CN105632015B - A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform - Google Patents
A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform Download PDFInfo
- Publication number
- CN105632015B CN105632015B CN201510992680.4A CN201510992680A CN105632015B CN 105632015 B CN105632015 B CN 105632015B CN 201510992680 A CN201510992680 A CN 201510992680A CN 105632015 B CN105632015 B CN 105632015B
- Authority
- CN
- China
- Prior art keywords
- image
- bill
- detected
- cloud platform
- sola
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
- G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
- G07D7/20—Testing patterns thereon
Abstract
The invention discloses a kind of bank money fingerprint characteristic anti false authentication method based on cloud platform includes the following steps: S1, the multispectral image of big bill of acquisition；S2 pre-processes the multispectral image of acquisition, obtains the sola bill image of standard, and the fingerprint characteristic of all sola bill images is stored to cloud platform；S3 acquires the image information of bill to be detected, in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage, is detected respectively to the image information of acquisition by double-colored fiber check and measure, phosphor pattern detection, the detection of infrared image blanking feature；S4 determines that bill to be detected is true, otherwise, it is determined that bill to be detected is false when three kinds of detections determine that bill to be detected is true.The drawbacks of carrying out authentication this method eliminates single bill authentication technique, effectively improves the accuracy of authentication.
Description
Technical field
The present invention relates to a kind of financial document anti false authentication methods more particularly to a kind of bank money based on cloud platform to refer to
Line feature anti false authentication method, belongs to anti-counterfeit and counterfeit distinguishing technical field.
Background technique
With the rapid development of the national economy, financial document using more and more extensive.But current financial document management,
It there is also some problems for the use of with false distinguishing, crime target has been directed pointing bank, financial ticket by the criminal in society
It happens occasionally according to swindle case, is resulted in significant economic losses to country.Existing financial document false distinguishing method relies primarily on people
Based on work qualitative analysis, the problems of be it is artificial identify intensity it is big, it is time-consuming long, and be easy to cause to miss due to fatigue or carelessness
Inspection.
Fiber is one of important anti-false sign of bill, and in different bill position of fibers random distribution it is different.
Therefore, in existing bill anti-counterfeit discrimination method, bill anti-counterfeit identification usually is carried out using extraction fiber characteristics.In ultraviolet light
Under irradiation, fiber has fluorescent effect, convenient for extracting anti-counterfeiting characteristic.But in this anti-fake discrimination process, the fibre of bill surface
Dimensional feature extraction is committed step and problem.Since the background of bill surface is extremely complex, including text, frame, fluorized marking
With shoulder etc., intensity profile range is very wide, and fiber target is smaller, although gray scale is substantially distributed in highlight bar, with
There is no apparent boundaries for background intensity profile, distinguish relatively difficult.And it requires in practice for same bill structure
Similar machine extracts fiber characteristics, and result is consistent, and this requires the adaptivity of feature extracting method and stability all
It is better.
To solve the above-mentioned problems, one kind is proposed in the Chinese invention patent of Patent No. ZL201110362933.1
Bill anti-counterfeit discrimination method based on fiber individualized feature, include the following steps: the bill images that video camera is got into
Row processing, obtains standardized bill images；Image filtering is carried out using maximum filter, by the multiclass pair in bill images
As being converted to two class object of shoulder background and fiber target；Using the two-dimensional entropy segmentation bill images of optimization,
Detect fiber target；The anti-counterfeiting characteristic of fiber target is extracted, anti-counterfeiting characteristic is in center-of-mass coordinate, area, curvature and moment characteristics
It is one or more；Characteristic matching is carried out based on anti-counterfeiting characteristic, identifies the true and false of bill.This method has merged very big value filtering and has changed
Into the fiber small target deteection technology of two-dimensional entropy, there is good adaptivity and stability.
Although the above method can improve the accuracy of anti-fake identification to a certain extent.But with the hair of science and technology
Exhibition and progress, law-breaker make the horizontal appearance also with new edition bill of fictitious bill and change, fake it is horizontal also increasingly
It is high.Single bill authentication technique is difficult to ensure the accuracy that authentication is carried out to bill.
Summary of the invention
In view of the deficiencies of the prior art, technical problem to be solved by the present invention lies in provide a kind of silver based on cloud platform
Row bill fingerprint characteristic anti false authentication method.
For achieving the above object, the present invention uses following technical solutions:
A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform, includes the following steps:
S1, the multispectral image of big bill of acquisition；
S2 pre-processes the multispectral image, obtains the sola bill image of standard, by all sola bills
The fingerprint characteristic of image is stored to cloud platform；
S3 acquires the image information of bill to be detected, in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage, leads to
Cross double-colored fiber check and measure, phosphor pattern detects, the detection of infrared image blanking feature respectively detects described image information；Its
In, the image information of acquisition is detected by phosphor pattern detection, includes the following steps: to acquire bill to be detected respectively
The ultraviolet reflectance image BIMG of positive ultraviolet reflectance image FIMG and the back side, and convert thereof into the picture of HSV format；To just
The V component of the ultraviolet reflectance image FIMG in face carries out wavelet transformation, obtains transformed low-frequency information FIMG0；To the low frequency
The V component of information FIMG0 carries out image subdivision using OTSU Threshold Segmentation Algorithm and cuts；From the single-ticket of cloud platform extraction standard
According to the fingerprint characteristic of image, the phosphor pattern of the sola bill image of the phosphor pattern feature and standard of the image after subdivision is cut
Feature is matched, and judges the true and false of bill by the value of related coefficient；
S4 determines that bill to be detected is very, otherwise, it is determined that be checked when three kinds of detections determine that bill to be detected is true
It is false for surveying bill.
Wherein more preferably, in step s 2, the multispectral image is pre-processed, obtains the sola bill figure of standard
Picture includes the following steps:
S21 carries out distortion correction to the multispectral image of big bill of acquisition according to distortion factor and distortion model；
S22 carries out scaling cutting and obtains the sola bill image of standard to the image obtain after distortion correction.
Wherein more preferably, it in step S22, to the image obtained after distortion correction, carries out scaling cutting and is marked
Quasi- sola bill image, includes the following steps:
RGB color is transformed into hsv color space for opening the white light reflectance image WIMG of bill greatly by S221；
S222, using fixed threshold dividing method, by H, S component in white light reflectance image WIMG respectively in the first fixed area
Interior pixel obtains binary segmentation image as target point；
S223 is scanned image using rectangular window identical with the water-soluble linear dimension of bill, counts the mesh in rectangular window
Punctuate number detects water-soluble line position if target point data exceeds the 1/3 of rectangular window size, records rectangle window center
Position is (x0, y0)；
S224 carries out sola bill image according to bill size with coordinate points (x0-w1, y0-h 1) for bill starting point
It cuts, obtains the sola bill image of standard, wherein w1 is distance of the water-soluble line regional center of bill on the left of bill, h1
For distance of the water-soluble line regional center of bill on the upside of bill.
