CN108230535B - Facing identification method and device for paper money - Google Patents
Facing identification method and device for paper money Download PDFInfo
- Publication number
- CN108230535B CN108230535B CN201611196640.XA CN201611196640A CN108230535B CN 108230535 B CN108230535 B CN 108230535B CN 201611196640 A CN201611196640 A CN 201611196640A CN 108230535 B CN108230535 B CN 108230535B
- Authority
- CN
- China
- Prior art keywords
- image
- characteristic
- feature
- crown word
- binarization
- 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
Images
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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Inspection Of Paper Currency And Valuable Securities (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a facing identification method and device of paper money. The method comprises the following steps: determining a first characteristic image, a second characteristic image, a third characteristic image and a fourth characteristic image in a complete image of the paper money according to a preset position, wherein one of the four characteristic images comprises a crown word number; carrying out binarization on the four characteristic images to obtain a first characteristic binarization image, a second characteristic binarization image, a third characteristic binarization image and a fourth characteristic binarization image; and detecting the serial number according to the serial number features and the four feature binarization images, and determining the face of the paper money according to the detection result. By the technical scheme, the problem of identifying the paper currency faces of different versions and different currency values is solved, and the accuracy of identifying the paper currency faces is improved.
Description
Technical Field
The embodiment of the invention relates to an image processing technology, in particular to a facing identification method and device of paper money.
Background
The face recognition of the paper money is a basic recognition part for performing true and false recognition of the paper money, and the face recognition result directly influences the effect of the subsequent true and false recognition of the paper money. If the identification result is wrong, the false detection of the subsequent authenticity identification can be caused.
There are two main types of methods for performing identification-oriented in the prior art. One type is an intelligent learning algorithm based on the paper money gray image features, such as an identification-oriented algorithm based on a machine learning method such as a neural network algorithm or a support vector machine. The algorithm needs a large amount of banknote sample data, and the quality of the selected sample data directly influences the identification precision of the algorithm. Another method is to count the gray value of a specific region of the banknote, such as calculating the mean or sum of the pixel values in the specific region, and then comparing the statistical results of the selected specific regions, thereby performing face-to-face recognition.
Disclosure of Invention
The embodiment of the invention provides a facing identification method and device of paper money, which are used for realizing facing identification of paper money with different versions and different currency values and improving the facing identification accuracy of the paper money.
In a first aspect, an embodiment of the present invention provides a method for identifying facing of a banknote, including the following steps:
determining a first characteristic image, a second characteristic image, a third characteristic image and a fourth characteristic image in a complete image of the paper money according to a preset position, wherein one of the four characteristic images comprises a crown word number;
carrying out binarization on the four characteristic images to obtain a first characteristic binarization image, a second characteristic binarization image, a third characteristic binarization image and a fourth characteristic binarization image;
and detecting the serial number according to the serial number features and the four feature binarization images, and determining the face of the paper money according to the detection result.
In a second aspect, an embodiment of the present invention further provides a banknote face recognition apparatus, including:
the characteristic image determining module is used for determining a first characteristic image, a second characteristic image, a third characteristic image and a fourth characteristic image in the complete image of the paper money according to a preset position, wherein one of the four characteristic images contains a crown word number;
the characteristic binarization image determining module is used for binarizing the four characteristic images to obtain a first characteristic binarization image, a second characteristic binarization image, a third characteristic binarization image and a fourth characteristic binarization image;
and the orientation determining module is used for detecting the crown word number according to the crown word number characteristics and the four characteristic binary images and determining the orientation of the paper money according to the detection result.
According to the embodiment of the invention, four characteristic images containing the crown word number in the complete image of the paper currency are obtained, binarization processing is carried out on the four characteristic images to obtain corresponding four characteristic binarization images, crown word number detection is carried out on the four characteristic binarization images according to the crown word number characteristics, and the paper currency orientation is determined according to the detection result. The problem of the paper currency face recognition of different editions and different currency values is solved, the influence of the brightness degree of the image is weakened based on binarization processing and unique crown word feature detection, and the accuracy of the paper currency face recognition is improved.
Drawings
FIG. 1 is a flow chart of a method for identifying a banknote facing the face of a banknote in accordance with a first embodiment of the present invention;
FIG. 2a is a complete reflectance gray scale image of four facing notes, front-side forward, front-side reverse, back-side forward and back-side reverse, of a 2005 edition of the RMB 100 in a first embodiment of the present invention;
FIG. 2b is a complete infrared transmission image of four facing notes, front-side forward, front-side reverse, back-side forward and back-side reverse, of a 2005 edition of the RMB 100 in the first embodiment of the present invention;
FIG. 3a is a diagram of four feature images corresponding to the front and back directions in FIG. 2a according to a first embodiment of the present invention;
FIG. 3b is a diagram of four feature images corresponding to the front-facing direction in FIG. 2b according to a first embodiment of the present invention;
FIG. 4a is a diagram of a binarized image of four features corresponding to FIG. 3a according to a first embodiment of the present invention;
FIG. 4b is a diagram of a binarized image of four features corresponding to FIG. 3b according to a first embodiment of the present invention;
FIG. 5 is a flowchart of a banknote face recognition method according to a second embodiment of the present invention;
FIG. 6 is a flow chart of a banknote face recognition method according to a third embodiment of the present invention;
FIG. 7 is a flowchart of a banknote face recognition method according to a fourth embodiment of the present invention;
fig. 8 is a schematic diagram of a fourth embodiment of the present invention, illustrating a method for obtaining a preset segmentation position template;
FIG. 9 is a schematic structural view of a banknote face-recognition apparatus according to a fifth embodiment of the present invention;
fig. 10 is a schematic structural view of a banknote recognition apparatus according to a sixth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a banknote recognition method according to an embodiment of the present invention, which may be implemented by a banknote recognition device, which may be implemented by software and/or hardware, and which may be integrated into any financial device requiring banknote recognition, such as a banknote validator, a banknote counter, a sorter, or the like. The method specifically comprises the following steps:
s100, determining a first characteristic image, a second characteristic image, a third characteristic image and a fourth characteristic image in the complete image of the paper currency according to a preset position.
