WO2018019041A1 - Pasted paper money detection method and device - Google Patents

Abstract

A pasted paper money detection method and device. The method comprises: collecting a paper money image (S1); performing coarse positioning on an area probably for pasting in the paper money image to acquire a coarse positioning area image (S2); performing grey level histogram statistics on the coarse positioning area image, performing grading on the coarse positioning area image in accordance with a statistical result according to a grey level value interval, and dividing the coarse positioning area image into a plurality of grey level images (S3); respectively performing binarization processing on each of the grey level images, and performing envelopment fitting on the binarized images to acquire an image envelope of each of the grey level images (S4); and performing straight line detection on the image envelope, and if it is detected that the image envelope includes two straight lines not belonging to paper money itself, marking the paper money corresponding to the paper money image as pasted paper money (S5). According to an image feature of pasted paper money shown in a paper money image, with regard to patterns with different grey level conditions probably formed by a pasted tape, the detection rate of the pasted paper money can be effectively improved, the solution is easy to realize, and the equipment costs are economized.

Description

Method and device for detecting banknotes Technical field

The present invention relates to the field of image processing technologies, and in particular, to a method and apparatus for detecting banknotes.

Background technique

With the widespread use of financial equipment, the demand for equipment has risen rapidly. Banknote (banknote) receiving equipment is often used in environments such as ATM equipment, station ticket vending machines, retail stores, and vending machines of various financial institutions. Because people do not understand the use of financial equipment, it is easy to receive a variety of poor quality banknotes during the operation of the equipment, such as the old banknotes that are spliced by the banknotes after the banknotes are broken, or the illegal elements are used to splicing the counterfeit banknotes with real banknotes. The spliced banknotes are collectively referred to herein as spliced notes. If it cannot be detected, it will have a great impact on the reliability and safety of the equipment.

At present, the sticking of the banknote is mainly identified by detecting the thickness signal or the reflective signal of the banknote adhesive tape. However, the detection of the banknote thickness signal requires high precision of the device, and the detection is difficult. It is necessary to install a plurality of thickness detecting sensors, and the device cost is high. The detection method of the banknote reflection signal utilizes the reflective property of the adhesive tape, and it is not only required to add a light sensor, but also when the adhesive tape used does not have the reflective property, the device cannot detect it.

Summary of the invention

It is an object of the present invention to provide a method and apparatus for detecting a banknote, which is based on an image processing technique for effective identification of a banknote.

In order to achieve the above object, an embodiment of the present invention provides a method for detecting a banknote, comprising:

Collecting banknote images;

Perform coarse positioning on the area of the banknote image that may be pasted, and obtain an image of the coarse positioning area;

Performing gray scale histogram statistics on the coarse positioning area image, classifying the coarse positioning area image according to a gray value interval according to a statistical result, and dividing the coarse positioning area image into a plurality of gray level images;

Performing binarization processing on each gray level image, and performing envelope fitting on the binarized image to obtain an image envelope of each gray level image;

The image envelope is detected by a straight line. If it is detected that the image envelope contains two straight lines of non-banknotes, the banknote corresponding to the banknote image is marked as a banknote.

Preferably, the gray-scale histogram statistics are performed on the coarse positioning area image, and the coarse positioning area image is classified according to a gray value interval according to a statistical result, and the coarse positioning area image is divided into multiple gray levels. Level image, including:

Obtaining a gray histogram of the image of the coarse positioning area;

And connecting respective vertices of the gray histogram into a curve, and smoothing the curve to obtain an envelope curve function of the gray histogram;

Deriving the envelope curve function and calculating an extreme point of the envelope curve function;

Calculating a midpoint of two adjacent extreme points, classifying the coarse positioning area image by using the midpoint as a boundary point of the gray value interval, and dividing the coarse positioning area image into multiple gray levels image.

Preferably, the respective gray level images are separately binarized, and the binarized image is subjected to envelope fitting to obtain an image envelope of each gray level image, including:

Performing binarization processing on each grayscale image separately, and performing an opening operation on the binarized image;

The image after the opening operation is retrieved in a preset step size in both the horizontal and vertical directions to obtain a boundary point; and the boundary point is fitted to obtain an image envelope.

