CN101290689A - Banknote quality checking and recognition method - Google Patents

Abstract

The invention provides a method for detecting and identifying the quality of a bank note, relating to a method for detecting and identifying the quality of the bank note, which can be applied to various currency count machines, sorting machines and fields for identifying the quality of the bank note. A contact image sensor CIS is used to detect the gray level of a feature region of the bank note in order to overcome aging of a light source and environment effect, a gray signal is converted into a corresponding digital quantity through an A/D converter, and is performed with feature extraction, arrangement and combination to realize the identification to bank notes with different qualities. The method is characterized by detecting changes of compensation coefficients in the CIS feature region to find a rule to the changes of different compensation coefficients; and then, adopting a specific algorithm to correct the compensation coefficients in the CIS feature region to realize the correction of CIS output, thereby achieving the aim of accurately identifying the quality of the bank note.

Description

A kind of detection of banknote quality and recognition methods

Technical field

The present invention relates to a kind of detection and recognition methods of banknote quality, be applicable to various paper money counters, cleaning-sorting machine and the field that the bank note quality is discerned.

Technical background

The detection of banknote quality and recognition technology are very important technology in the financial machine and tool industry.Financial industry is proposing the requirement of " circulation of RMB certificate will reach seventy percent new more than " under the support of country, therefore the cash currency letter sorting that requires to steam again of financial department such as bank is three levels of audit quality: the ATM banknote, circulation banknote and damaged banknote, and realize that the gordian technique of quality letter sorting is exactly the detection and the recognition technology of banknote quality (new and old).Adopt the detection of quality and the banknote that recognition technology is carried out treatment classification, not only satisfied simultaneously ATM with paper money and circulating banknote again to the different requirements of banknote quality, also reached the purpose of rejecting and reclaim worn coin, the speed that cash drops into circulation again be can accelerate, the security and the reliability of throwing in coin again improved greatly.

Summary of the invention

The present invention adopts contact-type image sensor CIS (Contact Image Sensor) can detect gray scale (also the claiming brightness) signal of banknote characteristic area, convert this grey scale signal to the corresponding digital amount by A/D converter, and it is carried out the identification that feature extraction, permutation and combination just can realize different quality banknotes.But, because the CIS illuminating source is along with Effect of Environmental such as aging and electromagnetic interference (EMI) take place inevitably in the passing of service time, it is inaccurate to cause sensor to read in signal, and this is the detection of banknote quality and the place of the key issue that recognition methods need solve.

Test is found after deliberation, be subjected to the sensor of such environmental effects to read in grey scale signal rising or the reduction that the inaccurate performance of signal is not whole output, but produce more shallow or darker strip shade in the reconstructed image, this has just caused difficulty to compensation, but its regularity is arranged: the some output variation of identical penalty coefficient has consistance, can come according to the variation of characteristic area penalty coefficient the penalty coefficient of workspace is revised, thereby can realize the correction that workspace CIS exports.

The technical solution used in the present invention is: adopt contact-type image sensor CIS, signal amplifier, high-speed a/d converter, first-in first-out buffer memory FIFO, steering logic CPLD and high speed digital signal processor DSP to form.One end of the detection zone of CIS got 25mm stick white specimen page, and the parcel black tape to be preventing printing opacity that can extract and characteristic area that detecting sensor output penalty coefficient changes thereby formed, remaining part has just become to detect the workspace (as shown in Figure 1) of identification.Detect by penalty coefficient variation the CIS characteristic area, find out the rule that different penalty coefficients change, adopt specific algorithm that the penalty coefficient of CIS workspace is proofreaied and correct then, thereby reach effective correction, realize the accurate detection and the identification of banknote quality CIS output.This compensation method may further comprise the steps:

A. measure the bright output Vpi and dark output Vdi of every bit, obtain maximum bright output Vpmax and minimum dark output Vdmin (as shown in Figure 4);

B. calculate penalty coefficient β i=(Vpmax-Vdmin)/(Vpi-Vdi), thus the compensation after output valve Wi=β i * (Vi-Vdi), wherein Vi is the actual output voltage value of each point.

