CN107742357A - A kind of recognition methods of paper money number and device - Google Patents
A kind of recognition methods of paper money number and device Download PDFInfo
- Publication number
- CN107742357A CN107742357A CN201710936146.0A CN201710936146A CN107742357A CN 107742357 A CN107742357 A CN 107742357A CN 201710936146 A CN201710936146 A CN 201710936146A CN 107742357 A CN107742357 A CN 107742357A
- Authority
- CN
- China
- Prior art keywords
- word number
- crown word
- character
- characteristic area
- interference region
- 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.)
- Pending
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
- G07D7/2016—Testing patterns thereon using feature extraction, e.g. segmentation, edge detection or Hough-transformation
Abstract
The embodiment of the invention discloses a kind of recognition methods of paper money number and device, this method includes：Obtain the crown word number characteristic area of banknote image to be identified；After the interference region of the crown word number characteristic area is removed, crown word number boundary alignment is carried out to the crown word number characteristic area, obtains the boundary alignment result of each character in crown word number；The boundary alignment result of each character in the crown word number is encoded, extracts the pixel variation characteristic of each character, and according to the pixel variation characteristic and matching template, identifies the content of each character in the crown word number.The embodiment of the present invention can accurately be identified to the crown word number in bank note, in actual applications, even if identification of the pattern of banknote image to crown word number interferes, also can exclusive PCR region recognition go out the content of crown word number, the accuracy rate of crown word number identification is improved, the subsequent treatment for being easy to crown word number to identify.
Description
Technical field
The present embodiments relate to recognition methods and the device of image recognition technology, more particularly to a kind of paper money number.
Background technology
The crown word number of bank note is used to record emission of banknotes sequence, has control note issue quantity and guiding against false of paper currency
Effect.Crown word number can be understood as the identity card of each bank note, and bank or financial self-service equipment can be to flowing in or out
Bank note carries out crown word number record, in order to manage the flow direction of evidence obtaining and tracking bank note.ATM or access integral machine etc.
Financial self-service equipment the true and false of bank note can also be identified according to the crown word number of bank note.As can be seen here, accurately identify
Paper money number is particularly significant.
In the prior art, the crown word number position of part bank note and the floral designs of bank note are closer to the distance, the knowledge to crown word number
Difficulty is not caused, reduces the accuracy of crown word number identification.
The content of the invention
The embodiment of the present invention provides recognition methods and the device of a kind of paper money number, to solve in the prior art to bank note
The problem of identification of crown word number is not accurate enough.
In a first aspect, the embodiments of the invention provide a kind of recognition methods of paper money number, including：
Obtain the crown word number characteristic area of banknote image to be identified；
After the interference region of the crown word number characteristic area is removed, crown word number is carried out to the crown word number characteristic area
Boundary alignment, the boundary alignment result of each character in crown word number is obtained, wherein, it is special that the interference region includes the crown word number
The crown word number and the pixel region with the pixel value difference of the crown word number within a preset range are closed in sign region；
The boundary alignment result of each character in the crown word number is encoded, the pixel for extracting each character becomes
Change feature, and according to the pixel variation characteristic and matching template, identify the content of each character in the crown word number.
Further, after the crown word number characteristic area of banknote image to be identified is obtained, including：
Gray processing processing is carried out to the crown word number characteristic area of the banknote image to be identified, obtains the crown word number feature
The gray level image in region；
According to given threshold, binary conversion treatment is carried out to the gray level image of the crown word number characteristic area, exports the hat
The binaryzation result of font size characteristic area.
Further, the interference region of the crown word number characteristic area, including：
The binaryzation result of the crown word number characteristic area is handled using Hough transformation line detection method, with fixed
The boundary straight line of the position interference region；
Using the boundary straight line as line of demarcation, the hat will not be included in the binaryzation result of the crown word number characteristic area
The region of font size is as interference region.
Further, crown word number boundary alignment is carried out to the crown word number characteristic area, obtains each character in crown word number
Boundary alignment result, including：
The crown word number characteristic area for removing the interference region is scanned by column, determines each word in the crown word number
The left and right border of symbol；
The crown word number characteristic area for removing the interference region is progressively scanned, determines each word in the crown word number
The upper and lower border of symbol.
Further, the crown word number characteristic area for removing the interference region is scanned by column, determines the prefix
The left and right border of each character in number, including：
A row are obtained in the crown word number characteristic area for removing the interference region and are used as currently processed target；
Calculate the row pixel value summation of each pixel of the currently processed target, and by the row pixel value summation and the
One boundary threshold is compared；
If it is determined that the row pixel value summation is more than first boundary threshold, it is determined that is object boundary；
Return to perform and a row are obtained in the crown word number characteristic area for removing the interference region as currently processed
The operation of target, until complete to remove the interference region the crown word number characteristic area in target complete border really
It is fixed；
Sequence notation is carried out to the target complete border according to direction initialization, according to the result of sequence notation, determines institute
State the left margin and right margin of crown word number.
Further, the crown word number characteristic area for removing the interference region is progressively scanned, determines the prefix
The upper and lower border of each character in number, including：
A line is obtained in the crown word number characteristic area for removing the interference region as currently processed target；
Calculate the row pixel value summation of each pixel of the currently processed target, and by the row pixel value summation and the
Two boundary thresholds are compared；
If it is determined that the row pixel value summation is more than the second boundary threshold value, it is determined that is object boundary；
Return to perform and a line is obtained in the crown word number characteristic area for removing the interference region as currently processed
The operation of target, until complete to remove the interference region the crown word number characteristic area in target complete border really
It is fixed；
Sequence notation is carried out to the target complete border according to direction initialization, according to the result of sequence notation, determines institute
State coboundary and the lower boundary of crown word number.