The preset position is the position of a preset first characteristic image in the complete image of the paper currency. The preset positions can be expressed in absolute length form, for example, by setting the length of the left side position of the preset position from the left side start position in the long side direction of the bill to a1, the length of the right side position from the left side start position in the long side direction of the bill to a2, the length of the upper position from the upper start position in the short side direction of the bill to b1, and the length of the lower position from the upper start position in the short side direction of the bill to b2, the preset positions can be expressed as [ a1, a2] and [ b1, b2 ]. The default positions may also be expressed in the form of row-column coordinates, which depend on the resolution of the image, and the row-column coordinate values in the form of row-column coordinates may be different for different image sensors, which is only described by taking a sensor with a resolution of 200DPI in the horizontal direction and 150DPI in the vertical direction as an example, and taking the row-column coordinate of the first pixel point at the upper left corner of the image as row 1 and column 1, the default positions may be preset to (310,20), (310,290), (370,290) and (370,20), i.e. row 310 to 370 and column 20 to 290 in the complete image of the banknote.
The complete image of the banknote refers to the image of the entire banknote, which may be a gray scale image related to reflectance, an infrared transmission image, or a color image obtained by direct scanning or photographing. For a reflectivity gray image of a paper currency or a color image obtained by direct scanning or photographing, one paper currency has two complete images of the front surface and the back surface, the complete image of the upward surface of the paper currency is called a first surface image, and correspondingly, the image of the downward surface of the paper currency is called a second surface image. For example, in the process of feeding paper money, sensors are respectively arranged above and below the paper money, when the paper money passes through the sensors, the sensors positioned above the sensors can acquire the reflectivity gray level image or the color image of the upward surface of the paper money, namely the first surface image, and the sensors positioned below the sensors can acquire the reflectivity gray level image or the color image of the downward surface of the paper money, namely the second surface image. For infrared transmission images of a note, a note has only one complete image. For example, in the process of feeding paper money, only one infrared transmission sensor is arranged above the paper money, and when the paper money passes through the infrared transmission sensor, the infrared transmission sensor acquires the transmission image of the paper money.
Specifically, a banknote will have four faces: face forward, face reverse, back forward and back reverse, and the position of the crown size in the complete image of the note facing differently is different. For example, the positions of the prefix numbers in the reflectance grayscale image of the different orientation shown in fig. 2a and the infrared transmission image of the different orientation shown in fig. 2b are respectively: the lower left corner of the front side forward reflectance grayscale image 201 or the front side forward infrared transmission image 201 ', the upper right corner of the front side reverse reflectance grayscale image 202 or the front side reverse infrared transmission image 202', the lower right corner of the back side forward reflectance grayscale image 203 or the back side forward infrared transmission image 203 ', and the upper left corner of the back side reverse reflectance grayscale image 204 or the back side reverse infrared transmission image 204'. If the facing of a paper currency is to be recognized, the sub-images of the four specific areas in the complete image of the paper currency, namely the first feature image, the second feature image, the third feature image and the fourth feature image, can be acquired, then the crown word number is detected in the four sub-images, the specific position of the sub-image containing the crown word number is determined, and the facing of the paper currency can be determined according to the positions of the crown word number in different facing.
For the reflectivity gray scale image of the paper money or the color image obtained by direct scanning or photographing, the four characteristic images respectively correspond to the sub-image of the lower left corner area in the first surface image, the sub-image of the upper right corner area in the second surface image and the sub-image of the lower left corner area in the second surface image. Taking the front-side reflectance gray-scale image 201 of the banknote shown in fig. 2a as the first-side image, the second-side image is the back-side reverse reflectance gray-scale image 204 of the banknote shown in fig. 2 a. Assuming that the total number of columns and the total number of rows of the complete banknote in the figure is S and L for the same image resolution, the positions of the remaining three feature images can be deduced from the relationship between the preset positions (i.e. the row and column number coordinates (310,20), (310,290), (370,290) and (370,20)) and the positions of the four feature images. For example, the second characteristic image position is centrosymmetric with the first characteristic image position about the center of the banknote, and then the second characteristic image positions are (L-370, S-290), (L-370, S-20), (L-310, S-20) and (L-310, S-290); the third feature image position is a position in the second face image corresponding to the second feature image, namely, (L-370, S-290), (L-370, S-20), (L-310, S-20), and (L-310, S-290); the fourth feature image position is a position in the second face image corresponding to the first feature image, i.e., (310,20), (310,290), (370,290), and (370, 20). According to the positions of the four characteristic images, corresponding sub-images are respectively cut out from the front positive reflectivity gray scale image 201 and the back negative reflectivity gray scale image 204 of the banknote shown in fig. 2a, and as shown in fig. 3a, the sub-images respectively correspond to a first characteristic image 301, a second characteristic image 302, a third characteristic image 303 and a fourth characteristic image 304.
And for the infrared transmission image of the paper money, the four characteristic images respectively correspond to sub-images of four areas of the lower left corner, the upper right corner, the lower right corner and the upper left corner of the transmission image. Accordingly, taking the infrared transmission image 201' of the front side of the banknote shown in FIG. 2b as an example, at the same image resolution, the first characteristic image positions can be determined as (310,20), (310,290), (370,290) and (370,20) and the second characteristic image positions can be determined as (L-370, S-290), (L-370, S-20), (L-310, S-20) and (L-310, S-290) according to the preset positions; the third characteristic image position and the first characteristic image position are in axial symmetry about the central line of the long side direction of the paper currency, and then the third characteristic image positions are (310, S-290), (310, S-20), (370, S-20) and (370, S-290); the fourth feature image positions are axisymmetrical with the first feature image positions with respect to the center line in the short side direction of the bill, and then the fourth feature image positions are (L-370,20), (L-370,290), (L-310,290), and (L-310, 20). Based on the positions of the four feature images, corresponding sub-images are respectively cut out from the front forward infrared transmission image 201 ' of the banknote shown in fig. 2b, and correspond to the first feature image 301 ', the second feature image 302 ', the third feature image 303 ' and the fourth feature image 304 ' as shown in fig. 3 b.
