A Comparative Study of Vehicle Number Plate Recognition Systems

The traffic management based on vehicle number plate recognition in Nigeria has not recorded the muchexpected result because it is manually done. Having studied the existing solution, it is opined that every nationhas its unique vehicle number plate, and off – the – shelf automatic number plate recognition system developedfor one nation is not likely to work optimally for another nation. Despite the fact that the new Nigerian numberplate system was announced in 2011, it is observed that quite a large number of vehicles on Nigerian roads stillhave the old number plate system. However, the system that will detect and recognize both Nigerian numberplate systems has not been announced. Hence, the need to develop a system to detect and recognize bothNigerian number plate systems. Therefore, the aim of this paper is to carry out a comparative study of existingvehicle number plate recognition systems, especially for Nigerian roads and also to carry out experimentalstudies on Nigerian number plate recognition systems. The methodology used includes the acquisition of 934sample images of new Nigerian number plates and 567 sample images of old Nigerian number plates. Then preprocessingof the acquired images, extraction of the identification on the number plate via charactersegmentation, character normalization (extracted characters reduced to 42 x 24 pixels), feature extraction andrecognition of the extracted characters using template matching. From the study and analysis of the test,individual character recognition accuracy of 86% was gotten from the dataset, which shows that 791 sampleimages of new Nigerian number plates and 499 old Nigerian number plates were successfully recognized. Due tothe errors encountered during implementation, it is recommended to create new character template with the samefont as that on Nigerian number plate for accuracy.