As time goes on, because the CIS illuminating source takes place aging inevitably and Effect of Environmental such as electromagnetic interference (EMI), it is inaccurate to cause sensor to read in signal, need compensate CIS output, at this moment:

C. the bright output of measurement features district every bit and dark output, obtain characteristic area penalty coefficient δ i and with the variation difference DELTA δ i (Δ δ i=δ i-β i) of former characteristic area penalty coefficient β i, set up the functional relation of characteristic area former penalty coefficient β i and its variation difference DELTA δ i:

$F\left(\mathrm{\βi}\right)\↔\mathrm{\Δ\δi}$

D. change functional relation according to the characteristic area penalty coefficient and adjust the workspace penalty coefficient.

The detection of banknote quality and the characteristics of recognition methods: the inventive method is reasonable in design, be applicable to that various paper money counters, cleaning-sorting machine etc. to the field that the bank note quality detects and discerns, can effectively compensate the output bias of contact-type image sensor CIS and proofread and correct.

Description of drawings

The invention will be further described below with reference to accompanying drawing.

Fig. 1 is the synoptic diagram of the contact-type image sensor CIS that adopts of the present invention.

Fig. 2 is the synoptic diagram that CIS working sensor district of the present invention and characteristic area distribute.

Fig. 3 is the diagrammatic cross-section in CIS sensor characteristics of the present invention district.

Fig. 4 is the block scheme of banknote quality detection of the present invention and recognition methods.

Fig. 5 is the DSP compensation data algorithm flow chart of banknote quality detection of the present invention and recognition methods.

Fig. 6 is the constitutional diagram of the real output value of CIS sensor dialogue specimen page of the present invention.

Fig. 7 is the bright output of CIS sensor dialogue specimen page of the present invention and the constitutional diagram that penalty coefficient distributes.

Fig. 8 is the constitutional diagram of the compensation output valve of the CIS sensor dialogue specimen page after of the present invention the working long hours.

Fig. 9 is the constitutional diagram of the CIS sensor compensation output valve of dialogue specimen page after the penalty coefficient correction after of the present invention the working long hours.

Wherein: 101, contact-type image sensor CIS; 102, the light tight adhesive tape of black; 103, white specimen page; L, detection zone; L0, workspace; L1, characteristic area; Banknote in A, the motion; 201, signal amplifier; 202, high-speed a/d converter; 203, dual-port pushup storage FIFO (1); 204, dual-port pushup storage FIFO (2); 205, programmable logic controller (PLC) CPLD; 206, digital signal processor DSP; 207, LCD MODULE LCM; 301, beginning; 302, the bright output in acquisition characteristics district Vpi; 303, the acquisition characteristics district secretly exports Vdi; 304, calculated characteristics district penalty coefficient δ i; 305, the characteristic area penalty coefficient changes ∑ | δ i-β i| 〉=Δ c? 30 δ, obtain characteristic area penalty coefficient changes delta δ i; 307, the workspace penalty coefficient is adjusted δ i=β i+ Δ δ i; 308, finish.

Specific embodiments

With reference to the accompanying drawings 1～3, the schematic appearance of contact-type image sensor CIS: imageing sensor CIS 101 is divided into workspace and characteristic area two parts, the length L 1 〉=25mm of characteristic area; Characteristic area is by white specimen page 103, light tight adhesive tape 102 parcels of black; The characteristic area L1 output that the changing value of coefficient comes compensation work district L0 that affords redress.

With reference to the accompanying drawings 4, the detection of banknote quality and recognition methods: the banknote in the motion is when contact-type image sensor CIS 101, banknote quality is converted into the correspondent voltage signal, and under the control of programmable logic controller (PLC) CPLD 205, voltage signal is read from sensor; Its signal peak reaches the signal voltage that can detect to this voltage signal through signal amplifier 201 biasings with after amplifying, the signal voltage that can detect is 1～5V, convert digital signal to by high-speed a/d converter 202 and deposit among dual-port pushup storage FIFO (1) 203 and the dual-port pushup storage FIFO (2) 204, two FIFO are used alternatingly; Programmable logic controller (PLC) CPLD 205 is responsible for control and coordinates bus timing, and will sending digital signal processor DSP 206 to after the data rough handling among the dual-port pushup storage FIFO, LCD MODULE LCM 207 is used for the result of display digit signal processor to banknote quality identification.

With reference to the accompanying drawings 5, banknote quality detects the DSP compensation data algorithm flow chart with recognition methods: the bright output Vpi by acquisition characteristics district dialogue specimen page reaches the penalty coefficient δ i that dark output Vdi calculates characteristic area, and judge the overall variation amount ∑ of δ i to former penalty coefficient β i | whether δ i-β i| exceeds a certain threshold value Δ c: if exceed, then the variable quantity according to the characteristic area penalty coefficient compensates the workspace; If do not exceed, then need not compensate.

