CN106898085B - Identify bank note towards method and device - Google Patents
Identify bank note towards method and device Download PDFInfo
- Publication number
- CN106898085B CN106898085B CN201710100698.8A CN201710100698A CN106898085B CN 106898085 B CN106898085 B CN 106898085B CN 201710100698 A CN201710100698 A CN 201710100698A CN 106898085 B CN106898085 B CN 106898085B
- Authority
- CN
- China
- Prior art keywords
- region
- image
- area
- template
- pixel
- 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
- G07D7/2008—Testing patterns thereon using pre-processing, e.g. de-blurring, averaging, normalisation or rotation
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses identification bank note towards method and device.This method comprises: choosing the central point central symmetry of first area and second area, second area and first area about the first face of bank note to be identified image on the first face image of bank note to be identified;Third region is chosen on the second face image of setting in the first face image and the fourth region, third region and the fourth region are identical in the position of the first face image as first area and second area respectively in the position of the second face image;First area, second area, third region and the corresponding image template of the fourth region are matched with the reference template in pre-set image region respectively;According to matching result determine bank note to be identified towards.The embodiment of the present invention determined according to the matching result of the image template in four regions of bank note to be identified and reference template towards, image template and reference template include bank note pattern form, by relatively being identified towards accuracy rate height with template.
Description
Technical field
The present embodiments relate to image recognition technology more particularly to it is a kind of identification bank note towards method and device.
Background technique
Bank note is the basis of banknote validation towards identification, only determine bank note it is whole towards rear, money-checking could be allowed to set
The standby each anti-fake region for finding bank note, and verified for each anti-fake region.
It is a kind of identification bank note towards the prior art be: choose bank note dual-side image on several specific regions, by this
The gray value mean value of several specific regions is compared to each other, according to mutual comparison result determine bank note towards.
The above-mentioned prior art has the disadvantage that placement of the bank note in money-checking equipment is inevitably offset, and the region of selection is not
It is enough accurate;The information extracted from bank note is very few, and gray value mean value cannot characterize pattern form, just with gray value mean value this
One seldom information, bank note towards recognition result inaccuracy.
Summary of the invention
The embodiment of the present invention provide identification bank note towards method and device, recognition efficiency can be improved and identification is accurate
Rate.
In a first aspect, the embodiment of the invention provides it is a kind of identification bank note towards method, comprising:
Choose first area and second area on the first face image of bank note to be identified, the second area and described the
Central point central symmetry of one region about the first face of bank note to be identified image;
Third region and the fourth region are being chosen on the second face image of setting with first face image, described the
Three regions and the fourth region exist with the first area and the second area respectively in the position of second face image
The position of first face image is identical;
By the corresponding image template in the first area, the corresponding image template of the second area, the third region
Corresponding image template and the reference with pre-stored pre-set image region respectively of the corresponding image template of the fourth region
Template is matched;
According to matching result determine the bank note to be identified towards.
Second aspect, the embodiment of the invention also provides it is a kind of identification bank note towards device, comprising:
Module is chosen in first and second region, for choosing first area and the on the first face image of bank note to be identified
Two regions, the second area and center dot center pair of the first area about the first face of bank note to be identified image
Claim;
Third and the fourth region choose module, choose module with first and second described region and are connected, be used for it is described
First face image chooses third region and the fourth region, the third region and the described 4th on the second face image of setting
Region is in the position of second face image respectively with the first area and the second area in first face image
Position is identical;
Matching module chooses module with the third and the fourth region and is connected, is used for the corresponding figure in the first area
As template, the corresponding image template of the second area, the corresponding image template in the third region and the fourth region pair
The image template answered is matched with the reference template in pre-stored pre-set image region respectively;
Towards determining module, it is connected with the matching module, for determining the bank note to be identified according to matching result
Towards.
The embodiment of the present invention is according to the image template in four regions of bank note to be identified and the matching result of reference template come really
It is fixed towards relative to the pattern for using gray value mean value identification bank note towards, image template and reference template including bank note
Shape, by with template relatively to identify it is higher towards accuracy rate.
Detailed description of the invention
Fig. 1 be the embodiment of the present invention one provide a kind of identification bank note towards method flow diagram;
Fig. 2 be it is provided by Embodiment 2 of the present invention it is a kind of identification bank note towards method flow diagram;
Fig. 3 be the embodiment of the present invention three provide a kind of identification bank note towards method flow diagram;
Fig. 4 be the embodiment of the present invention four provide a kind of identification bank note towards method flow diagram;
Fig. 5 be the embodiment of the present invention five provide a kind of identification bank note towards method flow diagram;
Fig. 6 be the embodiment of the present invention six provide a kind of identification bank note towards apparatus structure schematic diagram.
Specific 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 limiting the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
Embodiment one
Fig. 1 be a kind of identification bank note that provides of the embodiment of the present invention one towards method flow diagram, the present embodiment is applicable
In finance device identifies bank note the case where, this method can by identification bank note towards device execute, which can be by soft
Part and/or hardware mode realize that the device can integrate in any equipment for providing paper money recognition, such as typically money-checking
Machine or automatic teller machine etc..This method comprises:
S101, first area and second area, second area and first are chosen on the first face image of bank note to be identified
Central point central symmetry of the region about the first face of bank note to be identified image.
Wherein, the first face image can be the front or back of bank note to be identified.
Imaging sensor is generally equipped on paper money identification equipment, the image on the two sides for acquiring bank note to be identified.To know
Other bank note is towards selection first area and second area, first area and the secondth area on the first face image of bank note to be identified
Domain can be the identical arbitrary shape of shape, generally choose rectangle for convenience, second area and first area are about wait know
The a certain square in the lower left corner on the central point central symmetry of other the first face of bank note image, such as the optional first face image in first area
Shape, the rectangle that second area is centrosymmetric with regard to the upper right corner on the image of optional first face and first area, in this way no matter bank note why
Sample placement can accurately choose the two regions.
