Study On Identification Method Of RMB Cash Registers Image

With the rapid development of China’s economy, the currency in circulation is alsoincreasing rapidly. Thus the needs of banks and financial companies for banknote sorter,cash regisisters and other financial equipment is growing as well. As the core techniques of abanknote sorter and other inspection equipment, image processing techniques to identify themodule level throughout the cash registers is of particularly importance.This paper firstly describes the issue of banknote image preprocessing. The thresholdvalue of banknote image is obtained utilizing iterative method, and the image is segmentedwith the background after being processed utilizing binarizing. The tilt angle is caculatedbased on boundary point coordinate after positioning the point of the boundary. To rotatethus correcting banknotes is processed according to the tilt angle. Secondly, afterpreprocessing the banknote image, the serial number, Braille, infrared and ultraviolet featureregion and version identification feature region are located. Finally, to achieve the variousbanknote recognition, structure-based number identification method already existed in thelaboratory is utiltized in terms of number identification; braille-based identification methodis uitilized to achieve denomination banknote recognition in terms of denomination banknoterecognition; the results to test magnetic signals is utilized in terms of discrimination in thedirection of how bill is placed on cash registers. Every recognition algorithm proposed inthis paper is of high speed, small footprint and high recognition rate.Banknote image recognition is an active topic in the field of pattern recognition inrecent years, which has a very broad application prospects. Thereby banknote sortertechnology developed by banks and other financial systems are playing an increasinglyimportant role. Therefore, there exists great research value and application value in this topic.Number identification, authenticity identification, denomination recognition and otherfunctions should be essential functions of cash registers, sorter and other financial machinery.In this stituation, this paper designed images identification method of RMB banknotesaccording to the images characteristics of the fifth set of RMB banknotes in circulationa,which is applied to actual device.In this paper, banknote images collected via the cash registers has been tested on PC.And algorithm program has been embedded to the DSP and for real-time test, which achieved a higher recognition rate. Test results idicated that the algorithm adopted in thispaper could meet real-time requirement.