Preferably, the banknote image comprises a reflected image and a transmitted image.

Preferably, the image envelope is linearly detected. If it is detected that the image envelope contains two straight lines of the non-banknotes, the banknote corresponding to the banknote image is marked as a banknote, including:

Obtaining a paste score of the banknote image according to a result of the line detection; wherein

When the banknote image is a reflected image, the obtained pasting score is a first pasting score;

When the banknote image is a transmission image, the obtained pasting score is a second pasting score;

When the banknote image is a transmission image, detecting whether there is a bright point above and/or below the area where the image envelope is located, or a gray point whose value is greater than a preset gray level threshold; or detecting whether the gray level is present in the image of the coarse positioning area a bright line whose value is greater than a preset grayscale threshold; acquiring a third pasting score of the banknote image according to the detection result of the bright spot or the bright line;

Performing a weighted summation on the first pasting score, the second pasting score, and the third pasting score, and calculating a comprehensive pasting score; if the comprehensive pasting score is greater than a preset pasting threshold, Then, the banknote is determined to be a banknote; otherwise, the banknote is determined to be a non-stick banknote.

Correspondingly, an embodiment of the present invention further provides a device for detecting a banknote, comprising:

An image acquisition module for collecting banknote images;

a coarse positioning module, configured to perform rough positioning on an area that may be pasted in the image of the banknote, and obtain an image of the coarse positioning area;

a gray level grading module, configured to perform gray level histogram statistics on the coarse positioning area image, classify the coarse positioning area image according to a gray value interval according to a statistical result, and divide the coarse positioning area image into multiple Grayscale image;

An envelope fitting module is configured to separately perform binarization processing on each grayscale image, and perform envelope fitting on the binarized image to obtain an image envelope of each grayscale image;

The line detection module is configured to perform line detection on the image envelope. If it is detected that the image envelope includes two straight lines that are not in the banknote, the banknote corresponding to the banknote image is marked as a banknote.

Preferably, the gray leveling module comprises:

a histogram acquisition unit, configured to acquire a gray histogram of the image of the coarse positioning area;

An envelope function obtaining unit, configured to connect each vertex of the gray histogram into a curve, perform smoothing on the curve, and acquire an envelope curve function of the gray histogram;

An extreme point calculation unit configured to derive the envelope curve function and calculate an extreme point of the envelope curve function;

a gray grading unit, configured to calculate a midpoint of two adjacent extreme points, and classify the coarse positioning area image by using the midpoint as a boundary point of the gray value interval, and the coarse positioning area image Divided into multiple grayscale images.

The envelope fitting module includes:

a binarization unit for respectively performing binarization processing on each grayscale image, and performing an opening operation on the binarized image;

The boundary fitting unit is configured to retrieve the image after the opening operation in a preset step size in both horizontal and vertical directions to obtain a boundary point; and fit the boundary point to obtain an image envelope.

Preferably, the banknote image comprises a reflected image and a transmitted image.

Preferably, the line detection module comprises:

a pasting score calculation unit, configured to acquire a pasting score of the banknote image according to a result of the straight line detection; wherein

When the banknote image is a reflected image, the obtained pasting score is a first pasting score;

When the banknote image is a transmission image, the obtained pasting score is a second pasting score;

When the banknote image is a transmission image, detecting whether there is a bright point above and/or below the area where the image envelope is located, or a gray point whose value is greater than a preset gray level threshold; or detecting whether the gray level is present in the image of the coarse positioning area a bright line whose value is greater than a preset grayscale threshold; acquiring a third pasting score of the banknote image according to the detection result of the bright spot or the bright line;

a comprehensive judging unit, configured to perform weighted summation on the first pasting score, the second pasting score, and the third pasting score, and calculate a comprehensive pasting score; if the comprehensive pasting score is greater than a pre- If the pasting threshold is set, the banknote is determined to be a banknote; otherwise, the banknote is determined to be a non-sticker.

