Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method
Automatic License Plate Recognition (ALPR) system is a mass surveillance method that uses optical character recognition on images to read the license plates on vehicles. This system has been used widely overseas. However, the different forms of Malaysian license plates still a problem that makes thi...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2010
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/40711/1/FK%202010%206R.pdf |
| _version_ | 1825949652702199808 |
|---|---|
| author | Al-Faqheri, Wisam Salah |
| author_facet | Al-Faqheri, Wisam Salah |
| author_sort | Al-Faqheri, Wisam Salah |
| collection | UPM |
| description | Automatic License Plate Recognition (ALPR) system is a mass surveillance method that uses optical character recognition on images to read the license plates on vehicles. This system has been used widely overseas. However, the different forms of Malaysian license plates still a problem that makes this system harder to be applied locally.
The proposed license plate recognition algorithm is aimed to recognize the different Malaysian license plates by employing two methods: Fuzzy Logic to recognize standard license plate (the plates which consist of characters and numbers), and Template Matching to recognize non-standard plates (the plates which consist of non-standard word and numbers).
Mathematical Morphology is the first preprocessing step used to enhance Malaysian license plate image quality, by removing noise from the binarized image. The second
step is to remove license plate borders by implementing Mathematical Morphology process with conditional statements. The third preprocessing step is a new Skew
Detection and Correction (SDC) method proposed to correct the skewness of license plate image. License plate level testing follows the preprocessing step in order to
check if the license plate is one or two rows (the license plate elements are in one or two rows). The standard and non-standard test is performed by checking if the input
image is representing a standard or a non-standard plate. Vertical scanning (VS) and horizontal scanning (HS) have been used to segment license plate image elements.
Segmentation process is the step where license plate elements are segmented. The next step is to forward the extracted characters and numbers to the Fuzzy Logic
system to be recognized in case of standard license plates input, while forward nonstandard words images to the Template Matching in order to be recognized in case of
non-standard license plates input. The output of recognition step will be a string of numbers and characters which represent the recognized license plate.
The proposed M-LPR algorithm has shown an impressive result to recognize different Malaysian license plate forms. Fuzzy Logic system has been tested on standard license plate shows 92.16% recognition accuracy and 0.88 second
processing time. The Template Matching shows 92% recognition accuracy and 1.06 second processing time when it is tested on non-standard license plate. The proposed
SDC method has been evaluated by comparing with different other existing SDC methods such as Hough Transform, Projection Profile, Mathematical Morphology and Bounding Box methods. |
| first_indexed | 2024-03-06T08:47:51Z |
| format | Thesis |
| id | upm.eprints-40711 |
| institution | Universiti Putra Malaysia |
| language | English |
| last_indexed | 2024-03-06T08:47:51Z |
| publishDate | 2010 |
| record_format | dspace |
| spelling | upm.eprints-407112015-09-28T00:52:29Z http://psasir.upm.edu.my/id/eprint/40711/ Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method Al-Faqheri, Wisam Salah Automatic License Plate Recognition (ALPR) system is a mass surveillance method that uses optical character recognition on images to read the license plates on vehicles. This system has been used widely overseas. However, the different forms of Malaysian license plates still a problem that makes this system harder to be applied locally. The proposed license plate recognition algorithm is aimed to recognize the different Malaysian license plates by employing two methods: Fuzzy Logic to recognize standard license plate (the plates which consist of characters and numbers), and Template Matching to recognize non-standard plates (the plates which consist of non-standard word and numbers). Mathematical Morphology is the first preprocessing step used to enhance Malaysian license plate image quality, by removing noise from the binarized image. The second step is to remove license plate borders by implementing Mathematical Morphology process with conditional statements. The third preprocessing step is a new Skew Detection and Correction (SDC) method proposed to correct the skewness of license plate image. License plate level testing follows the preprocessing step in order to check if the license plate is one or two rows (the license plate elements are in one or two rows). The standard and non-standard test is performed by checking if the input image is representing a standard or a non-standard plate. Vertical scanning (VS) and horizontal scanning (HS) have been used to segment license plate image elements. Segmentation process is the step where license plate elements are segmented. The next step is to forward the extracted characters and numbers to the Fuzzy Logic system to be recognized in case of standard license plates input, while forward nonstandard words images to the Template Matching in order to be recognized in case of non-standard license plates input. The output of recognition step will be a string of numbers and characters which represent the recognized license plate. The proposed M-LPR algorithm has shown an impressive result to recognize different Malaysian license plate forms. Fuzzy Logic system has been tested on standard license plate shows 92.16% recognition accuracy and 0.88 second processing time. The Template Matching shows 92% recognition accuracy and 1.06 second processing time when it is tested on non-standard license plate. The proposed SDC method has been evaluated by comparing with different other existing SDC methods such as Hough Transform, Projection Profile, Mathematical Morphology and Bounding Box methods. 2010-01 Thesis NonPeerReviewed application/pdf en http://psasir.upm.edu.my/id/eprint/40711/1/FK%202010%206R.pdf Al-Faqheri, Wisam Salah (2010) Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method. Masters thesis, Universiti Putra Malaysia. Fuzzy logic Automobiles - Licenses - Malaysia Automobile license plates - Malaysia |
| spellingShingle | Fuzzy logic Automobiles - Licenses - Malaysia Automobile license plates - Malaysia Al-Faqheri, Wisam Salah Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| title | Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| title_full | Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| title_fullStr | Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| title_full_unstemmed | Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| title_short | Real-time Malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| title_sort | real time malaysian automatic license plate recognition using hybrid fuzzy logic with skew detection and correction method |
| topic | Fuzzy logic Automobiles - Licenses - Malaysia Automobile license plates - Malaysia |
| url | http://psasir.upm.edu.my/id/eprint/40711/1/FK%202010%206R.pdf |
| work_keys_str_mv | AT alfaqheriwisamsalah realtimemalaysianautomaticlicenseplaterecognitionusinghybridfuzzylogicwithskewdetectionandcorrectionmethod |