Further, the boundary alignment result of each character in the crown word number is encoded, extracts each word
The pixel variation characteristic of symbol, and according to the pixel variation characteristic and matching template, identify each word in the crown word number
The content of symbol, including：
Using Huffman encoding to being encoded to the boundary alignment result of each character in the crown word number, to extract hat
The pixel light and shade variation characteristic of each character in font size；
The pixel light and shade variation characteristic of each character in the crown word number is compared with the matching template, according to institute
The pixel light and shade variation characteristic of each character in crown word number and the similarity of the matching template are stated, is identified in the crown word number
The content of each character.
Further, the boundary alignment result of each character encodes in the crown word number, and extraction is described each
The pixel variation characteristic of character, and according to the pixel variation characteristic and matching template, identify each in the crown word number
Before the content of character, including：
Binary conversion treatment is carried out to the boundary alignment result of each character in the crown word number using grey stretching method, obtained
Binaryzation result；
The binaryzation result is handled by Mathematical Morphology Method, to fill each character in the crown word number
Breakaway poing and hole；
The edge of each character in the crown word number is extracted using edge detection operator.
Further, gray processing processing is carried out in the crown word number characteristic area to the banknote image to be identified, obtains institute
After the gray level image for stating crown word number characteristic area, including：
Filter out the noise of the gray level image of the crown word number characteristic area.
Second aspect, the embodiment of the present invention additionally provide a kind of identification device of paper money number, including：
Characteristic area acquisition module, for obtaining the crown word number characteristic area of banknote image to be identified；
Crown word number locating module, for after the interference region of the crown word number characteristic area is removed, to the prefix
Number characteristic area carries out crown word number boundary alignment, obtains the boundary alignment result of each character in crown word number, wherein, the interference
Region includes closing on the crown word number in the crown word number characteristic area and with the pixel value difference of the crown word number in preset range
Interior pixel region；
Crown word number identification module, for being encoded to the boundary alignment result of each character in the crown word number, extraction
The pixel variation characteristic of each character, and according to the pixel variation characteristic and matching template, identify the prefix
The content of each character in number.
The embodiment of the present invention obtains prefix by carrying out localization process to the crown word number characteristic area of banknote image to be identified
In number after the border of each character, and feature extraction is carried out to each character in crown word number, identify the specific interior of crown word number
Hold.The embodiment of the present invention can accurately be identified to the crown word number in bank note, strong adaptability, in actual applications, even if
Identification of the pattern of banknote image to crown word number interferes, also can exclusive PCR region recognition go out the content of crown word number, carry
The accuracy rate of crown word number identification is risen, the subsequent treatment for being easy to crown word number to identify.
Brief description of the drawings
Fig. 1 is a kind of flow chart of the recognition methods of paper money number in the embodiment of the present invention one；
Fig. 2 a are a kind of flow charts of the recognition methods of paper money number in the embodiment of the present invention two；
Fig. 2 b are a kind of gray level images of Cuba's coin crown word number characteristic area in the embodiment of the present invention two；
Fig. 3 is a kind of flow chart of the recognition methods of paper money number in the embodiment of the present invention three；
Fig. 4 is a kind of flow chart of the recognition methods of paper money number in the embodiment of the present invention four；
Fig. 5 is a kind of structural representation of the identification device of paper money number in the embodiment of the present invention five.
Embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention, rather than limitation of the invention.It also should be noted that in order to just
Part related to the present invention rather than entire infrastructure are illustrate only in description, accompanying drawing.
It also should be noted that for the ease of description, illustrate only in accompanying drawing part related to the present invention rather than
Full content.It should be mentioned that some exemplary embodiments are described before exemplary embodiment is discussed in greater detail
Into the processing or method described as flow chart.Although operations (or step) are described as the processing of order by flow chart,
It is that many of which operation can be implemented concurrently, concomitantly or simultaneously.In addition, the order of operations can be by again
Arrange.The processing can be terminated when its operations are completed, it is also possible to the additional step being not included in accompanying drawing.
The processing can correspond to method, function, code, subroutine, subprogram etc..
Embodiment one
Fig. 1 is a kind of flow chart of the recognition methods for paper money number that the embodiment of the present invention one provides, and the present embodiment can
The situation of paper money number identification is carried out suitable for the banknote image by interception, this method can be by a kind of paper money number
Identification device perform, the device can be realized by the way of software and/or hardware, and can typically be integrated in self-service finance
In equipment, for example, ATM, access integral machine, cash inspecting machine etc..As shown in figure 1, the method for the present embodiment specifically includes：
S110, the crown word number characteristic area for obtaining banknote image to be identified.
Specifically, the crown word number of bank note specifically refers to Rome number or letter adds the combination of Arabic numerals, variety classes
The particular content of paper money number differ, the position of crown word number also differs on bank note, for example, the prefix of rmb paper currency
Number it is made up of letter and number, the position of crown word number is in the lower left corner of RMB, and the crown word number of Cuba's coin is consist of only digits,
The position of crown word number is in the upper right corner of Cuba's coin.For different types of banknote image to be identified, its corresponding crown word number is obtained
Characteristic area.
Inventor is had found by studying, and when crown word number region is intercepted in the banknote image of collection, is inevitably wrapped
The part floral designs for the bank note that crown word number position is close in banknote image are included, exemplarily, the crown word number of Cuba's coin is set
It is closer to the distance with the floral designs of Cuba coin in the upper right corner of Cuba's coin, when obtaining the crown word number characteristic area of Cuba's coin, meeting
Inevitably the floral designs of introducing portion Cuba coin, the identification to crown word number interfere.Therefore, crown word number characteristic area
Typically not only include crown word number, also including interference region.
S120, after the interference region of the crown word number characteristic area is removed, the crown word number characteristic area is carried out
Crown word number boundary alignment, the boundary alignment result of each character in crown word number is obtained, wherein, the interference region includes the hat
The crown word number and the pixel region with the pixel value difference of the crown word number within a preset range are closed in font size characteristic area.
