This seminar title is discussed about 3D Scanning Technology and 3D Modeling. At the end, I will also choose the most suitable technique to use in my Final Year Project.
Nowadays, 3D technology was approaching us step by step. However, most of the people still do not know that, what is 3D and how they do it? Well, as a good example what kind of 3D that appear in our real life and is so famous recently is 3D movie. I think that most of the youngster or movie chaser watch 3D movie before for sure. For example Avatar 3D, Ice Age 2 3D, Resident Evil: After Life 3D and many others.
However, most of the people do not understand that how does 3D works, and how they create 3D model. In my seminar I would like to introduce different types of 3D modeling, can comparison between them. But for sure, different type of 3D modeling technique are uses in different area. So, I will introduce most of them in this seminar.
Almost all 3D models can be divided into two categories.
Solid - These models define the volume of the object they represent. These are more realistic, but more difficult to build. Solid models are mostly used for nonvisual simulations such as medical and engineering simulations, for CAD and specialized visual applications such as ray tracing and constructive solid geometry
Shell/boundary - These models represent the surface. These are easier to work with than solid models. Almost all visual models used in games and film are shell models.
Technology was improved and changes rapidly day by day. A new technology call 3D Digitizing which will use special equipment such as Terrestrial Laser Scanner, 3D Scanner and others was invented. These machines can provides highly accurate, three-dimensional images enabling designers to experience and work directly with real-world conditions by viewing and manipulating rich point-clouds in computer-aided design software. In the chapter below we will go deeper about this technology which was widely used in those fields that involved 3D.
Even many new and advanced type of modeling technique was found, many designer, construction, and others will still use the basic modeling technique, such as Polygonal modeling, NURBS, Splines & Patches modeling, Primitives modeling and Sculpt modeling. In the chapter below, we will explore more about these techniques that are still widely used for those designers and the reason.
Main Body
3D Digitizing
3D Digitizing can be defining as transform the real object (such as a building) to a digitized format. It was like a graphic input system that will record coordinate for x, y and z then form a 3D model.
The main device for 3D Digitizing is 3D Scanner which was so popular and most common device that use for 3D Digitizing. 3D scanner is a device that will scan and analyze the object or image. The data that gain will be construct a possible same 3D model which the scanned object or image.
There were 3 types of 3D Scanner:
- Contact
- Non-contact Active
- Non-contact Passive.
Contact 3D Scanner
As the definition of "contact" which means the 3D Scanner that have physical contact with the object that need to be scan. A CMM (coordinate measuring machine) is a most common example for contact 3D scanner. CMM are widely use in industrial due to its precision.
CMM is composed of three axes, an X, Y and Z. These axes are orthogonal to each other in a typical three dimensional coordinate system. Each axis has a scale system that indicates the location of that axis. The machine will read the input from the touch probe, as directed by the operator or programmer. The machine then uses the X, Y, Z coordinates of each of these points to determine size and position. Typical precision of a coordinate measuring machine is measured in Microns.
The disadvantage of CMMs is it requires having physical contact with the object being scanned. Thus, there will be a possibility to make damage or modify the object during scanning process. It was a very serious problem when need to scan with those valuable antique or those artifacts. The other disadvantage of CMMs was time consuming. It may take a long time to finish scan a large or refinement object.
Figure 1 - CMM to scan small object Figure 2 CMM to scan big object
Critical Evaluation on Contact 3D Scanner
Basically, the biggest problem for Contact 3D Scanner is it requires having physical contact with the object during scanning process, and sometimes it might damage the object wanted to scan. Furthermore, if the object was an artifact or antique, once have damage there is no other replacement or will be very costly.
Another problem for Contact 3D scanner, CMM as mention above is the scanning speed. CMM need to have physical contact with the object and most of the object has different shape and different surface. So scanning the object using CMM was a very time consuming task.
However, thanks to the disadvantages of CMM which is having physical touch with object, the advantage of CMM is it can produce a very precise and accurate result.