The invention performs rough positioning on the collected banknote image, obtains a coarse positioning area image in which the pasting situation may exist, performs gray scale histogram statistics on the coarse positioning area image, and classifies according to the gray value interval according to the statistical result, The image of each level is binarized and envelope-fitting, the image envelope of each level of image is obtained, and the image envelope is detected by a straight line, and the image envelope is detected to include two non-banknotes. In the case of a straight line (especially two parallel straight lines), the banknote corresponding to the banknote image is marked as a banknote. According to the image feature of the pasted banknote in the banknote image, the pattern of different gray scale conditions which may be formed by the adhesive tape can correctly position the tape and detect the tape in the case of dirt or multi-layer chromatic aberration. Effectively improve the detection rate of the attached banknotes, the solution is easy to implement, and the equipment cost is saved.

DRAWINGS

1 is a flow chart of a method for detecting an embodiment of a method for detecting a banknote according to the present invention;

2 is a schematic view showing a reflection image of a sticking banknote provided by the present invention;

3 is a schematic view showing a transmission image of a sticking banknote provided by the present invention;

4 is a schematic view of an image of a rough positioning area of a banknote provided by the present invention;

Figure 5 is a gray histogram of the coarse positioning area image A1 shown in Figure 4;

Figure 6 is a schematic diagram showing the division of the gray value interval of the gray histogram shown in Figure 5;

Figure 7 is a schematic diagram of an image envelope of the second-level image a2 shown in Figure 4;

Figure 8 is a block diagram showing the structure of an embodiment of the stick detecting device of the present invention.

detailed description

The technical solutions in the embodiments of the present invention will be clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. The described embodiments are only a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

Referring to Fig. 1, a flow chart of a method of an embodiment of a method for detecting a banknote of the present invention is provided.

As shown in FIG. 1, the method includes steps S1 to S5:

S1, collecting banknote images.

The banknote image includes a reflected image and a transmitted image, the reflected image being an image in which the light source is reflected in front of the banknote and the light is reflected on the surface of the banknote, as shown in FIG. The transmission image is a dark background image formed by the light source on the back side of the banknote and the light passing through the banknote, as shown in FIG.

S2: Perform coarse positioning on the area of the banknote image that may be pasted, and obtain an image of the coarse positioning area.

Due to the use of the tape and the uncertainty of the stitching position, it is difficult to position the tape to the precise position. After the banknote image is collected, the image of the coarse positioning area where the special color block is located can be quickly and coarsely located on the collected original image, as shown in A1 of FIG. 4, and the positioning of the special color block can be compared with the standard banknote image. Obtain. In addition, the position where the banknotes are folded more (such as 1/2, 1/3, and 1/4 positions) is liable to cause damage. The adhesive tape is often in these positions, and these positions can be preferentially checked during rough positioning to improve the detection. effectiveness.

S3, performing grayscale histogram statistics on the coarse positioning area image, classifying the coarse positioning area image according to a gray value interval according to a statistical result, and dividing the coarse positioning area image into a plurality of gray level images.

In a specific implementation, the step S3 includes:

S31. Acquire a gray histogram of the image of the coarse positioning area.

Before the gray histogram statistics are performed, the coarse positioning area image may be linearly enhanced to enhance the contrast of the coarse positioning area image. The gray histogram of the image of the coarse positioning area shown by A1 in Fig. 4 is as shown in Fig. 5.

S32, connecting each vertex of the gray histogram into a curve, and smoothing the curve to obtain an envelope curve function of the gray histogram.

Each vertical stripe in the gray histogram represents the number of pixels corresponding to the gray value, and the vertices of the stripe are connected to obtain a curve, and then the curve is smoothed to obtain the envelope curve of the gray histogram. function. The envelope curve function obtained is a binary quadratic equation:

y=f(x), x∈(0,255)

S33, deriving the envelope curve function, and calculating an extreme point of the envelope curve function.

Deriving the envelope curve function obtained in step S2, and obtaining an extremum point x _i and a corresponding extremum y _i at a derivative value of 0,

For the number of extreme points in the gray value of 0 to 255, the analytical expression of the envelope curve function can be calculated:

S34. Calculate a midpoint of two adjacent extreme points, and classify the coarse positioning area image by using the midpoint as a boundary point of the gray value interval, and divide the coarse positioning area image into multiple grays. Degree image.

