|
The bill is the main form of the general equivalent market , it plays an irreplaceable role in the social life of the people . Along with the rapid development of the market economy , many social problems began. You can see now the most vicious crimes related to cash , has brought great harm to the social order , and seriously disrupted the harmonious development of society and the majority of the people live a normal life . The number of banknotes is unique, any two real legitimate banknotes will never re- number , according to the nature of the bill number , registration number of banknotes banknotes before storage . In the circulation of money market automatic identification devices to provide a bill number , bill number identification data into computer processing , have the same number or the number does not exist , you can confirm the current circulation coins for the illicit circulation . This article describes a self-designed based on ARM and FPGA technology acquisition bill number , image acquisition system , a real-time acquisition of high - speed image information and image preprocessing ; collected using the Contact Image Sensor (CIS) CIS image signal after shading output compensation , second order filter , modulus conversion, binarization, pre - conditioning, is saved in the static synchronous memory (SRAM) , for further image processing for the ARM ; CIS sensor system timing signal by FPGA design , the ARM is responsible for the system program running and image recognition . The system will be collected to an image signal for image format conversion, the gradation of an image , the image binarization , normalized to the basic algorithm of a process , also studied the banknote recognition system key algorithms , including: p location, character segmentation , number tilt correction of the geometric method , the number of vertical and horizontal segmentation algorithm , number recognition , neural networks, artificial immune algorithm . The experiments show that the binarization image acquisition to clear ; recognition speed of up to 10 frames per second , the high reliability of the image recognition , and the results of this study can be generalized to the entry system of the number of securities , notes , and other related areas of character recognition system .
|