So, as my final evaluation on Contact 3D Scanner, it is a very precise and accurate 3D scanner yet, I suggest that avoid from scan on those artifact and valuable object. Nowadays, speed also is a key point of success, the speed for Contact 3D Scanner was a biggest stone for it. So, if can eliminate the speed problem of Contact 3D Scanner, it will be a best scanning machine ever.
Non-contact Active Scanner
Non-contact active scanners transmit some kind of radiation or light such as light, ultrasound or x-ray to detect its reflection in order to probe an object or environment. Non-contact Active Scanner does not have any physical touch with the object to scan.
There is few type of Non-Contact Active Scanner, such as
Time-of-flight
Triangulation
Structured Light
Each of them is use in different field, but their function was same. The example I list was only a few, it might be a lot of other Non-contact Active Scanner exist.
Time-of-Flight
Time-of-flight 3D laser scanner uses laser light to scan the object. Laser rangefinder is the main part of Time-of-Flight Scanner. The laser rangefinder works like echo, but in light not in sound. It will emit a light to an object that want to scan, and finds the distance of a surface by calculate the time needed for the light to travel in a round-trip.
Laser rangefinder will only emit one point once. So to have a complete scan an object, it requires to scan many times by changing the laser rangefinder direction. If use Time-of-Flight scanner, this is the biggest stone that can't be avoid.
For those who were hard to understand how Time-of-Flight works, I strongly recommend that people go search a song call House of Card by Radiohead and watch their MV. This MV was using Time-of-Flight technique to create. Hopefully I can help people to more understanding how it works.
Figure 3- This is the design for Time-of-Flight device
Triangulation
The triangulation 3D laser scanner uses laser light to detect the environment. Triangulation will transmit a laser that will shines on the subject and a camera will use to record the exactly location for the laser dot. The laser dot, the camera and the laser emitter form a triangle, and that's how this technique named.
After scan is complete, the length of the triangle, the distance between the camera and the angle laser emitter corner will found. By checking the location of the laser dot in the camera's field of view, the angle of the camera corner can be found. Then by combining all the data gets, it will show the model with almost same shape and size. Figure 4 show how Triangulation 3D laser scanner works.
Figure 4
Structured Light 3D Scanner
Structured light 3D scanners project a pattern of light on the object and detect the deformation of the pattern on the object. They are basically non-contact optical systems, based almost entirely on the principles of photogrammetry in order to transform image pairs to surface information. They are able to achieve information of very high density and of very high accuracy. [1]
Speed is the biggest advantage for Structured Light Scanner. Structured Light Scanner scan multiple points or whole object in one time only, so it was very fast to digitize the scanned image in computer. This will reduces or eliminates the problem of distortion from motion. But the technology was so advance today there were some existing systems are able to scan a moving object in a real-time.
Figure 5- Way Structured Light 3D Scanner work.
Critical Evaluation on Non-contact Active 3D Scanner
Time-of-Flight
The disadvantage of Time-of-Flight scanners is that they can only scan an object one point at a time. Moreover, the accuracy of Time-of-Flight scanner was not good enough. This is because it is using light and the speed of light was extremely fast, it will be very difficult the calculate the round trip for the light travels and this will cause the accuracy of the distance between object and scanner become very low.
Yet, it still has its advantage. As we know, speed of light was very fast so Time-of-Flight scanner can operate over a very long distance.
As a result, Time-of-Flight was ideal to use during scan object like hill, castle or others.
Triangulation
Unlike Time-of-Flight scanners, Triangulation scanners were just like the opposite for Time-of-Flight scanners. Pro of Triangulation scanners is the accuracy when scanning the object is very high. Furthermore, the pro of triangulation scanners is their ability to complete low resolution scans in a matter of seconds.
Yet, the con of Triangulation Scanner is limited range. Different with Time-of-Flight scanners, Triangulation scanners use camera instead of using laser finder, so the range to scan object was short.
As a conclusion, Triangulation was good in scanning smaller object like cup, table or others.
Structured Light 3D Scanner
Structured Light 3D scanners able to scan multiple points or the entire object at a same time, as a result of that, the speed for Structured Light 3D Scanner was fast.
