Documente Academic
Documente Profesional
Documente Cultură
ASSEMBLY www.assemblymag.com
Vision-Guided
Robotics
pg 52
052_asb0905rvisrep.qxd 11|17|05 17:22 Page 52
ASSEMBLY Robotics
Guided by Vision
Vision-guided robots can substantially increased the flexibility
of today’s automated assembly systems.
F
ingly high priority for robots with machine vision more prac-
many U.S. manufactur- tical. With user-friendly software, man-
ers, and the vision-guid- ufacturing engineers can make product
ed robot has become a changes, adjust tools and create new
powerful tool to meet measurements without calling a vision
this demand. Vision-guided consultant.
robotic (VGR) systems can be Today’s VGR systems represent a
quickly adapted from one prod- good value for manufacturers, not only
uct to the next, facilitating new because the cost of robots and vision
product introductions. VGR systems has decreased considerably,
systems can enhance assembly, but also because the tools for both have
packaging, test and inspection vastly improved. Robot manufacturers
processes, and they can per-
form in environments that are What was formerly
hazardous to humans.
Robots have been at the an art has been
forefront of flexible automa-
tion for many years, but until
reduced almost to a
recently robot vision was lim-
ited. In the past, vision sys-
science.
tems required a PC with a cus- now include high-powered vision algo-
tom frame grabber to capture rithms with their software, in addition to
or “grab” images from cam- standard measurement and object-
eras, and custom software was recognition tools. Ethernet connections
needed to analyze the images. allow robots to communicate quickly
These systems were too com- with other robots, remote controllers
plex for manufacturers to and factory networks. Most robot con-
maintain with their own staff. trollers have built-in I/O modules with
Instead, vision experts had to programmable logic, allowing them to
be brought in to maintain the control peripheral devices. This makes
system and to create the using a PLC or PC as the main con-
Vision guidance allows this robot to pick sophisticated algorithms necessary for troller optional.
randomly oriented footballs from a moving image analysis. This was very costly. Vision-tracking routines are
conveyor belt and correctly place them in boxes.
Now, all that has changed. What was available with some robots as an
formerly the art of integrating vision integrated package that includes both
systems has been reduced almost to a hardware and software. Cameras can be
By Richard G. Zens Jr. science. Cost-effective vision systems connected directly to the robot
Vice President are readily available, with greater accu- controller to enable on-the-fly picking
Abacus Automation racy and a much wider range of capa- of parts or products from moving
Bennington, VT bilities than earlier systems. Today’s conveyors.
52 ASSEMBLY / September 2005 www.assemblymag.com
052_asb0905rvisrep.qxd 11|17|05 17:22 Page 53
Robotics
These precision-ground plates had mirror-like finishes that created hot A wheel, mounted on the press ram, follows a groove cut in a metal
spots for the camera, washing out critical groove detail. Indirect lighting plate. Sophisticated vision software generated the path for the
solved the problem. robot to follow.
settings were optimized, care had to be the grooves were defined and displayed angle to minimize distortion and fit in a
taken to prevent factory lighting from on the operator interface. Acquiring factory environment.
affecting the images. Outside windows data for 24 separate grooves took less Even so, there was still a great deal
near the system created potential than 5 seconds. Groove data for each of distortion due to the angle from cen-
problems not only between day and plate was stored on a hard disk to be ter below the camera to the outside
night, but also due to the different used by the robot. edges of the magnetic chuck. However,
positions of the sun throughout the year. We were close, but we still needed to the high-end vision system had a new
Fully enclosing the camera and lights improve the system’s accuracy to meet tool that made it possible to correct for
ensured the settings would not be the 0.02-inch tolerance requirement. parallax and lens distortion. Using a
disrupted by external light sources. The high-resolution camera did not precision checkerboard grid that cov-
With the lighting defined, the next have sufficient pixel counts to measure ered the magnetic chuck, calibration
task was to find the grooves in the an object with an accuracy of 0.02 inch software automatically corrected the
plates. The magnetic chuck was made over a 3 foot by 4 foot area. Parallax dis- field of view for any distortion. The
of precision-ground steel and looked tortion was also a concern. To have zero variable thicknesses of the plates creat-
similar to the plates, making it difficult parallax error, the camera would ideal- ed additional parallax, as the angles
to distinguish between the two. An ly be mounted infinitely far from the became greater when the plates were
experienced vision consultant wrote work. Experiments determined that a placed nearer the outside edges of the
custom software to find the grooves by distance of 8 feet would give the best chuck. Based on the positions and
analyzing the field of view. However, combination of lighting and viewing heights of the grooved plates, offsets
the program took hours to run were added to yield precise
and could find only partial Press With Vision-Guided points for the robot controller
grooves. Four Axis Robot to use.
An in-house controls engi- Camera The groove data-points
neer solved this problem by were sent to an array to guide
creating arrows and attaching the robot to follow each
them to 3/4-inch diameter 8' groove. The last challenge
Cartesian robot
magnetic discs that were was to correlate the groove
placed in front of the grooves. X-axis data to the robot coordinates.
The camera was then pro- magnetic Using the calibration grid, the
grammed to search for the chuck robot was moved to various
arrows and thus find the start points on the grid, and the X
of each groove. From there, and Y coordinates were
the grooves were followed recorded. When a point var-
using edge tools included with ied slightly from the grid
the vision system. Groove- position, the robot was moved
Press arm
position data was taken every manually to align the press
0.05 inch. Then, using averag- Sophisticated lighting enabled engineers to automate high-precision, ram to the exact location on
press-fit assembly of metal tubes into grooved steel blocks.
ing and smoothing routines, the grid. Offsets were record-
54 ASSEMBLY / September 2005 www.assemblymag.com
052_asb0905rvisrep.qxd 11|17|05 17:22 Page 56
Robotics
ed and a table was created of all the off-
sets. When the robot was commanded to
go to a new position, these offsets were
averaged and added to the groove data.
This created a precise path for the robot
and met the 0.02-inch tolerance require-
ment.
Camera Mounting
For robot guidance, machine vision
cameras can be mounted to fixed
positions in the cell, or they can be
mounted directly to the end-of-arm
tooling. The latter position enables the
camera to view multiple locations while
guiding the arm. This is useful when
small objects need to be viewed over
large areas, for inspection or to
determine precise position. The vision
tools correlate the arm’s position with
the field of view to specify the robot’s
next move.
In one such system, boxes of glass
tubes are fed to an unload station. The
boxes are tilted to 30 degrees to keep the
tubes upright as they are removed from
the boxes. Quick-change probe tools and
selectable programs allow tubes of vary-
ing diameters and heights to be trans-
ferred at this station. The camera locates
the ends of the glass tubes and guides the
probes to engage them. Air pressure is
channeled through the probes to expand
their O-rings and grip the tubes. The
robot then transfers the tubes to the
assembly machine, where they are
processed. This system addresses flexi-
ble production requirements, reduces
breakage, cuts loading costs and mini-
mizes glass-handling hazards.
As flexibility requirements increase,
vision-guided robots are meeting the
challenge. Robot manufacturers are
providing stronger and faster arms, and
software engineers are developing
new capabilities and making existing
tools easier to integrate. The next big
challenge is to improve lighting tech-
niques to keep pace with advancements
in robotics and vision. Great strides
have been made to accommodate varia-
tions in lighting, but it still remains an
art. Even so, robotic vision systems are
well-positioned to fulfill the needs of
manufacturing in the years to come. A