Automatic license plate detection system based on the point weighting and template matching

License plate recognition is one of the important security measurements, reinforcing the transportation laws and car-tracing. An effective license plate recognition must deal with problems such as environmental noise, damages to the license plate, light changes, weather conditions and car movements. In this paper, some attributes of the license plates have been extracted, in different conditions in which they are more stable. We proposed an accurate and flexible license plate recognition; applying morphological methods, and a novel method for weighting points. In this process, we face two stages: 1) locating the license plate, 2) reading its characters. We used novel pattern recognition techniques and methods in both sections. The results show that the proposed model is capable of faster recognizing the license plate in the fixed distance range in comparison with other techniques being used currently in the industry. Beside the 93% accuracy rate of this method, it has the ability to work in environments with strong noise. The license plate reader algorithm can also extract, process, and identify the characters from the image of the license plate when we cannot locate the exact location of the license plate through the image segmentation method.to evaluate our algorithm, we applied it to a database comprised of 120 vehicle images with different backgrounds, brightness, distances and viewing angles.