|
|
Time-of-flight
3D camera technology provides multi-dimensional image solutions
Leveraging the benefits of Time-of-Flight camera technology
in 3D machine vision applications
Overview
Time-of-flight cameras are a recent innovation made possible through improvements
to semi-conductor technology and are a type of 3D range-imaging camera
that use a collection of techniques to produce a multi-dimensional image
showing the distance to points in a scene from a specific point. The subsequent
image captured has pixel values which correspond to the distance, where
brighter values mean shorter distances. These pixel values can then be
translated into physical units of measurement.
How does a time-of-flight camera work?
Time-of-flight cameras capture 3D information by shining infrared light
against an object and measuring the time delay between the emission and
detection of the light. These cameras are similar to laser range scanners,
except that the whole scene is captured simultaneously in real time and
with no moving parts by an area scan sensor. Image acquisition is done
without the usual need for scanning point-by-point as happens with a laser
scanner, but, instead, the entire scene is captured with each light-pulse.
This method of capture results in fast imaging and high lateral resolution
combined with the depth information of the scene being grabbed.
Advantages
of Time-of-Flight cameras |
Simplicity |
Compared with other systems, the time-of-flight solution
is very simple and compact, comprising no moving
parts and with built-in illumination placed adjacent to the lens. |
Efficiency |
Due to the ease with which distance information is extracted when
using time-of-flight cameras, only a small amount of processing power
is used. |
Speed |
In contrast with laser scanners that move and measure point-by-point,
time-of-flight cameras measure a complete scene with one shot with
up to 100 frames-per-second making ToF cameras much faster then their
laser alternatives. |
Read more application-specific applications:
Case study #1 - Mesa SR 4000 Time-of-Flight cameras
in logistics
Case study #2 - Mesa SR4000 Time-of-Flight camera in
security/area control
While Time-of-Flight
cameras, have decided advantages over other forms of image capture
for certain applications, users should be aware of factors that, unless
managed, might compromise the camera’s output.
Areas of caution |
Environment |
The degree of the object's reflective properties, colour,
gloss and complexity of the scene has considerable effects on the
accuracy and co ToF cameras output a confidence map with each 3D map |
Background light |
Background light, whether artificial or sunlight, while being suppressed,
may need additional management. However modern ToF cameras are overcoming
these issues with more advanced designs |
Multiple reflections |
Unlike laser scanning systems, ToF cameras illuminate an entire
scene. Due to multiple reflections, the light may reach the objects
along several paths, consequently care should be taken with multiple
path reflections. |
Interference |
If several ToF cameras are running at the same time, each may disturb
each others' measurements. However, there are simple fixes for this
situation |
Applications
for Time-of-Flight Cameras
Optimised for use in industrial applications, a time-of-flight camera
can be deployed in a range of environments:
Machine Vision
More general machine vision applications require highly stable
and reliable sensory systems. ToF cameras provide a reliable set
of depth data, increasing the robustness and flexibility of many
surveillance, inspection and logistics systems
Robotics
The ability of technology to mimic the complexity of human eyes,
in real time, has been anticipated for some time. TOF cameras overcome
the problems of alternative robotic vision solutions such as stereoscopic
cameras or laser scanners which are bulky and don’t always
provide accurate information. A single ToF camera allows a single
compact sensor to serve as a highly capable input device in many
robotic applications such as:
• Autonomously guided vehicles, with improved obstacle identification
and avoidance, for map building, localization, and path planning.
• Service robots in both industrial and consumer applications.
• Industrial robots in assembly, quality control monitoring,
material handling, bin picking and automation
Security / Surveillance
Segmentation of 2D images is highly dependent on illumination conditions
and is difficult to achieve in many cases. The third dimension that
is delivered by a ToF camera allows for much easier segmentation
of images. This allows, for example, determining the presence, location,
and number of persons or objects in a target zone and to address
efficiently the following applications in the field of security
and surveillance:
• Person-counting at secured gates for detection of tail-gating
or undesired intrusions.
• Person-counting at airports and in mass-transit systems
for security or statistical purposes.
• Object surveillance or detection in sensitive areas.
Medical / Biometrics
The ability of the ToF camera to provide immediate, cost-effective
images enables a diverse set of emerging medical and biometric applications,
including:
• Human-machine interface for geriatric rehabilitation computer
games.
• Clinical use of ToF for patient-position and patient-movement
monitoring to increase the effectiveness of various therapy systems
including radio-surgery.
• Laproscopic surgeons being provided with surface dimensions
to assist in endoscopic surgery procedures, enabling the marriage
of the real-time image to previously-generated scanned images
|
CASE
STUDY #1: ToF cameras in logistics
The
challenge: Pick-up of pallets in automated warehouses
Today’s warehouses are highly automated in
order to increase the throughput capacity and reduce both transit times
and the number of staff.
Conventional,
manually operated forklifts are replaced by automatic guided vehicles
(AGVs), which are controlled by the warehouse management system and typically
travel on pre-defined routes using markers
for positioning and navigation.
It remains a significant challenge for AGVs to pick up pallets with reliability
as, even though the location of a pallet is well-known, the exact coordinates
to which the fork must be moved to pick up the pallet are not known with
sufficient accuracy. Therefore, a sensor is needed to detect the pallet
and locate the exact position of the slots or pockets into which the
forks are to be inserted. (Figure, right, shows the position
of the camera in red).