Structured Light 3D Scanner was a good choice to implement in 3D fields .It can give the result in very fast. But, I don't think that it was good when we need to scan an large image or object, if use Structured Light 3D Scanner to scan large object, it might will lose the precision and the accuracy.
Structured Light 3D Scanner is using same theory or principle with Triangulation Scanner. In facts, Structured Light 3D Scanner was only suitable to those smaller object but not for large object.
Non-contact Passive 3D Scanner
The main different between Active and Passive Scanners is Passive Scanners do not emit any kind of radiation themselves, but instead it detect changes in reflected ambient energy such as visible or infrared light. Passive methods can be cheap because they do not require emitters.
Basically, there was 2 type of Non-contact Passive 3D Scanner, such as
Stereoscopic system
Photometric system
All the system and technique above only need camera to process, so it will be cheaper compare to Non-contact Active 3D Scanner.
Stereoscopic System
Stereoscopic usually use two video cameras, and look at the same image or object. By analyzing the slight differences between the images seen by each camera, it calculates the distance at each point in the images. This method was very familiar for those people who love to watch 3D movie.
This system actually was based on the same principle with Human Stereoscopic Vision. Cameras in Stereoscopic System were act like human eyes. As we know, human eyes can judge the distance of the object by seeing the object and human left eye and right eye will produce different image even is looking a same image. A very good example is, close your right eye and see an object after try to close your left eye and see the object again. You will get a different image between left eye and right eye.
Just the same way how the stereoscopic glass and cameras works, two cameras put at different position and they will see different thing. By combine the image that recorded by the two cameras and judging the distance in between it will generate a 3D image.
Figure 6- Stereoscopic Image
Photometric Systems
Instead of using two cameras in Stereoscopic, this system uses a single camera, but takes multiple images under different direction of light source. These techniques attempt to invert the image formation model in order to recover the surface orientation at each pixel.
Figure 7 Figure 8
Figure 7 show the camera will take the image but in different direction light source. The image taken will be different which show in Figure 8, then after that they will use system to compute all images taken from different direction of light source and generate a 3D image.
Critical Evaluation on Non-contact Passive Scanner
Basically, Stereoscopic and Photometric system does not emit any radiation; instead they normally will use a digital camera. Stereoscopic System functions just like when we go watch 3D movie. However, Stereoscopic system requires special glasses such as the 3D glasses when watching 3D movie to have the 3D effects.
Photometric System was good because by providing a different direction of light sources, the surface of the object wanted to scan will be capture in details, as a result of that, a precise and perfect 3D object can be present.
Last but not less, the most important advantage of Non-contact Passive Scanner is cost. It does not require any devices or machines but only need camera. So, it might be a good choice if the budget was tight but wanted to try how to create 3D object.
3D Modeling
3D Scanning was an advance technology to create 3D object, but nothing is perfect even with all those high tech devices. For example in movie field, people will critics those object or images even they already using most high tech technology and technique. So to reduce those critics, we have to minimize those defects. By using 3D modeling, we can create 3D object by own hand and also can make the object as perfect as possible. This is because 3D modeling techniques can draw or change the detail of the object. User may also can maximize or zoom in the object to change even a small particle.
3D modeling refers to the creation of 3D objects that are defined mathematically and geometrically. There were some example or template can found in that 3D modeling software such as 3D Max, Lightwave, MAYA and others.
There was few type of modeling techniques, such as
Polygonal Modeling
NURBS Modeling
Primitive Modeling
Polygonal Modeling
Polygons are straight-sided shapes, defined by three-dimensional points and the straight line that connect them. The interior region of the polygon is called the face. Vertices, Edges and faces are the basic components of polygons as show in Figure 9. User can select and modify polygons using these basic components.
An individual polygon commonly call as Face, and it was define as the area of three or more vertices and link with the edges. When many faces are connected together, it will form as polygon mesh. Polygon mesh will share the vertices and edges that link with another polygon mesh. These polygons mesh that share vertices and edges are known as shared vertices or shared edges. Figure 10 show sample for share vertices.
Figure 9 Figure 10
NURBS Modeling
NURBS was known as Non-uniform Rational B-spline, it uses mathematical representations of 3-D geometry that can accurately describe any shape from a simple 2-D line, circle, arc, or curve to the most complex 3-D organic free-form surface or solid. Because of their flexibility and accuracy, NURBS models can be used in any process from illustration and animation to manufacturing.