S102, third region and the fourth region, third are being chosen on the second face image of setting with the first face image
Region and the fourth region the position of the second face image respectively with first area and second area the first face image position phase
Together.
Wherein, the first face image is corresponding with the front of bank note to be identified, reverse side with the second face image.When the first face image
For bank note to be identified front when, then the second face image be bank note to be identified reverse side, when the first face image be bank note to be identified
Reverse side when, then the second face image be bank note to be identified front.
Positive, positive reversed, reverse side forward direction and the reverse side towards respectively front of bank note are reversed, and bank note may be with any surface
To being placed in finance device.In order to guarantee bank note be in it is any towards when can obtain the image of specific region, with the first face
Image chooses third region and the fourth region, third region and the fourth region in the second face figure on the second face image of setting
The position of picture is identical in the position of the first face image as first area and second area respectively.Such as the optional bank note in third region
The rectangle identical with first area position in the two faces image lower left corner, the fourth region optional the second face of bank note image upper right corner and the
Three regional locations are identical.
S103, by the corresponding image template in first area, the corresponding image template of second area, the corresponding figure in third region
As template and the corresponding image template of the fourth region are matched with the reference template in pre-stored pre-set image region respectively.
The reference template in pre-stored pre-set image region can select it is a certain towards bank note first area, second
The corresponding image template in a certain region in region, third region or the fourth region, so pre-stored pre-set image region
Reference template be known.The first area obtained from the first face image and the second face image of bank note to be identified is corresponding
The corresponding image template of image template, second area, the corresponding image template in third region and the corresponding image mould of the fourth region
Plate is unknown.
By the corresponding image template in first area, the corresponding image template of second area, the corresponding image mould in third region
Plate and the corresponding image template of the fourth region are matched with the reference template in pre-stored pre-set image region respectively, according to
Region recognition bank note where matching is best towards.
Image template and reference template include the pattern form of bank note, identify bank note relative to using gray value mean value
Towards, by with template relatively to identify it is higher towards accuracy rate.
S104, determined according to matching result bank note to be identified towards.
The corresponding image template in first area, the corresponding image template of second area, the corresponding image template in third region
And one in the corresponding image template of the fourth region this four image templates is likely to occur and pre-stored pre-set image area
The similar image of the reference template in domain, the image in which specific region and the reference template in pre-stored pre-set image region
It is similar depending on bank note to be identified towards four towards respectively corresponding four kinds of situations.
For 2005 editions hundred yuans, there is phonetic " YUAN " in the lower left corner for the side for being printed on the Great Hall of the People
Word, if using the rectangular area comprising " YUAN " word as pre-set image region.Bank note to be identified towards it is unknown in the case where,
" YUAN " word is possible to appear in one in first area, second area, third region or the fourth region, respectively corresponds four kinds
Towards." YUAN " word occur region be matching result, according to matching result can determine bank note to be identified towards.For example, if
" YUAN " word appears in first area and then shows that bank note to be identified is that reverse side is positive, the table if " YUAN " word appears in second area
Bright bank note to be identified is that reverse side is reversed, shows that bank note to be identified is reversed for front if " YUAN " word appears in third region, if
" YUAN " word appears in the fourth region and then shows that bank note to be identified is positive for front.
The present embodiment determines face according to the matching result of the image template in four regions of bank note to be identified and reference template
To, relative to using gray value mean value identification bank note include towards, image template and reference template bank note pattern form,
By with template relatively to identify it is higher towards accuracy rate.
Embodiment two
Fig. 2 be it is provided by Embodiment 2 of the present invention it is a kind of identification bank note towards method flow diagram, embodiment two is with above-mentioned
Based on embodiment, before choosing first area and second area on the first face image of bank note to be identified, selection is increased
Pre-set image region is handled to obtain reference template to pre-set image region.Correspondingly, the method for the present embodiment includes:
S201, any one following region is chosen on the direct picture of reference bank note as pre-set image region: reversed
It is the first area of placement, the positive first area placed, oppositely positioned second area, the positive second area placed, reversed
The third region of placement, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed;Or
Person chooses any one following region as pre-set image region: oppositely positioned first on the verso images of reference bank note
Region, the positive first area placed, oppositely positioned second area, the positive second area placed, oppositely positioned third
Region, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed.
Optional one with reference in four regions on bank note front or back image in pre-set image region it is positive place or
Oppositely positioned image, regular shape and sharp-edged image are generally selected in pre-set image region, in this way convenient for paper to be identified
The image template of coin is matched, and also can guarantee towards the accurate of recognition result.
S202, pre-set image region is handled to obtain reference template.
If do not handled pre-set image region and directly by the image template of bank note to be identified and reference template
It can be larger with operand.The reference template obtained after handling pre-set image region, compared to the pre-set image before processing
Region, image is simplified and feature is more obvious, improves recognition efficiency and recognition accuracy.
Reference template is obtained from reference bank note, when bank note to be identified is matched with reference template, can be identified
Towards can also play the role of discerning the false from the genuine.If four regions of bank note to be identified are low with the matched optimum value of reference template
In the threshold value of genuine note, then illustrate that bank note to be identified is counterfeit money or is not corresponding currency type.
S203, first area and second area, second area and first are chosen on the first face image of bank note to be identified
Central point central symmetry of the region about the first face of bank note to be identified image.
S204, third region and the fourth region, third are being chosen on the second face image of setting with the first face image
Region and the fourth region the position of the second face image respectively with first area and second area the first face image position phase
Together.
S205, by the corresponding image template in first area, the corresponding image template of second area, the corresponding figure in third region
As template and the corresponding image template of the fourth region are matched with the reference template in pre-stored pre-set image region respectively.
S206, determined according to matching result bank note to be identified towards.
The present embodiment pre-set image region and is handled to obtain reference template by choosing to pre-set image region, will be to
Identify that the image template of bank note and reference template are compared to determine towards can accurately identify bank note towards can also play mirror
The effect of the other true and false.Because reference template is handled by pre-set image region, the efficiency towards identification is improved.
