Development Of A Nigerian Vehicle Plate Character Recognition System Using Template Matching

ABSTRACT

Tracking of vehicles, manage on-site car parking spaces, and warehouse traffic are problem that need real-time solution and can be solve by using character recognition in vehicle plate which can be obtain by using different classifier e.g. artificial neural network, fuzzy logic, Neural-fuzzy hybridization, and generic algorithm, template matching which can solve regression and binary classification problems. This project aims is to develop a system that will detect and recognize plate number images of Nigerian vehicle using template matching. The proposed system will use the following phase to achieve the result image acquisition by using digital camera, image preprocessing by greyscale, remove noise, image enhancement and thresholding, then image segmentation by multiple thresholding, dilation, enrode and dissection. Recognition by template matching for the classification and the system will be evaluated. The result of test carried out on the system been developed shows that the prototype have 18% segmentation error, RGB color recognition model have is 88.8% accuracy and 77.7% accuracy of the template matching character recognition. The system at the end of the test shows the usefulness and effectiveness of template matching in character recognition and the system have several advantages in transportation sector.