3D SCANNER SYSTEM
INTRODUCTION
We live in
a three-dimensional world. Yet the images we see in magazines and on websites
are just two-dimensional. They literally give us only part of the picture. Flatbed scanners are commonplace in many home and office environments,
representing the third largest segment of the PC peripherals market. But they
can presently only work in two dimensions. Flat objects such as photos, house
plans, and letters are 'scanned' and displayed as a two-dimensional digital
image file.
Creating 3D
models manually is time consuming and creates a bottleneck for many practical
applications. It is both difficult to model complex shapes and to recreate
complex object appearance using standard parametric reflectance models. Not
surprisingly, techniques to create 3D models automatically by scanning real
objects have greatly increased in significance. An ideal system would acquire
an object automatically and construct a detailed shape and appearance model
sufficient to place the synthetic object in an arbitrary environment with new
illumination.
As a part of our B.E. syllabus for project
work, we are working on developing a laser based 3D scanner i.e. a scanner that
uses laser to scan 3D objects and creates an exact view of the scanned object. This presentation will consist of two
sections; an explanation of the design of a laser based 3D scanner, followed by
a description of applications that make use of such technology. An insight into
the design of a laser based 3D scanner system will be presented, from the
initial concept through to scanned 3D image. The first part focuses on the
construction and working of the 3D scanner. The second part of the presentation
will focus on applications possible today using such technology. Finally,
advantages, limitations and future developments of the laser based 3D scanner
will be discussed.
The laser
based 3D scanner is capable of generating a three-dimensional image, but one
that doesn't require special glasses or other equipment to view. Displaying
jewellery and other small products for sale over the Internet or managing
museum collections of plant or animal specimens could be revolutionized by the
laser based 3D scanner for scanning and displaying 3D objects.
The 3D
scanner will allow users to scan small objects and obtain a digital image file,
which is three-dimensional, and so gives extra information about the object's
surface hape and texture. The file could be emailed to a friend, put on a
website, published in a print catalogue or displayed in an art gallery.
SCOPE OF SYSTEM
Comparison of
laser based scanners with electromechanical scanners
Electromechanical scanner has an electromechanical arm that moves
along the surface of the object in x and then in y direction and gets the all
the co-ordinates of the 3D object and reconstructs the object using these
co-ordinates to give the exact 3D view of this object.
Differences
between laser based scanners and electromechanical scanners
§
Laser based scanners can precisely
scan soft objects where as, electromechanical scanners are not precise to scan
soft object as they change the shape of the object while scanning it.
§
Laser based scanners are lighter
than electromechanical scanners.
§
Laser based scanners are safer than
electromechanical scanners as they do not come in contact with the object to be
scanned and so cannot cause any harm to the object.
§
Laser based scanners are cheaper as
compared to the electromechanical scanners.
§
Scanning concave surfaces is
difficult with laser based scanners where as, electromechanical scanners can do
it precisely.
LITERATURE REVIEW
Keywords
§
Laser based 3D scanner
§
Vibrator
§
Laser profiles
§
Profile extraction
§
Wire mesh
§
Navigation tool
§
Surface rendering
§
Electromechanical scanner
Requirements for
construction of the 3D scanner
Hardware
requirements
The
hardware components required are rotating platform, laser gun,
vibrator, video camera and a stepper motor. The object to be scanned is placed
on a rotating platform. The platform is rotated using a stepper motor, if
necessary. A
laser gun is used to create a laser profile on the object to be scanned. A
vibrator vibrates this laser gun which forms a vertical plane of the laser on
the object to be scanned. A video camera is used to capture the rotating object
along with the laser profile on it. This method results in an uncomplicated
hardware design that is both affordable and reliable.
Software Requirements
The software components required are as follows:
§
Microsoft Visual Basics 6.0
- Video Edit software utility
- Graphics library
- Microsoft Access
- Navigation Tool
Construction of
vibrator
A laser
beam is to be used to scan the 3D object. The laser beam is vibrated using a
vibrator circuit to form a vertical plane. The vibrator circuit is constructed
using a 555-timer circuit, which gives a frequency of approximately 10Hz. But,
we require a higher frequency as our video camera captures frames at the rate
of 25 frames/sec and in order to obtain the accurate curve of the profile we
need at least 20 points in each frame. Thus the desired frequency is greater
than or equal to 20 points X 25 frames = 500Hz. So an amplifier is used to
amplify the frequency obtained as the output of the timer circuit. Output of
the amplifier i.e. high frequency is applied to a 6-volt relay.
