Information Technology Reference
In-Depth Information
the search speed for finding planes. Finally, the performance of the algorithm for
real range data obtained from a stereo imaging system has been verified.
system
3D
data
Plane
detection
Keywords Digital
imaging
range
Randomized Hough transform
Mobile robot
List of Abbreviations
CHT
Combinatorial Hough transform
DGHT
Dynamic generalized Hough transform
HT
Hough transform
IRHT
Iteractive randomized Hough transform
KIAT
Korea Institute for Advancement of Technology
LUT
Look up table
MEST
Ministry of Education, Science Technology
NRF
National Research Foundation of Korea
PDC
Plane detection cell
RHT
Randomized Hough transform
11.1 Introduction
Finding a planar region in 3D space is important for safely driving and operating a
mobile robot or a moving system that can freely drive. A robot operating in an
unknown 3D space must identify its surroundings before the system can conduct
its assigned tasks. Biped or wheel-based robots should be able to recognize
obstacles within their area of motion and avoid or drive over detected obstacles
where possible. In the presence of a staircase, a robot should be able to recognize
the step planes and traverse them. In order to operate without falling in an
unknown 3D space, it must be able to use the available information about 3D depth
to recognize planar regions without obstacles in the direction of movement.
Planar space detection has many fields of application. A robot moving around
indoors should be able to recognize the planar floor region on which it will drive.
Also, the inclination of this plane with respect to the direction of gravity should not
be too great, to prevent the robot falling while it is moving. Planar space recog-
nition can be used in recognizing obstacles, such as walls, a table on which objects
are located, or the plane of a table. It can also be used to find the centerline of road
on which a robot can drive. Recognizing the step planes is an important application
for the movement of humanoid biped robots [
1
].
Recently, a method was proposed which detects regions in real-time for driving
mobile robots. Based on the Hough Transform (HT), the method was applied to
biped robots such as ASIMO of Honda, which walked on steps and avoided
obstacles successfully. Using the HT, the X-Y-Z data in 3D space is transformed
Search WWH ::
Custom Search