AUTOMATED NUMBER PLATE RECOGNITION SYSTEM

Abstract

ABSTRACT The strategy of the project is to design automatic system for recognizing the number plate of the car from image taken by the traffic surveillance system. Thus, the project is to investigate and construct an application whereby the system will recognize the car number plate from the image captured of the car. The system will be based on a personal computer and software packages available, MATLAB program and a digital camera, which help to capture the images of the car. The general algorithm involves the following steps: Ø Image Processing: The image captured is preprocessed and reduction in the contrast. Ø Conversion of RGB image to Gray image: Processed image is converted into a gray image. Ø Erosion and Dilation: Unwanted noise is removed. Ø Localization of number plate using Histogram approach: Thresholding is applied to both the row and column difference vectors obtained. Ø Smoothening and Filtering: Unwanted rows and columns are removed using averaging method. Ø Final extraction of number plate: Final number plate is extracted by selecting the area having maximum intensity variation between pixels and having maximum row and column sum of pixel intensity. Ø Number plate recognition based on OCR method.