When a
9-volt power supply is applied to the circuit, the coil of the relay gets
electromagnetically induced. Due the magnetic field generated by the coil, the
metallic flap in front of the coil starts vibrating with the desired frequency.
As the laser gun is too heavy to be vibrated, a thin mirror of very small size
is vibrated on which the laser is projected to form the laser beam. The mirror
to be vibrated is mounted on the metallic flap of the relay. Thus, the vibrator
forms a vertical plane of the laser of the required frequency.
Applications
Because 3D modeling using laser scanning is both fast and
cost-effective, it is well suited for many applications. The potential
applications of this device look very interesting.
Visualization
of the photo realistic appearance
One of the persistent problems in 3D graphics today is
how to automatically generate photo realistic 3D content that can be put on the
web or used in movies, broadcast media, and computer games. Most of the
existing 3D content providers sell 3D models that only describe the object's
shape, because the object's appearance is too difficult to acquire, process,
and display. The 3D scanner technology allows for simple acquisition,
description, and fast visualization of the photo realistic appearance of
objects.
Educational
Application
Electronic educational material providers are able to
enhance their products, such as e-books or electronic encyclopedias, with
realistic, interactive 3D illustrations that better explain new concepts and
ideas.
E-commerce
E-commerce generally takes on a new look; companies
are able to make their online presence more appealing by placing natural
looking 3D models of their products on their websites.
Mechanical
Engineering
Laser scanning is ideal for mechanical applications because the software solutions available convert point data into CAD primitives quickly and accurately. Laser scanning also opens the door for many firms that initially prefer to sculpt objects in traditional mediums to retain the tactile and visual advantages that CAD systems lack.
§ Construction
Design
Construction
design is one of the largest areas for 3D modeling development. Applications
include roadway, bridge, and building design and rehabilitation. Designing
construction projects using 3D modeling has been found to have many benefits
such as efficient generation of multiple views, minimization of coordination
issues with virtual design and construction.
§
Industrial Design
For rapid prototyping, reverse engineering, CAD/CAM, 3-D
modeling.
§ 3-D
Game Software Development
Quickly and easily scan and digitize character models.
§ Animation
& Virtual Reality
The market that offers a huge potential is the animation market. That could be animation in movies or animation in computer games. This is such a fast moving industry that software houses are always striving to find ways to make their games more and more realistic. Guiding yourself down a ski slope or through a war zone is probably about as realistic as it can get without actually doing it!
§ Research
For
analyzing 3-D data fields like human interfacing and robotic visioning, FEA and
mold flow analysis.
§ Product
Prototyping
Prototype of the product can be made and approved,
scanned, and a mould can be made of any proportion quickly and easily; all of
this happening in a matter of days and also reducing the total cost. Thus, the
desired product can be sculpted easily and then scanned to insure the intended
result.
§ Internet shopping
Laser scanning to scan human body can be used to enhance Internet shopping sites in a number of ways. First of all from a visualization point of view, it would be preferable for a customer to see them selves dressed in the clothing to see what they look like from all angles before making a purchasing decision. This application exists today in the form of aviators, but it won't be long before you can actually dress yourself instead of a virtual mannequin. Additionally, by having the body scan of a customer the Internet retailer knows their exact measurements so that clothing can be supplied in the correct size first time.
§ Automobile
Design
Designing a car in AutoCAD is difficult as exact shape
and curves of the car cannot be obtained so a clay model of the car can be
scanned using 3D scanner.
§ Design
collaboration from remote location
In today's world, manufacturing processes are carried out by
multiple parties, often from different locations around the globe. The client
and the design process can be in one place, while the manufacturing occurs in
another. The synergistic effect of having several people collaborating on the
development of an idea substantially broadens the scope of the design and
manufacturing process. Once a prototype has been scanned, the engineering,
analysis, quality control, and various other functions that used to take place
consecutively, can take place concurrently before committing to manufacturing.
All parties involved with the project can work from the same digital file. The
result is a shortened development cycle, improved product performance, and
greater flexibility -- positive ramifications at every level.
