The Research Of License Plate Recognition System Based On Android Platform

With the development of the social economy of our country, the industry Internet of things plays an increasingly enormous influence. As one of the leading-edge applications of its integration into the daily life, Intelligent transportation has brought us unprecedented convenience. The technology research of License Plate Recognition is always the popular direction in the field of intelligence transportation. With the increasing popularity of smart phones, more convenient and multifunction of License plate recognition system will be the development trends of its next.In this paper, the operations of preprocessing, plate location, character segmentation and character recognition were discussed and then a kind of license plate recognition system based on Android mobile phone is proposed in this paper. The main contents are the follows:(1) According to the actual environment, the plate image may be uneven illuminated or license plate too dark, and so on. This paper presents a reasonable method to improve the preprocessing. Combined with the hardware limitations in mobile platform, this paper proposes a fast and effective method of locating processing based on edge detection, morphological processing and projection analysis. The method use sobel vertical operator detect image edge and a series of processing operations isolate the license plate connected region, complete the preliminary location. Then combined with vertical projection and horizontal projection, we cut the image accurately in the edges so that locate the plate area accurately. Finally, considering of the low-quality photographs took by mobile phones for the reason of the different position we stand, an improved method based on Gaussian scale transformation was been proposed.(2) This paper presents a tilt correction method by extracting the boundary of license plate characters. We calculate the tilt angle of the license plate image by extracting the character boundary, this method perform well in the condition of missing license plate frame. At the same time it presents a fast and effective character segmentation processing method combination of the priori knowledge and the vertical projection which can successfully segment the seven isolated areas of characters. In view of the actual situation, the phenomenon of blocking characters has a corresponding improvement measures.(3) Due to the simple and fast characteristics, we use a modified template matching methods. By adding Gaussian noise into the original template, performance of the recognition algorithm improved a lot. In the case of some similar characters may cause to errors, the paper also proposed an improved method by analysing the different areas.(4) Make use of each part of the algorithm in Android platform, and implement a complete portable license plate recognition system.