Wherein more preferably, in step s3, the image information for acquiring bill to be detected, by double-colored fiber check and measure to acquisition
Image information detected, include the following steps:
S301 acquires the positive ultraviolet reflectance image FIMG of bill to be detected and the ultraviolet reflectance image at the back side respectively
BIMG, and convert thereof into the picture of HSV format；
S302, it is the characteristics of according to indigo plant fiber red in bill, H, S component in the picture of HSV format is fixed second respectively
Pixel in section is carried out H, S component coarse segmentation, is carried out on V component using OTSU Threshold Segmentation Algorithm as target point
Image subdivision is cut；
S303 carries out fiber target extraction to the figure after segmentation；
S304 obtains same fiber respectively in the ultraviolet reflectance image of positive ultraviolet reflectance image FIMG and the back side
The fiber presented in BIMG carries out the fiber in the ultraviolet reflectance image BIMG of positive ultraviolet reflectance image FIMG and the back side
Matching, if two fibers match, bill to be detected passes through double-colored fiber check and measure；Otherwise, bill to be detected is false tickets.
Wherein more preferably, when carrying out the true and false of double-colored fiber check and measure bill to be detected, further sentenced by double-colored characteristic
The true and false of bill certainly to be detected.
Wherein more preferably, the true and false for adjudicating bill to be detected by double-colored characteristic includes the following steps:
For same fiber, it is extracted respectively in the ultraviolet reflectance image of positive ultraviolet reflectance image FIMG and the back side
The minimum value of H component, average value and maximum value, are denoted as respectively in BIMG
The double-colored characteristic decision function of building are as follows:
Judge the value of F；If F is not zero, the fiber meets double-colored characteristic, and bill to be detected is true；Otherwise, described
Fiber is unsatisfactory for double-colored characteristic, and bill to be detected is false.
Wherein more preferably, in step s3, the image information for acquiring bill to be detected, in conjunction with the single-ticket of cloud platform storage
According to the fingerprint characteristic of image, the image information of acquisition is detected by phosphor pattern detection, is included the following steps:
S311 acquires the positive ultraviolet reflectance image FIMG of bill to be detected and the ultraviolet reflectance image at the back side respectively
BIMG, and convert thereof into the picture of HSV format；
S312 carries out wavelet transformation to the V component of positive ultraviolet reflectance image FIMG, obtains transformed low frequency letter
Cease FIMG0；
S313 carries out image subdivision using OTSU Threshold Segmentation Algorithm and cuts to the V component of the low-frequency information FIMG0；
S314, from the fingerprint characteristic of the sola bill image of cloud platform extraction standard, the fluorescence of the image after subdivision is cut
Pattern characteristics are matched with the phosphor pattern feature of the sola bill image of standard, judge bill by the value of related coefficient
It is true and false.
Wherein more preferably, in step s3, the calculation formula of the related coefficient are as follows:
Wherein, r is related coefficient；(x, y) is pixel；FIMG0 (x, y) is positive ultraviolet reflectance image FIMG in V
Low-frequency information on component；MIMG (x, y) is the two-value template image of the sola bill image of standard；W is in pixel (x, y)
The maximum value of x；H is the maximum value of y in pixel (x, y)；When related coefficient is greater than phosphor pattern detection threshold value, determine
Phosphor pattern meets anti-fake requirement；Bill is true；Otherwise, it is determined that phosphor pattern is unsatisfactory for anti-fake requirement, bill is false
Wherein more preferably, in step s3, the image information for acquiring bill to be detected, in conjunction with the single-ticket of cloud platform storage
According to the fingerprint characteristic of image, the image information of acquisition is detected by the detection of infrared image blanking feature, including is walked as follows
It is rapid:
S321 acquires the positive infrared external reflection image RIMG and white light reflectance image WIMG of bill to be detected respectively, and
Be converted to gray level image；
S322, to the grayscale image of positive the infrared external reflection image RIMG and white light reflectance image WIMG of bill to be detected
Picture carries out wavelet transformation respectively, obtains transformed low-frequency information RIMG0 and WIMG0；
S323 constructs corresponding threshold binary image to low-frequency information RIMG0 and WIMG0 respectively, carries out threshold Image Segmentation, obtain
To corresponding bianry image RIMG1 and WIMG1；
S324 merges bianry image RIMG1 and WIMG1, obtains bianry image RWIMG1:
S325, from the two-value template image of the infrared image of the sola bill image of cloud platform extraction standard, using correlation
Matching algorithm matches the bianry image with the two-value template image, if successful match, with detection bill
Infrared blanking characteristic meets anti-fake requirement, and bill is true；Otherwise, it is determined that being unsatisfactory for the infrared blanking characteristic for detecting bill anti-fake
It is required that bill is false.
Wherein more preferably, in step S323, threshold binary image is constructed to low-frequency information RIMG0, is included the following steps:
For any pixel point (x, y), the gray average M1 in its neighborhood window is calculated；
Judge whether inequality R IMG0 (x, y) > M1 is true, if set up, the threshold binary image TIMG (x, y) at point (x, y)
=M1；
Otherwise, optimal segmenting threshold t, threshold binary image TIMG (x, y)=t at point (x, y) are calculated.
Bank money fingerprint characteristic anti false authentication method provided by the present invention based on cloud platform, acquires bill to be detected
Image information pass through double-colored fiber check and measure, phosphor pattern inspection in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage
It surveys, the detection of infrared image blanking feature respectively detects the image information of acquisition；When three kinds of detections determine ticket to be detected
When according to being true, determine bill to be detected be it is true, when three kinds of detections have it is a kind of determine that bill to be detected is fictitious time, determine ticket to be detected
According to being false.The drawbacks of single bill authentication technique carries out authentication is eliminated, the accurate of authentication is effectively improved
Property.
Detailed description of the invention
Fig. 1 is the flow chart of the bank money fingerprint characteristic anti false authentication method provided by the present invention based on cloud platform；
Fig. 2 is the circuit connection diagram for the multi-optical spectrum image collecting system that bill is opened greatly by bank；
Fig. 3 is in anti false authentication method provided by the present invention, and using high-definition network camera acquisition, draft is opened greatly by bank
Multispectral image structural schematic diagram；
Fig. 4 is in anti false authentication method provided by the present invention, and the position distribution of separate unit video camera and light source, draft is illustrated
Figure；
Fig. 5 is pixel grey scale distribution histogram in anti false authentication method provided by the present invention.
Specific embodiment
Detailed specific description is carried out to technology contents of the invention in the following with reference to the drawings and specific embodiments.