Embodiment three
Fig. 3 be the embodiment of the present invention three provide a kind of identification bank note towards method flow diagram, embodiment three is with above-mentioned
Based on embodiment, by the corresponding image template in first area, the corresponding image template of second area, the corresponding figure in third region
As template and the corresponding image template of the fourth region are matched with the reference template in pre-stored pre-set image region respectively
Process be optimized, in the image template and reference template for the first time after it fails to match of bank note to be identified, to image template
It is overturn and is matched again with reference template.Correspondingly, the method for the present embodiment includes:
S301, any one following region is chosen on the direct picture of reference bank note as pre-set image region: reversed
It is the first area of placement, the positive first area placed, oppositely positioned second area, the positive second area placed, reversed
The third region of placement, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed;Or
Person chooses any one following region as pre-set image region: oppositely positioned first on the verso images of reference bank note
Region, the positive first area placed, oppositely positioned second area, the positive second area placed, oppositely positioned third
Region, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed.
S302, pre-set image region is handled to obtain reference template.
S303, first area and second area, second area and first are chosen on the first face image of bank note to be identified
Central point central symmetry of the region about the first face of bank note to be identified image.
S304, third region and the fourth region, third are being chosen on the second face image of setting with the first face image
Region and the fourth region the position of the second face image respectively with first area and second area the first face image position phase
Together.
S305, in the corresponding image template in first area, the corresponding image template of second area, the corresponding figure in third region
Picture template and the corresponding image template of the fourth region carry out in first time matching process with reference template respectively, if there is any one area
It fails to match for the image template in domain, then carries out following at least one turning operation to the image of failure: being flipped up, turn over downwards
Turn, overturn and overturn to the right to the left;Image template after overturning is matched with reference template.
For 2005 editions hundred yuans, using positive " YUAN " word as reference template, table 1 show one kind towards with
The relationship of each area image template gives four kinds of original graphs towards corresponding each region of paper money identification equipment acquisition in table
Picture, it can be seen that " YUAN " word that some regions occur turns around, so image template and reference template in each region
It is necessary to carry out turning operation to region when it fails to match for the first time.The present embodiment using the region in the first face image lower left corner as
First area, no matter bank note to be identified with which kind of towards being placed, can one in four regions find just or
" YUAN " word upsided down, according to " YUAN " word which region occur can recognize towards." YUAN " word goes out when bank note front is positive
Present the fourth region, " YUAN " word appears in third region when bank note front is reversed, and " YUAN " word occurs when bank note reverse side forward direction
In first area, " YUAN " word appears in second area when bank note reverse side is reversed.Occur when bank note reverse side is reversed in second area
" YUAN " word and when bank note front positive all turned around in " YUAN " word that the fourth region occurs, with reference template
Positive " YUAN " word cannot be matched directly, when it fails to match for the first time for the image template of bank note to be identified and reference template, then
Image template match again with reference template after overturning up and down, can guarantee that one in four regions finds in this way
Positive " YUAN " word.
Table 1 is towards the relationship with each area image template
It should be pointed out that table 1 in bank note towards for convenience, be subject to the first face image towards.According to paper
Coin identifies that equipment obtains the difference of image mode, the direction relations of the first face image and the second face image in each column, it may be possible to
Forward direction another reversed situation in table 1, it is also possible to it is positive simultaneously or same the first face image or the second face image occur
The case where Shi Fanxiang, either way may recognize that bank note towards.
S306, determined according to matching result bank note to be identified towards.
The present embodiment bank note to be identified image template and reference template for the first time after it fails to match, to image template into
Row overturning is matched with reference template again, enable bank note to be identified difference towards when always have the image template in a region
With reference template accurate match, ensure that towards the accurate of recognition result.
Example IV
Fig. 4 be the embodiment of the present invention four provide a kind of identification bank note towards method flow diagram, example IV is with above-mentioned
Based on embodiment, pre-set image region will be handled obtain reference template and be optimized for, pre-set image region will be dropped
Sampling and binary conversion treatment, the boundary in pre-set image region is determined according to the image after binaryzation, obtains reference template.Also to root
Determine that the boundary in pre-set image region has carried out further optimization according to the image after binaryzation.Correspondingly, the side of the present embodiment
Method includes:
S401, any one following region is chosen on the direct picture of reference bank note as pre-set image region: reversed
It is the first area of placement, the positive first area placed, oppositely positioned second area, the positive second area placed, reversed
The third region of placement, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed;Or
Person chooses any one following region as pre-set image region: oppositely positioned first on the verso images of reference bank note
Region, the positive first area placed, oppositely positioned second area, the positive second area placed, oppositely positioned third
Region, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed.
S402, down-sampled and binary conversion treatment is carried out to pre-set image region.
Pre-set image region includes multiple pixels, if whole pixels, which is carried out matching, will increase operand, can also be dropped
Low matching speed, down-sampled includes that adjacent multiple pixels are synthesized a pixel, and the gray value of the pixel after synthesis takes conjunction
At preceding multiple pixels gray value mean value or in which a gray value.Down-sampled be equivalent to breaks mosaic, i.e., by pre-set image
Region blur, to reduce matched operand and improve matching speed.
Pre-set image region gray scale of each pixel before carrying out binary conversion treatment changes between 0-255, and gray value is not
It is unified also to will increase operand and reduce matching speed.To the binary conversion treatment that pre-set image region carries out, can be, by gray scale
The gray value for being worth higher pixel is uniformly set to identical high gray value, and the gray value of the lower pixel of gray value is uniformly set to
Identical low ash angle value is convenient for calculating and improving efficiency in this way.
S403, the boundary that pre-set image region is determined according to the image after binaryzation, obtain reference template.
Minor shifts up and down are likely to occur when bank note is placed in finance device.Generally compare in pre-set image region
Reference template wants larger, even if minor shifts occur for bank note, also can determine pre-set image region according to the image after binaryzation
Boundary, and obtain reference template.Determine that the boundary in pre-set image region can be by the background decorative pattern or figure of reference template and surrounding
Case distinguishes, and improves matched speed and accuracy.
