![](bg1.png)
TRAFFIC LIGHT RECOGNITION USING CIRCULAR
SEPARABILITY FILTER
Shodai Horima and Kazunori Onoguchi
Graduate School of Science and Technology, Hirosaki University, 3 Bunkyo-cho, Hirosaki, Aomori, 036-8561, Japan
Keywords: Traffic light detection, Traffic light recognition, Color identification, Circular Separability Filter, ITS.
Abstract: This paper proposes the camera-based approach to recognize the traffic light for driver assistance. The
circular separability filter applied to RGB images extracts the area of the traffic light. The separability has
large value in the boundary where the intensity between two areas changes like the step and it doesn't
depend on the intensity difference (height of the step). Scanning the circular mask in each RGB image, the
separability is calculated. The separability becomes large in an area where a color is homogeneous and a
shape is similar to the circle. Therefore, the pixel with large separability is selected as the candidate of the
traffic light. Unlike the conventional method which calculates the circularity from the binarized region, the
proposed method can identify the traffic light whose outline is indistinct and whose radius is small. At first,
the proposed method removes the region where the saturation is low and the brightness is extremely low or
high because there is few possibility that the traffic light is included in these regions. Next, the circular
mask is scanned in each RGB image captured from the on-vehicle color camera and the separability
between the inside circle and the outside ring is calculated. The maximum value of separability calculated in
RGB images is selected as the separability of each pixel. Pixels with large separability are detected as the
candidate region of the traffic light. Finally, the candidate region around which inactive traffic lamps exist is
identified as the traffic light. Experiments recognizing various traffic lights under various weathers and time
show the effectiveness of the proposed method.
1 INTRODUCTION
More than 700,000 traffic accidents a year still occur
in Japan though the number of traffic accidents tend
to decrease recently. Because older drivers will also
increase, it is expected that the risk of the traffic
accident will rise in the future. To deal with this
situation, many driving support technologies have
been developed as part of Intelligent Transport
Systems (ITS). It is important to decrease the traffic
accident in the intersection because more than half
of traffic accidents occur in intersections. In the
intersection, overlooking or misidentifying a traffic
light caused the serious accident. Therefore, the
driving support system which rouses the attention or
avoids danger by showing an aspect of the traffic
light to the driver is very useful.
It requires large cost and large time to construct
the road-to-vehicle communication system
transmitting an aspect of the traffic light to vehicles
by the telecommunication facility. Therefore, a lot of
methods to recognize the color of the active traffic
light in images captured from the on-vehicle camera
have been proposed. Because an active traffic light
is usually a red, yellow or green bright region, most
of conventional methods first convert the RGB color
space to some color spaces so as to detect candidate
regions with specific colors of traffic lights. Then,
candidate regions are detected in the converted
image by the binarization and the morphological
operation. Finally, traffic lights are identified by
verifying information around candidate regions, e.g.,
their contours. M. R. Yelal et al. (M. R. Yelal, 2006)
proposed the method using the La*b* color space.
This method detects only traffic lights with simple
background, e.g., clear sky because traffic lights are
identified by verifying edge information around
candidate regions. L. Tsinas et al. (L. Tsinas, 1996)
proposed the method using the HSI color space. This
method causes a lot of false detection because
candidate regions of traffic lights are verified by
only the size of the region. Several methods
identifying the traffic light from the circularity of the
candidate region were proposed because the outline
277
Horima S. and Onoguchi K. (2012).
TRAFFIC LIGHT RECOGNITION USING CIRCULAR SEPARABILITY FILTER.
In Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods, pages 277-283
DOI: 10.5220/0003741402770283
Copyright
c
SciTePress