As shown in Figure 1, the bank money fingerprint characteristic anti false authentication method provided by the present invention based on cloud platform, including
Following steps: firstly, passing through the multispectral image of big bill of video camera acquisition；Secondly, the mostly light of the big bill to acquisition
Spectrogram picture is pre-processed, and obtains the sola bill image of standard, and the fingerprint characteristic of all sola bill images is stored
To cloud platform；Then, the image information for acquiring bill to be detected, the fingerprint in conjunction with the sola bill image of cloud platform storage are special
Sign, by double-colored fiber check and measure, phosphor pattern detection, the detection of infrared image blanking feature respectively to the bill to be detected of acquisition
Image information is detected；When three kinds of detections determine that bill to be detected is true, determine that bill to be detected is very, to examine when three kinds
Survey have it is a kind of determine that bill to be detected is fictitious time, determine that bill to be detected is false.Wherein, in embodiment provided by the present invention
In, the fingerprint characteristic of bank money includes double-colored fiber characteristics, phosphor pattern feature and the infrared image blanking of bank money
Feature etc..Detailed specific description is done to this process below.
S1 passes through the multispectral image of big bill of video camera acquisition.
In embodiment provided by the present invention, a bill is made of 20 Standard Bank bills greatly, is adopted by video camera
The multispectral image of big bill of collection is the multispectral image using big draft of 4 high-definition network camera acquisition banks.
Or the multispectral image using big bill of CMOS line-scan digital camera mould group acquisition.In embodiment provided by the present invention, with 4
It is illustrated for the multispectral image of big draft of platform high-definition network camera acquisition bank.Wherein, with 4 high-definition networks
The multispectral image of big bill of video camera acquisition bank is carried out using the multi-optical spectrum image collecting system that bill is opened greatly by bank
It realizes.The circuit connection that the multi-optical spectrum image collecting system of bill is opened greatly by bank is as shown in Figure 2.
Hardware device mainly includes three ARM control panel, light source module, high-definition network camera (video camera) parts.Its
In, ARM control panel is mainly used for power supply control, according to front-end computer order to different light source (infrared light, ultraviolet light and
White light) and 4 high-definition network cameras power on.ARM editions use 2131 chip of LPC, using 2131 chip of LPC
LPC2131 microcontroller is the ARM7TDMI-SCPU that real-time simulation and Embedded Trace are supported based on one.
When selecting high-definition network camera, in order to acquire the bank money image of high-resolution, using ten million Pixel-level
High-resolution video camera.In addition to this, due to needing to acquire the infrared spectroscopy feature of image, the 700nm of removal camera front end is needed
Low pass colour filter.
In order to acquire multispectral bank money image, light source module uses three kinds of LED light sources, including infrared, ultraviolet source
And white light, three kinds of light sources are integrated on a light source board, are respectively provided with independent 12V power supply, are controlled by ARM control panel, it can be independent
Or it powers on simultaneously.
The multispectral image for opening bill greatly using 4 high-definition network camera acquisition banks is illustrated below.Acquisition
Region signal is as shown in Figure 3.Wherein 2 video cameras in top respectively acquire 4 width bill images, and the video camera of lower section 2 respectively acquires 6 width tickets
According to image.
Light source includes infrared light, ultraviolet light and white light, is divided into two groups up and down, the position of separate unit video camera and light source, bill
Distribution is as shown in figure 4, one group of light source plate provides light source above, for acquiring bill reflection spectrum images, following set light source board
Light source is provided, for acquiring bill transmitted spectrum image.By the transformation of not sharing the same light of infrared light, ultraviolet light and white light, can adopt
Image information under the light source irradiation that collection needs.
S2 pre-processes the multispectral image of big bill of acquisition, obtains the sola bill image of standard, and will
The fingerprint characteristic of all sola bill images is stored to cloud platform.
After collecting the multispectral image of big bill by video camera, since generally there are abnormal for camera lens
It needs to be checked with the fingerprint characteristic of sola bill when becoming, and carrying out the detection of bill to be detected, so needing to acquisition
The multispectral image of big bill pre-processed, obtain the sola bill image of standard, and extract sola bill image
Fingerprint characteristic, by the fingerprint characteristic storage of all sola bill images to cloud platform, data when in case of subsequent being veritified
It provides.Wherein, the multispectral image of big bill of acquisition is pre-processed, obtains the sola bill image of standard, specifically
Include the following steps:
S21 carries out distortion correction to the multispectral image of big bill of acquisition according to distortion factor and distortion model.
Generally there are distortion for camera lens, it is therefore desirable to carry out distortion school to the multispectral image of big bill of acquisition
Just.The distortion of big bill acquisition camera lens is based on radial distortion, and distortion degree is little.It therefore is distortion with first order radial distortion
Model indicates are as follows:
Wherein,For ideal image coordinate；xu,yuFor fault image coordinate, ruIndicate point (xu,yu) former with coordinate
The distance of point, i.e.,k0、k1For distortion factor.
Above formula can be rewritten as:
It enablesThen distortion model may be expressed as:
Y=k0+k1x
There are two unknown distortion parameters for above formula.In order to seek the two parameters, video camera is marked using gridiron pattern
It is fixed, n is obtained to control point { (xi,yi)；I=1,2 ..., n }, In conjunction with distortion model,
Using cumulative and method, constructor group:
Wherein, subscript (1) and (2) respectively indicate 1 rank and the accumulation of 2 ranks and the definition of k rank accumulation sum is
AndIndicate k rank accumulate substantially with, calculation formula is
IfDistortion factor can be obtained are as follows:
According to distortion factor and distortion model, distortion correction is carried out.For fault image coordinate (xu,yu), corresponding correction
Coordinate (xd,yd) be
Wherein, ruFor point (xu,yu) at a distance from coordinate origin, i.e.,The multispectral figure of bill is opened greatly
As the distance of coordinate to coordinate origin,rdIndicate point (xd,yd) at a distance from coordinate origin, i.e., calibration coordinate arrives coordinate origin
Distance；
rdMeet following condition:
For the coordinate (x after correctiond,yd), pass through (x on image after bilinear interpolation method calculating correctionu,yu) point
Gray value.
S22 carries out scaling cutting and obtains the sola bill image of standard to the image obtain after distortion correction.
In view of the water-soluble line position of bill is relatively fixed, and it is easy to detect.Therefore, the present invention is according to the water-soluble line position of bill
It is calibrated, carries out image cropping according to the priori knowledge of bill size.To the image obtained after distortion correction, determined
Mark cuts and obtains the sola bill image of standard, specifically comprises the following steps:
RGB color is transformed into hsv color space for opening the white light reflectance image WIMG of bill greatly by S221, is turned
It is as follows to change formula:
Wherein, RGB is a kind of color standard of industry, is by red (R), green (G), blue (B) three Color Channels
Variation and their mutual superpositions obtain miscellaneous color, RGB be represent red, green, blue three it is logical
The color in road, this standard almost include all colours that human eyesight can perceive, and are current with most wide cie system of color representation cie
One of system.HSV (Hue, Saturation, Value) is a kind of color space, also referred to as hexagonal pyramid model (Hexcone
Model), the parameter of color is respectively in this model: tone (H), saturation degree (S), brightness (V).V1And V2For the centre of brightness
Variable.
