License plate segmentation in images

Vehicle License Plate Recognition Systems are used for automated recognition of vehicles. This system can be efficiently employed for traffic monitoring of roads, automated toll collection, entry surveillance of hospitals, military bases, official buildings, etc. The first and the important step in license plate recognition is the license plate segmentation. In this thesis, license plate segmentation systems for the vehicles which have license plate that conforms to Turkish License Plate Standards have been discussed.The first method introduced in this thesis is based on The Optical Character Recognition (OCR) system. In this method, recognizing the segmented characters in the images are aimed and histogram equalization and smearing algorithms are used for implementation of this system.The second method introduced in this thesis is based on determining high frequency coefficients within the image to find out license plate by wavelet transform. Performances of Haar and Daubechies filters are compared on that basis.The last method introduced in this thesis is based on edge detection method. Edge detection is done by Sobel operator and a moving frame is used to determine the candidate areas with a dimension similar with license plate. Finally, to find out the license plate, the area having the highest edge pixels is taken.