§ Medicine
If doctor is not available for an operation, the
patient’s body can be scanned using 3d scanner and the scanned image can be
sent to the doctor. Virtual operation theatre can be created where the doctor
carries out the operation and gives command to carry out the operation
similarly on the patient.
Advantages
§ One
of the great advantages of the 3D scanning technology is that only the laser
light makes contact with the surface allowing even very vulnerable surfaces to
be scanned without risk. The lasers used in laser scanning are the same type of
low-power red light laser used in barcode readers and CD players and are not
capable of generating enough light emission to cause damage.
§ Laser
scanning is precise even with soft objects.
§ One of the most
obvious benefits to 3D scanning is the tremendous increase in speed with which
a prototype can be reproduced.
§ Product verification
is another advantage of 3D scanning. After a product has been produced, it can
be scanned and the resulting data compared to the CAD drawing. Deviations from
the specifications can then be accurately determined. This allows for greatly
improved quality control, and helps to detect flaws in the manufacturing
process.
§ As the object to be
scanned is never touched physically, it is not harmed in any way.
IMPLEMENTATION
Working
principle of the 3D scanner
Following are the steps in the operation the 3D scanner:
Hardware Setup
The 3D scanner shines a safe, low-intensity laser on an object to
create a lighted profile. A high-quality video sensor placed at a
pre-calculated angle with respect to the laser captures this profile. The angle is
selected so that there is sufficient relief and the hidden areas can be
efficiently managed. At 0° the camera could not detect the relief, at 90° the
relief would be optimum, but the slightest variation would prevent camera from
detecting the laser trace.
The system can digitize
thousands of these profiles in a few seconds to capture the shape of the entire
object.
Scanning
the object
The object
to be scanned is placed on the rotating platform. The laser beam vibrates to
form a vertical plane, which falls on the object to trace a curve on the
object. When initially projected, the lines are straight and vertical,
but upon hitting a curved surface (e.g. a human body) the lines distort and
bend. This is the principle behind the laser based 3D scanner. By analyzing and comparing
a series of such patterns, it is possible to build up an accurate 3D image of
the surface. The video camera captures the
object along with the laser profile on the object. The object keeps on rotating
and the camera captures the entire scene of complete rotation of the object
with all laser profiles. The obtained video is in AVI (Audio Video Interleave)
format. A software utility called ‘Video Edit’ converts this video into frames
i.e. BMPs. A desired number of frames are selected from all the frames. Let’s
say we want 18 vertical profiles i.e. one profile at every 20 degrees and considering
the time to scan the entire object to be 18 seconds, then we want one frame
after every 1 second and our video camera captures frames at the rate of 25
frames/second. So, we want every 25th frame i.e. if we select 1st
frame then the next frames to be selected would be 26,51,76,…..401,426. All
these 18 frames would be saved and the rest discarded.
Profile
extraction, point identification and storage into database
To understand color scanning, we must
first understand what it is we are trying to capture. What is color and how is
it interpreted by the camera and the human eye? As we know, white light is made
up from equal parts of red, green and blue light, which gives its ‘RGB’ value.
Surfaces appear to have different colors because they reflect or absorb certain
parts of the white light. For example a perfectly red surface appears to be red
when a pure white light is shone on it because it absorbs all of the blue and
green and reflects only the red component.
Other
colors are produced from a combination of different levels of red, green and
blue. The two
extremes are black and white surfaces where the black absorbs all colors and
reflects none, whereas the white reflects all colors and absorbs none. Humans
have the ability to interpret colors because at the back of the eye are cells that
react to the different red, green and blue components of light. The brain then
translates the level of reaction from each of the cells to produce what we
understand to be color. Digital cameras operate using a similar principal
except instead of biological cells, there are CCD cells that produce an
electronic signals are processed by a computer before being displayed on a
screen.
Now all the
selected frames are scanned one at a time to extract the laser profile from the
frame. Each frame is scanned for its ‘RGB” values. Using the above principle,
the point at which ‘R’ value is maximum will be the point on the laser profile.
The corresponding x and y values of that point will be stored in the database.
But if we consider the thickness of the laser line then we will see that for
same y value there are 4 to 5
x values i.e. one laser point covers 4
to 5 pixels. Now two methods can be used to take only one x value
for corresponding y value.
Either consider the first pixel and leave all
other pixels with same y values and store this x and y values in the database
or take the mean of x values of all the pixels with same y values and then
store the x and y values in the database.
