License Plate Automatic Recognition based on edge detection

In this paper, we present an Automatic License Plate Recognition System (ALPRS) to identify license plates which is an application of image processing. The main process of ALPRS is divided into four steps: The noise in the image is removed by using FMH filter. A simple algorithm is used for background subtraction. Canny edge detection is used to localize the license plate location. Finally, letters and digits are extracted through template matching technique. The proposed algorithms have two advantages: First, the method has strong robustness against noise. Second, it can deal with license plates with different colors. The performance of the algorithm is tested in a real-time video stream. Based on the result, our algorithm shows the missing rate is almost 16% from 70 vehicle images.