WO2003081542A1

WO2003081542A1 - A multiple testing apparatus for bank notes having high-technology

Info

Publication number: WO2003081542A1
Application number: PCT/CN2003/000211
Authority: WO
Inventors: Dongshan Bao
Original assignee: Dongshan Bao
Priority date: 2002-03-25
Filing date: 2003-03-24
Publication date: 2003-10-02
Also published as: AU2003236081A1

Abstract

A testing apparatus for bank notes include a pick-up head(A), an image processing unit(B), a matching process algorithm unit(C), a threshold analysis algorithm and resulting display unit(D), a unit of continuous input bank notes and a clock synchronization unit(E), in which the pick-up head(A), the image processing unit(B), the matching process algorithm unit(C), the threshold analysis algorithm and resulting display unit(D), the unit of continuous input bank notes are all connected with the clock synchronization unit(E), the pick-up head(A) is connected with the image processing unit(B), the image processing unit(B) is connected with the matching process algorithm unit(C), and the matching process algorithm unit(C) is connected with the threshold analysis algorithm and resulting display unit(D). The testing apparatus for bank notes is used for precisely detecting counterfeit bank notes.

Description

Compound high-tech banknote detector

Technical field

The present invention relates to the following four fields: the field of optics and electronics, the field of electronic information processing technology, the field of fuzzy mathematics and electronic information processing technology, and the field of machinery.

Background technique

At present, the anti-counterfeiting technology of banknotes in various countries is mainly concentrated on special paper (including plastic paper) and anti-counterfeiting plates and printing. At present, the large denominations of the euro, the US dollar and the new version of the RMB mainly use color-changing inks and holographic technology. This is the most difficult feature of counterfeiters, and even if they work hard, errors will occur (such as the imitation of color-changing inks is extremely difficult, even with modern electronic color matching technology, it is difficult to imitate the color-changing effects of all different angles). At present, the world's existing banknote detection technology mainly uses the human eye or optical or chemical aids to identify optical images (its essence is optical image recognition), which is more accurate than modern digital image processing technologies (essentially computer signal processing). There is a great difference in the observation effect, and it is difficult to realize automatic continuous operation.

Summary of the Invention

The purpose of the present invention is to provide a compound high-tech banknote detector which can use high-tech technology and traditional technology to accurately determine the authenticity of banknotes.

The purpose of the present invention is achieved as follows: A compound high-tech banknote detector includes a banknote counting device, and further includes a camera (A), an image processing unit (B), a matching processing algorithm (C), a threshold analysis algorithm unit, and The result shows (D), continuous banknote counting unit (F) and clock synchronization control unit (E); where the camera (A) is connected to the image processing unit (B) and the clock synchronization control unit (E), and the image processing unit (B) Connected to the matching processing algorithm (C) and the clock synchronization control unit (E), and the matching processing algorithm (C) is connected to the threshold analysis algorithm unit and the result display (D), the clock synchronization control unit (E), and the threshold analysis. The algorithm unit and result display (D) are connected to the matching processing algorithm (C) and the clock synchronization control unit (D), and the continuous money counting unit (F) is connected to the clock synchronization control unit (E).

——The number of said cameras (A) is 2-4.

——The image processing unit (B) is composed of an acquisition image buffer area and a plastic sampling software storage area (4 MB RAM), a plastic sampling processing unit, an I / O interface or RS232 and a bus.

——The matching processing algorithm (C) consists of a graphic matrix buffer area and processing software storage area (4 MB RAM), a pre-stored graphic matrix storage area (2 MB ROM), which accepts B module data. I / O or RS232, matching processing unit, I / O or RS232 for outputting data to D module, and bus.

——The threshold analysis algorithm unit and the result display (D) are the graphics matrix buffer area and processing software storage area (4 MB RAM), pre-stored threshold vector storage area (2 MB ROM), and accept C module data I / O or RS232, threshold processing and result display unit, I / O outputting data to computer terminal Or RS232 and bus composition.

