Number Identification In Complex Background And Its Application

In this paper, not only study on number identification technology of complex background and its applications in some areas, but also summarize the application background and development status of the identification number. According to present various implementation methods on two aspects of the number of techniques in the paper currency number recognition and vehicle license plate locating are reviewed. Research on the complex background of the system implementation on these two aspects and developed the corresponding software routine to realize the algorithm.In the aspect of paper currency recognition system, division and localization of character are based on the arrangement rule of the paper currency number. Then based on the requirements on the speed and the precision of the paper currency number recognition, compared with the each method of threshold value division in the process of pretreats. In view of the characteristic of character, proposed that recognition algorithm based on the template matching, this algorithm can aim at the accurate recognition paper money the series number.In the license plate recognition systems, classify vehicles image in the context of complex background including that the situation of light changes, shadows interfere with the night , and multi-vehicle distance change. In response to these complex situations, using of the car license plate character constant density of the vertical edge features, present a method based on improved cumulative absolute difference method to detect edge and threshold of license plate, and then search the horizontal and vertical projection, finally eliminate pseudo-region, positioning the plate portion. Compared with the traditional method, it can achieve very nice effect. Then use of projection method to segment at the character. Finally in order to achieve character recognition, use of neural network to train numbers and letters.The experimental results show that the number technology in this paper used in the field of the above two aspects, obtained better result, the rate of positioning and recognition has reached the level of practical application.