S222, using fixed threshold dividing method, by H, S component in white light reflectance image WIMG respectively in the first fixed area
Interior pixel obtains binary segmentation image as target point；Wherein, in white light reflectance image WIMG H, S component first
Fixed interval is respectively [270,360], [0.0,0.2].
S223 is scanned image using the rectangular window having a size of 950 × 160, wherein the bill of acquisition system acquisition
In image, the width in water-soluble line region is about 950, height about 160.The target point number in rectangular window is counted, if target
Point data exceed rectangular window size 1/3, then it is assumed that detect water-soluble line position, at this time record rectangular window center be (x0,
y0)；
S224, the cutting of sola bill image is carried out according to the priori knowledge of bill size, and bill starting point coordinate is (x0-
W1, y0-h1), cutting picture size is 2048 × 1536.Wherein, w1 is the water-soluble line regional center of bill on the left of bill
Distance, h1 is distance of the water-soluble line regional center of bill on the upside of bill, in embodiment provided by the present invention, w1=
760, H1=590.In the bill images of acquisition system acquisition, the width in sola bill region is about 2048, height about 1536.
According to above-mentioned steps, carry out being syncopated as all sola bill figures in the image obtained after distortion correction from big bill
Picture, and by all sola bill image and it includes fingerprint characteristic store to cloud platform, be effectively guaranteed standard
The accuracy of sola bill image and comprehensive, improves the comprehensive of anti-counterfeiting detection data source, and then improve detection
Accuracy stores to cloud platform and has been effectively saved local memory space.
S3 acquires the image information of bill to be detected, in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage, leads to
Double-colored fiber check and measure, phosphor pattern detection, the detection of infrared image blanking feature are crossed respectively to the image with detection bill of acquisition
Information is detected.
When carrying out authenticity verification to bill to be detected, double-colored fiber check and measure, phosphor pattern detection, infrared image is respectively adopted
The detection of blanking feature detects the image information of acquisition, only when three kinds of detections determine that bill to be detected is true, sentences
Fixed bill to be detected is very, to be fictitious time when three kinds of detections have a kind of detection judgement bill to be detected, determine that bill to be detected is false.
The sequence of three kinds of detections, which can according to need, to be adjusted, and is illustrated separately below.
The image information for acquiring bill to be detected is detected by image information of the double-colored fiber check and measure to acquisition, tool
Body includes the following steps:
S301, control light source, acquire respectively bill to be detected positive ultraviolet reflectance image FIMG and the back side it is ultraviolet
Reflected image BIMG, and convert thereof into the picture of HSV format.
Light source is controlled, acquires the positive ultraviolet reflectance image FIMG of bill to be detected and the ultraviolet reflectance figure at the back side respectively
As BIMG, the image of acquisition is RGB24 format, since RGB component is influenced vulnerable to light-source brightness, therefore is converted into hsv color
Space.
S302, it is the characteristics of according to indigo plant fiber red in bill, H, S component in the picture of HSV format is fixed second respectively
Pixel in section is carried out H, S component coarse segmentation, is carried out on V component using OTSU Threshold Segmentation Algorithm as target point
Image subdivision is cut.
Fiber in 2010 editions bills has red blue two kinds of colors, and statistics discovery, H component is respectively in section [270,360]
In [160,240], S component is all in section [0.3,0.6].For this purpose, fixed threshold method is used here, by the section of H component
The section [0.3,0.6] of [270,360] and [160,240] and S component is used as the second fixed interval, to image FIMG and BIMG
HS component carry out coarse segmentation.
Specifically, any pixel (i, j) of zone of fiber meets the following conditions:
After image segmentation, target pixel value is denoted as 1, and background pixel value is denoted as 0, similarly hereinafter.
During HS component coarse segmentation, threshold value setting is more relaxed, although in this way will not missing inspection fiber pixel, part is non-
Fiber pixel can be by erroneous detection.
Statistics discovery, on V component, the general brightness of zone of fiber is larger.Although in entire image, part background area
Pixel intensity it is also larger, but in the neighborhood near fiber, the brightness of the brightness of fiber pixel and its neighbouring background pixel
It differs greatly.For this purpose, here in each zone of fiber that coarse segmentation obtains, for V component, using segmentation effect stabilization, certainly
Adaptable OTSU Threshold Segmentation Algorithm carries out image subdivision and cuts, to reject part by the non-fiber pixel of erroneous detection.
Wherein, the mathematical description of OTSU algorithm is as follows: setting Probability p (i)=ni/N (i that brightness value i occurs in image-region
=0,1 ..., 255), N is sum of all pixels in image-region, and ni is the pixel number that brightness value is i.For threshold value t by image
Pixel is divided into two class of target (M) and background (B) in region, and target brightness value is greater than background value.
Note:
Then inter-class variance σ (t) can be indicated are as follows:
σ (t)=wB(t)wM(t)(μB(t)-μM(t))2；
All gray values are traversed, selection makes the maximum threshold value t of σ (t), it may be assumed that
Generally, in similar acquisition structure, fiber brightness all 100 or more, disappears to reduce accidentally segmentation and time
Consumption, the lower limit of setting fiber brightness are 100, at this time threshold value t value are as follows:
According to threshold value t, zone of fiber is finely divided and is cut, pixel of the brightness value greater than t is target point, rejects part quilt
The non-fiber pixel of erroneous detection.
S303 carries out fiber target extraction to the figure after segmentation.
For the image after segmentation, the method labeled fibers target combined using median filtering and mathematical morphology, step
It is rapid as follows:
S3031 eliminates isolated noise point using median filtering for the image after segmentation；
S3032, using in mathematical morphology expansion and etching operation, remove " hole " of image；
S3033 is identified the bianry image that detected using the adjacent connection method of 8-.In addition, due to fiber mesh
The pixel number that mark includes (generally 40~200, unit: pixel) in a limited range, can be picked using dual-threshold voltage
Except false target, bottom threshold is set as 40, and the upper limit is set as 200.Wherein, median filtering, expansion and corrosion in mathematical morphology
Operation and the adjacent connection method of 8- are approach well known, are just repeated no more herein.
S304 obtains the fiber that same fiber is presented in FIMG and BIMG respectively, to the fiber in FIMG and BIMG
It is matched, if two fibers match, bill to be detected passes through double-colored fiber check and measure；Otherwise, bill to be detected is false tickets.