S200, binarizing the four feature images to obtain a first feature binarized image, a second feature binarized image, a third feature binarized image and a fourth feature binarized image.
Specifically, as can be seen from the description of step S100, the acquired four feature images may be a reflectance grayscale image, a color image, or an infrared transmission image. In order to facilitate the detection of the subsequent crown word number, it is necessary to filter out useless information in the feature image containing the crown word number, i.e., all background texture information except the crown word number. In the embodiment, a binarization method is adopted for information elimination so as to highlight the required useful information. The binarization method can be a P parameter method, a maximum inter-class variance method and the like, and the P parameter method is preferably adopted, so that the binarization processing can be completed more quickly. In addition, the binarization threshold in the binarization method for the four feature images in the embodiment is determined according to the threshold of the feature image containing the prefix number, that is, the feature image containing the prefix number may be binarized in advance, and the binarization threshold may be obtained and used as the binarization threshold of the four feature images. And then, carrying out binarization on the four characteristic images by using the binarization threshold value and the binarization method to obtain four corresponding characteristic binarization images. Referring to fig. 4a, binarizing the four feature images of the reflectivity gray scale image in fig. 3a may obtain corresponding feature binary images, i.e., a first feature binary image 401, a second feature binary image 402, a third feature binary image 403, and a fourth feature binary image 404; accordingly, referring to fig. 4b, binarizing the four feature images of the infrared transmission image in fig. 3b may obtain corresponding feature binary images, i.e., a first feature binary image 401 ', a second feature binary image 402', a third feature binary image 403 'and a fourth feature binary image 404'.
S300, crown word number detection is carried out according to the crown word number features and the four feature binarization images, and the face of the paper money is determined according to the detection result.
The crown word number features refer to some features of the crown word number in the paper money, for example, the upper part of the crown word number is completely blank, the crown word number is 10 characters which are regularly arranged, the total length of the crown word number is fixed, the maximum height of the crown word number is fixed, and the like.
Specifically, as can be seen from fig. 4a and 4b, only one of the four feature binary images contains a crown word number, and although the lengths of the crown word numbers in the reflectance grayscale image and the infrared transmission image are different, the crown word number features are very obvious, so that the crown word number is easily identified in the four feature binary images according to the crown word number features, and the feature binary image with the crown word number is found. Then, according to the description of step S100, the facing of the banknote can be determined according to the position of the feature binary image in the complete banknote image.
According to the technical scheme of the embodiment, four feature images containing the crown word numbers in two faces of the paper money are obtained, binarization processing is carried out on the four feature images, corresponding four feature binarization images are obtained, crown word number detection is carried out on the four feature binarization images according to the crown word number features, and the face of the paper money is determined according to the detection result. The problem of the paper currency face recognition of different versions and different currency values is solved, the influence of the brightness degree of an image is weakened based on binarization processing and unique crown word feature detection, and the speed and accuracy of the paper currency face recognition are improved.
Example two
Fig. 5 is a flowchart of a banknote face recognition method according to a second embodiment of the present invention, which is optimized based on the second embodiment, and step S300 is optimized as step S310 and step S320. Wherein, the same steps as those in the above embodiments are denoted by the same reference numerals, and explanations of the same or corresponding terms as those in the above embodiments are omitted. Correspondingly, the method of the embodiment includes the following operations:
s100, determining a first characteristic image and a second characteristic image in a first side image of the paper currency according to a preset position, and determining a third characteristic image and a fourth characteristic image in a second side image of the paper currency.
S200, binarizing the four feature images to obtain a first feature binarized image, a second feature binarized image, a third feature binarized image and a fourth feature binarized image.
S310, crown word number detection is carried out on the four feature binary images according to the first crown word number feature and/or the second crown word number feature in the crown word number features, and the feature binary images containing the crown word numbers are determined.
The first crown word number characteristic is that the area above the crown word number is blank, and the area where the crown word number is located is the characteristic of a large number of black dots, and the characteristic can be judged according to whether the black dots exist in the characteristic binary image or not. The second crown character feature is a feature that 10 crown characters are regularly arranged, and the feature can be determined according to the fixed distance between the characters or whether the characters exist in a sliding window with a fixed size.
Specifically, one of a first crown character feature, a second crown character feature and a combination of the first crown character feature and the second crown character feature is selected to identify the crown character number in the four feature binary images. For example, (1) a first crown word number feature is selected, and then a feature binary image without black points in a continuous region is searched in the four feature binary images, and the feature binary image containing the crown word number is determined. (2) And selecting a second crown word size characteristic, counting the character characteristics of the crown word size in advance, acquiring the distance between the characters, namely the width of a blank area between the characters, or acquiring the width of each character, calculating the average width of the characters in the crown word size, then taking the distance or the average width of the characters as the size of a sliding window, identifying the crown word size in the four characteristic binary images according to the average width of the characters or the sliding step length of the distance, and determining the image with the identified crown word size as the characteristic binary image containing the crown word size. (3) And selecting a combination of the first crown word number characteristic and the second crown word number characteristic to identify the crown word number, namely firstly using the first crown word number characteristic to identify the crown word number for the first time to determine a crown word number identification result image, then using the second crown word number characteristic to identify the crown word number for the second time on the crown word number identification result image, and if the crown word number is determined to be contained in the crown word number identification result image again in the second time of crown word number identification, determining the crown word number identification result image as a characteristic binary image containing the crown word number. Therefore, the crown word number can be more accurately identified so as to accurately determine the characteristic binary image containing the crown word number.