The present invention uses multi-directional high-tech means to continuously and simultaneously simulate human eyes, or the artificial person recognizes these features by using auxiliary means (such as light identification watermarking, changing angles to identify features of color-changing inks and holographic images, and detecting points and Line, detection of holographic images by polarized light sources), not only greatly improves identification efficiency and reduces labor costs, but also the identification accuracy is several orders of magnitude higher than the human eye and existing banknote detection equipment (such as color-changing inks and holographic images at different fixed angles). Color features, if image processing technology is used, the digitization level of their features will be nearly a thousand times higher than that of human eyes and polarized light detection equipment), counterfeit effects that are difficult to observe by human eyes and existing equipment, and detection in composite high-tech banknote detectors The means will be easily exposed.

The present invention is the first time in the world to use digital information processing methods to test banknotes, and its design ideas are derived from the terrain matching technology used by similar strategic missiles or cruise missile end guidance, and similar strategic early warning radars or airborne attack radars are used to determine alert targets And moving target determination technology used for attack targets. The technical solution is to convert the color-changing ink image and holographic optical image that are most difficult for counterfeiters in various currencies (such as large face value USD, RMB, Euro, etc.) from optical images into digital images that can be processed by computers, Feature matching images of banknotes and standard banknotes are processed for image matching, and then fuzzy mathematical methods are used to set analysis and judgment rules (algorithms) to eliminate the interference of system errors, environmental errors and human errors, and focus on accurately determining the authenticity of banknotes.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of the overall structure of the present invention;

2 is a schematic diagram of collecting information on a banknote to be inspected by a camera according to the present invention;

FIG. 3 is a schematic diagram of graphic processing of an image processing unit (B) of the present invention; FIG.

4 is a structural diagram of a matching processing algorithm (C) of the present invention;

5 is a threshold analysis algorithm unit and a result display (D) structure diagram of the present invention;

6 is a clock synchronization control logic diagram of a clock synchronization control unit (D) of the present invention;

Figure Ί is a schematic diagram of the product structure of the present invention;

In Figure 7, 71—the information processing unit circuit board; 72—the computer interface and the power supply interface; 73—the parallel light source; 74—the banknote into the slot; 75—the banknote counter; 76—the banknote out of the slot; 77—the transmission device; 78— Focus on the camera.

FIG. 8 is a perspective structural view of a money counter in the present invention.

In Fig. 8, 1 banknote device; 11 banknote inlet; 111 cam; 112-sensing switch; 12 丄 banknote outlet; 121-impeller; 13-drive motor; 131-first belt pulley; 14-belt; 15—the second batch of pulleys; 151—the second gear; 16—the roller; 161—the first gear; 17—the currency channel; 18—the impeller motor.

detailed description

1. From A, convert existing low-cost cameras on the market (such as domestic 36-megapixel cameras) into several aggregation cameras, which are placed at different positions and different angles for different currencies, and holograms of banknotes Optical images and color-changing ink feature points are recorded, and the feature images to be identified are converted into basic digital signals (USB signals or A / D-converted RGB signals) processed by a controllable computer. The schematic is shown in Figure 2. 2. From B, through graphic processing of shaping and sampling techniques, form a vector signal or graphic matrix that retains the image features of the banknote feature points and compresses the data amount (where the vector signal or graphic matrix is defined according to different currencies and different image characteristics It is one of the key points of the present invention), and a schematic diagram of specific graphic processing is shown in FIG. 3. Through the shaping process of the graphics processing software, the unnecessary edges are first removed, so that the vector to be processed becomes the required (based on different currencies and denominations) standard matrix. For example, a 100 yuan banknote is formed from a 164X 220 dot matrix (36 thousand pixels) into an 81 X 181 dot matrix. After sampling, it becomes a 41 X 91 dot matrix.