Here we
store first x value for a corresponding y value and the y value itself in the
database. Thus we save all x and y values for all the curves in the database.
Construction
of wire mesh diagram
The
database contains x and y values of all the extracted laser profiles in matrix
form where each column gives x and y values of all points on one curve. So to
construct the wire mesh diagram we connect the points in the database.
On
connecting the points in the database column wise we get the longitudinal
curves and on connecting the points row wise we get latitudinal curves. Thus we
get the wire mesh diagram of the scanned object.
Viewing
and hidden surface removal
A
navigation tool is created which allows us to view the wire mesh diagram from
any direction. Graphical transformations like translation, scaling and rotation
can also be applied to the wire mesh diagram. As we view the object from any
one direction the backside of the object is not visible. So to make the back
portion of the object invisible we use algorithm for hidden surface removal. In
the hidden surface removal algorithm, small polygons formed by the wire mesh
diagram are scanned in anticlockwise direction to get normal to the polygon. If
the normal points to the viewer i.e. dot product of the normal with the viewers
direction is greater than zero then the polygon is visible and it is drawn else
if the dot product is zero or less than zero than the polygon is made invisible
i.e. it is not drawn. This algorithm is applied to all the polygons of the wire
mesh diagram.
Surface
rendering
Surface
rendering is done to apply surface and colors to the wire mesh diagram. As
color information in 3D digitizing makes available nearly all the information a
graphics application needs to fully describe an object. In addition to
enhancing realism in graphic models, color denotes boundaries that are not
obvious from shape alone. Color indicates surface texture and reflectance. And
by marking an object's surface before digitizing, one can use color to transfer
ideas from the object to the graphic model. In specialized applications, color
can reveal characteristics such as skin discoloration, the locations of
landmarks, or other features. Working in the infrared region, a customized
color subsystem could even detect surface temperature.
Some of the rendering techniques are scalar graphs, is
surfaces, cutting planes, orthogonal slices, vector glyphs, streamlines, streak
lines & particle advection and textures which give the exact view of the
scanned object. We are not sure which of these techniques we would be using,
but if time permits we will be creating our own rendering software.
FUTURE
SCOPE AND CONCLUSION
Future developments
The
public has very high expectation levels that are set by exposure to television,
computer games and digital photography. If 3D internet applications are to be
accepted by the public then the images have to be as real as possible in terms
of both geometry and color. For virtual try-on, the person on the screen has to
look like the subject otherwise the application will not be adopted and will
eventually fail. Recent advances in lighting positions and techniques
improvements in surface quality and integrity are bringing us even closer to
this goal. One day using a 3D colour image of your self will be as common place
as using a digital camera is today. Trying on clothes over the internet or
appearing as yourself in a computer game will be the norm. Improving the
quality of the images to meet the public’s expectation levels is a key factor
in making this possible.
Transparent
and translucent objects have multiple reflections which makes it difficult to
classify the points. So, in future we plan to work on solving this problem so
that these objects can also be scanned.
Conclusion
Quantifying physical abnormalities,
guiding corrective and plastic surgery, manufacturing clothing,
three-dimensional CAD and other related fields all benefit from the increasing
use of 3D scanners. 3D scanners can give a very precise reconstruction of the
shape of a real object
This laser based
3D scanner is easy to use and enables us to scan opaque objects accurately. Its
major advantage is that it is less expensive as compared to other 3D scanners
available. As it is light-weight and smaller in size it can be easily carried
to any place. Laser based 3D scanning can
provide a measurable difference for improved quality and accelerated time to
market, while reducing costs for new products.
Our experience of working on this project is good as the project is
very interesting and challenging. The difficulties encountered during the
development of the scanner helped us in learning in many new concepts which we
could use to improve the scanner.
We are working to make this scanner more efficient and
reliable.
REFERENCE
§
Mastering Visual Basics 6 -
Evangelos Petroutsos
§
Computer Graphics -
Steven Harrington
§
Advanced Electronics -
Ramakant Gaikwad
§
For information on Laser Based 3D
Scanner
http:://www.3dscanner.co.uk/
http:://www.muellerr.ch/engineering/laserscanner/default.htm
http:://www.mwm.com/guide/image/analysis.htm
http:://biocomp.stanford.edu/3dreconstruction