And S320, determining the facing of the paper currency according to the feature binarization image containing the crown word number.
Specifically, according to the description in step S100, if the feature binary image containing the crown word number is determined, the face of the banknote can be determined according to the position of the image in the complete banknote image.
Exemplarily, step S320 may include: if the characteristic binary image containing the crown word number is a first characteristic binary image, determining that the face of the paper currency is a positive face; if the characteristic binary image containing the crown word number is a second characteristic binary image, determining that the face of the paper currency is a front face reverse direction; if the feature binarization image containing the crown word number is a third feature binarization image, determining that the face of the paper currency is a reverse face and a forward face; and if the characteristic binary image containing the crown word number is a fourth characteristic binary image, determining that the face of the paper currency is reverse.
Specifically, according to the description in step S100, the first feature binary image corresponds to the first feature image at the lower left corner in the full banknote image, so that the feature binary image containing the crown word number is the first feature binary image, and it can be determined that the banknote is facing in the front-face forward direction. Accordingly, the second feature binarized image corresponds to the second feature image which is the upper right corner in the full banknote image, so if the feature binarized image containing the crown word size is the second feature binarized image, it can be determined that the banknote is the face side reverse. The third feature binarized image corresponds to the third feature image which is in the lower right corner of the full banknote image, so if the feature binarized image containing the crown word number is the third feature binarized image, it can be determined that the banknote is facing in the reverse-side forward direction. The fourth feature binarized image corresponds to the fourth feature image which is the upper left corner in the full banknote image, so if the feature binarized image containing the crown word size is the fourth feature binarized image, it can be determined that the banknote is reverse in orientation.
According to the technical scheme, four feature images containing the crown word numbers in two faces of the paper money are obtained, binarization processing is carried out on the four feature images, corresponding four feature binarization images are obtained, crown word number detection is carried out on the four feature binarization images according to the first crown word number feature and/or the second crown word number feature, the feature binarization images containing the crown word numbers are determined, and the face of the paper money is determined according to the feature binarization images. The problem of the paper currency face recognition of different editions and different currency values is solved, and the speed and the accuracy of the paper currency face recognition are improved based on the detection of unique crown word number characteristics.
EXAMPLE III
Fig. 6 is a flowchart of a banknote face recognition method according to a third embodiment of the present invention, which is optimized based on the third embodiment, and step S310 is optimized to step S311 to step S314. Wherein, the same steps as those in the above embodiments are denoted by the same reference numerals, and explanations of the same or corresponding terms as those in the above embodiments are omitted. Correspondingly, the method of the embodiment includes the following operations:
s100, determining a first characteristic image and a second characteristic image in a first side image of the paper currency according to a preset position, and determining a third characteristic image and a fourth characteristic image in a second side image of the paper currency.
S200, binarizing the four feature images to obtain a first feature binarized image, a second feature binarized image, a third feature binarized image and a fourth feature binarized image.
S311, whether a first preset line number inner line and a first statistical line number smaller than the first line and a threshold exist in the current feature binarization image or not is determined, wherein the first statistical line number is larger than or equal to the first line number threshold.
The current characteristic binary image refers to a certain characteristic binary image which is processed in the current cycle process in the four characteristic binary images. The first preset line number is a preset line number with consecutive line numbers, and is used for judging whether a certain area in the feature binarized image is blank, for example, the first preset line number may be set to 20 lines with consecutive line numbers. The row and column refer to the total number of black dots in a row. The first row and the threshold are the total number of black dots in the row direction set in advance, and are used to determine whether or not a row of a certain row number is blank, and the row and the threshold are set to exclude the influence of individual noise, and may be set to 10 black dots, for example. The first statistical row number refers to a total row number satisfying a condition within the first preset row number, for example, the first statistical row number refers to a row number having at least a first row number threshold within the first preset row number, and is smaller than the first row number and the threshold. The first line number threshold corresponds to a first preset line number, and is also used for determining whether a certain area in the feature binarized image is blank, but the line number of the line in the first line number threshold may not be continuous, and the first line number threshold is generally smaller than the first preset line number, for example, the first line number threshold may be set to 15 lines.
Specifically, the rows and columns of each row are computed for the current feature binarized image. Then, taking the range covered by the first preset line number as a sliding window, starting from the first line of the current feature binarization image, performing sliding window statistics in the current feature binarization image according to a set step length (such as one line), performing statistics on the line in the sliding window and the total line number smaller than the first line and a threshold value each time, and comparing the total line number with the first line number threshold value. And if the total line number counted in the sliding window appearing for the first time in the whole current feature binarization image is larger than or equal to the first line number threshold value, determining the total line number as the first counting line number, and stopping the sliding window counting. On the contrary, if the total line number obtained by each sliding window statistics in the whole current feature binarization image is less than the first line number threshold value, it is determined that the first statistical line number does not exist in the current feature binarization image.
For example, when the first binary feature image 401 shown in fig. 4a is the current binary feature image, the first line of the image is taken as the starting line, 20 consecutive lines and the total number of lines less than 10 are counted line by line, the total number of lines counted by the first sliding window is 18 lines, and the total number of lines is greater than the first threshold 15 lines, then the first count number of lines is determined to be 18 lines, and the sliding window counting is stopped. When any one of the second, third and fourth feature binarized images 402, 403 and 404 shown in fig. 4a is the current feature binarized image, the line counted by the sliding window every time and the total number of lines greater than the first line and the threshold value are all smaller than the first line threshold value by 15 lines, and then the first counted number of lines does not exist in the three feature binarized images.
S312, whether a second preset line number inner line and a second statistical line number larger than a second line number threshold exist in the current feature binarization image or not is determined, wherein the second statistical line number is larger than or equal to the second line number threshold.