NURBS use spline curves, which are influenced by weighted Control Points. The curve are follows the point by increasing the weight for a point will pull the curve closer to that point. NURBS are truly smooth surfaces, not approximations using small flat surfaces, and so are particularly suitable for organic modeling.
Control Points is the main part for NURBS modeling, it determine the shape of the curve. Each point of the curve is computed by taking a weighted sum of a number of control points. The smoothness is determined by the degree of the polynomial.
The knot vector is a sequence of parameter values that determines where and how the control points affect the NURBS curve. A knot can be inserted into a NURBS curve without changing the shape of the curve. The desired side effect of this operation is an additional control point that provides finer control of the related region of the NURBS curve or surface.
Figure 11- Control Point and Control Polygon when using NURBS
Primitive Modeling
Another well know technique in 3D modeling. Cubes, Pyramid, Cone, Sphere, Cylinder and Torus are known as 3D primitive objects. Primitive object is a base or basic to create more complex objects.
All basic primitive objects could be extended to other shape and known as "extended" 3D primitives. By using extended 3D primitive objects it can create more complex object. Another method to create complex object is combine two or more basic primitive object. Figure 13 show the example of complex object that created from basic primitive object.
Figure 12- Basic Primitive Objects for Primitive Modeling
Figure 13- Complex Objects
Critical Evaluation on 3D Modeling
Polygonal Modeling
The main advantage for Polygonal Modeling is the Details or Branched model is very easy. As we know Polygonal Modeling 3 main components are vertices, edge and face, it was very easy to link or create the model. For example when hand need to merge with the fingers, they just need to link or break the point that is connected so that it can be merge. Another advantage is all the polygonal models are being recognized by almost all of the 3D animation software or rendering software, so it can transfers the model to any software.
Data size of a complex polygon model will be larger compare with Primitive models, and this is the main disadvantages of Polygon Modeling. Polygonal Modeling only use polygon to create an object, unlike Primitive it have many other basic object that can change into complex object. For example, create a square box by using Polygonal Modeling will needed more data size compare Primitive Modeling, because Polygonal Modeling need to put maybe a thousand of polygon (or face) and combine in one square box.
NURBS Modeling
Well, NURBS Modeling was so hard to control. it is because user need to be handle the control point. The control point is determine the curve of the object, it the control point does not calculate or handle in well, it will cause the whole object become fake or not nice.
However, if the control point was control and calculate with perfect, the model created will be super smooth. Furthermore, the resolution for NURBS modeling is independence. No matter how close the camera zoom in to NURBS model, it's surface will still remain smooth.
Primitive Modeling
Primitive Modeling can be very quick and easy to construct and the forms are define by mathematically. So it can be very precise and also can create different type of models by using the basic model available.
But, Primitive Modeling does not suitable to modeling organic structure. It is because from the basic model that available, is hard to change it to organic structure such as human body.
Conclusion
3D Scanners was an advanced technology to scan the image, but the cost to implement 3D Scanners might be a little bit high. In the other hand, 3D modeling was a low cost but time consuming to create a 3D model.
Yet, even the 3D model scan by using 3D scanner also need to edit or modify using 3D modeling software to make the model perfect and nice. Unlike 3D Scanner, 3D Modeling can't be able to create a precise and perfect model if without a set of data that about the model that need to be modeling.
3D Scanners have their own pros and cons just like 3D Modeling. As for my Final Year Project I had to choose 3D Modeling technique to create the 3D models. This is because 3D Scanners techniques was too advance for me, even it can create a 3D model with very fast but I think that 3D modeling is just enough for me. Furthermore, 3D Scanners are more tend to industrial field.
Yet, if the budget is allow, I would be very pleasure if can use 3D scanners in my Final Year Project. It not only can save the time, it can also provide a more precise of 3D model.
So, for the final conclusion, I will choose 3D modeling to be use in my Final Year Project in modeling the Sport Complex of College Tunku Abdul Rahman.