3. From C, set a series of specific algorithms to perform matching processing on timely collected and pre-stored vector signals or graphic matrices (here is a process of initializing pre-storage for different currencies, such as USD, EUR, RMB, etc.) . This series of algorithms is another key point of the present invention, and its essence is to compare the difference between the RGB signal of the banknote under test and the standard banknote with a digital recognition ability that is thousands of times or even tens of thousands of times higher than the resolution of the human eye or existing recognition equipment. , That is, the difference in the color separation levels of red, green, and blue. Using software programmed by a specific matching algorithm will generate an error signal that can be processed by the threshold. The schematic diagram is shown in Figure 4. As can be seen from FIG. 4, in each clock cycle t, the graphics matrix buffer area first stores the RGB feature vector output by the B module, and this vector is matched with the pre-stored RBG feature vector pre-existing in the graphics matrix storage area (ROM) as a match. After processing (matrix subtraction, etc.), it is refreshed into the buffer area and input to the D module through the output interface.

4. From D, set an algorithm, that is, a threshold determination rule, perform a threshold analysis on the results of the matching process (comprehensive analysis and determination of the result matrix elements), and then display the determination conclusion. It should be emphasized that this system is oriented to the actual application environment. It is precisely because the actual application environment will face environmental errors such as banknote positioning, banknote old and new, flat, folded, temperature difference, humidity, and human errors, as well as system errors of this system. Make the result of the matching process have a normal error range, that is, even the result of the matching process of real banknotes is error. How to define this range with fuzzy mathematics is the goal of threshold analysis. Using the software compiled by the specific threshold analysis algorithm to process the error signal, the final recognition result of the corresponding count will be generated and displayed on the terminal screen through a display software. The schematic diagram is shown in Figure 5. It can be seen from FIG. 5 that in each clock cycle t, the buffer area of the graphic matrix after the matching process is first stored in the error (RGB) feature vector output by the C module, and this vector is compared with the pre-stored threshold vector, and then blurred. The discriminative software of the mathematical method processes it to obtain the final recognition result, which is driven by the display software and output to the computer terminal.

5. F, extract the system's synchronization trigger signal and counting signal from the continuous counting unit, and supply it to the D and E modules.

6. The E module provides synchronous control to ensure the reliable identification of each banknote. The corresponding implementation method is shown in the clock synchronization control logic (Figure 6). The specific implementation is implemented by Intel 586-133 and z stored in a memory (2 MB RAM). The display software inside is complete. The time T from the focusing camera to the end of the threshold processing is based on a reasonable design (selection of fast processing devices), which is of course less than the interval t of the continuous counting unit of the banknote counting unit, and the role of clock synchronization control and overall control is to ensure the signal of the system The processing part and the mechanical rotation part, and the absolute display of the results are synchronized.

7. The product schematic diagram of the compound high-tech banknote detector is shown in Figure 7. Connect the power supply and computer, and the compound high-tech banknote detector can start to work. First set the face value of the banknote through the computer and keyboard, and then put the banknote to be tested through the banknote into the slot. After the banknote counting is completed, remove the banknote from the banknote out of the slot, and then put the opposite side of the banknote into the banknote into the slot. There is a reverse counting function). The computer display will show the counting result and the corresponding banknote checking result.

Description of specific embodiments

1. The present invention focuses on new thinking methods, working principles and processing modes, and also has economic considerations. Therefore, some low-grade devices (software environment is Windows 98) 3 years ago are selected to illustrate that its principles and structure are practical. feasible. Therefore, the current prototype of the compound high-tech banknote detector is slightly different from the final product in the future large-scale promotion.

2. The processing unit of each module of B, C, D, and E of the compound high-tech banknote detector uses the same CPU (Intel 586—133 MHz). At present, the price of Intel 586-133 MHz is less than 50 yuan. If you use Intel 586-166 MHz, Intel 586-200 MHz or Intel 586-233 MHz, the speed will be faster, and the price is not much expensive.