Find the midpoint of each adjacent extreme point

The series is divided by ρ _i as the boundary point of each level. The number of levels is recorded as

(The maximum number of levels allowed can be set according to the situation). The gradation of the gray histogram in Fig. 5 is as shown in Fig. 6, and the gradation values of 0 to 255 are divided into three sections of Level 1, Level 2, and Level 3. Correspondingly, the coarse positioning area image A1 in FIG. 4 is divided into the first level image a1, the second level image a2, and the third level image a3 according to the gray value interval, and each gray level image only contains the gray value. The pixels within the corresponding gray value interval.

S4: Perform binarization processing on each grayscale image, and perform envelope fitting on the binarized image to obtain an image envelope of each grayscale image.

In a specific implementation, the step S4 includes:

S41, performing binarization processing on each grayscale image, and performing an opening operation on the binarized image.

After binarization processing is performed on each grayscale image, a valid interval image can be obtained, and each valid interval image is opened. After corrosion and expansion, the boundary interference and small dirty spots are eliminated. A clearer binarized image is beneficial to improve the accuracy of subsequent envelope fitting.

S42: Search for the opened image in a preset step size in two directions, horizontally and vertically, obtain a boundary point, and fit the boundary point to obtain an image envelope.

The binarized images of each level after the opening operation are searched in the horizontal and vertical directions respectively in a specific step size, and several boundary points can be found, and the fitting can be performed under the condition that a certain deviation value e is allowed, and the fitting can be obtained. The image envelope of each level of image, as shown in Figure 7, is the image envelope of the second level image a2.

S5, performing linear detection on the image envelope, if it is detected that the image envelope contains two non-banknotes having straight lines (the line referred to herein includes any approximate straight line, not strictly geometrical The straight line) marks the banknote corresponding to the banknote image as a banknote.

The image envelope of each level of image is searched for linear traces step by step, and some dirty blocks or interference blocks can be directly excluded during the detection process. After the straight line shape appears, the line is searched for translation. It is detected whether there is another straight line parallel to (or nearly parallel to), and the areas formed by the two straight lines are of the same color level (color intervals having the same or similar gray value). If there are two straight lines at the same time, and the spacing between the two lines meets the preset spacing interval (ie tape width), it is marked as a paste. In a specific implementation, it is also possible to determine the presence of the adhesive tape by comparing with the standard banknote image template according to whether there is a straight line trace (such as an original pattern or a security thread on the banknote) that the banknote itself has in the selected area.

Although the reflection image is taken as an example in the above embodiment, it is known to those skilled in the art that the same operation can be performed on the transmission image to perform the detection of the sticker. In a specific implementation, the above operations may be performed on the reflected image and the transmitted image of the banknote, respectively, to improve the accuracy of the detection of the deposited banknote. For example, when a straight line having two non-banknotes is detected in the reflected image, and a corresponding straight line is also detected at the same position of the transmitted image, it is confirmed that the position is pasted with a tape-like foreign matter.

Specifically, the comprehensive analysis and judgment can be performed by combining the reflected image and the transmitted image in the following manner:

In a specific implementation, the step S5 includes:

S51. Acquire a paste score of the banknote image according to the result of the line detection. among them,

When the banknote image is a reflected image, the acquired pasting score is the first pasting score score1.

When the banknote image is a transmission image, the acquired pasting score is the second pasting score score2.

When the banknote image is a transmission image, detecting whether there is a bright point above and below the area where the image envelope is located, and the gradation value is greater than the first gradation threshold and is similar to the line arrangement; or detecting the coarse positioning area image Whether there is a gray value greater than the first gray threshold a bright line; according to the detection result of the bright spot or the bright line, the third pasting score score3 of the banknote image is obtained.

The specific calculation rules for pasting the score can be set as needed. For example, the closer the two lines are to parallel or the clearer the boundary, the higher the paste score; the closer the width between the two lines is to the width of the common adhesive tape, the paste score The higher. A person skilled in the art can select a calculation rule that conforms to the characteristics of the banknote according to actual needs, which is not limited by the present invention. In a specific implementation manner, the first pasting score score1 and the second pasting score score2 are calculated as follows: two straight lines are found in the image after binarizing the grayscale image, if the two straight lines are close to each other Parallel, the paste score S1 is given; if the two lines are not parallel, but the area of the image where the two lines are located reaches the set value, the paste score S2 is given; if there is a line boundary and a curve boundary, it can be recorded The score S3 is pasted; wherein the score S1>S2>S3 is pasted.

