A new method for license plate validation using the orientations algorithm

A license plate identification system comprises three main parts: isolation of the license plate, extraction of the characters, and recognition of the characters. This paper is a contribution for solving the problems related to the first part, more specifically the license plate verification. We present a new license plate locating approach inspired from the context of the fingerprint with two steps. First we partition the input image into a set of homogeneous zones according to their color similarity and spatial proximity using, respectively, the algorithm of color detection and the algorithm of associated regions. Second we classify the regions into plate regions and not plate regions, using the dimensions of the plate and the orientations algorithm, which is compared to two well known algorithms from the literature. The proposed algorithm was tested and compared to the two other algorithms using 100 images of vehicles taken at different distances, angles and under different illuminations. The obtained results show that the algorithm is the best in terms of success rate.