License Plate Recognition (LPR)

Abstract- License Plate Recognition (LPR) plays a major role in this busy world. In this paper we have developed algorithms and MATLAB programs to efficiently recognize the license plate number. License Plate Recognition (LPR) system basically consists of three main processing steps such as Pre-processing of the number plate, Segmentation of the number plate and Recognition of the characters. Among this, character segmentation is the most challenging task as the accuracy of the character recognition relies on the accuracy of character segmentation. INTRODUCTION LPR (License Plate Recognition) is an application image processing used to identify vehicles by the license plate numbers. The number on the license plate is a unique identification

Edge Detection is the most common method for finding the discontinuities in the image. This is a very vital step in the pre-processing of the image. In this the brightness of the image becomes sharp. We have used this step after thresholding. In the image we have used, also the brightness has changed sharply. And wherever the image brightness has changed sharply, that points are arranged in to a set of curved lines. And these are known as Edges. Similarly the discontinuities that are found in 1D signal are known as step detection. And the phenomena where the signal discontinuities are found over the time is called change detection. When we apply edge detector to an image we will get a set of connected curves which will point out the object boundaries. Therefore the amount of data should be processed is reduced and thus it will filter out the less important data. When the edge detection is successful then the following step of information interpretation becomes simple. The gradient function is defined

OCR is usually used to recognize the characters in the image. For example like Number plate in the vehicles. Connecting some random points to make a character or letter is known as the Pattern recognition. It is used mainly to observe the trends and to make the future predictions. In a character image, the characters are classified as three. Character set A, character set B and character set C. Similar type of characters comes under one set. For example take characters as z 7 = o b y T. Here z, 7 and = comes under character set A. O and b comes under B and y and T comes under C. In A between the points in the x-y plane a regression line is drawn. Then the distance from each points to the regression line is measured. If each points on either side of the points of the regression line maintain almost equal distances then the character is recognised as ‘=’. If almost all the points on either side of the regression line maintain equal distances and there are some points present near the regression line, then it is recognised as’z’. If both the conditions failed between both lines and the angle between them is 50-70, then the character is