With reference to the accompanying drawings 6～9, bright output Vpi and dark output Vdi by the survey sensor every bit obtain maximum bright output Vpmax and minimum dark and export Vdmin, can calculate penalty coefficient

βi＝(Vpmax-Vdmin)/(Vpi-Vdi)，

Thereby the output valve after the compensation

Wi＝βi×(Vi-Vdi)。

When after after a while, because factor affecting such as light source led is aging, environment electromagnetics interference need compensate CIS output, at this moment:

The bright output of measurement features district every bit and dark output, obtain characteristic area penalty coefficient δ i and with former spy

Levy the variation difference DELTA δ i (Δ δ i=δ i-β i) of district penalty coefficient β i, it is former to set up characteristic area

The functional relation of penalty coefficient β i and its variation difference DELTA δ i:

$F\left(\mathrm{\βi}\right)\↔\mathrm{\Δ\δi}$

Thereby change functional relation according to the characteristic area penalty coefficient and adjust the workspace penalty coefficient.

The output terminal of sensor (101) and signal amplifier (201) input end join, signal amplifier (201) output terminal and A/D converter (202) join, convert digital signal to by A/D converter (202), this digital signal is temporarily stored into dual-port pushup storage FIFO (203 earlier, 204), to make things convenient for programmable logic controller (PLC) CPLD (205) to read, data-signal among the FIFO is read by CPLD and carries out delivering to digital signal processor DSP (206) after the preliminary Filtering Processing, DSP can will discern banknote quality, and the result of identification delivers to LCD MODULE LCM (207) and shows.

Claims (8)

1, a kind of detection of banknote quality and recognition methods, form by contact-type image sensor CIS (101), signal amplifier (201), high-speed a/d converter (202), dual-port pushup storage FIFO (203,204), programmable logic controller (PLC) CPLD (205), digital signal processor DSP (206), LCD MODULE LCM (207); It is characterized in that: the output terminal of sensor (101) and signal amplifier (201) input end join, signal amplifier (201) output terminal and A/D converter (202) join, convert digital signal to by A/D converter (202), this digital signal is temporarily stored into dual-port pushup storage FIFO (203 earlier, 204), to make things convenient for programmable logic controller (PLC) CPLD (205) to read, data-signal among the FIFO is read by CPLD and carries out delivering to digital signal processor DSP (206) after the preliminary Filtering Processing, DSP can will discern banknote quality, and the result of identification delivers to LCD MODULE LCM (207) and shows.

2, the detection of banknote quality according to claim 1 and recognition methods, it is characterized in that by detecting the variation of contact-type image sensor CIS (101) characteristic area L1 penalty coefficient, find out the rule that different penalty coefficients change, come the output calibration of correction work district L0 penalty coefficient and then realization sensor, thereby realize accurate identification banknote quality.

3, the detection of banknote quality according to claim 1 and recognition methods is characterized in that contact-type image sensor CIS (101) is divided into workspace L0 and characteristic area L1; Characteristic area L1 wraps up with white specimen page (103) and the light tight adhesive tape of black (102) outside.

4, the detection of banknote quality according to claim 1 and recognition methods is characterized in that length 〉=25mm of characteristic area L1.

5, the detection of banknote quality according to claim 1 and recognition methods is characterized in that detected characteristics district penalty coefficient variable quantity sum ∑ | the magnitude relationship of δ i-β i| and a certain threshold value Δ c.

6, the detection of banknote quality according to claim 1 and recognition methods is characterized in that the characteristic area penalty coefficient variation identical with the workspace has consistance.

7, the detection of banknote quality according to claim 1 and recognition methods is characterized in that programmable logic controller (PLC) CPLD (205) provides contact-type image sensor CIS (101), high-speed a/d converter (202), dual-port pushup storage FIFO (203,204), required sequential control, the address decoding of digital signal processor DSP (206) co-ordination.

8, the detection of banknote quality according to claim 1 and recognition methods is characterized in that dual-port pushup storage FIFO (1) (203), dual-port pushup storage FIFO (2) (204) alternately receive the data of high-speed a/d converter (202).

Images