Detection Of Vehicle License Plates In Complex Scenes

Automatic detection and recognition of vehicle license plates plays an important role in traffic surveillance systems. Such systems, which are applied in parking areas, highways, bridges and tunnels, can help a human operator and improve the overall quality of a service. Any situation requiring the automatic control of the presence and identification of a vehicle provided with a license number may represent a potential application. Being a special computer vision system in the real-time case, the LPR system mainly includes the subsystem of license plate detection and character recognition. The LPR system involves numerous discipline domains, such as Pattern Recognition and Artificial Intelligence, Computer Vision, Digital Image Processing, etc. The detection of license plates is the key of LPR system. Because of the complex of image background, the uncertainty of plate position and image quality, the location of plates is not satisfied. Therefore, the study on the algorithm of license plate location is always the hotspot problem.In this paper, we have summarized the latest research achievements and development of license plate location and segmentation and discussed the intrinsic characteristic of license plate. On the basis of previous work, firstly, we studied the image characteristic of license plate and put forward our algorithms based on image analysis. We address the issue of a novel method to extract vehicle license plates from a complex scene by considering both the distributive regulation of the characters in a license plate and the geometrical features of a license plate. In our approach, we first present a segmenting algorithm, looking for candidate regions that probably contain characters in a range of size. Then we give each candidate region a confidence value to measure its likelihood to be a license plate and combine these regions according to some rules to get a higher confidence value. At last the vehicle license plate can be found on the base of its confidence value.Experimental results show that the algorithm is robust in dealing with different conditions. such as poor illumination and distortion of a license plate generated by different visual angle. However, the algorithm can only detect those license plates whose size is limited to a certain range.As the part of vehicle license plates detection and recognition, the algorithms proposed in this paper provide the location of the plates. But only the statistical characteristics of the plates are used, so the location of the plates is inaccurate. We'llmake further research on this problem utilizing the arrangement of characters of the plates in the plate recognition algorithms.