Red blue contrast effect can be presented in same fiber in FIMG and BIMG.Therefore, before carrying out double-colored Characteristics Detection,
It first has to match the fiber in FIMG and BIMG, that is, finds out that certain fiber in FIMG is corresponding in BIMG to be
Which bar fiber.
Assuming that the maximum enclosure rectangle of certain fiber in image FIMG are as follows:
Wherein, (x, y) is any pixel point on the fiber.
Similarly, it is assumed that the maximum enclosure rectangle of certain fiber in image BIMG are as follows:
Here two functions are defined:
Wherein, whether function Bound is for describing point (x, y) in rectangle R1=(x, y) | x1≤x≤x2,y1≤y≤y2}
It is interior；Function Cross is for describing rectangle R2=(x, y) | xl≤x≤xr,yt≤y≤ybAnd rectangle R1=(x, y) | x1≤x≤
x2,y1≤y≤y2It whether is in right-angled intersection state.
Since same fiber meets horizontal conditions mirror in the position of FIMG and BIMG, the following judgement of construction here
Function:
Wherein, W indicates the width of image.
If f is not zero, show that two fibers match, bill to be detected passes through double-colored fiber check and measure；Otherwise, table
Bright two fibers mismatch.
It, can be with when carrying out the true and false of double-colored fiber check and measure bill to be detected in embodiment provided by the present invention
The true and false of bill to be detected is adjudicated by double-colored characteristic.For same fiber, its H component in FIMG and BIMG is extracted respectively
Minimum value, average value and maximum value, be denoted as respectively
Since the depth of fiber insertion paper is different, therefore the H value of each pixel also has deviation on fiber.Therefore, judgement is double
When color characteristic, all pixels on fiber cannot be required all to meet double-colored characteristic, all fibres in bill can not be required all
Meet double-colored characteristic.Here the double-colored characteristic decision function constructed are as follows:
If F is not zero, show that the fiber meets double-colored characteristic, bill to be detected is true；Otherwise, show that the fiber is discontented
The double-colored characteristic of foot, bill to be detected are false.
In the image of bill to be detected, all position of fibers for meeting double-colored characteristic and shape are detected, after can be used as
Continue the foundation of further Fibre sorting.
The image information for acquiring bill to be detected passes through in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage
Phosphor pattern detection detects the image information of acquisition, specifically comprises the following steps:
S311, control light source, acquire respectively bill to be detected positive ultraviolet reflectance image FIMG and the back side it is ultraviolet
Reflected image BIMG, and convert thereof into the picture of HSV format.
S312 carries out wavelet transformation to the V component of positive ultraviolet reflectance image FIMG, retains transformed low frequency letter
Cease image information of the FIMG0 as subsequent processing.
There are two aspects for the contribution of wavelet transformation, first is that reducing image to be processed under the premise of retaining image main information
Size, to reduce the operand that subsequent image handles each stage；Second is that reducing the interference of illumination and noise, enhance fingerprint
The robustness of characteristic detection method.In embodiment provided by the present invention, the formula of wavelet transformation used are as follows:
Wherein, W (j, m, n) coefficient is the approximation of the image f (x, y) at scale j, and φ is Haar wavelet scaling function, is used
Formula indicates are as follows:
S313 carries out image subdivision using OTSU Threshold Segmentation Algorithm and cuts to the V component of low-frequency information FIMG0.
To the V component of FIMG0, image subdivision is carried out using OTSU Threshold Segmentation Algorithm above-mentioned and is cut.Equally, similar
It acquires in structure, all 100 or more, in order to reduce accidentally segmentation and time loss, brightness is arranged in the brightness of each pixel of phosphor pattern
Lower limit is 100, at this time threshold value t value are as follows:
In segmentation, brightness value is greater than the pixel of t for target point, in the bianry image after dividing according to threshold value t, not only
It comprising pattern, and include few fibers.Since fiber target is very small for pattern target, pattern characteristics are detected
As a result very little is influenced, so there is no need to reject fiber target.
S314, from the fingerprint characteristic of the sola bill image of cloud platform extraction standard, the fluorescence of the image after subdivision is cut
Pattern characteristics are matched with the phosphor pattern feature of the sola bill image of standard, judge bill by the value of related coefficient
It is true and false.
From the fingerprint characteristic of the sola bill image of cloud platform extraction standard, it is assumed that the sola bill of standard under identical size
The two-value template image of the phosphor pattern of image is MIMG, in embodiment provided by the present invention, using Image Matching
Carry out template matching, the calculation formula of related coefficient are as follows:
Since image to be matched (subdivision cut after image) and template image (the sola bill image of standard) are all two-values
Image, background and target are used 0 and 1 indicate respectively.Therefore r can quickly be calculated with following formula:
Wherein, & indicates AND operation.
Phosphor pattern detection threshold value is set, in embodiment provided by the present invention, phosphor pattern detection threshold value is set as 70,
If r > 70, determine that phosphor pattern meets anti-fake requirement；Otherwise, it is determined that phosphor pattern is unsatisfactory for anti-fake requirement, bill is false tickets.
The image information for acquiring bill to be detected passes through in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage
The detection of infrared image blanking feature detects the image information of acquisition, specifically comprises the following steps:
S321 controls light source, acquires the positive infrared external reflection image RIMG and white light reflection figure of bill to be detected respectively
As WIMG, and it is passed through into formula:Be converted to gray level image.
Light source is controlled, acquires the positive infrared external reflection image RIMG and white light reflectance image of bill to be detected respectively
WIMG.The image of acquisition is RGB24 format, is converted into gray level image, and formula is as follows:
Wherein, RGB respectively represents the color in three channels of red, green, blue, and V indicates the brightness (V) in hsv color space.
S322, to the grayscale image of positive the infrared external reflection image RIMG and white light reflectance image WIMG of bill to be detected
Picture carries out wavelet transformation respectively, retains image information of the transformed low-frequency information RIMG0 and WIMG0 as subsequent processing.
To the gray level image of positive the infrared external reflection image RIMG and white light reflectance image WIMG of bill to be detected, respectively
Wavelet transformation is carried out, image information of the transformed low-frequency information RIMG0 and WIMG0 as subsequent processing is retained.Wavelet transformation
Effect and implementation method it is identical as the description of front.It just repeats no more herein.
S323 constructs corresponding threshold binary image TIMG to low-frequency information RIMG0 and WIMG0 respectively, carries out threshold binary image point
It cuts, obtains corresponding bianry image RIMG1 and WIMG 1.