The second preset line number is a line number with a continuous line number, and is used for judging whether a certain area in the feature binary image contains enough useful information, for example, the second preset line number can be set to 20 lines with continuous line numbers. The second line and the threshold are the total number of black dots in the preset line direction, and are used for judging whether the line of a certain line number contains enough useful information, for example, the second line and the threshold can be set to 50 black dots. The second statistical line number refers to a total number of lines satisfying a condition within the second preset line number, for example, the second statistical line number refers to a line number of lines having at least a second line number threshold within the second preset line number, and is greater than the second line number threshold. The second line number threshold corresponds to a second preset line number, and is also used for determining whether a certain area in the feature binary image contains enough useful information, but the line number of the line in the second line number threshold may not be continuous, and the second line number threshold is generally smaller than the second preset line number, for example, the first line number threshold may be set to 15 lines.
Specifically, the range covered by the second preset line number is set as a sliding window, starting from the first line of the current feature binarization image, or when the first statistical line number exists in step S311, starting from the last line of the first statistical line number, performing sliding window statistics in the current feature binarization image according to a set step length, performing statistics on the line in the sliding window and the total line number greater than the second line number and the threshold value each time, and comparing the total line number with the second line number threshold value. And if the total line number counted in the sliding window appearing for the first time in the whole current feature binarization image is larger than or equal to the second line number threshold value, determining the total line number as the second counting line number, and stopping the sliding window counting. On the contrary, if the total line number obtained by each sliding window statistics in the whole current feature binarization image is less than the second line number threshold value, it is determined that the second statistical line number does not exist in the current feature binarization image.
For example, when the first binary feature image 401 shown in fig. 4a is the current binary feature image, the first line of the image is taken as the starting line, the number of consecutive 20 lines and the total number of lines greater than 50 are counted line by line, and if the total number of lines counted by a certain sliding window is 20 lines and is greater than the second line threshold value 15 lines, the second statistical line number is determined as 20 lines, and the sliding window counting is stopped. When any one of the second, third and fourth feature binarized images 402, 403 and 404 shown in fig. 4a is the current feature binarized image, the line counted by the sliding window every time and the total number of lines greater than the second line and the threshold value are all less than the first line number threshold value 15 lines, and then the second statistical line number does not exist in the three feature binarized images.
It should be understood that, according to the above description, the steps S311 and S312 may be executed in sequence, or in reverse order, that is, the step S312 is executed first, and then the step S311 is executed, except that the starting line of the sliding window statistics in the step S312 is set as the first line of the entire current feature binary image.
S313, judging whether the first statistical line number and the second statistical line number exist.
Specifically, for the current feature binarized image, if both the first statistical line number in step S311 and the second statistical line number in step S312 exist, step S314 is performed. Otherwise, if any one of the first statistical line number and the second statistical line number does not exist, step S311 is executed to perform a loop judgment of the next feature binarized image.
And S314, determining the characteristic binary image containing the crown word number.
Specifically, in step S313, it is determined that the first statistical line number and the second statistical line number both exist, that is, a part of the area in the current feature binarized image is blank, and a part of the area is a region containing a large number of black dots, and the image feature conforms to the feature of the first crown word number, so that it may be determined that the current feature binarized image contains the crown word number, that is, the current feature binarized image may be determined as the feature binarized image containing the crown word number.
And S320, determining the facing of the paper currency according to the feature binarization image containing the crown word number.
According to the technical scheme of the embodiment, four characteristic images containing crown word numbers in two faces of the paper money are obtained, binarization processing is carried out on the four characteristic images to obtain corresponding four characteristic binarization images, and then a-c is carried out on each characteristic binarization image in the four characteristic binarization images in a circulating mode, namely a, a first statistic line number is determined according to lines of lines which are at least provided with a first line number threshold value in a first preset line number and are smaller than the first line and the threshold value; b. determining a second statistical line number according to the line number of the lines with at least a second line number threshold value in a second preset line number and the line number larger than the second line number threshold value; c. and judging whether the first statistical line number and the second statistical line number exist or not until the characteristic binarization image containing the crown word number is determined, and determining the paper currency orientation according to the characteristic binarization image containing the crown word number. The problem of the different currency values of different editions towards discernment is solved, based on the detection of unique first hat number characteristic, the rate and the rate of accuracy that the paper currency faces towards the discernment have been improved.
Example four
Fig. 7 is a flowchart of a banknote face recognition method according to a fourth embodiment of the present invention, which is optimized based on the above-described embodiments, and the steps S311 to S313 are optimized to steps S311 'to S313'. Wherein, the same steps as those in the above embodiments are denoted by the same reference numerals, and explanations of the same or corresponding terms as those in the above embodiments are omitted. Correspondingly, the method of the embodiment includes the following operations:
s100, determining a first characteristic image and a second characteristic image in a first side image of the paper currency according to a preset position, and determining a third characteristic image and a fourth characteristic image in a second side image of the paper currency.
S200, binarizing the four feature images to obtain a first feature binarized image, a second feature binarized image, a third feature binarized image and a fourth feature binarized image.
S311', cutting a subarea image from the current feature binarization image.
Specifically, the range covered by the set line number (for example, 25 continuous lines) is set as the size of the sliding window, and then the line sum of all lines in the sliding window is counted line by line starting from the first line of the current feature binary image. In the above manner, the sliding window statistics are performed on the whole current feature binary image, for example, the whole current feature binary image has 60 lines as described in step S100, and there are 36 lines and a sum. And comparing the lines and the sums to determine the largest line and sum, determining the range corresponding to the largest line and sum as the range of the sub-region image, and cutting out the sub-region image from the current characteristic binary image according to the line number.
Briefly, the process is to determine a region where the maximum value of the row sum of the continuous 25 rows in the whole current feature binary image is located under the image resolution of 200DPI in the transverse direction and 150DPI in the longitudinal direction, and then determine the region as a required sub-region. This is because the number of lines occupied by the prefix number in the feature binarized image is about 25 lines, and it can be known from the first prefix number feature that if the prefix number exists in the image, the line sum of the region where the prefix number exists is the largest, so as to determine the region corresponding to the maximum value of the line sum of the continuous 25 lines, the sub-region possibly containing the prefix number can be determined. In the first binary feature image 401 shown in fig. 4a, it can be determined that the area image between the 30 th row and the 55 th row is a sub-area image, which includes the prefix "E4X 0053771". The purpose of setting the subarea image is to reserve only the area that may contain the prefix number, so as to avoid the interference of information other than the prefix number in the subsequent process.
S312', according to the preset segmentation position template, determining the segmentation position in the subarea image.
The preset segmentation position template is a position of an interval between the characters of the crown word number, which is determined by performing feature analysis on the crown word number in advance. As shown in fig. 8, the crown word number in the feature image 800 for obtaining the preset segmentation position template is "AD 206262288" and 9 character intervals are set, segmentation lines are set between every two characters of the crown word number, and then, starting from the first left column of the feature image 800, the first segmentation line 801 between the character "a" and the character "D", the second segmentation lines 802, … … between the character "D" and the character "2", the eighth segmentation line 808 between the character "2" and the character "8", and the ninth segmentation line 809 between the character "8" and the character "8" are determined to have column numbers of X, X +10, … …, X +70, and X +80 in the feature image, respectively, so that the preset segmentation position templates are 0, 10, … …, 70, and 80. For the crown word size in the first feature binary image 401' of fig. 4b, since the length is 4 characters shorter than that of the crown word size shown in fig. 8, the preset segmentation position template has 4 segmentation line positions less than that of the crown word size shown in fig. 8, i.e. the preset segmentation position templates are 0, 10,20, 30 and 40.
Specifically, taking the reflectance grayscale image as an example, a preset segmentation position template is first set in a first column on the left side of the subregion image, at this time, each segmentation line in the preset segmentation position template corresponds to one column, and a column sum of 9 segmentation line-corresponding columns in the template is obtained as a first group column sum and is recorded as a minimum column sum group. Then, starting from the first column, sliding the preset segmentation position template according to a set step (for example, one column), counting a group of column sums each time sliding, and comparing the group sums with the minimum column sum value. If the value of the set of column sums of the new statistic is less than the value of the minimum column sum set, then the set of column sums of the new statistic is set as the minimum column sum set. And repeating the steps until each column sum in the minimum column sum group is smaller than the set column sum (for example, 3), and the position of the preset segmentation position template is the segmentation position in the sub-region image. For example, when the preset division position template is set in the first column on the left side of the sub-region image, the column sums of the 1 st, 11 th, … … th, 71 th and 81 th columns are counted as the first group column sum (0,21,25,19,10,22,18,23,19), and the group column sum is set as the minimum column sum group. Then, the preset segmentation position template is slid to the right by one column in the subregion image, the column sums of the 2 nd column, the 12 th column, … …, the 72 th column and the 82 nd column are counted, a second group of column sums (0,19,28,17,15,26,18,15,17) are obtained, the size of the minimum column sum group and the second group of column sums is compared, for example, the sum of which array is smaller is compared, and the second group of column sums is set as the minimum column sum group. And repeating the steps until each column sum in the minimum column sum group is determined to be smaller than the set column sum 3, wherein the position corresponding to the preset segmentation position template is the segmentation position in the sub-region image. The segmentation location determination process for ir-reflectance images is similar, except that the number of elements in each set of column sums is 5.
S313', whether the sum of the columns of all the columns between every two adjacent segmentation positions in the subarea image is larger than a column sum threshold value or not is judged.
The column sum threshold is a total number of black dots within a predetermined column number set in advance, and is used to determine whether a character with a crown number exists in the predetermined column number, for example, if the predetermined column number is set as a number between two dividing lines in a predetermined dividing position template, then the column sum threshold may be set as a total number of columns and sums of all columns between the two dividing lines, and in order to better determine whether a character exists, a total number of black dots 40 of a character "1" is set as a column sum threshold.
Specifically, the column sum total of every two adjacent division positions, i.e., all columns between every two adjacent division lines, in the division positions determined in step S312' can be obtained to obtain 8 or 4 column sum total values, and the 8 or 4 column sum total values are compared with the column sum threshold values, respectively. If the 8 or 4 column sum total values are greater than the column sum threshold, then step S314 is performed; otherwise, it is stated that all the characters do not exist between the segmentation positions, the second crown word number feature is not met, the crown word number does not exist in the sub-region image, that is, the crown word number does not exist in the current feature binarized image to which the sub-region image belongs, and then step S311' is executed to perform the loop judgment of the next feature binarized image.
And S314, determining the characteristic binary image containing the crown word number.
And S320, determining the facing of the paper currency according to the feature binarization image containing the crown word number.
According to the technical scheme of the embodiment, four feature images containing the crown word numbers in two faces of the paper money are obtained, binarization processing is carried out on the four feature images to obtain corresponding four feature binarization images, and then a-c is carried out on each feature binarization image in the four feature binarization images in a circulating mode, namely a, a subregion image is cut from the current feature binarization image; b. determining the segmentation position in the subregion image according to a preset segmentation position template; c. and judging whether the sum of the columns of all the columns between every two adjacent segmentation positions in the sub-region image is greater than the columns and the threshold value or not until the feature binary image containing the crown word number is determined, and determining the face of the paper money according to the feature binary image containing the crown word number. The problem of the paper currency face of different currency values of different editions discerns is solved, based on the detection of unique second hat shop characteristic, improved the speed and the rate of accuracy that paper currency face discerned.
EXAMPLE five
Fig. 9 is a schematic structural diagram of a banknote recognition device according to a fifth embodiment of the present invention, and the same or corresponding terms as those in any of the above embodiments are not repeated herein. The apparatus may include:
the feature image determining module 910 is configured to determine a first feature image, a second feature image, a third feature image, and a fourth feature image in a complete image of a banknote according to a preset position, where one feature image of the four feature images includes a crown word number.
