阅读说明：本技术 一种图像识别点钞机 (Image recognition currency count machine ) 是由 李汝会 张通庆 何涛 于 2018-06-29 设计创作，主要内容包括：一种图像识别点钞机,属于图像识别技术领域,是主要由传感器组、低速AD、CIS数据采集模块、高速AD、逻辑电路、DSP模块、计数器、托钞板和捻钞收钞控制电机组成的,其特征在于：所述传感器组和低速AD相连接,CIS数据采集模块和高速AD相连接,低速AD和高速AD均与逻辑电路相连接,DSP模块与逻辑电路相连接,计数器和托钞板均与逻辑电路相连接,捻钞收钞控制电机设置在逻辑电路输出端,本发明的图像识别点钞机,更体现了数字化、智能化,使用的图像识别方法,是通过图像二值化,目标表达和描述,模板匹配,并运用数学形态学中的一些基本运算实现普通票面面额识别,用于点钞机中,可以实现对不同版本不同面额纸币进行“自由混点”。(The utility model provides an image recognition counting currency machine, belongs to image recognition technology field, is mainly by sensor group, low-speed AD, CIS data acquisition module, high-speed AD, logic circuit, DSP module, counter, hold in the palm the paper money board and twist with fingers paper money and receive the control motor and constitute, its characterized in that: the image recognition currency count machine of the invention further embodies digitization and intellectualization, and the used image recognition method is realized by image binarization, target expression and description and template matching and by using some basic operations in mathematical morphology, and can realize 'free mixing point' on different versions of paper money with different denominations.)

1. The utility model provides an image recognition counting currency machine, is mainly by sensor group, low-speed AD, CIS data acquisition module, high-speed AD, logic circuit, DSP module, counter, holds in the palm the paper money board and twists with fingers paper money and receive the control motor and constitute, its characterized in that: the sensor group is connected with the low-speed AD, the CIS data acquisition module is connected with the high-speed AD, the low-speed AD and the high-speed AD are both connected with the logic circuit, the DSP module is connected with the logic circuit, the counter and the bill supporting plate are both connected with the logic circuit, and the bill twisting and collecting control motor is arranged at the output end of the logic circuit.

2. An image recognition banknote counter according to claim 1 wherein: the CIS data acquisition module adopts a Contact Image Sensor (CIS) SV255A4HV, and can finish feature selection and feature extraction of paper money so as to finish the identification of denomination and authenticity; high speed A/D converter TLC5510 was used.

3. An image recognition banknote counter according to claim 1 wherein: the DSP module adopts TMS320C6203, and the image acquisition speed can reach nearly 3000 lines/second.

Technical Field

The invention relates to an image recognition currency count machine, and belongs to the technical field of image recognition.

Background

The appearance of the cash register can effectively improve the working efficiency of cash business, realize the automation of cashier business, liberate cashiers from the boring and tasteless cash counting work, reduce the contact chance of the cashiers and the cash on the other hand, reduce the harm of various germs attached to the surface of the coin to the body of the cashiers to a certain extent, and improve the sanitary condition of the working environment. In addition, the mode of combining the mechanical anti-counterfeiting and the manual anti-counterfeiting of the cash register can effectively reduce the circulation of counterfeit coins and bring convenience to the stable development of economy, the maintenance of financial order and the life of people in China, the cash register is the most widely used cash processing equipment used by all financial systems, the market demand is the largest, and banks are provided with the most electronic equipment and also widely used in occasions with larger cash processing capacity, such as markets, restaurants and the like. Therefore, the currency counter is indispensable in the daily business of banks, and the usage amount of the currency counter is increased year by year. The invention develops an image recognition currency count machine.

Disclosure of Invention

Aiming at the defects, the invention provides an image recognition currency count machine.

The invention is realized by the following technical scheme: the utility model provides an image recognition counting currency machine, is mainly by sensor group, low-speed AD, CIS data acquisition module, high-speed AD, logic circuit, DSP module, counter, holds in the palm the paper money board and twists with fingers paper money and receive the control motor and constitute, its characterized in that: the sensor group is connected with the low-speed AD, the CIS data acquisition module is connected with the high-speed AD, the low-speed AD and the high-speed AD are both connected with the logic circuit, the DSP module is connected with the logic circuit, the counter and the bill supporting plate are both connected with the logic circuit, and the bill twisting and collecting control motor is arranged at the output end of the logic circuit.

The CIS data acquisition module adopts a Contact Image Sensor (CIS) SV255A4HV, and can finish feature selection and feature extraction of paper money so as to finish the identification of denomination and authenticity; high speed A/D converter TLC5510 was used.

