REAL TIME NUMBER PLATE LOCALIZATION ALGORITHMS Balázs Enyedi

Segmentation procedures are complicated and incorporate diverse problems, among which number plate localization, which is the subject of the current paper, constitutes a special case. Number plate identification comprises two well distinguishable fields: localization of number plates in the image and recognition of characters within the located areas. Neither is an easy task, ie both are time consuming procedures, but certain measurements indicate that localization and separation of the characters may last even 25 times longer than recognition. Number plate localization strategies that require remarkably smaller computational resources shall be introduced in the following. Algorithms based on fast and classical image processing methods such as filtering, edge finding and adaptive thresholding, which do not incorporate any learning procedure, shall be discussed. The first part of the current article introduces algorithms that are optimized for nearly horizontal number plates, while the second part focuses on differently located license plates. The section presenting the results reveals under what circumstances and conditions the individual procedures can be used effectively, considering also reliability and computational requirements.