A Fast Vertical Edge Detection Algorithm for Car License Plate Detection
Recently, License Plate Detection (LPD) has been used in many applications especially in transportation systems. Many methods have been proposed in order to detect license plates, but most of them worked under restricted conditions, such as fixed illumination, stationary background, and high reso...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2009
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/7316/1/FK_2009_28a.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Recently, License Plate Detection (LPD) has been used in many applications
especially in transportation systems. Many methods have been proposed in order to
detect license plates, but most of them worked under restricted conditions, such as
fixed illumination, stationary background, and high resolution images. LPD plays an
important role in Car License Plate Recognition (CLPR) system because it affects the
system's accuracy.
This thesis aims to propose a fast vertical edge detector using Vertical Edge Detection
Algorithm (VEDA) and to build a Car License Plate Detection (CLPD) method.
Pre-processing step is performed in order to enhance and initialize the input image for
the next steps. This step is divided into three processes: First, the color image
conversion to a gray scale image. Second, an adaptive thresholding is used in order to constitute a binarized image. Third, Unwanted Lines Elimination Algorithm (ULEA)
is used in order to enhance the image. The next step is to extract the vertical edges
from the 352x288 resolution image by using VEDA. This algorithm is based on the
contrast between the values in the binarized image. VEDA is proposed in order to
enhance the CLPD method computation time. After the vertical edges have been
extracted by VEDA, a morphological operation is used to highlight the vertical details
in the image. Next, candidate regions are extracted. Finally, the license plate area is
detected.
This thesis shows that VEDA is faster than Sobel operator; the results reveal that
VEDA is faster than Sobel by about 5-9 times, this range depends on the image
resolution. The proposed CLPD method can efficiently detect the license plate area.
The method shows the total time of processing one 352x288 image is 47.7 ms, and it
meets the requirement of real time processing. Under the experiment datasets, which
were taken from real scenes, 579 from 643 images are successfully detected. The
average accuracy of car license plate detection is 90%. For more evaluation and
comparison purposes, the proposed CLPD method is compared with a similar
Malaysian license plate detection method, which is CAR Plate Extraction Technology
(CARPET). This comparison reveled that the CLPD method is more efficient than
CARPET and also has more detection rate. |
|---|