Extraction and Recognition of License Plates of Motorcycles and Vehicles on
Highways
Hsi-Jian Lee, Si-Yuan Chen and Shen-Zheng Wang
Department of Computer Science and Information Engineering,
National Chiao-Tung University, Hsinchu, Taiwan 30050 R.O.C.
Abstract
In this paper, a recognition system is proposed to
extract and recognize license plates of motorcycles and
vehicles on highways. In the first stage, a block-difference
method is used to detect moving objects. According to the
variance and the similarity of the MxN blocks defined on
two diagonal lines, the blocks are categorized as three
kinds: low-contrast, stationary and moving blocks. In the
second stage, a screening method based on the projection
of edge magnitudes is used to find two peaks in the
projection histograms to bound license plates. The
scanning lines with low counts can be removed. In the
third stage, character images are segmented and
recognized. In our experiments, we tested 180 pairs of
images. The block-difference method has a 98% success
rate and can remove 88% of pixels from an image on
average. The screening method has a 94.4% success rate
and the character recognition method has a 95.7%
precision rate.
1. Introduction
Traffic problems are significant in a developing or
developed country. Intelligent Transportation Systems
(ITS) combining electronics, information, communication,
network technologies and so on are developed to improve
traffic problems. Most of applications applied in ITS need
to identify vehicles first. One of effective and useful
identification methods is license-plate recognition through
digital image processing. Although riding motorcycles on
highways is dangerous and against the traffic law, it is a
frequently-occurred risky behavior and cause tremendous
traffic problems. This paper aims to develop methods to
extract and recognize license plates of motorcycles. The
methods developed can also detect and recognize the
license plates of vehicles.
Most license-plate recognition systems focus on the
processing of images with only one vehicle [1-5].
However, an input image may contain several
motorcycles and vehicles. The situation becomes more
difficult when the target objects are occluded by others.
An example is shown in Fig. 0. There are four target
objects in the image, but only three target objects have
whole license plates. Furthermore, the license plates on
the target objects may have different sizes due to the
various distances from the camera to the vehicles as
shown in Fig. 0. Other difficulties include that license
plates in target objects are skew because the target objects
are skew. The motor license plates may have similar gray
levels in background and the target objects. These
problems will also be handled in this paper.
Extracting moving objects from an image sequence is a
major approach to various applications. But motion
detection techniques [6-9] usually rely on pixel-level
classification. The computations of these technologies are
time-consuming. In this paper, we propose an efficient
block-difference method to detect the moving objects.
The blocks are categorized into three kinds: low-contrast
block, stationary block and moving block. We are
interested in the moving blocks mainly.
2. Moving Object Detection
The size of an input image may be very large and the
image usually contains complex backgrounds. Without
pre-processing, it will spend much time to identify the
Fig. 0. Different sizes of license plates due to
the various distances.
Fig. 0. More than one license plate in occluded
movin
ob
ects in an input ima
e.
Proceedings of the 17th International Conference on Pattern Recognition (ICPR’04)
1051-4651/04 $ 20.00 IEEE