In addition, since there is a crack in the attached banknote itself, it is difficult to completely repair the gap after artificially splicing, and under the transmission image, light leakage is likely to occur, forming a bright spot or a bright line. The present invention comprehensively considers the above situation, and uses it as a detection item for sticking banknotes, which is advantageous for further improving the accuracy of detecting the banknotes.

S52: Perform weighted summation on the first pasting score score1, the second pasting score score2, and the third pasting score score3, and calculate a comprehensive pasting score _final . If the composite pasting score _{final is} greater than the preset pasting threshold _set , the banknote is determined to be a banknote; otherwise, the banknote is determined to be a non-sticker.

Comprehensively analyze the shape of the pattern searched by the reflected image and the transmitted image, and weight the sum of the used items to obtain the comprehensive pasting score:

Score _final =α*score1+β*score2+γ*score3

Wherein, α, β, and γ are the weights of the first pasting score score1, the second pasting score score2, and the third pasting score score3, respectively. When the score _final > score _set , it can be determined that the banknote is a paste banknote.

According to the image features of the pasted banknotes in the reflected image and the fluoroscopic image, the invention comprehensively considers the pattern of different gray scale conditions that the tape may form, and excludes non-adhesive interference according to the characteristics thereof, in the presence of dirt or multi-layer chromatic aberration. In this case, the position of the tape can be correctly positioned and the tape can be detected, which can effectively improve the detection rate of the stuck banknote, and the solution is easy to implement and saves equipment cost.

FIG. 8 is a structural diagram of an apparatus for detecting a banknote detecting apparatus according to the present invention. The principle of the embodiment is the same as that of the embodiment shown in FIG. 1. For details in this embodiment, reference may be made to FIG. A related description in the embodiment is shown.

As shown in FIG. 8, the stick detecting device includes:

An image acquisition module 71 is configured to collect an image of a banknote;

The coarse positioning module 72 is configured to perform coarse positioning on the area of the banknote image that may be pasted, and obtain an image of the coarse positioning area;

The gray level grading module 73 is configured to perform gray level histogram statistics on the coarse positioning area image, classify the coarse positioning area image according to a gray value interval according to a statistical result, and divide the coarse positioning area image into multiple Grayscale image;

The envelope fitting module 74 is configured to separately perform binarization processing on each gray scale image, and perform envelope fitting on the binarized image to obtain an image envelope of each gray scale image;

The line detection module 75 is configured to perform line detection on the image envelope. If it is detected that the image envelope includes two straight lines of the non-banknotes, the banknote corresponding to the banknote image is marked as a banknote.

The gray level grading module 73 includes:

a histogram obtaining unit 731, configured to acquire a gray histogram of the image of the coarse positioning area;

An envelope function obtaining unit 732, configured to connect each vertex of the gray histogram into a curve, and smooth the curve, Obtaining an envelope curve function of the gray histogram;

An extreme point calculation unit 733, configured to derivate the envelope curve function, and calculate an extreme point of the envelope curve function;

a gray leveling unit 734, configured to calculate a midpoint of two adjacent extreme points, and classify the coarse positioning area image by using the middle point as a boundary point of the gray value interval, and the coarse positioning area is The image is divided into a plurality of grayscale images.

The envelope fitting module 74 includes:

The binarization unit 741 is configured to separately perform binarization processing on each gray scale image, and perform an open operation on the binarized image;

The boundary fitting unit 742 is configured to retrieve the image after the opening operation in a preset step size in both horizontal and vertical directions to acquire a boundary point; and fit the boundary point to obtain an image envelope .

The banknote image preferably includes a reflected image and a transmitted image.

In a specific implementation, the line detection module 75 includes:

a pasting score calculation unit 751, configured to acquire a pasting score of the banknote image according to a result of the straight line detection; wherein

When the banknote image is a reflected image, the obtained pasting score is a first pasting score;

When the banknote image is a transmission image, the obtained pasting score is a second pasting score;

When the banknote image is a transmission image, detecting whether there is a bright point above and/or below the area where the image envelope is located, or a gray point whose value is greater than a preset gray level threshold; or detecting whether the gray level is present in the image of the coarse positioning area a bright line whose value is greater than a preset grayscale threshold; acquiring a third pasting score of the banknote image according to the detection result of the bright spot or the bright line;

The comprehensive judging unit 752 is configured to perform weighted summation on the first pasting score, the second pasting score, and the third pasting score, and calculate a comprehensive pasting score; if the comprehensive pasting score is greater than The preset paste is divided into thresholds, and then the banknote is determined to be a banknote; otherwise, the banknote is determined to be a non-sticker.

In summary, the method and device for detecting the banknotes provided by the present invention obtains a coarse positioning area image in which a pasting situation may exist by performing coarse positioning on the collected banknote image, and performs grayscale histogram statistics on the image of the coarse positioning area. According to the statistical result, the gray value interval is used to classify, and the image of each level is binarized and envelope-fitting, and the image envelope of each level image is obtained, and the image envelope is detected by the straight line, and the image is detected. When the image envelope contains two straight lines (especially two parallel straight lines) which are not in the banknote itself, the banknote corresponding to the banknote image is marked as a banknote. According to the image feature of the pasted banknote in the banknote image, the pattern of different gray scale conditions which may be formed by the adhesive tape can correctly position the tape and detect the tape in the case of dirt or multi-layer chromatic aberration. Effectively improve the detection rate of the attached banknotes, the solution is easy to implement, and the equipment cost is saved.

It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical. Units can be located in one place or distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, and specifically, one or more communication buses or signal lines can be realized. Those of ordinary skill in the art can understand and implement without any creative effort.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general hardware, and of course, dedicated hardware, dedicated CPU, dedicated memory, dedicated memory, Special components and so on. In general, any function performed by a computer program can be easily implemented with the corresponding hardware, and The specific hardware structure to achieve the same function can also be various, such as analog circuits, digital circuits or dedicated circuits. However, for the purposes of the present invention, software program implementation is a better implementation in more cases. Based on the understanding, the technical solution of the present invention, which is essential or contributes to the prior art, can be embodied in the form of a software product stored in a readable storage medium, such as a floppy disk of a computer. , U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., including a number of instructions to make a computer device (may be A personal computer, server, or network device, etc.) performs the methods described in various embodiments of the present invention.

The above is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the appended claims.

Claims (10)

1. A method for detecting a banknote, characterized in that it comprises:

   Collecting banknote images;

   Perform coarse positioning on the area of the banknote image that may be pasted, and obtain an image of the coarse positioning area;

   Performing gray scale histogram statistics on the coarse positioning area image, classifying the coarse positioning area image according to a gray value interval according to a statistical result, and dividing the coarse positioning area image into a plurality of gray level images;

   Performing binarization processing on each gray level image, and performing envelope fitting on the binarized image to obtain an image envelope of each gray level image;

   The image envelope is detected by a straight line. If it is detected that the image envelope contains two straight lines of non-banknotes, the banknote corresponding to the banknote image is marked as a banknote.
2. The method of detecting a banknote according to claim 1, wherein the gray-scale histogram statistics are performed on the image of the coarse positioning area, and the image of the coarse positioning area is classified according to a gray value interval according to a statistical result. Dividing the coarse positioning area image into a plurality of gray level images, including:

   Obtaining a gray histogram of the image of the coarse positioning area;

   And connecting respective vertices of the gray histogram into a curve, and smoothing the curve to obtain an envelope curve function of the gray histogram;

   Deriving the envelope curve function and calculating an extreme point of the envelope curve function;

   Calculating a midpoint of two adjacent extreme points, classifying the coarse positioning area image by using the midpoint as a boundary point of the gray value interval, and dividing the coarse positioning area image into multiple gray levels image.
3. The method for detecting a banknote according to claim 1, wherein said each grayscale image is separately binarized, and the binarized image is subjected to envelope fitting to obtain respective grayscale levels. The image envelope of the image, including:

   Performing binarization processing on each grayscale image separately, and performing an opening operation on the binarized image;