A feature binary image determining module 920, configured to binarize the four feature images determined by the feature image determining module 910 to obtain a first feature binary image, a second feature binary image, a third feature binary image, and a fourth feature binary image.
A facing determination module 930, configured to perform crown word number detection according to the crown word number features and the four feature binarized images obtained by the feature binarized image determination module 920, and determine the facing of the banknote according to the detection result.
The facing identification device for the paper currency solves the problem of facing identification of the paper currency with different versions and different currency values, and improves the accuracy of facing identification of the paper currency.
The facing identification device for the paper currency, provided by the embodiment of the invention, can execute the facing identification method for the paper currency, provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
EXAMPLE six
Fig. 10 is a schematic structural diagram of a banknote recognition device according to a sixth embodiment of the present invention, which is specifically described and optimized based on the foregoing embodiments. Wherein, the figure units same as the above embodiments are provided with the same reference numerals, and the explanation of the same or corresponding terms as any of the above embodiments is not repeated herein. The apparatus may include:
the feature image determining module 910 is configured to determine a first feature image, a second feature image, a third feature image, and a fourth feature image in a complete image of a banknote according to a preset position, where one feature image of the four feature images includes a crown word number.
A feature binary image determining module 920, configured to binarize the four feature images determined by the feature image determining module 910 to obtain a first feature binary image, a second feature binary image, a third feature binary image, and a fourth feature binary image.
The orientation determination module 930 includes a crown-number feature binarization image determination sub-module 931 and an orientation determination sub-module 932.
Optionally, the crown word number feature binarization image determining sub-module 931 is configured to perform crown word number detection on the four feature binarization images according to a first crown word number feature and/or a second crown word number feature in the crown word number features, and determine a feature binarization image containing the crown word number.
Further, the crown word number feature binarized image determining sub-module 931 includes a first crown word number feature binarized image determining unit operable to: a. determining whether a first preset line number inner line and a first statistical line number smaller than the first line and a threshold exist in the current characteristic binary image, wherein the first statistical line number is larger than or equal to the first line number threshold; b. determining whether a second preset line number inner line and a second statistical line number larger than a second line number threshold exist in the current characteristic binary image, wherein the second statistical line number is larger than or equal to the second line number threshold; c. if the first statistical line number and the second statistical line number exist, determining that the crown word number is contained in the current feature binarization image; and a-c is carried out on each feature binarization image in the four feature binarization images until the feature binarization image containing the crown word number is determined.
Further, the crown word number feature binarized image determining sub-module 931 further includes a second crown word number feature binarized image determining unit configured to: a. cutting a subregion image from the current characteristic binarization image; b. determining the segmentation position in the subregion image according to a preset segmentation position template; c. if the sum of the columns of all the columns between every two adjacent segmentation positions in the sub-region image is greater than the column sum threshold, determining that the crown word number is contained in the current feature binarization image; and a-c is carried out on each feature binarization image in the four feature binarization images until the feature binarization image containing the crown word number is determined.
Optionally, the orientation determining sub-module 932 is configured to determine the orientation of the banknote according to the feature binarized image containing the crown number determined by the crown number feature binarized image determining sub-module 931.
Further, the orientation determination submodule 932 is specifically configured to: if the characteristic binary image containing the crown word number is a first characteristic binary image, determining that the face of the paper currency is a positive face; if the characteristic binary image containing the crown word number is a second characteristic binary image, determining that the face of the paper currency is a front face reverse direction; if the feature binarization image containing the crown word number is a third feature binarization image, determining that the face of the paper currency is a reverse face and a forward face; and if the characteristic binary image containing the crown word number is a fourth characteristic binary image, determining that the face of the paper currency is reverse.
The facing identification device for the paper currency solves the problem of facing identification of the paper currency with different versions and different currency values, and improves the accuracy of facing identification of the paper currency.
The facing identification device for the paper currency, provided by the embodiment of the invention, can execute the facing identification method for the paper currency, provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (4)
1. A method for identifying a face of a banknote, comprising:
determining a first characteristic image, a second characteristic image, a third characteristic image and a fourth characteristic image in a complete image of the paper money according to a preset position, wherein one of the four characteristic images comprises a crown word number;
carrying out binarization on the four characteristic images to obtain a first characteristic binarization image, a second characteristic binarization image, a third characteristic binarization image and a fourth characteristic binarization image;
a. determining whether a first preset line number inner line and a first statistical line number smaller than the first line and a threshold exist in the current characteristic binary image, wherein the first statistical line number is larger than or equal to the first line number threshold;
b. determining whether a second preset line number inner line and a second statistical line number larger than a second line number threshold exist in the current characteristic binary image, wherein the second statistical line number is larger than or equal to the second line number threshold;
c. if the first statistical line number and the second statistical line number exist, determining that the crown word number is contained in the current feature binarization image;
performing a-c on each feature binarization image in the four feature binarization images until the feature binarization image containing the crown word number is determined;
determining the face of the paper currency according to a feature binarization image containing the crown word number;
wherein the row is the total number of black dots in a row.
2. The method according to claim 1 wherein said determining the facing of the note from the feature binarized image containing the crown word size comprises:
if the characteristic binary image containing the crown word number is a first characteristic binary image, determining that the face of the paper currency is a positive face;
if the characteristic binary image containing the crown word number is a second characteristic binary image, determining that the face of the paper currency is a front face reverse direction;
if the feature binarization image containing the crown word number is a third feature binarization image, determining that the face of the paper currency is a reverse face and a forward face;
and if the characteristic binary image containing the crown word number is a fourth characteristic binary image, determining that the face of the paper currency is reverse.
