Automatic License Plate Recognition in Real Time Videos using Visual Surveillance Techniques

Abstract

In this paper we present full-featured vehicle detection, tracking and licence plate recognition system. It consists of vehicle detection, licence plate extraction and a character recognition module. Here first foreground estimation is done by Gaussian mixture model then proposing a real time and robust method of license plate extraction based on block variance technique. License plate extraction is an important stage in license plate recognition for automated transport system. The Extracted license plates are segmented into individual characters by using a region-based approach. The recognition scheme combines adaptive iterative thresholding with a template matching algorithm. The method is invariant to illumination and is robust to character size and thickness, skew and small character breaks. The major advantages of our system are its real-time capability and that it does not require any additional sensor input (e.g. from infrared sensors) except a video stream. We evaluate our system on a large number of vehicle images. Experimental results demonstrate the great robustness and efficiency of our method.