The boundary that pre-set image region is determined according to the image after binaryzation includes: to calculate row according to the image after binaryzation
The gray value of pixel sums it up and the gray value of column pixel adduction;There is into the position being mutated in the gray value adduction of row pixel
The row bound as pre-set image region is set, and the position that being mutated is occurred in the gray value of column pixel adduction is as pre-set image
The column boundary in region;Or, corresponding row is as pre-set image region when the gray value for the pixel for setting line number is summed it up maximum
Row bound, will set the pixel of columns gray value adduction it is maximum when column side as pre-set image region of corresponding column
Boundary, setting line number are less than the line number in pre-set image region, and setting columns is less than the columns in pre-set image region.
For determining the row bound in pre-set image region, the gray scale of row pixel is first calculated according to the image after binaryzation
Value adduction, at the row bound in pre-set image region, the gray value adduction of row pixel can mutate, therefore can be by row pixel
Row bound of the gray value adduction position that occurs being mutated as pre-set image region.Because setting line number is less than pre-set image area
The line number in domain, so choosing setting line number from pre-set image region, there are many selection, the corresponding setting rows of different start-stop rows
The gray value of several pixels sums it up difference, should cover point most on pattern with the row of description selection when maximum, corresponding
Most uplink and most downlink can be used as the row bound in pre-set image region.Determine that the column boundary in pre-set image region is similar.
The boundary that pre-set image region is determined according to the image after binaryzation includes: to be calculated respectively according to the image after binaryzation
The lateral difference of pixel;By the absolute value of the pixel transverse direction difference of every row it is cumulative obtain row difference and, and to row difference and progress
Difference, by the row bound at the maximum absolute value of the difference of row difference sum as pre-set image region;Or, the row that line number will be set
Row bound of the difference with row corresponding when cumulative maximum as pre-set image region, setting line number are less than the row in pre-set image region
Number;Longitudinal difference of each pixel is calculated according to the image after binaryzation;The absolute value of the pixel longitudinal direction difference of each column is cumulative
And, and to column difference and difference is carried out to column difference, and pre-set image area will be used as at the maximum absolute value of the difference of column difference sum
The column boundary in domain;Or, column boundary of the column difference of columns with column corresponding when cumulative maximum as pre-set image region will be set,
Set the columns that columns is less than pre-set image region.
For determining the row bound in pre-set image region, the transverse direction for first calculating each pixel according to the image after binaryzation is poor
Point, by the absolute value of the pixel transverse direction difference of every row it is cumulative obtain row difference and, during determining row bound row difference and with
The effect of the gray value adduction of front row pixel is similar, the difference of row difference sum at the row bound in pre-set image region
Maximum absolute value, therefore can be by the row bound at the maximum absolute value of the difference of row difference sum as pre-set image region.If
Calculate row difference sum difference when be with the row difference of downlink and subtract neighboring upstream row difference and, then be at difference maximum value
Coboundary is lower boundary at difference minimum value.Because the line number that line number is less than pre-set image region is set, from pre-set image
There are many selections for selection setting line number in region, and the row difference and accumulated value of the corresponding setting line number of different start-stop rows are different,
This and it is maximum when description selection row cover point most on pattern, corresponding most uplink and most downlink can be used as pre-set image
The row bound in region.Determine that the column boundary in pre-set image region is similar.
The lateral difference and longitudinal difference that each pixel is calculated according to the image after binaryzation, highlight the difference of foreground and background
Different, gained boundary is more accurate, and anti-noise ability is strong.
First area, second area, third region or the corresponding image template of the fourth region of bank note to be identified can also be into
The processing such as the down-sampled, binaryzation of row and determining boundary, process is similar with the processing to pre-set image region, the figure of bank note to be identified
As template can be matched preferably after identical processing with reference template.
S404, first area and second area, second area and first are chosen on the first face image of bank note to be identified
Central point central symmetry of the region about the first face of bank note to be identified image.
S405, third region and the fourth region, third are being chosen on the second face image of setting with the first face image
Region and the fourth region the position of the second face image respectively with first area and second area the first face image position phase
Together.
S406, by the corresponding image template in first area, the corresponding image template of second area, the corresponding figure in third region
As template and the corresponding image template of the fourth region are matched with the reference template in pre-stored pre-set image region respectively.
S407, determined according to matching result bank note to be identified towards.
The present embodiment reduces operand and improves face by carrying out down-sampled and binary conversion treatment to pre-set image region
To the efficiency of identification, determines that the boundary in pre-set image region obtains reference template according to the image after binaryzation, improve towards knowledge
Other accuracy rate.The difference of foreground and background is highlighted according to the method that each pixel difference of image after binaryzation determines boundary,
Gained boundary is more accurate, and anti-noise ability is strong.
Embodiment five
Fig. 5 be the embodiment of the present invention five provide a kind of identification bank note towards method flow diagram, embodiment five is with above-mentioned
Based on embodiment, binary conversion treatment will be carried out to pre-set image region, be optimized for, according to the corresponding histogram in pre-set image region
The number of pixels percentage of figure determines gray threshold, and the gray scale primary system of the pixel of gray threshold will be lower than in pre-set image region
One is set to lowest gray value, and the gray value higher than the pixel of gray threshold is uniformly set to highest gray value.Correspondingly, this implementation
Example method include:
S501, any one following region is chosen on the direct picture of reference bank note as pre-set image region: reversed
It is the first area of placement, the positive first area placed, oppositely positioned second area, the positive second area placed, reversed
The third region of placement, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed;Or
Person chooses any one following region as pre-set image region: oppositely positioned first on the verso images of reference bank note
Region, the positive first area placed, oppositely positioned second area, the positive second area placed, oppositely positioned third
Region, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed.
S502, to pre-set image region carry out it is down-sampled, according to the number of pixels of the corresponding histogram in pre-set image region
Percentage determines gray threshold.