The DSP module adopts TMS320C6203, and the image acquisition speed can reach nearly 3000 lines/second.

The invention has the advantages that: the image recognition currency count machine of the invention embodies digitalization and intellectualization, the used image recognition method realizes the identification of the denomination of the common ticket by image binaryzation, target expression and description, template matching and some basic operations in mathematical morphology, and can realize the 'free mixed point' of the currency with different denominations of different versions when being used in the currency count machine.

Drawings

FIG. 1 is a schematic diagram of the system of the present invention.

In the figure, the device comprises a sensor group 1, a sensor group 2, a low-speed AD, a low-speed CIS data acquisition module 4, a high-speed AD, a high-speed logic circuit 5, a logic circuit 6, a DSP module 7, a counter 8, a bill supporting plate 9 and a bill twisting and collecting control motor.

Detailed Description

The utility model provides an image recognition counting currency machine, is mainly by sensor group 1, low-speed AD2, CIS data acquisition module 3, high-speed AD4, logic circuit 5, DSP module 6, counter 7, hold in the palm paper money board 8 and twist with fingers paper money and receive paper money control motor 9 and constitute, its characterized in that: the sensor group 1 is connected with a low-speed AD2, the CIS data acquisition module 3 is connected with a high-speed AD4, the low-speed AD2 and the high-speed AD4 are both connected with a logic circuit 5, the DSP module 6 is connected with the logic circuit 5, the counter 7 and the bill supporting plate 8 are both connected with the logic circuit 5, and the bill twisting and collecting control motor 9 is arranged at the output end of the logic circuit 5.

The CIS data acquisition module 3 adopts a Contact Image Sensor (CIS) SV255A4HV, and can complete feature selection and feature extraction of paper money so as to complete denomination and authenticity identification; high speed A/D converter TLC5510 was used.

The DSP module 6 adopts TMS320C6203, and the image acquisition speed can reach nearly 3000 lines/second.

In the system, on one hand, the system needs to complete more functions and needs certain control capability, and on the other hand, the speed required in the aspects of data acquisition and processing is higher. The DSP has the characteristics of high data acquisition and processing speed and strong functions, but the control capability of the DSP on peripheral equipment is weak, if all the functions are completed by one digital signal processor, the requirements are difficult to achieve, and the performance of the whole system is influenced, so that the structure of the D SP + CPLD is adopted. The logic circuit is designed by adopting the CPLD, so that the design is simple, the modification is convenient, the control is flexible, and the debugging is easy. The CIS is used for completing data acquisition of the banknote image, and the high-speed A/D is used for completing an analog-to-digital conversion function of the image data. The infrared sensor, the purple light sensor and the magnetic sensor are respectively used for anti-counterfeiting detection, and the low-speed A/D is used for completing the quantification of data acquired by the anti-counterfeiting sensor. The external data memory is used for storing the whole image data acquired by the CIS, temporarily storing the data in an intermediate state in the image data preprocessing and identifying process, and also storing characteristic quantity matching template data which is transferred from the FALSH after power-on, so that a larger storage space is required. The image scanning and the image recognition are carried out in sequence, a certain time is reserved between two adjacent image scanning for the recognition of the previous image, and the interval between the two adjacent image scanning is generally required to be a part more than the recognition time so as to ensure that the recognition, the communication after the recognition and the hardware control can be completed in time. The interval time between the first recognition and the next image scanning is controlled by a clutch, if the interval between two paper currencies is too small, the recognition outputs a continuous or incomplete signal, so the paper currency interval needs to be effectively pulled, and on the other hand, the specified rotating speed needs to be ensured, so that the whole recognition process of the paper currencies needs to have strong real-time performance, otherwise, the recognition result loses significance. For image data with the gray scale of 0-255, each pixel point is stored by 8bits, namely each pixel point corresponds to one byte. The resolution was 67DPI (dots per inch) in both length and width directions; the width of the paper money is 42-92 mm, and the length of the paper money is 110-170 mm; the paper money counter basically requires 600-1000 pieces per minute, namely 10 pieces per second. The face denomination recognition method based on the mathematical morphology realizes the face denomination recognition of the common ticket by image binarization, target expression and description and template matching and some basic operations in the mathematical morphology, is used in a cash counter, can realize the 'free mixing point' of different versions of different face notes, and can recognize non-notes mixed in the notes at the same time, so that the notes have a certain counterfeit identification function.

It will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in the embodiments described above without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims.