Improve The Algorithm Of Banknotes Counting And Identifying And Realize Its Program

In recent years, with the rapid development of our socialist market economy and the continuous deepening of foreign trade, the circulation of domestic banknotes is increasing. Financial and banking sector is increasingly plagued by mountains of paper money, there is still using the manual count and the nominal value of identifiable within their departments. The method is labor-intensive, single nature of the labor, low efficiency, and the accuracy and security is difficult to guarantee. In this case, it is imperative to develop a detection system which can automatically count recognition and the nominal value for banknotes-card machine.The main contents of this paper include structural improvements of machine prototype, pre-processing of banknotes image research, study and improvement of banknotes counting algorithm in ideal and actual conditions, study and improvement of the denominations recognition algorithm of complex conditions, and realize programmed of the whole process.In the improvement of card machine prototype structure, we will set the instruments-ray source center, which make the gray level distribution of image is more uniform and symmetry, in order to get much better and more conducive banknotes images to the later processing.The ultimate goal of image preprocessing is to make the image get later more propitious to counting and identification. In order to better protect the image information isn’t to ignore or filter out, median filtering algorithm for banknotes image adaptive is proposed in this paper; On the other hand, by comparing the various threshold of segmentation to extract the target striped effect, adaptive threshold segmentation method for banknotes image is put forward.In the card counts, image should be tilt corrected before calculate the fringe spacing, to eliminate the error caused to count the picture tilt. Calculation methods which combine the centroid distance of the image adjacent with the target stripe image bank notes outline edge is proposed in actual conditions.In aspects of identify nominal value, proposed a gray-scale image segmentation curve fitting method to calculate the true length of the paper money, combine the paper money its own gray-scale and the location information to achieve the nominal value identification. It improves the recognition accuracy of the nominal value, simplify the notes containers and extend the system identification function. The system has a certain practical significance.In the machine system functional design. It makes the new arithmetic procedures and makes the process of counting and recognition accurate. The program realizes the functions of the number statistics and display, realizes the functions of the nominal value recognition and display, realizes the function of alarm when the hanks not allow or banknotes mixed. It also realizes the function of printing payment slips, and other information management functions.