Research On Counting And Identifying Of Banknotes Based On Digital Image Processing

In recent years, with the rapid development of China’s national economy, the circulation of money become faster and faster, market circulation of banknotes needed are growing. The notes handing of bank is very heavy. However, in the bank, the banknotes counting and identifying is still using manual counting currently. This method is not only a drab, onerous, repeatability higher physical labor, but also the accuracy and security are difficult to ensure. Given the banking predicament, it is necessary to develop an automatically count on the note detection system which will make the banknotes counting mechanization and automation.This paper contains two key parts, one is the counting of banknotes, the other is the identifying or Recognizing of banknotes.The so-called banknotes cards to count, is to identify X-ray scanning and statistical banknotes onto the imaging agent binding stripe imaging to determine the number of banknotes tied to the number and the bundle number. The so-called banknotes cards to identify, is identify the kind of the RMB in the bag. The main content includes the structural design and improvements, the development of imaging agent, the banknotes image preprocessing, the notes image recognition and counting in the ideal and actual condition, the actual conditions of the judgment of the banknotes.The banknotes image preprocessing mainly includes the extraction of target stripe. For the characteristics of the same banknotes and mixed banknotes, comparing of several common image thresholding method, this paper is based on the imaging characteristics of the target stripe adaptive threshold method. This method is simple and the results are very good for mixed banknotes.For the part of cunting, it can recognize and count accurately by counting the number of the target stripe centroids in the ideal condition. But in actual working conditions, Because of the emergence of tied stripe、the position of the tied stripe and the missing of the tied stripe with lacking of the imaging agent, the system can’t count the actual number of bundles. In response, according to the characteristics of the image itself, This paper presents an image centroid distance between adjacent target stripe and the edge of the image notes outline the combination of multi-information fusion algorithm, to make up for missing the stripe, so the results are still accurate count.For the part of the recogition of the banknotes,mixed and nomixed of banknotes both can be judged. According to characteristics of banknotes and X ray imaging principle, For different thickness of banknotes it can recognize the kind of the banknotes by grey value; For the same thickness of the banknotes it can identify the kind of banknotes by calculating the average length of bundles,To sum up,it can identify the banknotes through three parameters (gray value,the length of the banknotes and coordinates)which ensure the accuracy.