Binary conversion treatment is carried out to pre-set image region to need first to determine gray threshold, it can be by each pixel in pre-set image region
Gray scale obtain grey level histogram, the percentage as shared by low gray-scale pixels can determine gray threshold.If using " YUAN " word as
The pixel of reference template, the covering of " YUAN " word accounts for 25% in reference template, so pre-set image region pair can be used as by 25%
The number of pixels percentage for the histogram answered." YUAN " word can be highlighted when determining gray threshold in this way.
Determine that gray threshold includes: by gray value according to the number of pixels percentage of the corresponding histogram in pre-set image region
It is ascending that the pixel for meeting number of pixels percentage is chosen from pre-set image region;It is selected in the pixel low from gray value
Take maximum gradation value as gray threshold, or from residual pixel point choose be greater than maximum gradation value, and with maximum gradation value phase
Adjacent gray value is as gray threshold.
Residual pixel point refers to unselected pixel.Assuming that reference template has 100 pixels, percentage 25%,
Lower 25 pixels of gray value are then selected from 100 pixels, can using maximum gradation value in this 25 pixels as
Gray threshold, can also be using gray value adjacent with maximum gradation value in remaining 75 pixels as gray threshold.
S503, the gray value for being lower than the pixel of gray threshold in pre-set image region is uniformly set to lowest gray value,
Gray value higher than the pixel of gray threshold is uniformly set to highest gray value.
Such as can be by pre-set image region, the gray value lower than the pixel of gray threshold is uniformly set to 0, is higher than gray scale
The gray value of the pixel of threshold value is uniformly set to 255, and this completes the binary conversion treatments to pre-set image region.Default figure
As there are many pixels on region, after carrying out binary conversion treatment according to gray threshold, the image mould of bank note to be identified can be reduced
Operand when plate is matched with reference template improves recognition efficiency.
S504, the boundary that pre-set image region is determined according to the image after binaryzation, obtain reference template.
S505, first area and second area, second area and first are chosen on the first face image of bank note to be identified
Central point central symmetry of the region about the first face of bank note to be identified image.
S506, third region and the fourth region, third are being chosen on the second face image of setting with the first face image
Region and the fourth region the position of the second face image respectively with first area and second area the first face image position phase
Together.
S507, by the corresponding image template in first area, the corresponding image template of second area, the corresponding figure in third region
As template and the corresponding image template of the fourth region are matched with the reference template in pre-stored pre-set image region respectively.
If the pixel of " YUAN " word covering can be described as foreground point, in reference template using " YUAN " word as reference template
Remaining pixel can be described as background dot.The status requirement bank note to be identified of foreground point in reference template can be used in matched mode
Corresponding position be also foreground point, background dot meets certain threshold in the status requirement of background dot bank note to be identified in reference template
Value, can protrude the matching of foreground point.
Similarity is calculated relative to using related coefficient, matching way provided in this embodiment reduces calculation amount.
S508, determined according to matching result bank note to be identified towards.
The present embodiment determines gray threshold according to the number of pixels percentage of the corresponding histogram in pre-set image region, will be pre-
If the gray value of the pixel in image-region lower than gray threshold is uniformly set to lowest gray value, higher than the pixel of gray threshold
The gray value of point is uniformly set to highest gray value, can be reduced the operation when image template of bank note to be identified is matched with reference template
Amount improves recognition efficiency.
Embodiment six
Fig. 6 be a kind of identification bank note that provides of the embodiment of the present invention six towards apparatus structure schematic diagram, which is used for
Execute above-described embodiment in identification bank note towards method.The device includes: that module 601, the are chosen in first and second region
Three and the fourth region choose module 602, matching module 603 and towards determining module 604.
Module 601 is chosen in first and second region, for choosing first area on the first face image of bank note to be identified
And the central point central symmetry of second area, second area and first area about the first face of bank note to be identified image.
Third and the fourth region choose module 602, choose module 601 with first and second region and are connected, are used for the
One face image chooses third region and the fourth region on the second face image of setting, and third region and the fourth region are second
The position of face image is identical in the position of the first face image as first area and second area respectively.
Matching module 603 chooses module 602 with third and the fourth region and is connected, is used for the corresponding image in first area
The corresponding image template of template, second area, the corresponding image template in third region and the corresponding image template point of the fourth region
It is not matched with the reference template in pre-stored pre-set image region.
Towards determining module 604, it is connected with matching module 603, for determining the face of bank note to be identified according to matching result
To.
Further, identification bank note towards device further include that module and processing module are chosen in pre-set image region.
Module is chosen in pre-set image region, makees for choosing any one following region on the direct picture of reference bank note
For pre-set image region: oppositely positioned first area, the positive first area placed, oppositely positioned second area, forward direction
The second area of placement, oppositely positioned third region, the positive third region placed, oppositely positioned the fourth region and forward direction
The fourth region of placement;Alternatively, choosing any one following region on the verso images of reference bank note as pre-set image area
Domain: first area, oppositely positioned second area, positive the secondth area placed of oppositely positioned first area, positive placement
Domain, oppositely positioned third region, the positive third region placed, oppositely positioned the fourth region and positive the 4th area placed
Domain.
Processing module, chooses module with pre-set image region respectively and first and second region chooses module and is connected, and is used for
Pre-set image region is handled to obtain reference template.
Further, matching module includes overturning submodule and again matched sub-block.
Overturn submodule, corresponding image template, the corresponding image template of second area, third region pair in first area
The corresponding image template of image template and the fourth region answered carries out in first time matching process with reference template respectively, appoints if having
Anticipating, it fails to match for the image template in a region, then carries out following at least one turning operation to the image of failure: be flipped up, to
Lower overturning overturns to the left and overturns to the right.
Matched sub-block again is connected with overturning submodule, matches to the image template after overturning with reference template.
Further, processing module further includes that binary conversion treatment submodule and boundary determine submodule.
