Introduction
Machine vision plays a key
role in broadening the range of
applications for industrial robots.
The application of vision-guided
robot systems to underbody sealing
(UBS) lines of automobile factories
can increase productivity,
improve job quality and reduce
work-related musculoskeletal
disorders which frequently
plague workers engaged in
repetitive tasks like UBS.
The position-measuring
vision system for industrial
robots, newly developed by
Hyundai Electro-Mechanical
Research Institute (HEMRI), can
measure six-degree-of-freedom
(6-DOF) position and orientation
of a car body on a continuously
moving conveyor.
Vision System for Sealing
Robots
HEMRI¡¯s vision-guided
robotic sealing system consists of
a continuously moving overhanging
conveyor, two sealing
robots and three stereo visions,
as shown below.
Procedure for measuring 6-
DOF position and orientation of
car body:
1. User Coordinate Setup
Set up a common coordinate
system between left-hand side
and right-hand side robots.
2. Identifying the Position of
Gauge Holes of Reference Car
Body and Vision Calibration
Boards
Select three gauge holes to be
measured on the reference car
body, and find their threedimensional
position data in the
common coordinate system by
moving the robot tip to the center
of each hole. Then, teach the
robots for each camera to get the
corresponding gauge hole image
in its field-of-view.
Attach the camera calibration
board near the gauge hole and
move the robot tip to the predefined
three points on the board
to obtain three-dimensional position
data of each hole. Repeat the
procedure for all selected gauge
holes.
3. Camera Calibration
In the camera calibration step,
the camera matrices from left and
right cameras are obtained by
using the calibration pattern.
4. 3-D Coordinate Reconstruction
from 2-D Stereo
Image Coordinates
In this step, the image
coordinate-based threedimensional
coordinate
values are reconstructed
by using the triangulation
from the stereo matching
points and the calculated
camera matrices.
5. Calculation of 6-DOF Position
and Orientation of Reference
Car Body
A calibration board coordinate
system is created for each
hole by using position data,
obtained in Step 2, of three points
on the calibration board.
By transforming the reconstructed
three-dimensional coordinates
obtained in Step 4 into
calibration board coordinates,
three-dimensional positions of
gauge holes in the robot coordinate
system are obtained.
6-DOF position and orientation
of the reference car body is
calculated from three-dimensional
positions of three gauge holes.
Following Steps 1 to 5, a
preparatory stage is completed
for production.
6. Measuring and Robot Sealing
in Production Stage
In the production stage, 6-
DOF position and orientation of
every incoming car is estimated
by measuring the position of the
gauge holes with the vision system.
The robots perform sealing
work along the trajectory corrected
by the amount of difference in
position and orientation between
the reference and the incoming
car body.
Conclusion
Machine vision is a key factor
in intelligent robot application.
The need for vision-guided
robotic automation systems will
increase rapidly in the near
future. Hyundai Heavy
Industries will continue its cutting-
edge research in this field to
develop robot and vision solutions
that meet our customers'
needs.
|