Detection and Recognition of Broken Character in Car Plate Image

In Malaysia, there are many cars that have a broken character on their car plates. This case affects the process of recognition of the car plate especially when any unexpected and dangerous situation such as accident and car robbery. In literature, there are some researches on broken character reconstruction and reconnection on a different kind of application which are vital for character recognition system. Therefore, this study is concerned to develop a tool with recognition algorithm that able to detect the broken character on car plate by using Template Matching. Two type of approaches which are correlation and similarity measure are measured in Template Matching. In this study, a car plate image with broken character is segmented using Connected Component Labelling. Then, the process of character relocation is completed before the detection and recognition process using Template Matching. Results for both approaches are shown for measurement purposes. As conclusion, the tool using the recognition algorithm is suggested to be an alternative in the area of character recognition in car plate to assist in the recognition of the broken character. This study is significant due to its importance in various real world applications such as Road enforcement system, police traffic control, parking ticketing system, or hometown security management.