Binary conversion treatment submodule, for carrying out down-sampled and binary conversion treatment to pre-set image region.
Boundary determines submodule, is connected with binary conversion treatment submodule, default for being determined according to the image after binaryzation
The boundary of image-region, obtains reference template.
Further, boundary determines that submodule is specifically used for: the gray scale of row pixel is calculated according to the image after binaryzation
The gray value adduction of value adduction and column pixel;There is into the position being mutated as default figure in the gray value of row pixel adduction
As the row bound in region, and the position that being mutated is occurred in the gray value of column pixel adduction is as the column side in pre-set image region
Boundary;Or, corresponding row, will as the row bound in pre-set image region when the gray value for the pixel for setting line number is summed it up maximum
The gray value for setting the pixel of columns sums it up column boundary of the corresponding column as pre-set image region when maximum, and it is small to set line number
Line number in the pre-set image region, setting columns are less than the columns in the pre-set image region.
Boundary determines that submodule is also used to: the lateral difference of each pixel is calculated according to the image after binaryzation;By every row
The absolute value of pixel transverse direction difference is cumulative obtain row difference and, and to row difference and difference is carried out, by the difference of row difference sum
Row bound at maximum absolute value as pre-set image region;Or, by corresponding when setting the row difference of line number with cumulative maximum
Row bound of the row as pre-set image region, setting line number are less than the line number in the pre-set image region;
Longitudinal difference of each pixel is calculated according to the image after binaryzation;
By the absolute value of the pixel longitudinal direction difference of each column it is cumulative obtain column difference and, and to column difference and carry out difference, will
Column boundary at the maximum absolute value of the difference of column difference sum as pre-set image region;Or, by set columns column difference and
Column boundary of the corresponding column as pre-set image region when cumulative maximum, setting columns are less than the column in the pre-set image region
Number.
Further, binary conversion treatment submodule includes gray threshold determination unit and gray value setting unit.
Gray threshold determination unit, for being determined according to the number of pixels percentage of the corresponding histogram in pre-set image region
Gray threshold.
Gray value setting unit is connected with gray threshold determination unit, for will be lower than gray scale threshold in pre-set image region
The gray value of the pixel of value is uniformly set to lowest gray value, and the gray value higher than the pixel of gray threshold is uniformly set to highest
Gray value.
Further, gray threshold determination unit is specifically used for: choosing from pre-set image region and meets number of pixels hundred
The pixel for dividing the gray value of ratio low;In the pixel low from gray value choose maximum gradation value as gray threshold, or from remain
It is chosen in afterimage vegetarian refreshments and is greater than maximum gradation value, and the gray value adjacent with maximum gradation value is as gray threshold.
Identification bank note provided by the embodiment of the present invention towards device can be used for executing any embodiment of that present invention and mentioned
The identification bank note of confession towards method, have and execute the corresponding function of this method and beneficial effect.
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The invention is not limited to the specific embodiments described herein, be able to carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out by above embodiments to the present invention
It is described in further detail, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, also
It may include more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (16)
1. a kind of identification bank note towards method, comprising: first area and the are chosen on the first face image of bank note to be identified
Two regions, the second area and center dot center pair of the first area about the first face of bank note to be identified image
Claim;
Third region and the fourth region, the third area are being chosen on the second face image of setting with first face image
Domain and the fourth region are in the position of second face image respectively with the first area and the second area described
The position of first face image is identical;
It is characterised by comprising:
The corresponding image template in the first area, the corresponding image template of the second area, the third region is corresponding
Image template and the corresponding image template of the fourth region reference template with pre-stored pre-set image region respectively
It is matched;
According to matching result determine the bank note to be identified towards.
2. the method according to claim 1, wherein choosing the firstth area on the first face image of bank note to be identified
Before domain and second area, further includes:
Any one following region is chosen on the direct picture of reference bank note as the pre-set image region: oppositely positioned
It is first area, the positive first area placed, oppositely positioned second area, the positive second area placed, oppositely positioned
Third region, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed;Alternatively, joining
It examines and chooses any one following region on the verso images of bank note as the pre-set image region: the firstth oppositely positioned area
Domain, the positive first area placed, oppositely positioned second area, the positive second area placed, oppositely positioned third area
Domain, the positive third region placed, oppositely positioned the fourth region and the positive the fourth region placed;
The pre-set image region is handled to obtain the reference template.
3. method according to claim 1 or 2, which is characterized in that by the corresponding image template in the first area, second
The corresponding image template in region, the corresponding image template in third region and the corresponding image template of the fourth region respectively with deposit in advance
The reference template in the pre-set image region of storage carries out matching
Corresponding image template, the corresponding image template of second area, the corresponding image mould in third region in the first area
Plate and the corresponding image template of the fourth region carry out in first time matching process with the reference template respectively, if there is any one area
It fails to match for the image template in domain, then carries out following at least one turning operation to the image of failure: being flipped up, turn over downwards
Turn, overturn and overturn to the right to the left;
Image template after overturning is matched with the reference template.
4. according to the method described in claim 2, it is characterized in that, being handled to obtain the ginseng to the pre-set image region
Examining template includes:
Down-sampled and binary conversion treatment is carried out to the pre-set image region;
The boundary that the pre-set image region is determined according to the image after binaryzation, obtains the reference template.
5. according to the method described in claim 4, it is characterized in that, determining the pre-set image area according to the image after binaryzation
The boundary in domain includes:
The gray value adduction of row pixel and the gray value adduction of column pixel are calculated according to the image after the binaryzation;
There is into the position being mutated as the row bound in the pre-set image region in the gray value of row pixel adduction, and will
Column boundary of the position that being mutated occurs in the gray value adduction of the column pixel as the pre-set image region;Or, will setting
The gray value of the pixel of the line number row bound of corresponding row as pre-set image region when summing it up maximum, will set columns
The gray value of pixel when summing it up maximum column boundary of the corresponding column as the pre-set image region, the setting line number it is small
Line number in the pre-set image region, the setting columns are less than the columns in the pre-set image region.