3. Based on the current design pattern of the prototype, the final product in the future large-scale promotion has two trends. One is the direction of the software structure. In fact, from Figure 1, Figure 3, Figure 4, and Figure 5, we can easily find that the C, D, and E modules are a computer together. Considering that the CPU speed of computers will become faster and faster in the future (currently Intel? 6111:; 1 and-] ^ has exceeded 2. 2 & ¾), the storage capacity is getting larger and larger (and its corresponding PC's memory is larger than 256 MB) and other considerations For customers who have advanced cabinet terminals, it is not necessary to set up hardware for the compound high-tech banknote detector. With the cabinet terminal hardware, only the complete set of software for the required functions is stored in the PC as the cabinet terminal, and then connected The synchronization signal line and trigger signal line of the money counting unit can work.

The second is the direction of the hardware structure. For customers who are not convenient to use cabinet terminals or lower-level cabinet terminals, it is a good choice to use low-cost (such as outdated) CPUs and small-capacity storage chips, because these two types of chips will be very cheap in the future. In large-scale production, you can even produce dedicated chips or use programmable controllers. Of course, in this case, a simple liquid crystal display unit should be considered.

4. At present, due to funding problems, the aggregation camera used by the A module can only be modified to achieve it. However, in mass production, custom dedicated focusing cameras will be used.

5. The mechanical transmission part of the banknote counting unit of the compound high-tech banknote detector is a preferred intermediate product purchased from the market, so it will not be described in detail in FIG. 7. There are two main choices, one is American products, which are characterized by high quality and high price; the other is domestic products, which are low in price but second in quality. The trade-off will depend on the customer (eg domestic use or export).

6. According to Figure 1, the A, B, C, D, and E modules are combined to form a single banknote processing system for counter operators and general cashiers.

7. According to the circulation of RMB, USD, Euro, etc., the compound high-tech banknote detector will have special models for different currencies (mainly different recording structures); but for the models of the same currency, it can be set in advance Party Method to process banknotes of different denominations.

8. For the same currency, because there are positive and negative points (the front and back points have been considered in the structural design), during the specific operation, the front point will be clicked once, and then the reverse point will be clicked once, but the display software will display according to the corresponding number of counts. Money check result.

9. The circuit board as the information processing unit can also be inserted on the PC motherboard as a cabinet terminal.

Claims

Claim

1. A compound high-tech banknote detector, comprising a banknote counting device, characterized in that it further comprises a camera (A), an image processing unit (B), a matching processing algorithm (C), a threshold analysis algorithm unit and a result display ( D), the continuous money counting unit (F) and the clock synchronization control unit (E); wherein the camera (A) is connected to the image processing unit (B) and the clock synchronization control unit (E), and the image processing unit (B) is matched with The processing algorithm (C) and the clock synchronization control unit (E) are connected, and the matching processing algorithm (C) is connected to the threshold analysis algorithm unit and the result display (D), the clock synchronization control unit (E), the threshold analysis algorithm unit and The results show that (D) is connected to the matching processing algorithm (C) and the clock synchronization control unit (D), and the continuous money counting unit (F) is connected to the clock synchronization control unit (E).

2. The compound high-tech banknote detector according to claim 1, wherein the number of said cameras (A) is 2-4.

3. The compound high-tech banknote detector according to claim 1, characterized in that: said image processing unit (B) is composed of a captured image buffer area and a plastic sampling software storage area (4 MB RAM), and plastic sampling processing Unit, I / O interface or RS232 and bus.

4. The compound high-tech banknote detector according to claim 1, characterized in that: said matching processing algorithm (C) is a graphic matrix buffer area and a processing software storage area (4 MB RAM) collected in a timely manner. , Pre-stored graphics matrix storage area (2 MB ROM), I / O or RS232 receiving module B data, matching processing unit, I / O or RS232 outputting data to D module, and bus configuration.

5. The compound high-tech banknote detector according to claim 1, characterized in that: the threshold analysis algorithm unit and the result display (D) are a graphic matrix buffer area and a processing software storage area after matching processing ( 4 MB RAM), pre-stored threshold vector storage area (2 MB ROM), I / O or RS232 that accepts C module data, threshold processing and result display unit, I / O or RS232 that outputs data to the computer terminal, and bus configuration.