   The image after the opening operation is retrieved in a preset step size in both the horizontal and vertical directions to obtain a boundary point; and the boundary point is fitted to obtain an image envelope.
4. The method of detecting a banknote according to any one of claims 1 to 3, wherein the banknote image comprises a reflected image and a transmitted image.
5. The method for detecting a banknote according to claim 4, wherein said image envelope is linearly detected, and if it is detected that said image envelope contains two straight lines of non-banknotes, The banknote corresponding to the banknote image is marked as a banknote, including:

   Obtaining a paste score of the banknote image according to a result of the line detection; wherein

   When the banknote image is a reflected image, the obtained pasting score is a first pasting score;

   When the banknote image is a transmission image, the obtained pasting score is a second pasting score;

   When the banknote image is a transmission image, detecting whether there is a bright point above and/or below the area where the image envelope is located, or a gray point whose value is greater than a preset gray level threshold; or detecting whether the gray level is present in the image of the coarse positioning area a bright line whose value is greater than a preset grayscale threshold; acquiring a third pasting score of the banknote image according to the detection result of the bright spot or the bright line;

   Performing a weighted summation on the first pasting score, the second pasting score, and the third pasting score, and calculating a comprehensive pasting score; if the comprehensive pasting score is greater than a preset pasting threshold, Then, the banknote is determined to be a banknote; otherwise, the banknote is determined to be a non-stick banknote.
6. A sticking banknote detecting device, comprising:

   An image acquisition module for collecting banknote images;

   a coarse positioning module, configured to perform rough positioning on an area that may be pasted in the image of the banknote, and obtain an image of the coarse positioning area;

   a gray level grading module, configured to perform gray level histogram statistics on the coarse positioning area image, classify the coarse positioning area image according to a gray value interval according to a statistical result, and divide the coarse positioning area image into multiple Grayscale image;

   An envelope fitting module is configured to separately perform binarization processing on each grayscale image, and perform envelope fitting on the binarized image to obtain an image envelope of each grayscale image;

   The line detection module is configured to perform line detection on the image envelope. If it is detected that the image envelope includes two straight lines that are not in the banknote, the banknote corresponding to the banknote image is marked as a banknote.
7. The affixing detecting device of claim 6, wherein the gradation grading module comprises:

   a histogram acquisition unit, configured to acquire a gray histogram of the image of the coarse positioning area;

   An envelope function obtaining unit, configured to connect each vertex of the gray histogram into a curve, perform smoothing on the curve, and acquire an envelope curve function of the gray histogram;

   An extreme point calculation unit configured to derive the envelope curve function and calculate an extreme point of the envelope curve function;

   a gray grading unit, configured to calculate a midpoint of two adjacent extreme points, and classify the coarse positioning area image by using the midpoint as a boundary point of the gray value interval, and the coarse positioning area image Divided into multiple grayscale images.
8. The adhesive banknote detecting device according to claim 6, wherein the envelope fitting module comprises:

   a binarization unit for respectively performing binarization processing on each grayscale image, and performing an opening operation on the binarized image;

   The boundary fitting unit is configured to retrieve the image after the opening operation in a preset step size in both horizontal and vertical directions to obtain a boundary point; and fit the boundary point to obtain an image envelope.
9. The stick detecting device according to any one of claims 6 to 8, wherein the banknote image comprises a reflected image and a transmitted image.
10. The stick detecting device of claim 9, wherein the line detecting module comprises:

    a pasting score calculation unit, configured to acquire a pasting score of the banknote image according to a result of the straight line detection; wherein

    When the banknote image is a reflected image, the obtained pasting score is a first pasting score;

    When the banknote image is a transmission image, the obtained pasting score is a second pasting score;

    When the banknote image is a transmission image, detecting whether there is a bright point above and/or below the area where the image envelope is located, or a gray point whose value is greater than a preset gray level threshold; or detecting whether the gray level is present in the image of the coarse positioning area a bright line whose value is greater than a preset grayscale threshold; acquiring a third pasting score of the banknote image according to the detection result of the bright spot or the bright line;

    a comprehensive judging unit, configured to perform weighted summation on the first pasting score, the second pasting score, and the third pasting score, and calculate a comprehensive pasting score; if the comprehensive pasting score is greater than a pre- If the pasting threshold is set, the banknote is determined to be a banknote; otherwise, the banknote is determined to be a non-sticker.

Images