Character classification for license plate recognition system based on image processing using MATLAB

A License Plate Recognition System (LPRS) is one of the most important systems used for monitoring and controlling transportation and traffic in many countries. The LPRS is used for many purposes such as toll collection, traffic monitoring and control, smart parking, speed limiting Because of its importance, LPRS should be continually studied and improved by doing a lot of studies to solve each problem that can reduce the performance of LPRS. One of the problems is the noise in the captured images, caused by rain and haze, which lead the system to incorrectly recognize the characters. To address this problem, multiple filters to reduce the noise inside the images, especially the noises which is caused by haze and rain, have been investigated. Studies attempt to enhance the effectiveness of the system that is used to detect and recognize car plate characters and numbers and to find the accurate algorithm most suitable for specific countries, depending on the country’s standard car plate specifications. Because of our study done in Malaysia, and for Malaysian car plates we should know more about Malaysian car plate design. Malaysia has specific car plates designed with black background and white font at a fixed size. Several applications have been developed in Malaysia to identify these plates and recognize the characters and numbers. This research is mainly focused on comparing two such applications. Each application uses a different algorithm and each algorithm will be tested with the same proposed filters and dataset, which is taken under bad weather and illumination conditions to test each algorithm’s performance in the most challenging cases.

Preview Text FSKTM 2016 30 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item