From the point of view of intensity profile, the intensity profile of the targets such as wire frame, printing type face, hand-written script is variant, with check paper
Etc. backgrounds gray difference it is not significant enough, it is difficult to effectively divide the mesh such as wire frame, font according to one or several threshold values
Mark.In embodiment provided by the present invention, propose a kind of new threshold Image Segmentation algorithm, can effectively divide wire frame,
The targets such as font.Specifically comprise the following steps:
S3231 constructs corresponding threshold binary image TIMG to low-frequency information RIMG0 and WIMG0 respectively.Provided by the present invention
Embodiment in, for low-frequency information RIMG0 or WIMG0 (being abbreviated as f here), construct corresponding threshold binary image TIMG, step
It is as follows:
S32311 calculates the gray average M1 in its neighborhood window for any pixel point (x, y)；
S32312 judges whether inequality f (x, y) > M1 is true, if set up, is transferred to step S32313；Otherwise, it is transferred to
Step S32314；
S32313, threshold binary image TIMG (x, y)=M1 at point (x, y)；
S32314 calculates optimal segmenting threshold t, threshold binary image TIMG (x, y)=t at point (x, y)；It specifically includes as follows
Step:
Firstly, calculating neighborhood window (11 × 11) interior pixel grey scale distribution histogram p [i]=ni/N (i=0,1 ...,
255), ni is the pixel number that brightness value is i, and N is sum of all pixels in image-region, here N=11 × 11=121；
Then, first valley point M2 (as shown in Figure 5) of the direction search histogram according to gray value from M1 to 2；M2
Meet condition (p [M2+1] < p [M2-2] and p [M2-1] < p [M2+2])；
Judge whether M1's and M2 is equal, if M2=M1, TIMG (x, y)=M1；Otherwise, it is utilized between M1 and M2
OTSU algorithms selection optimal segmenting threshold t:
Wherein, it using having been addressed before OTSU algorithms selection optimal segmenting threshold t, just repeats no more herein.
TIMG (x, y)=t at this time.S3232 carries out threshold Image Segmentation using following formula, obtains corresponding binary map
As RIMG1 and WIMG1.
In this way, grey image R IMG0 and WIMG0 be after over-segmentation, corresponding bianry image be denoted as respectively RIMG1 and
WIMG1。
S324 merges bianry image RIMG1 and WIMG1, obtains bianry image RWIMG1:
Target point in RWIMG1 is mainly the frame and font of blanking under infrared image.
S325 is used from the two-value template image MIMG2 of the infrared image of the sola bill image of cloud platform extraction standard
Correlation matching algorithm matches bianry image RWIMG1 with two-value template image MIMG2, if successful match, band detection
The infrared blanking characteristic of bill meets anti-fake requirement, and bill is true；Otherwise, it is determined that the infrared blanking characteristic with detection bill is discontented
The anti-fake requirement of foot, bill are false.
The fingerprint characteristic of the sola bill image of standard is obtained from cloud platform, it is assumed that the sola bill of standard under identical size
The two-value template image of the infrared image of image is MIMG2, in embodiment provided by the present invention, using Image Matching
Template matching is carried out,
The calculation formula of related coefficient are as follows:
Phosphor pattern detection threshold value is set, in embodiment provided by the present invention, phosphor pattern detection threshold value is set as 70,
If r > 70, judge that infrared blanking characteristic meets anti-fake requirement, bill is true；Otherwise, it is determined that infrared blanking characteristic is unsatisfactory for preventing
Puppet requires, and bill is false.
S4 determines that bill to be detected is very, otherwise, it is determined that be checked when three kinds of detections determine that bill to be detected is true
It is false for surveying bill.
In conclusion the bank money fingerprint characteristic anti false authentication method provided by the present invention based on cloud platform, passes through
The multispectral image of big bill of video camera acquisition；The multispectral image of big bill of acquisition is pre-processed, is marked
Quasi- sola bill image, and the fingerprint characteristic of all sola bill images is stored to cloud platform, it is effectively guaranteed mark
The accuracy of quasi- sola bill image and comprehensive, improves the comprehensive of anti-counterfeiting detection data source, and then improve inspection
The accuracy of survey stores to cloud platform and has been effectively saved local memory space.In addition, acquiring the image letter of bill to be detected
Breath is detected by double-colored fiber check and measure, phosphor pattern in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage, is infrared
The detection of picture blanking feature respectively detects the image information of acquisition；When three kinds of detections determine that bill to be detected is true
When, determine bill to be detected be it is true, when three kinds of detections have it is a kind of determine that bill to be detected is fictitious time, determine that bill to be detected is
It is false.The drawbacks of carrying out authentication this method eliminates single bill authentication technique, effectively improves the standard of authentication
True property.
The bank money fingerprint characteristic anti false authentication method to provided by the present invention based on cloud platform has carried out in detail above
Thin explanation.For those of ordinary skill in the art, it is done under the premise of without departing substantially from true spirit
Any obvious change, the infringement for all weighing composition to the invention patent, will undertake corresponding legal liabilities.
Claims (9)
1. a kind of bank money fingerprint characteristic anti false authentication method based on cloud platform, it is characterised in that include the following steps:
S1, the multispectral image of big bill of acquisition；
S2 pre-processes the multispectral image, obtains the sola bill image of standard, by all sola bill images
Fingerprint characteristic storage to cloud platform；
S3 acquires the image information of bill to be detected, in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage, by double
Color fibre detection, phosphor pattern detection, the detection of infrared image blanking feature respectively detect described image information；Wherein,
The image information of acquisition is detected by phosphor pattern detection, includes the following steps: acquiring bill to be detected respectively just
The ultraviolet reflectance image FIMG in the face and ultraviolet reflectance image BIMG at the back side, and convert thereof into the picture of HSV format；To front
Ultraviolet reflectance image FIMG V component, carry out wavelet transformation, obtain transformed low-frequency information FIMG0；The low frequency is believed
The V component for ceasing FIMG0 carries out image subdivision using OTSU Threshold Segmentation Algorithm and cuts；From the sola bill of cloud platform extraction standard
The phosphor pattern of the fingerprint characteristic of image, the sola bill image of the phosphor pattern feature and standard of the image after subdivision is cut is special
Sign is matched, and judges the true and false of bill by the value of related coefficient；
S4 determines that bill to be detected is very, otherwise, it is determined that ticket to be detected when three kinds of detections determine that bill to be detected is true
According to being false.
2. the bank money fingerprint characteristic anti false authentication method based on cloud platform as described in claim 1, it is characterised in that
In step S2, the multispectral image is pre-processed, the sola bill image of standard is obtained, includes the following steps:
S21 carries out distortion correction to the multispectral image of big bill of acquisition according to distortion factor and distortion model；
S22 carries out scaling cutting and obtains the sola bill image of standard to the image obtain after distortion correction.
3. the bank money fingerprint characteristic anti false authentication method based on cloud platform as claimed in claim 2, it is characterised in that
In step S22, to the image obtained after distortion correction, carries out scaling cutting and obtain the sola bill image of standard, including
Following steps:
RGB color is transformed into hsv color space for opening the white light reflectance image WIMG of bill greatly by S221；
S222, using fixed threshold dividing method, by H, S component in white light reflectance image WIMG respectively in the first fixed interval
Pixel as target point, obtain binary segmentation image；
S223 is scanned image using rectangular window identical with the water-soluble linear dimension of bill, counts the target point in rectangular window
Number detects water-soluble line position if target point data exceeds the 1/3 of rectangular window size, records rectangular window center
For (x0, y0)；
S224 carries out the cutting of sola bill image according to bill size with coordinate points (x0-w1, y0-h1) for bill starting point,
Obtain the sola bill image of standard, wherein w1 is distance of the water-soluble line regional center of bill on the left of bill, and h1 is bill
Distance of the water-soluble line regional center on the upside of bill.
4. the bank money fingerprint characteristic anti false authentication method based on cloud platform as described in claim 1, it is characterised in that
In step S3, the image information of bill to be detected is acquired, is detected by image information of the double-colored fiber check and measure to acquisition, is wrapped
Include following steps:
S301 acquires the positive ultraviolet reflectance image FIMG of bill to be detected and the ultraviolet reflectance image BIMG at the back side respectively,
And convert thereof into the picture of HSV format；
S302, according in bill it is red indigo plant fiber the characteristics of, by H, S component in the picture of HSV format respectively in the second fixed interval
Interior pixel carries out H, S component coarse segmentation as target point, carries out image using OTSU Threshold Segmentation Algorithm on V component
Subdivision is cut；
S303 carries out fiber target extraction to the figure after segmentation；
S304, acquisition same fiber is respectively in the ultraviolet reflectance image BIMG at positive ultraviolet reflectance image FIMG and the back side
The fiber of presentation matches the fiber in the ultraviolet reflectance image BIMG of positive ultraviolet reflectance image FIMG and the back side,
If two fibers match, bill to be detected passes through double-colored fiber check and measure；Otherwise, bill to be detected is false tickets.
5. the bank money fingerprint characteristic anti false authentication method based on cloud platform as described in claim 1, it is characterised in that:
When carrying out the true and false of double-colored fiber check and measure bill to be detected, the true of bill to be detected is further adjudicated by double-colored characteristic
It is pseudo-.
6. the bank money fingerprint characteristic anti false authentication method based on cloud platform as claimed in claim 5, it is characterised in that institute
It states and is included the following steps: by the true and false that double-colored characteristic adjudicates bill to be detected
For same fiber, it is extracted respectively in the ultraviolet reflectance image BIMG at positive ultraviolet reflectance image FIMG and the back side
Minimum value, average value and the maximum value of middle H component, are denoted as respectively
The double-colored characteristic decision function of building are as follows:
Judge the value of F；If F is not zero, the fiber meets double-colored characteristic, and bill to be detected is true；Otherwise, the fiber
It is unsatisfactory for double-colored characteristic, bill to be detected is false.
7. the bank money fingerprint characteristic anti false authentication method based on cloud platform as described in claim 1, it is characterised in that:
In step s3, the calculation formula of the related coefficient are as follows:
Wherein, r is related coefficient；(x, y) is pixel；FIMG0 (x, y) is positive ultraviolet reflectance image FIMG in V component
On low-frequency information；MIMG (x, y) is the two-value template image of the sola bill image of standard；W is x in pixel (x, y)
Maximum value；H is the maximum value of y in pixel (x, y)；When related coefficient is greater than phosphor pattern detection threshold value, determine glimmering
Light pattern meets anti-fake requirement；Bill is true；Otherwise, it is determined that phosphor pattern is unsatisfactory for anti-fake requirement, bill is false.
8. the bank money fingerprint characteristic anti false authentication method based on cloud platform as described in claim 1, it is characterised in that
In step S3, the image information of bill to be detected is acquired, in conjunction with the fingerprint characteristic of the sola bill image of cloud platform storage, is passed through
The detection of infrared image blanking feature detects the image information of acquisition, includes the following steps:
S321, acquires the positive infrared external reflection image RIMG and white light reflectance image WIMG of bill to be detected respectively, and converts
For gray level image；
S322 divides the gray level image of positive the infrared external reflection image RIMG and white light reflectance image WIMG of bill to be detected
Not carry out wavelet transformation, obtain transformed low-frequency information RIMG0 and WIMG0；
S323 constructs corresponding threshold binary image to low-frequency information RIMG0 and WIMG0 respectively, carries out threshold Image Segmentation, obtains pair
The bianry image RIMG1 and WIMG1 answered；
S324 merges bianry image RIMG1 and WIMG1, obtains bianry image RWIMG1:
S325, from the two-value template image of the infrared image of the sola bill image of cloud platform extraction standard, using relevant matches
Algorithm matches the bianry image with the two-value template image, if successful match, with the infrared of detection bill
Blanking characteristic meets anti-fake requirement, and bill is true；Otherwise, it is determined that the infrared blanking characteristic with detection bill is unsatisfactory for anti-fake want
It asks, bill is false.
