Teo, Kein Yau (2015) Malaysia car plate recognition. Universiti Malaysia Sabah. (Unpublished)
|
Text
ae0000002748.pdf Download (854kB) | Preview |
Abstract
Automatic Number Plate Recognition (ANPR) has gained much popularity over the years. ANPR are used for automatic toll collection and management for parking areas. However, there is no international common standard for car plates which makes the task of automatic car plate’s recognition very challenging. In Malaysia, the Malaysian Road Transport Department (JPJ) is the government body in charge of issuing car plate licenses. Malaysian license plates consists of English alphabets and numbers and so designing ANPR for Malaysia license plates is straightforward and easy. However, there are a number of memorial plates, or plates with distinctive prefixes that are made available by the JPJ. These types of license plates are sold at a higher cost. These special plates are used to denote the manufacturer of the car such as Proton Malaysia introduced “Proton Waja� cars, a special event car plate BAMbee was issued in 2000 for Thomas and Uber Cup which was held in Kuala Lumpur in that particular year. Hence, designing an accurate ANPR system for such license plates is challenging. This research project involves autonomously localizes and recognizes non-standardized Malaysian’s car plates using conventional Backpropagation algorithm in combination with Feed-Forward Neural Network (BPNN). The experimental result is compared with the results obtained using simple Radial Basis Function Network (RBF). This research aims to solve four main issues; (1) localization of car plates that has the same colour with the vehicle colour, (2) detection and recognition of car plates with varying sizes, (3) detection and recognition of car plates with different font types, and (4) detection and recognition of non-standardized car plates. The proposed method involves two tasks, pre-processing and recognition. The captured car images are first binarized in order to remove unwanted small objects. Then, filtering is applied in order to remove larger objects. Next, a deblurring technique is proposed to create an area for bounding box. The bounding technique could segment the characters correctly. Lastly, BPNN as well RBF are used to train the segmented and extracted characters. The experimental results show that the combination of BPNN and RBF can be effectively used to solve these four issues. In BPNN, letters ‘J’ and ‘M’ and digit ‘7’ and ‘8’ achieved 90.91%, 85.71%, 97.22% and 97.14%, respectively. In RBF, letters ‘B’, ‘S’ and digit ‘0’ accuracy rates are 97.22%, 96.97% and 86.67%, respectively. Hence, it shows RBF performed better than BPNN.
| Item Type: | Academic Exercise |
|---|---|
| Keyword: | Automatic Number Plate Recognition (ANPR), car plates, conventional Backpropagation algorithm, Feed-Forward Neural Network (BPNN), Radial Basis Function Network (RBF) |
| Subjects: | ?? QA75 ?? |
| Department: | FACULTY > Faculty of Computing and Informatics |
| Depositing User: | ADMIN ADMIN |
| Date Deposited: | 09 Nov 2015 14:47 |
| Last Modified: | 27 Oct 2017 17:04 |
| URI: | https://eprints.ums.edu.my/id/eprint/12114 |
Actions (login required)
 |
View Item |