Specifically, in order to identify the crown word number of bank note exactly, it is necessary to remove the interference impacted to crown word number identification
Region, it is necessary first to identify the border of interference region.In the gray level image of bank note, the grey scale pixel value and hat of interference region
The grey scale pixel value of font size approaches, and within a preset range, preset range can be with for the Pixel gray difference of interference region and crown word number
Specifically set according to actual conditions, the present invention is without limitation.
In an optional embodiment of the present embodiment, Hough transformation line detection method location interference region is used
Border.Exemplarily, in the crown word number characteristic area of Cuba's coin image of acquisition, interference region is located at the top of crown word number,
The border of interference region is oriented by using Hough transformation line detection method.Skilled person will appreciate that for difference
The bank note of species, interference region are not limited to be located at the direction of crown word number not to interference region positioned at the top of crown word number, the present invention
It is limited.
After the interference region of crown word number characteristic area is removed, crown word number border is carried out for crown word number characteristic area
Positioning can position the left and right border of crown word number first, reposition the upper and lower border of crown word number, can also position prefix first
The left and right border of number each character, reposition the upper and lower border of each character of crown word number.By the border of target i.e. from background
Orient and.
In an optional embodiment of the present embodiment, the crown word number characteristic area for removing interference region is swept by column
Retouch, determine the left and right border of each character in the crown word number, the crown word number characteristic area for removing interference region is swept line by line
Retouch, determine the upper and lower border of each character in the crown word number.
S130, the boundary alignment result to each character in the crown word number encode, and extract each character
Pixel variation characteristic, and according to the pixel variation characteristic and matching template, identify each character in the crown word number
Content.
Specifically, being encoded to the boundary alignment result of each character in crown word number, Huffman encoding side can be used
Formula, but Huffman encoding mode is not limited to, by the coding to each character in crown word number, the light and shade for extracting each character becomes
Change feature, then light and shade variation characteristic is contrasted with matching template, by calculating each character and matching template in crown word number
Similarity, it is determined that identifying each character as recognition result with the immediate matching template of character light and shade variation characteristic
Content.
A kind of recognition methods for paper money number that the embodiment of the present invention one provides, the crown word number in bank note can be carried out
Accurately identification, strong adaptability, in actual applications, even if identification of the pattern of banknote image to crown word number interferes,
Can exclusive PCR region recognition go out the content of crown word number, improve the accuracy rate of crown word number identification, be easy to what crown word number identified
Subsequent treatment.
Embodiment two
Fig. 2 a are a kind of flow chart of the recognition methods for paper money number that the embodiment of the present invention two provides, and the present invention is implemented
Example two is optimized based on embodiment one, after the crown word number characteristic area specifically to obtaining banknote image to be identified
Operation further optimize, as shown in Figure 2 a, the embodiment of the present invention specifically includes：
S210, the crown word number characteristic area for obtaining banknote image to be identified.
S220, the crown word number characteristic area to the banknote image to be identified carry out gray processing processing, obtain the prefix
The gray level image of number characteristic area.
S230, according to given threshold, binary conversion treatment is carried out to the gray level image of the crown word number characteristic area, exports institute
State the binaryzation result of crown word number characteristic area.
Preferably, gray processing processing is being carried out to the crown word number characteristic area of the banknote image to be identified, described in acquisition
After the gray level image of crown word number characteristic area, including：
Filter out the noise of the gray level image of the crown word number characteristic area.
Specifically, after the crown word number characteristic area of banknote image to be identified is obtained, crown word number characteristic area is carried out
Gray processing processing, makes the gray value of each pixel in crown word number characteristic area between 0 to 255, wherein, the side of gray processing processing
Method includes histogram equalization, histogram specification etc., and the present invention is without limitation.Ash is being carried out to crown word number characteristic area
Degreeization processing after, processing can be filtered to gray level image using Wiener filter, with filter out gray level image it is steady with
Machine noise.Skilled person will appreciate that the method for filtering process is not limited to Wiener filter, the purpose of filtering is to filter out
The noise of gray level image, the quality of gray level image is lifted, be easy to ensuing image processing operations, the present invention is to filtering process
Method is not limited.After the noise of gray level image is filtered out, the gray level image using given threshold to crown word number characteristic area
Binary conversion treatment is carried out, given threshold does not determine in itself according to bank note, but according to the sensor color of financial self-service equipment
Come what is determined, given threshold can determine the luminance reference value of rank paper according to actual conditions, for example, given threshold can be set
For 20, exemplarily, if the gray value of pixel is more than 20, it is determined that be background, if the gray value of pixel less than or equal to 20,
Then it is defined as crown word number target.Because the purpose for carrying out binary conversion treatment to the gray level image of crown word number characteristic area is in order to right
Crown word number is positioned, so binary conversion treatment is carried out by the method for setting fixed threshold, to lift treatment effeciency.
S240, using Hough transformation line detection method to the binaryzation result of the crown word number characteristic area at
Reason, to position the boundary straight line of the interference region.
S250, using the boundary straight line as line of demarcation, will not include in the binaryzation result of the crown word number characteristic area
The region of the crown word number is as interference region.
Specifically, using Hough transformation line detection method to each picture in the binaryzation result of crown word number characteristic area
Vegetarian refreshments is handled, and Hough transformation line detection method is each pixel in the binaryzation result by crown word number characteristic area
It is mapped to after rectangular coordinate system, is transformed under polar coordinate system, the point belonged under rectangular coordinate system on same straight line, is joining
A plurality of straight line is formed under number coordinate system and is met at a bit.Can be before using Hough transformation line detection method, to crown word number
The binaryzation result of characteristic area carries out rim detection.Hough transformation line detection method is being used to crown word number characteristic area
After binaryzation result is handled, the boundary straight line of interference region is oriented, it is using boundary straight line as line of demarcation, crown word number is special
The region for not including crown word number in the binaryzation result in region is levied as interference region, interference is removed from crown word number characteristic area
Region.