9. the bank money fingerprint characteristic anti false authentication method based on cloud platform as claimed in claim 8, it is characterised in that
In step S323, threshold binary image is constructed to low-frequency information RIMG0, is included the following steps:
For any pixel point (x, y), the gray average M1 in its neighborhood window is calculated；
Judge whether inequality R IMG0 (x, y) > M1 is true, if set up, the threshold binary image TIMG (x, y) at point (x, y)=
M1；
Otherwise, optimal segmenting threshold t, threshold binary image TIMG (x, y)=t at point (x, y) are calculated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510992680.4A CN105632015B (en) | 2015-12-24 | 2015-12-24 | A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510992680.4A CN105632015B (en) | 2015-12-24 | 2015-12-24 | A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105632015A CN105632015A (en) | 2016-06-01 |
| CN105632015B true CN105632015B (en) | 2019-03-01 |
Family
ID=56046894
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510992680.4A Active CN105632015B (en) | 2015-12-24 | 2015-12-24 | A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105632015B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106952393B (en) * | 2017-03-28 | 2020-02-07 | 深圳怡化电脑股份有限公司 | Paper money identification method and device, electronic equipment and storage medium |
| CN107437293A (en) * | 2017-07-13 | 2017-12-05 | 广州市银科电子有限公司 | A kind of bill anti-counterfeit discrimination method based on bill global characteristics |
| CN109272006A (en) * | 2017-07-18 | 2019-01-25 | 北京柯斯元科技有限公司 | Anti-counterfeit sign with random texture pattern decision-making system and determination method |
| CN107464336B (en) * | 2017-07-26 | 2020-05-15 | 南昌印钞有限公司 | Colorless phosphorescent printed pattern quality online detection method and device |
| CN107527423B (en) * | 2017-08-03 | 2019-12-06 | 深圳怡化电脑股份有限公司 | Paper money anti-counterfeiting identification method and paper money identification device |
| CN108198353B (en) * | 2018-01-24 | 2021-02-02 | 深圳怡化电脑股份有限公司 | Control method and device for bill issuing equipment and bill issuing equipment |
| CN108492445A (en) * | 2018-02-06 | 2018-09-04 | 深圳怡化电脑股份有限公司 | The method and device of bank note classification |
| CN109064304A (en) * | 2018-08-03 | 2018-12-21 | 四川长虹电器股份有限公司 | Finance reimbursement bill automated processing system and method |
| CN112258724A (en) * | 2019-07-05 | 2021-01-22 | 深圳怡化电脑股份有限公司 | Bill authenticity identification method and device, terminal and storage medium |
| CN112862902B (en) * | 2021-02-24 | 2024-05-07 | 中国资源卫星应用中心 | Relative radiation correction method of space linear array camera |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0605208A2 (en) * | 1993-01-01 | 1994-07-06 | Canon Kabushiki Kaisha | Image processing apparatus and method, and image reading apparatus |
| CN1475960A (en) * | 2002-08-14 | 2004-02-18 | 北京兆日科技有限责任公司 | Image collection method of substance finger print |
| CN1510628A (en) * | 2002-12-25 | 2004-07-07 | 鲍东山 | Composite image processing banknote identifying system |
| CN101751713A (en) * | 2008-12-03 | 2010-06-23 | 北京金储自动化技术研究所 | Fake identification system of bank bills based on spectral image analysis |
| CN102254363A (en) * | 2010-05-31 | 2011-11-23 | 北京中科金财科技股份有限公司 | Spectrographic analysis-based bill anti-counterfeiting self-service telling method and device |
| CN102306415A (en) * | 2011-08-01 | 2012-01-04 | 广州广电运通金融电子股份有限公司 | Portable valuable file identification device |
| CN102324132A (en) * | 2011-05-25 | 2012-01-18 | 深圳市怡化电脑有限公司 | Bank-note detecting method by identifying ultraviolet-light image and system |
| CN102542655A (en) * | 2011-11-16 | 2012-07-04 | 中钞实业有限公司 | Note anti-counterfeiting discrimination method based on fiber personality characteristics |
| CN104298989A (en) * | 2014-08-20 | 2015-01-21 | 深圳大学 | Counterfeit identifying method and counterfeit identifying system based on zebra crossing infrared image characteristics |
| CN104574638A (en) * | 2014-09-30 | 2015-04-29 | 上海层峰金融设备有限公司 | Method for identifying RMB |
-
2015
- 2015-12-24 CN CN201510992680.4A patent/CN105632015B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0605208A2 (en) * | 1993-01-01 | 1994-07-06 | Canon Kabushiki Kaisha | Image processing apparatus and method, and image reading apparatus |
| CN1475960A (en) * | 2002-08-14 | 2004-02-18 | 北京兆日科技有限责任公司 | Image collection method of substance finger print |
| CN1510628A (en) * | 2002-12-25 | 2004-07-07 | 鲍东山 | Composite image processing banknote identifying system |
| CN101751713A (en) * | 2008-12-03 | 2010-06-23 | 北京金储自动化技术研究所 | Fake identification system of bank bills based on spectral image analysis |
| CN102254363A (en) * | 2010-05-31 | 2011-11-23 | 北京中科金财科技股份有限公司 | Spectrographic analysis-based bill anti-counterfeiting self-service telling method and device |
| CN102324132A (en) * | 2011-05-25 | 2012-01-18 | 深圳市怡化电脑有限公司 | Bank-note detecting method by identifying ultraviolet-light image and system |
| CN102306415A (en) * | 2011-08-01 | 2012-01-04 | 广州广电运通金融电子股份有限公司 | Portable valuable file identification device |
| CN102542655A (en) * | 2011-11-16 | 2012-07-04 | 中钞实业有限公司 | Note anti-counterfeiting discrimination method based on fiber personality characteristics |
| CN104298989A (en) * | 2014-08-20 | 2015-01-21 | 深圳大学 | Counterfeit identifying method and counterfeit identifying system based on zebra crossing infrared image characteristics |
| CN104574638A (en) * | 2014-09-30 | 2015-04-29 | 上海层峰金融设备有限公司 | Method for identifying RMB |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105632015A (en) | 2016-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105632015B (en) | A kind of bank money fingerprint characteristic anti false authentication method based on cloud platform | |
| CN109740478B (en) | Vehicle detection and identification method, device, computer equipment and readable storage medium | |
| US9245203B2 (en) | Collecting information relating to identity parameters of a vehicle | |
| CN109684987B (en) | Identity verification system and method based on certificate | |
| EP1271403B1 (en) | Method and device for character location in images from digital camera | |
| CN102542660B (en) | Bill anti-counterfeiting identification method based on bill watermark distribution characteristics | |
| CN111046703A (en) | Face anti-counterfeiting detection method and device and multi-view camera | |
| CN106203415B (en) | bank card number automatic identification device based on digital image processing | |
| CN105956578A (en) | Face verification method based on identity document information | |
| CN102902959A (en) | Face recognition method and system for storing identification photo based on second-generation identity card | |
| CN109740572A (en) | A kind of human face in-vivo detection method based on partial color textural characteristics | |
| Suresh et al. | Indian currency recognition and verification using image processing | |
| CN105844213B (en) | Green fruit recognition method | |
| Sharan et al. | Detection of counterfeit Indian currency note using image processing | |
| KR102038576B1 (en) | Method of detecting fraud of an iris recognition system | |
| CN107527423B (en) | Paper money anti-counterfeiting identification method and paper money identification device | |
| CN106875543B (en) | It is a kind of based on visually impaired people's bill acceptor system of RGB-D camera and recognition methods | |
| Amirsab et al. | An automated recognition of fake or destroyed Indian currency notes | |
| Lafuente-Arroyo et al. | Traffic sign classification invariant to rotations using support vector machines | |
| CN106447908A (en) | Paper money authentic identification method and device | |
| Gangopadhyay et al. | FACE DETECTION AND RECOGNITION USING HAAR CLASSIFIER AND LBP HISTOGRAM. | |
| CN107301718B (en) | A kind of image matching method and device | |
| CN110222660A (en) | A kind of signature false distinguishing method and system merged based on dynamic with static nature | |
| CN109753912A (en) | A kind of multi-light spectrum palm print matching process based on tensor | |
| CN105389805A (en) | Optical detection method of component shortage, wrong component and reverse component in plug-in operation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |