensure reliable operation.
In highly automated processes, the results of geometric measurements are transferred to machines that produce the corresponding objects. Due to friction and other mechanical processes, the digital model may differ slightly from the real object. In order to automatically capture and evaluate these deviations, the manufactured parts must be rescanned. For this, 3D scanners are used, which create a reference model with which the received data are compared.
The process of comparing 3D data and CAD model is called CAD comparison, and can be a useful method for determining mold and machine wear, final assembly accuracy, tear analysis, and the volumetric surface of a disassembled part.

Learn more

• 
  Recycle pla 3d printer
• 
  3D printers online india
• 
  Best 3d printer settings for pla
• 
  Solid 3d printer
• 
  How to use cad for 3d printing
• 
  3D body scanner fashion
• 
  Best gadgets to 3d print
• Joel 3d printing nerd
• 
  Resin dlp 3d printer
• 
  Scanner 3d professionnel
• 
  3D printer functions