S260, after the interference region of the crown word number characteristic area is removed, the crown word number characteristic area is carried out
Crown word number boundary alignment, the boundary alignment result of each character in crown word number is obtained, wherein, the interference region includes the hat
The crown word number and the pixel region with the pixel value difference of the crown word number within a preset range are closed in font size characteristic area.
S270, the boundary alignment result to each character in the crown word number encode, and extract each character
Pixel variation characteristic, and according to the pixel variation characteristic and matching template, identify each character in the crown word number
Content.
In the present embodiment, Fig. 2 b are a kind of gray-scale map for Cuba's coin crown word number characteristic area that the embodiment of the present invention two provides
Picture, by taking Cuba's coin as an example, referring to Fig. 2 b, interference region is located at the top of crown word number, and crown word number characteristic area filter out making an uproar
After sound and binary conversion treatment, the binaryzation result of crown word number characteristic area is obtained, uses Hough transformation line detection method
The binaryzation result of crown word number characteristic area is handled, it is special from crown word number after being positioned to the border of interference region
Sign removes interference region in region.
A kind of recognition methods for paper money number that the embodiment of the present invention two provides, passes through Hough transformation line detection method
The positioning on interference region border is carried out to the crown word number characteristic area of acquisition, can accurate location interference region, improve hat
The accuracy rate of font size identification.
Embodiment three
Fig. 3 is a kind of flow chart of the recognition methods for paper money number that the embodiment of the present invention three provides, and the present invention is implemented
Improvement is optimized in example three based on the various embodiments described above, to remove the crown word number characteristic area interference region it
Afterwards, crown word number boundary alignment is carried out to the crown word number characteristic area, obtains the boundary alignment result of each character in crown word number
Further illustrated, as shown in figure 3, the method for the embodiment of the present invention three specifically includes：
S310, the crown word number characteristic area for obtaining banknote image to be identified.
S320, the crown word number characteristic area for removing the interference region is scanned by column, determined in the crown word number
The left and right border of each character.
Preferably, the crown word number characteristic area for removing the interference region is scanned by column, determines the crown word number
In each character left and right border, including：
A row are obtained in the crown word number characteristic area for removing the interference region and are used as currently processed target；
Calculate the row pixel value summation of each pixel of the currently processed target, and by the row pixel value summation and the
One boundary threshold is compared；
If it is determined that the row pixel value summation is more than first boundary threshold, it is determined that is object boundary；
Return to perform and a row are obtained in the crown word number characteristic area for removing the interference region as currently processed
The operation of target, until complete to remove the interference region the crown word number characteristic area in target complete border really
It is fixed；
Sequence notation is carried out to the target complete border according to direction initialization, according to the result of sequence notation, determines institute
State the left margin and right margin of crown word number.
Specifically, in the crown word number characteristic area for removing interference region, the pixel value of each pixel in each row is calculated
And be used as row pixel value summation, by row pixel value summation compared with the first boundary threshold, if row pixel value summation is big
In the first boundary threshold, it is determined that be object boundary.Exemplarily, for the crown word number characteristic area of binaryzation, pixel value 0
Background can be represented, pixel value can represent crown word number target for 1, and the first boundary threshold could be arranged to 0, it can also be provided that
Close to 0 numerical value, to filter out the interference of noise.
In addition it is also possible to count remove interference region crown word number characteristic area each row in, pixel value be 1
Number, if statistics number exceedes boundary threshold, it is determined that be object boundary.
Sequence notation is carried out to the target complete border for removing the crown word number characteristic area of interference region according to direction initialization,
Exemplarily, direction initialization can be from left to right, according to first mark and last mark, determine crown word number a left side,
Right margin.It is determined that on the basis of the left and right border of crown word number, it may be determined that be each in crown word number when labeled as odd number
The left margin of character, it is the right margin of each character in crown word number when labeled as even number.
S330, the crown word number characteristic area for removing the interference region is progressively scanned, determined in the crown word number
The upper and lower border of each character.
Preferably, the crown word number characteristic area for removing the interference region is progressively scanned, determines the crown word number
In each character upper and lower border, including：
A line is obtained in the crown word number characteristic area for removing the interference region as currently processed target；
Calculate the row pixel value summation of each pixel of currently processed target, and by the row pixel value summation and the second side
Boundary's threshold value is compared；
If it is determined that the row pixel value summation is more than the second boundary threshold value, it is determined that is object boundary；
Return to perform and a line is obtained in the crown word number characteristic area for removing the interference region as currently processed
The operation of target, until complete to remove the interference region the crown word number characteristic area in target complete border really
It is fixed；
Sequence notation is carried out to the target complete border according to direction initialization, according to the result of sequence notation, determines institute
State coboundary and the lower boundary of crown word number.
Specifically, in the crown word number characteristic area for removing interference region, the pixel value of each pixel in calculating per a line
And be used as row pixel value summation, by row pixel value summation compared with the second boundary threshold value, if row pixel value summation is big
In the second boundary threshold value, it is determined that be object boundary.Exemplarily, for the crown word number characteristic area of binaryzation, pixel value 0
Background can be represented, pixel value can represent crown word number target for 1, and the second boundary threshold value could be arranged to 0, it can also be provided that
Close to 0 numerical value, to filter out the interference of noise.First boundary threshold can be equal with the second boundary threshold value, can also be with
Two boundary thresholds are unequal, and the present invention is without limitation.