6. according to the method described in claim 4, it is characterized in that, determining the pre-set image area according to the image after binaryzation
The boundary in domain includes:
The lateral difference of each pixel is calculated according to the image after the binaryzation;
By the absolute value of the pixel transverse direction difference of every row it is cumulative obtain row difference and, and to the row difference and carry out difference, will
Row bound at the maximum absolute value of the difference of the row difference sum as the pre-set image region;Or, line number will be set
Row bound of the row difference with row corresponding when cumulative maximum as the pre-set image region, the setting line number is less than described pre-
If the line number of image-region;
Longitudinal difference of each pixel is calculated according to the image after binaryzation;
By the absolute value of the pixel longitudinal direction difference of each column it is cumulative obtain column difference and, and to the column difference and carry out difference, will
Column boundary at the maximum absolute value of the difference of the column difference sum as the pre-set image region;Or, columns will be set
Column boundary of the column difference with column corresponding when cumulative maximum as the pre-set image region, the setting columns is less than described pre-
If the columns of image-region.
7. according to the described in any item methods of claim 4-6, which is characterized in that carry out binaryzation to the pre-set image region
Processing includes:
Gray threshold is determined according to the number of pixels percentage of the corresponding histogram in the pre-set image region;
The gray value for being lower than the pixel of the gray threshold in the pre-set image region is uniformly set to lowest gray value, it is high
Highest gray value is uniformly set in the gray value of the pixel of the gray threshold.
8. the method according to the description of claim 7 is characterized in that according to the picture of the corresponding histogram in the pre-set image region
Plain number percentage determines that gray threshold includes:
Meet the pixel of the number of pixels percentage by the ascending selection from the pre-set image region of gray value;
Maximum gradation value is chosen in the pixel low from the gray value to select as the gray threshold, or from residual pixel point
It takes and is greater than the maximum gradation value, and the gray value adjacent with the maximum gradation value is as the gray threshold.
9. a kind of identification bank note towards device, comprising: module is chosen in first and second region, for the of bank note to be identified
First area and second area are chosen on one face image, the second area and the first area are about the bank note to be identified
The central point central symmetry of first face image;
Third and the fourth region choose module, choose module with first and second described region and are connected, are used for described first
Face image chooses third region and the fourth region, the third region and the fourth region on the second face image of setting
In second face, the position of image is respectively with the first area and the second area in the position of first face image
It is identical;
It is characterised by comprising:
Matching module chooses module with the third and the fourth region and is connected, is used for the corresponding image mould in the first area
Plate, the corresponding image template of the second area, the corresponding image template in the third region and the fourth region are corresponding
Image template is matched with the reference template in pre-stored pre-set image region respectively;
Towards determining module, be connected with the matching module, for according to matching result determine the bank note to be identified towards.
10. device according to claim 9, which is characterized in that further include:
Module is chosen in pre-set image region, for choosing any one following region on the direct picture of reference bank note as institute
State pre-set image region: oppositely positioned first area, the positive first area placed, oppositely positioned second area, forward direction
The second area of placement, oppositely positioned third region, the positive third region placed, oppositely positioned the fourth region and forward direction
The fourth region of placement;Alternatively, choosing any one following region on the verso images of reference bank note as the default figure
As region: oppositely positioned first area, the positive first area placed, oppositely positioned second area, positive the placed
Two regions, oppositely positioned third region, the positive third region placed, oppositely positioned the fourth region and positive the placed
Four regions;
Processing module, chooses module with the pre-set image region respectively and first and second described region chooses module and is connected,
For being handled to obtain the reference template to the pre-set image region.
11. device according to claim 9 or 10, which is characterized in that the matching module includes:
Overturn submodule, corresponding image template, the corresponding image template of second area, third region pair in the first area
The corresponding image template of image template and the fourth region answered carries out in first time matching process with the reference template respectively, if
Having the image template in any one region, it fails to match, then carries out following at least one turning operation to the image of failure: to turning over
Turn, downwardly turn over, overturning to the left and overturning to the right;
Matched sub-block again is connected with the overturning submodule, carries out to the image template after overturning with the reference template
Matching.
12. device according to claim 10, which is characterized in that the processing module further include:
Binary conversion treatment submodule, for carrying out down-sampled and binary conversion treatment to the pre-set image region;
Boundary determines submodule, is connected with the binary conversion treatment submodule, for according to the image determination after binaryzation
The boundary in pre-set image region obtains the reference template.
13. device according to claim 12, which is characterized in that the boundary determines that submodule is specifically used for:
The gray value adduction of row pixel and the gray value adduction of column pixel are calculated according to the image after the binaryzation;It will
Row bound of the gray value adduction position that occurs being mutated of the row pixel as the pre-set image region, and by the column
Column boundary of the position that being mutated occurs in the gray value adduction of pixel as the pre-set image region;Or, line number will be set
The gray value of the pixel row bound of corresponding row as pre-set image region when summing it up maximum, will set the pixel of columns
The gray value of the point column boundary of corresponding column as pre-set image region when summing it up maximum, the setting line number is less than described
The line number in pre-set image region, the setting columns are less than the columns in the pre-set image region.
14. device according to claim 12, which is characterized in that the boundary determines that submodule is specifically used for:
The lateral difference of each pixel is calculated according to the image after the binaryzation;
By the absolute value of the pixel transverse direction difference of every row it is cumulative obtain row difference and, and to the row difference and carry out difference, will
Row bound at the maximum absolute value of the difference of the row difference sum as the pre-set image region;Or, line number will be set
Row bound of the row difference with row corresponding when cumulative maximum as the pre-set image region, the setting line number is less than described pre-
If the line number of image-region;
Longitudinal difference of each pixel is calculated according to the image after binaryzation;
By the absolute value of the pixel longitudinal direction difference of each column it is cumulative obtain column difference and, and to the column difference and carry out difference, will
Column boundary at the maximum absolute value of the difference of the column difference sum as the pre-set image region;Or, columns will be set
Column boundary of the column difference with column corresponding when cumulative maximum as the pre-set image region, the setting columns is less than described pre-
If the columns of image-region.