The solution (deploying
a MESA SR4000 Swiss Ranger camera)
The automatic pallet pick-up is enabled by a ToF camera from MESA Imaging.
The camera captures a 3D image that is analyzed to detect the exact position
of the fork pockets of the pallet. The camera is installed on the front
portion of an AGV and communicates via its Ethernet interface with the
control unit of the vehicle.
The advantages of ToF cameras in
logistics
Compared to alternative technologies, a TOF camera offers some key advantages:
It can scan an area in a single exposure within a fraction of a second
and achieve frame rates of up to 100 per second. Laser line scanners,
in contrast, only scan one line at a time and they must be mechanically
panned or translated to scan an area. Moreover, a TOF camera does not
rely on any complex image processing to acquire distance information.
In this respect it has a key advantage over a stereo vision system, where
features in the left and right image must be matched to compute the distance
information. This can result in long latency times, or in an incomplete
depth map for image regions where no matching is possible.
TOF cameras directly measure the distance for every pixel, the depth map
is, therefore, always complete and the latency time is very low.
Finally, an important advantage of TOF cameras is their ability to handle
varying light conditions and the fact that no image contrast is needed
to detect the fork pockets. This allows reliable detection in challenging
situations, e.g. in the presence of shadows, or when pallets are dirty
or damaged.
Key
benefits of using the Mesa SR 4000 in logistics deployments |
Reliability |
Very reliable fork pocket detection, eliminating need
for operator maintenance or trouble shooting and minimizing down time
of the system |
Speedy detection |
Very fast detection for short cycle times and maximum vehicle efficiency |
Robustness |
System can handle varying ambient light conditions, and is not troubled
by shadows. Reliable detection of dirty or damaged pallets. |
Small size |
With just 70x70x70mm easy to integrate in existing fork lift setup |
Flexibility |
Communication via Ethernet interface, platform independent, compatible
with any operating system |
Value-adding |
3D data from the camera can be used as complementary purposes, including
safety, collision avoidance, and navigation |
CASE STUDY #2: ToF
cameras in high-security access control systems
The challenge: Securing corporate protection against
criminal or terrorist activities in highly-sensitive areas![](/news/newsletter/201111-nov/Resources/mesa_security.jpg)
High security access control systems typically consist of an airlock,
i.e. a small room with 2 doors, where only one door can be opened at a
time.
In order to access a secured area, an authorized person enters the airlock
through the first door and is then identified by means of biometric sensor,
an RFID tag, etc. Once the system has verified the person’s identity
and access permission, the second door is unlocked, and the person can
enter into the secured area.
A significant challenge for access control systems is
to detect when an unauthorized person tries to enter the secured
area together with an authorized person, a so-called tailgating
situation (see figure below). Such a situation is very difficult
to detect with a conventional security camera. The situation
even more complex, when the person entering the secure area may carry
an object such as a briefcase or backpack). The system then needs to be
able to differentiate between people and various objects.
The solution (deploying a MESA
SR4000 Swiss Ranger camera)
The solution consists of using a single camera mounted on the ceiling
of the airlock (see figure right, above). The wide-angle objective allows
reliable, 24/7 monitoring of a large area even when the room height is
a constraint. With its compact size, discretion is assured and nor does
the camera impact on the structural design of the airlock. (Figure,
right, shows the position of the camera in red).
The camera communicates with the door controller unit via its Ethernet
interface. Image processing can be carried out in the door controller
or, as in the case of the Mesa SR 4000, in the camera itself on its embedded
processor.
The advantages of ToF cameras compared with conventional security
cameras
Compared to conventional security cameras that only see 2D contours of
an object, a ToF camera can sense the 3D shape, enabling very robust detection
of persons and objects, making it much more difficult to deceive the system.
Compared to stereo vision systems, a ToF camera needs no matching of image
features, and 3D data is available for every pixel in every frame in real
time at up to 50 frames-per-second (when using the Mesa SR 4000) and with
very low latency. Furthermore, TOF cameras do not suffer from occlusion.
ToF cameras are active imaging systems in illumination. They are; therefore,
not sensitive to variations in ambient light conditions and can even work
complete darkness.
How the
3D ToF camera sees and reports tailgating |
|
|
Black-and-white, low contrast image shows a normal (left) and tailgating
(right) situation |
|
The same situation using a colour-depth image |
![](/news/newsletter/201111-nov/Resources/mesa_tailgate2.jpg) |
Key
benefits of using the Mesa SR 4000 in security/surveillance operations |
Accuracy |
Very good detection accuracy of tailgating situations
meeting highest security requirements |
Flexibility |
Senses 3D shape of object, data quality allows differentiation between
persons and objects |
Robustness |
No image contrast required and there is no issue with varying ambient
light conditions or shadows |
Compactness |
Very compact device which is easy to integrate causing minimal impact
on structural design of room |
Simplicity of use |
Easy mechanical and electrical integration with embedded image-processing
capability |
Click
here to learn more about the Mesa Time-of-Flight camera available
through Adept Turnkey |
Adept Electronic Solutions are "The Machine Vision and Imaging
Specialists" and distributor of Mesa Imaging TOF products in Australia
and New Zealand. To find out more about any Mesa Imaging TOF products
please email us at: info@adeptturnkey.com.au
or call us at Perth (08) 92425411 / Sydney (02) 99792599 / Melbourne
(03) 95555621 or use our online contact
us page.
|