In addition it is also possible in counting every a line of the crown word number characteristic area for removing interference region, pixel value be 1 it is individual
Number, if statistics number exceedes boundary threshold, it is determined that be object boundary.
Sequence notation is carried out to the target complete border for removing the crown word number characteristic area of interference region according to direction initialization,
Exemplarily, direction initialization can be from top to bottom, according to first mark and last mark, determine crown word number it is upper,
Lower boundary.It is determined that on the basis of the upper and lower border of crown word number, it may be determined that be each in crown word number when labeled as odd number
The coboundary of character, it is the lower boundary of each character in crown word number when labeled as even number.
S340, the boundary alignment result to each character in the crown word number encode, and extract each character
Pixel variation characteristic, and according to the pixel variation characteristic and matching template, identify each character in the crown word number
Content.
A kind of recognition methods for paper money number that the embodiment of the present invention three provides, can fast and accurately position prefix
The border of each character in number, is easy to the identification of each character in follow-up crown word number.
Example IV
Fig. 4 is a kind of flow chart of the recognition methods for paper money number that the embodiment of the present invention four provides, and the present invention is implemented
Improvement is optimized in example four based on the various embodiments described above, for the boundary alignment knot to each character in the crown word number
Fruit is encoded, and extracts the pixel variation characteristic of each character, and according to the pixel variation characteristic and matching template,
Identify that the content of each character in the crown word number is further illustrated, as shown in figure 4, the side of the embodiment of the present invention four
Method specifically includes：
S410, the crown word number characteristic area for obtaining banknote image to be identified.
S420, after the interference region of the crown word number characteristic area is removed, the crown word number characteristic area is carried out
Crown word number boundary alignment, the boundary alignment result of each character in crown word number is obtained, wherein, the interference region includes the hat
The crown word number and the pixel region with the pixel value difference of the crown word number within a preset range are closed in font size characteristic area.
S430, using grey stretching method in the crown word number each character boundary alignment result carry out binaryzation at
Reason, obtain binaryzation result.
S440, by Mathematical Morphology Method the binaryzation result is handled, it is each in the crown word number to fill
The breakaway poing and hole of individual character.
S450, the edge for extracting using edge detection operator each character in the crown word number.
Specifically, the boundary alignment result of each character in crown word number is more become more meticulous using grey stretching method
Binary conversion treatment, then each character in crown word number is handled using expansion, the method for corrosion and holes filling method, with
Noise is filtered out, and fills the breakaway poing and hole of each character in the crown word number.Finally, extracted and be preced with using edge detection operator
The edge of each character in font size, edge detection operator can be Sobel operators, but be not limited to Sobel operators, and the present invention is to this
It is not limited.
S460, using Huffman encoding to being encoded to the boundary alignment result of each character in the crown word number, with
Extract the pixel light and shade variation characteristic of each character in crown word number.
S470, the pixel light and shade variation characteristic of each character in the crown word number is compared with the matching template,
According to the similarity of the pixel light and shade variation characteristic of each character in the crown word number and the matching template, the hat is identified
The content of each character in font size.
Specifically, being encoded to the boundary alignment result of each character in crown word number, Huffman encoding side can be used
Formula, but Huffman encoding mode is not limited to, by the coding to each character in crown word number, the light and shade for extracting each character becomes
Change feature, then light and shade variation characteristic is contrasted with matching template.Matching template, i.e. standard character storehouse, by by extraction
The light and shade variation characteristic of each character and the character light and shade variation characteristic in standard character storehouse are contrasted in crown word number, calculate both
Between similarity, it is determined that the light and shade variation characteristic with standard character storehouse similarity highest character, identify each in crown word number
The content of individual character.
A kind of recognition methods for paper money number that the embodiment of the present invention four provides, can extract each character in crown word number
Feature, feature is compared with matching template, to obtain the recognition result of character, not only improve identification crown word number effect
Rate, and improve the accuracy of identification crown word number.
Embodiment five
Fig. 5 is a kind of structural representation of the identification device of paper money number in the embodiment of the present invention five, and the device should
For carrying out the situation of paper money number identification by banknote image, the device can be by the way of software and/or hardware
Realize, and can typically be integrated in financial self-service equipment, for example, ATM, access integral machine, cash inspecting machine etc..Such as Fig. 5 institutes
Show, device includes：Characteristic area acquisition module 510, characteristic area acquisition module 520 and crown word number identification module 530, its
In：
Characteristic area acquisition module 510, for obtaining the crown word number characteristic area of banknote image to be identified；
Crown word number locating module 520, for after the interference region of the crown word number characteristic area is removed, to the hat
Font size characteristic area carries out crown word number boundary alignment, obtains the boundary alignment result of each character in crown word number, wherein, it is described dry
Disturbing region includes closing on the crown word number in the crown word number characteristic area and with the pixel value difference of the crown word number in default model
Enclose interior pixel region；
Crown word number identification module 530, for being encoded to the boundary alignment result of each character in the crown word number, carry
The pixel variation characteristic of each character is taken, and according to the pixel variation characteristic and matching template, identifies the hat
The content of each character in font size.
The embodiment of the present invention can accurately be identified to the crown word number in bank note, strong adaptability, in actual applications,
Even if identification of the pattern of banknote image to crown word number interferes, also can exclusive PCR region recognition go out the interior of crown word number
Hold, improve the accuracy rate of crown word number identification, the subsequent treatment for being easy to crown word number to identify.
On the basis of above-described embodiment, described device can include：
Gray processing unit, for carrying out gray processing processing to the crown word number characteristic area of the banknote image to be identified, obtain
Take the gray level image of the crown word number characteristic area；
First binarization unit, for according to given threshold, two to be carried out to the gray level image of the crown word number characteristic area
Value is handled, and exports the binaryzation result of the crown word number characteristic area.
On the basis of above-described embodiment, the crown word number locating module can include：
Interference region boundary alignment unit, for using Hough transformation line detection method to the crown word number characteristic area
Binaryzation result handled, to position the boundary straight line of the interference region；
Interference region determining unit, for using the boundary straight line as line of demarcation, by the two of the crown word number characteristic area
The region not comprising the crown word number is as interference region in value result.
On the basis of above-described embodiment, the crown word number locating module can also include：
First boundary alignment unit, for being scanned by column to the crown word number characteristic area for removing the interference region,
Determine the left and right border of each character in the crown word number；
The second boundary positioning unit, for being progressively scanned to the crown word number characteristic area for removing the interference region,
Determine the upper and lower border of each character in the crown word number.
On the basis of above-described embodiment, the first boundary alignment unit can include：
First currently processed target subelement, for being obtained in the crown word number characteristic area for removing the interference region
A row are taken as currently processed target；
First computation subunit, the row pixel value summation of each pixel for calculating the currently processed target, and will
The row pixel value summation is compared with the first boundary threshold；
First object border determination subelement, for if it is determined that the row pixel value summation is more than first border threshold
Value, it is determined that be object boundary；
First returns to subelement, is obtained for returning to perform in the crown word number characteristic area for removing the interference region
Operation of the row as currently processed target is taken, until in completing the crown word number characteristic area to removing the interference region
The determination on target complete border；
First mark subelement, for carrying out sequence notation to the target complete border according to direction initialization, according to suitable
The result of sequence mark, determine the left margin and right margin of the crown word number.
On the basis of above-described embodiment, the second boundary positioning unit can include：
Second currently processed target subelement, for being obtained in the crown word number characteristic area for removing the interference region
A line is taken as currently processed target；
Second computation subunit, the row pixel value summation of each pixel for calculating the currently processed target, and will
The row pixel value summation is compared with the second boundary threshold value；
Second object boundary determination subelement, for if it is determined that the row pixel value summation is more than the second boundary threshold
Value, it is determined that be object boundary；
Second returns to subelement, is obtained for returning to perform in the crown word number characteristic area for removing the interference region
Operation of a line as currently processed target is taken, until in completing the crown word number characteristic area to removing the interference region
The determination on target complete border；
Second mark subelement, for carrying out sequence notation to the target complete border according to direction initialization, according to suitable
The result of sequence mark, determine coboundary and the lower boundary of the crown word number.
On the basis of above-described embodiment, the crown word number identification module can include：
Coding unit, for using Huffman encoding to being carried out to the boundary alignment result of each character in the crown word number
Coding, to extract the pixel light and shade variation characteristic of each character in crown word number；
Comparing unit, for the pixel light and shade variation characteristic of each character in the crown word number to be entered with the matching template
Row compares, according to the similarity of the pixel light and shade variation characteristic of each character in the crown word number and the matching template, identification
Go out the content of each character in the crown word number.
On the basis of above-described embodiment, described device can also include：
Second binarization unit, for the boundary alignment result using grey stretching method to each character in the crown word number
Binary conversion treatment is carried out, obtains binaryzation result；
Mathematical morphology unit, for being handled by Mathematical Morphology Method the binaryzation result, with filling
The breakaway poing and hole of each character in the crown word number；
Edge detection unit, for extracting the edge of each character in the crown word number using edge detection operator.
On the basis of above-described embodiment, described device can also include：
Noise filtering unit, the noise of the gray level image for filtering out the crown word number characteristic area.
The identification device of paper money number provided in an embodiment of the present invention can perform the paper that any embodiment of the present invention provides
The recognition methods of coin crown word number, possess the corresponding functional module of recognition methods and beneficial effect for performing paper money number.
Pay attention to, above are only presently preferred embodiments of the present invention and institute's application technology principle.It will be appreciated by those skilled in the art that
The invention is not restricted to specific embodiment described here, can carry out for a person skilled in the art various obvious changes,
Readjust and substitute without departing from protection scope of the present invention.Therefore, although being carried out by above example to the present invention
It is described in further detail, but the present invention is not limited only to above example, without departing from the inventive concept, also
Other more equivalent embodiments can be included, and the scope of the present invention is determined by scope of the appended claims.
Claims (10)
- A kind of 1. recognition methods of paper money number, it is characterised in that including：Obtain the crown word number characteristic area of banknote image to be identified；After the interference region of the crown word number characteristic area is removed, crown word number border is carried out to the crown word number characteristic area Positioning, the boundary alignment result of each character in crown word number is obtained, wherein, the interference region includes the crown word number characteristic area The crown word number and the pixel region with the pixel value difference of the crown word number within a preset range are closed in domain；The boundary alignment result of each character in the crown word number is encoded, the pixel change for extracting each character is special Sign, and according to the pixel variation characteristic and matching template, identify the content of each character in the crown word number.
- 2. according to the method for claim 1, it is characterised in that obtaining the crown word number characteristic area of banknote image to be identified Afterwards, including：Gray processing processing is carried out to the crown word number characteristic area of the banknote image to be identified, obtains the crown word number characteristic area Gray level image；According to given threshold, binary conversion treatment is carried out to the gray level image of the crown word number characteristic area, exports the crown word number The binaryzation result of characteristic area.
- 3. according to the method for claim 2, it is characterised in that the interference region of the crown word number characteristic area, including：The binaryzation result of the crown word number characteristic area is handled using Hough transformation line detection method, to position State the boundary straight line of interference region；Using the boundary straight line as line of demarcation, the crown word number will not be included in the binaryzation result of the crown word number characteristic area Region as interference region.
- 4. the method according to claim 1 or 3, it is characterised in that crown word number side is carried out to the crown word number characteristic area Position is defined, obtains the boundary alignment result of each character in crown word number, including：The crown word number characteristic area for removing the interference region is scanned by column, determines each character in the crown word number Left and right border；The crown word number characteristic area for removing the interference region is progressively scanned, determines each character in the crown word number Upper and lower border.
- 5. according to the method for claim 4, it is characterised in that the crown word number characteristic area to removing the interference region Domain scans by column, and determines the left and right border of each character in the crown word number, including：A row are obtained in the crown word number characteristic area for removing the interference region and are used as currently processed target；Calculate the row pixel value summation of each pixel of the currently processed target, and by the row pixel value summation and the first side Boundary's threshold value is compared；If it is determined that the row pixel value summation is more than first boundary threshold, it is determined that is object boundary；Return to execution row of acquisition one in the crown word number characteristic area for removing the interference region and be used as currently processed target Operation, until complete to remove the interference region the crown word number characteristic area in target complete border determination；Sequence notation is carried out to the target complete border according to direction initialization, according to the result of sequence notation, determines the hat The left margin and right margin of font size.
- 6. according to the method for claim 4, it is characterised in that the crown word number characteristic area to removing the interference region Domain progressively scans, and determines the upper and lower border of each character in the crown word number, including：A line is obtained in the crown word number characteristic area for removing the interference region as currently processed target；Calculate the row pixel value summation of each pixel of the currently processed target, and by the row pixel value summation and the second side Boundary's threshold value is compared；If it is determined that the row pixel value summation is more than the second boundary threshold value, it is determined that is object boundary；Return to perform and a line is obtained in the crown word number characteristic area for removing the interference region as currently processed target Operation, until complete to remove the interference region the crown word number characteristic area in target complete border determination；Sequence notation is carried out to the target complete border according to direction initialization, according to the result of sequence notation, determines the hat The coboundary of font size and lower boundary.
- 7. according to the method for claim 1, it is characterised in that to the boundary alignment result of each character in the crown word number Encoded, extract the pixel variation characteristic of each character, and according to the pixel variation characteristic and matching template, know Do not go out the content of each character in the crown word number, including：Using Huffman encoding to being encoded to the boundary alignment result of each character in the crown word number, to extract crown word number In each character pixel light and shade variation characteristic；The pixel light and shade variation characteristic of each character in the crown word number is compared with the matching template, according to the hat The similarity of the pixel light and shade variation characteristic of each character and the matching template, is identified each in the crown word number in font size The content of character.
- 8. the method according to claim 1 or 7, it is characterised in that the border of each character is determined in the crown word number Position result is encoded, and extracts the pixel variation characteristic of each character, and according to the pixel variation characteristic and matching Template, identify in the crown word number before the content of each character, including：Binary conversion treatment is carried out to the boundary alignment result of each character in the crown word number using grey stretching method, obtains two-value Change result；The binaryzation result is handled by Mathematical Morphology Method, to fill the disconnected of each character in the crown word number Knick point and hole；The edge of each character in the crown word number is extracted using edge detection operator.
- 9. according to the method for claim 2, it is characterised in that in the crown word number characteristic area to the banknote image to be identified Domain carries out gray processing processing, after the gray level image for obtaining the crown word number characteristic area, including：Filter out the noise of the gray level image of the crown word number characteristic area.
- A kind of 10. identification device of paper money number, it is characterised in that including：Characteristic area acquisition module, for obtaining the crown word number characteristic area of banknote image to be identified；Crown word number locating module, it is special to the crown word number for after the interference region of the crown word number characteristic area is removed Levy region and carry out crown word number boundary alignment, obtain the boundary alignment result of each character in crown word number, wherein, the interference region Including closed in the crown word number characteristic area crown word number and with the pixel value difference of the crown word number within a preset range Pixel region；Crown word number identification module, for being encoded to the boundary alignment result of each character in the crown word number, described in extraction The pixel variation characteristic of each character, and according to the pixel variation characteristic and matching template, identify in the crown word number The content of each character.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710936146.0A CN107742357A (en) | 2017-10-10 | 2017-10-10 | A kind of recognition methods of paper money number and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710936146.0A CN107742357A (en) | 2017-10-10 | 2017-10-10 | A kind of recognition methods of paper money number and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107742357A true CN107742357A (en) | 2018-02-27 |
Family
ID=61237182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710936146.0A Pending CN107742357A (en) | 2017-10-10 | 2017-10-10 | A kind of recognition methods of paper money number and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107742357A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109584220A (en) * | 2018-11-15 | 2019-04-05 | 锐捷网络股份有限公司 | A kind of silk-screen localization method, silk-screen defect inspection method and relevant apparatus |
| CN110880002A (en) * | 2018-09-06 | 2020-03-13 | 深圳怡化电脑股份有限公司 | Account stain removal method and device and computer readable storage medium |
| CN111079735A (en) * | 2018-10-22 | 2020-04-28 | 莱芜钢铁集团电子有限公司 | Method and device for recognizing end face characters of casting blank |
| CN116486418A (en) * | 2023-06-19 | 2023-07-25 | 恒银金融科技股份有限公司 | Method and device for generating banknote crown word number image |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1685373A (en) * | 2003-03-14 | 2005-10-19 | 富士通株式会社 | Paper sheet identifying method and paper sheet identifying 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 |
| CN102779275A (en) * | 2012-07-04 | 2012-11-14 | 广州广电运通金融电子股份有限公司 | Paper characteristic identification method and relative device |
| CN103679918A (en) * | 2014-01-06 | 2014-03-26 | 成都术有科技有限公司 | High-speed banknote serial number extraction and identification method based on DSP (digital signal processor) |
| CN105006062A (en) * | 2015-07-29 | 2015-10-28 | 深圳怡化电脑股份有限公司 | Bill identification method and deposit and withdrawal device |
| CN105447956A (en) * | 2015-11-06 | 2016-03-30 | 东方通信股份有限公司 | Spliced banknote detection method |
| CN106447909A (en) * | 2016-09-23 | 2017-02-22 | 深圳怡化电脑股份有限公司 | Vertical serial number identification method and vertical serial number identification device |
| CN106600810A (en) * | 2016-12-05 | 2017-04-26 | 深圳怡化电脑股份有限公司 | Currency identification method and currency identification device |
| CN106710063A (en) * | 2016-12-27 | 2017-05-24 | 陕西科技大学 | RMB serial number recognition method based on computer vision |
| CN106845542A (en) * | 2017-01-18 | 2017-06-13 | 江苏阿瑞斯智能设备有限公司 | Paper money number intelligent identification Method based on DSP |
| CN107123188A (en) * | 2016-12-20 | 2017-09-01 | 北京联合众为科技发展有限公司 | Ticket of hindering based on template matching algorithm and edge feature is recognized and localization method |
-
2017
- 2017-10-10 CN CN201710936146.0A patent/CN107742357A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1685373A (en) * | 2003-03-14 | 2005-10-19 | 富士通株式会社 | Paper sheet identifying method and paper sheet identifying 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 |
| CN102779275A (en) * | 2012-07-04 | 2012-11-14 | 广州广电运通金融电子股份有限公司 | Paper characteristic identification method and relative device |
| CN103679918A (en) * | 2014-01-06 | 2014-03-26 | 成都术有科技有限公司 | High-speed banknote serial number extraction and identification method based on DSP (digital signal processor) |
| CN105006062A (en) * | 2015-07-29 | 2015-10-28 | 深圳怡化电脑股份有限公司 | Bill identification method and deposit and withdrawal device |
| CN105447956A (en) * | 2015-11-06 | 2016-03-30 | 东方通信股份有限公司 | Spliced banknote detection method |
| CN106447909A (en) * | 2016-09-23 | 2017-02-22 | 深圳怡化电脑股份有限公司 | Vertical serial number identification method and vertical serial number identification device |
| CN106600810A (en) * | 2016-12-05 | 2017-04-26 | 深圳怡化电脑股份有限公司 | Currency identification method and currency identification device |
| CN107123188A (en) * | 2016-12-20 | 2017-09-01 | 北京联合众为科技发展有限公司 | Ticket of hindering based on template matching algorithm and edge feature is recognized and localization method |
| CN106710063A (en) * | 2016-12-27 | 2017-05-24 | 陕西科技大学 | RMB serial number recognition method based on computer vision |
| CN106845542A (en) * | 2017-01-18 | 2017-06-13 | 江苏阿瑞斯智能设备有限公司 | Paper money number intelligent identification Method based on DSP |
Non-Patent Citations (1)
| Title |
|---|
| 梁杨等: "纸币冠字号预处理及组合特征识别方法", 《计算机工程与设计》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110880002A (en) * | 2018-09-06 | 2020-03-13 | 深圳怡化电脑股份有限公司 | Account stain removal method and device and computer readable storage medium |
| CN111079735A (en) * | 2018-10-22 | 2020-04-28 | 莱芜钢铁集团电子有限公司 | Method and device for recognizing end face characters of casting blank |
| CN111079735B (en) * | 2018-10-22 | 2023-12-22 | 莱芜钢铁集团电子有限公司 | Casting blank end face character recognition method and device |
| CN109584220A (en) * | 2018-11-15 | 2019-04-05 | 锐捷网络股份有限公司 | A kind of silk-screen localization method, silk-screen defect inspection method and relevant apparatus |
| CN116486418A (en) * | 2023-06-19 | 2023-07-25 | 恒银金融科技股份有限公司 | Method and device for generating banknote crown word number image |
| CN116486418B (en) * | 2023-06-19 | 2023-10-03 | 恒银金融科技股份有限公司 | Method and device for generating banknote crown word number image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106056751B (en) | The recognition methods and system of serial number | |
| CN108694393A (en) | A kind of certificate image text area extraction method based on depth convolution | |
| CN107742357A (en) | A kind of recognition methods of paper money number and device | |
| Abburu et al. | Currency recognition system using image processing | |
| WO2007127085A1 (en) | Generating a bitonal image from a scanned colour image | |
| CN107563380A (en) | A kind of vehicle license plate detection recognition method being combined based on MSER and SWT | |
| CN105046252A (en) | Method for recognizing Renminbi (Chinese currency yuan) crown codes | |
| CN107195069A (en) | A kind of RMB crown word number automatic identifying method | |
| CN103606220A (en) | Check printed number recognition system and check printed number recognition method based on white light image and infrared image | |
| CN110598566A (en) | Image processing method, device, terminal and computer readable storage medium | |
| Sharan et al. | Detection of counterfeit Indian currency note using image processing | |
| CN105654609A (en) | Paper money processing method and paper money processing system | |
| Arya et al. | Fake currency detection | |
| Suresh et al. | Indian currency recognition and verification using image processing | |
| Sawant et al. | Currency recognition using image processing and minimum distance classifier technique | |
| CN106340116B (en) | A kind of recognition methods of bank note and device | |
| CN107527423B (en) | Paper money anti-counterfeiting identification method and paper money identification device | |
| Rajan et al. | An extensive study on currency recognition system using image processing | |
| CN109543554B (en) | Bill detection method, device, terminal and computer readable storage medium | |
| CN106846609A (en) | It is a kind of based on perceiving the bank note face amount of Hash towards recognition methods | |
| KR102094234B1 (en) | Serial number recognition Apparatus and method for paper money | |
| CN111753842B (en) | Method and device for detecting text region of bill | |
| Shyju et al. | Indian currency identification using image processing | |
| CN108960222B (en) | Image binarization method, device, equipment and storage medium | |
| CN107240185A (en) | A kind of crown word number identification method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180227 |