15. the described in any item devices of 2-14 according to claim 1, which is characterized in that the binary conversion treatment submodule includes:
Gray threshold determination unit, for being determined according to the number of pixels percentage of the corresponding histogram in the pre-set image region
Gray threshold;
Gray value setting unit is connected with the gray threshold determination unit, for will be lower than institute in the pre-set image region
The gray value for stating the pixel of gray threshold is uniformly set to lowest gray value, higher than the gray value of the pixel of the gray threshold
Uniformly it is set to highest gray value.
16. device according to claim 15, which is characterized in that the gray threshold determination unit is specifically used for:
Meet the pixel of the number of pixels percentage by the ascending selection from the pre-set image region of gray value;
Maximum gradation value is chosen in the pixel low from the gray value to select as the gray threshold, or from residual pixel point
It takes and is greater than the maximum gradation value, and the gray value adjacent with the maximum gradation value is as the gray threshold.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710100698.8A CN106898085B (en) | 2017-02-23 | 2017-02-23 | Identify bank note towards method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710100698.8A CN106898085B (en) | 2017-02-23 | 2017-02-23 | Identify bank note towards method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106898085A CN106898085A (en) | 2017-06-27 |
CN106898085B true CN106898085B (en) | 2019-06-11 |
Family
ID=59184942
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710100698.8A Active CN106898085B (en) | 2017-02-23 | 2017-02-23 | Identify bank note towards method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106898085B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107705418B (en) * | 2017-10-10 | 2020-06-09 | 深圳怡化电脑股份有限公司 | Paper currency facing identification method, device, equipment and readable storage medium |
CN108389309B (en) * | 2018-02-06 | 2021-01-01 | 深圳怡化电脑股份有限公司 | Method and system for identifying authenticity of paper money |
CN108492447B (en) * | 2018-03-20 | 2020-09-15 | 深圳怡化电脑股份有限公司 | Dollar-oriented identification method, electronic device and computer-readable storage medium |
CN113408521B (en) * | 2021-07-16 | 2023-09-05 | 北京南昊科技股份有限公司 | Answer picture identification method, reading device and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102063758A (en) * | 2009-11-16 | 2011-05-18 | 中科院成都信息技术有限公司 | Banknote paper texture detection method and device |
CN104867225A (en) * | 2015-05-15 | 2015-08-26 | 深圳怡化电脑股份有限公司 | Banknote face orientation identification method and apparatus |
CN104992496A (en) * | 2015-07-08 | 2015-10-21 | 深圳怡化电脑股份有限公司 | Paper money face identification method and apparatus |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6534278B2 (en) * | 2015-03-31 | 2019-06-26 | グローリー株式会社 | Paper sheet discrimination apparatus and paper sheet discrimination method |
-
2017
- 2017-02-23 CN CN201710100698.8A patent/CN106898085B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102063758A (en) * | 2009-11-16 | 2011-05-18 | 中科院成都信息技术有限公司 | Banknote paper texture detection method and device |
CN104867225A (en) * | 2015-05-15 | 2015-08-26 | 深圳怡化电脑股份有限公司 | Banknote face orientation identification method and apparatus |
CN104992496A (en) * | 2015-07-08 | 2015-10-21 | 深圳怡化电脑股份有限公司 | Paper money face identification method and apparatus |
Non-Patent Citations (3)
Title |
---|
基于ARM的嵌入式系统开发与应用;刘刚;《中国优秀硕士学位论文全文数据库 信息科技辑》;20091116(第11期);全文 |
基于双核处理器的纸币清分系统研究与设计;朱帅;《中国优秀硕士学位论文全文数据库 信息科技辑》;20150516(第5期);全文 |
纸币清分机图象识别系统的研究与设计;丁瑞昕;《中国优秀硕士学位论文全文数据库 信息科技辑》;20080916(第9期);全文 |
Also Published As
Publication number | Publication date |
---|---|
CN106898085A (en) | 2017-06-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106898085B (en) | Identify bank note towards method and device | |
CN100580716C (en) | Verification method for determining areas within an image corresponding to monetary banknotes | |
CN106056751B (en) | The recognition methods and system of serial number | |
CN105989659B (en) | A kind of similar character recognition methods and paper money code recognition methods | |
CN104112128B (en) | Digital image processing system and method applied to bill image character recognition | |
RU2557461C2 (en) | Method of separating character string and device of highlighting character string | |
Youn et al. | Efficient multi-currency classification of CIS banknotes | |
JP5401695B2 (en) | Image identification device, image identification method, image identification program, and recording medium | |
CN105184957B (en) | A kind of paper money discrimination method and system | |
RU2658272C2 (en) | Authentication of security documents and mobile device to carry out authentication | |
CN104867225B (en) | A kind of bank note towards recognition methods and device | |
US7542593B1 (en) | Offline signature verification using high pressure regions | |
CN104992496A (en) | Paper money face identification method and apparatus | |
CN102324134A (en) | Valuable document identification method and device | |
CN106504407B (en) | A kind of method and device handling banknote image | |
CN103093185B (en) | Character recognition device, image processing apparatus and method thereof | |
CN107085883A (en) | A kind of method and apparatus of paper money recognition | |
JP2006301881A (en) | Currency identifying device, currency identifying method and currency identifying program | |
US20170256058A1 (en) | Apparatus and method for detecting circle | |
CN106780966B (en) | A kind of paper money discrimination method and device | |
CN105654609A (en) | Paper money processing method and paper money processing system | |
Prashanth et al. | Off-line signature verification based on angular features | |
CN101149837A (en) | Color processing method for identification of areas within an image corresponding to monetary banknotes | |
CN108460775A (en) | A kind of forge or true or paper money recognition methods and device | |
US20030091238A1 (en) | Detection of a pattern in a digital image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |