A Vehicle License Plate Recognition System Based On Computer Vision

A vehicle license plate recognition system(LPR) is presented in the paper, which is based on image processing and pattern recognition. LPR finds its increasingly wide and important application in entrance admission, security, parking control, harbour cargo control ,road trattic control, speed control and so no.Compared with traditional vehicle managing methods , it saves manpower and material resources, realizes scientific standard management and ensures traffic order.The innovations in this thesis are as follows(l)A novel license-plate localization is presented expoiting edge detected binarization and scanline.It is much more robust and faster than traditional methed.(2)A modified binarization is Presented based on mean-differential operator for extractsing characters in the plate. It is more robust and faster than traditional image segmentation metheds. (3)A fast and efficient algorithm is presented, which first thins letter images, then uses improved Hausdorff distance to measure the degree of similarity between templates and letter images. Experiments show that this algorithm is more robust than sample template matching method when recognized characters are slant, deformed or blurred.In Chapter One of this thesis, a vehicle license plate recognition system system is introduced.In Chapter Two are introduced the system architecture andthe functions of subsystem.In Chapter Three, the method of plate locating is first proposed according to different plate features. Then a vertical edge detection method and vertical edge based plate locating method are proposed.In Chapter Four, a character segmentation method is proposed.In Chapter Five, a fast and efficient algorithm is presented, which first thins letter images, then uses improved Hausdorff distance to measure the degree of similarity between templates and letter images.Chapter Six describes how a LPR System is applied in highway electronic toll collection.