3. A banknote face-recognition apparatus, comprising:
the characteristic image determining module is used for determining a first characteristic image, a second characteristic image, a third characteristic image and a fourth characteristic image in the complete image of the paper money according to a preset position, wherein one of the four characteristic images contains a crown word number;
the characteristic binarization image determining module is used for binarizing the four characteristic images to obtain a first characteristic binarization image, a second characteristic binarization image, a third characteristic binarization image and a fourth characteristic binarization image;
the determination-oriented module comprises a crown word feature binarization image determination submodule and a determination-oriented submodule;
the crown word number characteristic binarization image determining submodule is used for:
a. determining whether a first preset line number inner line and a first statistical line number smaller than the first line and a threshold exist in the current characteristic binary image, wherein the first statistical line number is larger than or equal to the first line number threshold;
b. determining whether a second preset line number inner line and a second statistical line number larger than a second line number threshold exist in the current characteristic binary image, wherein the second statistical line number is larger than or equal to the second line number threshold;
c. if the first statistical line number and the second statistical line number exist, determining that the crown word number is contained in the current feature binarization image;
performing a-c on each feature binarization image in the four feature binarization images until the feature binarization image containing the crown word number is determined;
the orientation determining submodule is used for determining the orientation of the paper money according to the feature binarization image containing the crown word number;
wherein the row is the total number of black dots in a row.
4. The apparatus of claim 3, wherein the orientation determination submodule is specifically configured to:
if the characteristic binary image containing the crown word number is a first characteristic binary image, determining that the face of the paper currency is a positive face;
if the characteristic binary image containing the crown word number is a second characteristic binary image, determining that the face of the paper currency is a front face reverse direction;
if the feature binarization image containing the crown word number is a third feature binarization image, determining that the face of the paper currency is a reverse face and a forward face;
and if the characteristic binary image containing the crown word number is a fourth characteristic binary image, determining that the face of the paper currency is reverse.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611196640.XA CN108230535B (en) | 2016-12-22 | 2016-12-22 | Facing identification method and device for paper money |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611196640.XA CN108230535B (en) | 2016-12-22 | 2016-12-22 | Facing identification method and device for paper money |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108230535A CN108230535A (en) | 2018-06-29 |
| CN108230535B true CN108230535B (en) | 2021-08-13 |
Family
ID=62656134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611196640.XA Active CN108230535B (en) | 2016-12-22 | 2016-12-22 | Facing identification method and device for paper money |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108230535B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112733834B (en) * | 2021-03-30 | 2021-07-06 | 恒银金融科技股份有限公司 | Character area positioning method based on non-matrix window mode |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010079754A (en) * | 2008-09-26 | 2010-04-08 | Secom Co Ltd | Notification device |
| JP2010176589A (en) * | 2009-01-30 | 2010-08-12 | Secom Co Ltd | Transaction monitoring device |
| CN104346858A (en) * | 2014-10-10 | 2015-02-11 | 新达通科技股份有限公司 | Paper currency face value identification method and device based on magnetic image |
| 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 |
| CN105160756A (en) * | 2015-08-18 | 2015-12-16 | 深圳怡化电脑股份有限公司 | Paper money facing direction recognition method and device |
| CN106022346A (en) * | 2016-05-24 | 2016-10-12 | 深圳怡化电脑股份有限公司 | Banknote number cutting method and device |
-
2016
- 2016-12-22 CN CN201611196640.XA patent/CN108230535B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010079754A (en) * | 2008-09-26 | 2010-04-08 | Secom Co Ltd | Notification device |
| JP2010176589A (en) * | 2009-01-30 | 2010-08-12 | Secom Co Ltd | Transaction monitoring device |
| CN104346858A (en) * | 2014-10-10 | 2015-02-11 | 新达通科技股份有限公司 | Paper currency face value identification method and device based on magnetic image |
| 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 |
| CN105160756A (en) * | 2015-08-18 | 2015-12-16 | 深圳怡化电脑股份有限公司 | Paper money facing direction recognition method and device |
| CN106022346A (en) * | 2016-05-24 | 2016-10-12 | 深圳怡化电脑股份有限公司 | Banknote number cutting method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108230535A (en) | 2018-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104992496A (en) | Paper money face identification method and apparatus | |
| CN106952393B (en) | Paper money identification method and device, electronic equipment and storage medium | |
| CN106920318B (en) | Method and device for identifying paper money | |
| CN103606220A (en) | Check printed number recognition system and check printed number recognition method based on white light image and infrared image | |
| CN106875545B (en) | Method and device for identifying paper money | |
| CN110599512A (en) | Medicine name identification method and system based on image identification | |
| CN106780966B (en) | A kind of paper money discrimination method and device | |
| KR20070011867A (en) | Method and apparatus for recognizing serial number of paper money | |
| CN115272305B (en) | Button hole defect detection method | |
| CN106204616B (en) | Method and device for identifying currency value of Iran paper money | |
| CN106898078B (en) | Port currency version identification method and device | |
| CN108734846B (en) | Method and device for denoising paper money image, terminal and storage medium | |
| CN106683264B (en) | Method and device for identifying stained crown word number | |
| KR102094234B1 (en) | Serial number recognition Apparatus and method for paper money | |
| CN101583978B (en) | Apparatus for media recognition and method for media kind distinction with the same | |
| CN107742357A (en) | A kind of recognition methods of paper money number and device | |
| CN107680246B (en) | Method and equipment for positioning curve boundary in paper money pattern | |
| CN108230535B (en) | Facing identification method and device for paper money | |
| CN106296975B (en) | method and device for identifying face value of dollar paper money | |
| CN106447908B (en) | Paper money counterfeit distinguishing method and device | |
| CN106683262B (en) | A kind of bank note towards recognition methods and device | |
| CN108073921B (en) | Method and device for identifying banknote information | |
| CN108171864B (en) | Method and device for identifying paper money version | |
| CN108256374A (en) | A kind of method for rapidly positioning and system of Maxicode bar codes | |
| CN108074322B (en) | Image identification method and device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |