CN114419644A - Banknote denomination recognition method and system - Google Patents
Banknote denomination recognition method and system Download PDFInfo
- Publication number
- CN114419644A CN114419644A CN202111658954.8A CN202111658954A CN114419644A CN 114419644 A CN114419644 A CN 114419644A CN 202111658954 A CN202111658954 A CN 202111658954A CN 114419644 A CN114419644 A CN 114419644A
- Authority
- CN
- China
- Prior art keywords
- banknote
- frequency domain
- recognized
- image
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 26
- 230000009466 transformation Effects 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 15
- 238000012706 support-vector machine Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 abstract description 8
- 230000000694 effects Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/14—Transformations for image registration, e.g. adjusting or mapping for alignment of images
- G06T3/147—Transformations for image registration, e.g. adjusting or mapping for alignment of images using affine transformations
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method and a system for identifying the denomination of a bank note, comprising the following steps: acquiring a bank note to be recognized, and extracting frequency domain characteristics of the bank note to be recognized; acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set; and determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized. The invention adopts a machine learning method to identify the denomination of the bank note, has the characteristics of strong universality and high execution efficiency, and particularly has better identification effect on the bank note which is deformed and stained.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to a banknote denomination recognition method and a banknote denomination recognition system.
Background
In the application of financial machine tools, the denomination recognition of bank notes is the core function, and all the functions of counting bank notes and identifying false are based on correctly recognizing the denomination of bank notes, and strictly speaking, the wrong denomination recognition is not allowed. The traditional image-based denomination recognition method has high requirements on the quality of the recognized image, and when the acquired image is deformed, the recognition method is difficult to effectively deal with, and particularly when the traditional image-based denomination recognition method faces plastic banknotes, the problem of inaccurate denomination recognition is more prominent. In addition, as the demand of the financial machine products for the bill counting speed increases day by day, more than 900 sheets/minute is generally required, so that the accuracy of the denomination recognition of the bills is required to be improved in a shorter time.
One of the existing identification technologies is a banknote denomination identification method based on image histogram matching, which is based on histogram information of image color components, and then is averaged to obtain a denomination identification template, wherein the training process comprises the following steps: acquiring a two-dimensional matrix of an image training sample; filtering each image training sample; extracting a first histogram vector of red, green and blue three-color components of each image training sample; respectively obtaining the average value of the histogram of each color component in a plurality of image training samples as a template of the paper money of the denomination; the identification process specifically comprises: acquiring a two-dimensional matrix of an image to be identified; filtering an image to be identified; and extracting a second histogram of red, green and blue three-color components with the identification image, respectively obtaining the distance between the second histogram and each denomination banknote template, and comparing the distances, wherein the denomination corresponding to the denomination banknote template with the minimum distance from the second histogram is judged as the denomination of the banknote to be identified. The scheme has the advantages of general identification accuracy, incapability of adapting to bank notes with abundant color information, general universality, need of calculating the characteristics of 3 color channels and general algorithm efficiency.
The training part can automatically extract the characteristics after collecting enough samples, and does not need human intervention, thereby greatly improving the recognition efficiency; the identifying part comprises acquiring graph edge information; finding a graph coordinate through edge information fitting, and performing plane transformation according to coordinate information to obtain a transformed graph coordinate; acquiring and calculating one-dimensional characteristic information of the image according to a preset method; and the banknote image is matched with the one-dimensional characteristic information of the standard image, whether the matched characteristic is abnormal or not is judged, and the problem of denomination recognition of one currency can be solved quickly, but when the banknote image is deformed and the time consumption requirement on an algorithm is extremely strict, the processing effect is poor.
Therefore, a detection algorithm which has higher identification accuracy, stronger universality and expansibility, can adapt to various banknote deformation and has higher detection time consumption requirement is provided for banknote denomination identification, and the problem to be solved is urgently needed.
Disclosure of Invention
The invention provides a banknote denomination recognition method and a banknote denomination recognition system, which are used for solving the defects of low recognition accuracy, low operation speed and poor expansibility commonly existing in the banknote denomination recognition in the prior art.
In a first aspect, the present invention provides a banknote denomination identification method comprising:
acquiring a bank note to be recognized, and extracting frequency domain characteristics of the bank note to be recognized;
acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set;
and determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
According to the banknote denomination recognition method provided by the invention, the steps of obtaining the banknote to be recognized and extracting the frequency domain characteristics of the banknote to be recognized comprise the following steps:
determining any front preset visible light channel image and any back preset visible light channel image of the bank note to be identified;
respectively carrying out linear planar affine transformation on any front preset visible light channel image and any back preset visible light channel image to obtain a front affine image and a back affine image with the same preset breadth size;
respectively extracting discrete cosine transform characteristics of the front affine image and the back affine image, and intercepting a characteristic array of a preset size area range from the discrete cosine transform characteristics to obtain a front characteristic array and a back characteristic array;
and connecting the front face characteristic array and the back face characteristic array to obtain the frequency domain characteristics.
According to the banknote denomination recognition method provided by the invention, the front face feature array and the back face feature array are connected to obtain the frequency domain features, and then the method further comprises the following steps:
and if the frequency domain features have abnormal data length, performing feature data filling on the frequency domain features based on a square matrix data form.
According to the banknote denomination identification method provided by the invention, any front preset visible light channel image and any back preset visible light channel image are the same visible light channel image.
According to the banknote denomination recognition method provided by the invention, the operator for acquiring the banknote classification features comprises the following steps:
acquiring a banknote classification sample image, and extracting frequency domain characteristics of the banknote classification sample image;
and performing multi-classification training on the frequency domain characteristics of the banknote classification sample images based on a Support Vector Machine (SVM) to obtain the banknote classification characteristic operator.
According to the banknote denomination recognition method provided by the invention, the length of the banknote classification characteristic operator is the frequency domain characteristic length of the banknote classification quantity multiplied by the image to be recognized.
According to the banknote denomination recognition method provided by the invention, the frequency domain feature of the banknote to be recognized and the banknote classification feature operator are subjected to preset operation to obtain a banknote category score set, and the method comprises the following steps:
and performing multiply-add operation on the frequency domain characteristics of the bank notes to be identified and the bank note classification characteristic operator to obtain a plurality of bank note category scores, and taking the plurality of bank note category scores as the bank note category score set.
In a second aspect, the present invention also provides a banknote denomination recognition system comprising:
the extraction module is used for acquiring the bank notes to be identified and extracting the frequency domain characteristics of the bank notes to be identified;
the calculation module is used for acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be identified and the banknote classification characteristic operator to obtain a banknote classification score set;
and the recognition module is used for determining that the banknote corresponding to the category with the highest score in the banknote category score set is classified as the recognition result of the banknote to be recognized.
In a third aspect, the present invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and operable on the processor, the processor when executing the program implementing the steps of the banknote denomination recognition method as described in any one of the above.
In a fourth aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, carries out the steps of the banknote denomination recognition method as described in any one of the above.
In a fifth aspect, the present invention also provides a computer program product comprising a computer program which, when executed by a processor, carries out the steps of the banknote denomination recognition method as claimed in any one of the above.
The banknote denomination recognition method and the banknote denomination recognition system provided by the invention have the characteristics of strong universality and high execution efficiency by adopting a machine learning method to recognize the denomination of the banknote, and particularly have a better recognition effect on deformed and stained banknotes.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a banknote denomination identification method provided by the present invention;
FIG. 2 is a schematic diagram of the implementation principle of the banknote denomination recognition method provided by the invention;
FIG. 3 is a schematic diagram of the configuration of a banknote denomination recognition system provided by the present invention;
fig. 4 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Aiming at the defects in the prior art, the invention provides a banknote denomination recognition method based on machine learning, fig. 1 is a schematic flow chart of the banknote denomination recognition method provided by the invention, and as shown in fig. 1, the method comprises the following steps:
step S1, acquiring the bank note to be recognized, and extracting the frequency domain characteristics of the bank note to be recognized;
firstly, the type of the bank note to be identified, such as RMB, dollar or Euro, is determined, and the frequency domain characteristics of the bank note are extracted through image transformation, and the frequency domain characteristics are generally extracted by adopting discrete cosine transformation characteristics.
Step S2, acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set;
before identifying the bank notes to be identified, a large number of bank note images need to be collected, the bank note images are classified according to different denomination version directions, then a visible light channel is selected, frequency domain features of the visible light channel are extracted, and then a Machine learning method SVM (Support Vector Machine) is used for training to obtain feature operators, wherein in the implementation principle diagram of FIG. 2, the feature operators are all completed in an upper computer in advance. Here, the method of extracting the frequency domain feature for the banknote sample image is the same as the method of extracting the frequency domain feature of the banknote to be recognized in step S1, except that the feature extraction and the operation of the banknote to be recognized are both completed at the financial machine side, that is, a DSP (Digital Signal processing) part.
And further carrying out preset operation on the frequency domain characteristics of the channel images corresponding to the banknotes to be recognized and the characteristic operators to obtain the score of each banknote classification, namely the banknote classification score set.
And step S3, determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
And taking the denomination version corresponding to the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
It should be noted that the banknote image features extracted by the present invention are not limited to frequency domain features, and may also be features of other domains.
The banknote denomination recognition method based on machine learning provided by the invention has the characteristics of high operation speed, strong compatibility, high recognition accuracy and stable operation; the invention solves the problems that the existing method can not be widely adapted to different bank notes and image quality, and also solves the problems of unstable detection and much time consumption; in the process of identifying the denomination of the bank note, only visible light images of a single channel are acquired, then frequency domain features are extracted, multiplication and addition operation is carried out on the frequency domain features and trained feature operators, classification results can be obtained, training and learning of the feature operators are completed on an upper computer, and the diversity of samples does not influence time consumption, but can improve accuracy.
Based on the above embodiment, the obtaining of the banknote to be recognized and the extracting of the frequency domain features of the banknote to be recognized include:
determining any front preset visible light channel image and any back preset visible light channel image of the bank note to be identified;
respectively carrying out linear planar affine transformation on any front preset visible light channel image and any back preset visible light channel image to obtain a front affine image and a back affine image with the same preset breadth size;
respectively extracting discrete cosine transform characteristics of the front affine image and the back affine image, and intercepting a characteristic array of a preset size area range from the discrete cosine transform characteristics to obtain a front characteristic array and a back characteristic array;
and connecting the front face characteristic array and the back face characteristic array to obtain the frequency domain characteristics.
Wherein, connect the front side characteristic array with the back side characteristic array, obtain the frequency domain characteristic, later still include:
and if the frequency domain features have abnormal data length, performing feature data filling on the frequency domain features based on a square matrix data form.
And the image of any front preset visible light channel and the image of any back preset visible light channel are the same visible light channel image.
Specifically, for the bank note to be recognized, the currency is selected as the renminbi, any visible light image of the front image of the bank note is selected at first, a green channel of a foreground image is taken as an example generally, and because the invention has higher requirement on the recognition operation efficiency, linear planar affine transformation of the image needs to be carried out first, and the foreground image is transformed to a proper breadth (the length is set as M and the width is set as N) which is small enough.
Here, the affine transformation is a linear change between two-dimensional coordinates, by which "straightness" of a two-dimensional image can be maintained, that is, straight lines remain straight lines after transformation, and "parallelism" can be maintained, that is, relative positional relationship between two-dimensional patterns remains unchanged, parallel lines remain parallel lines, and the positional order of points on the straight lines does not change, and a general expression of the radial transformation is:
then, extracting frequency domain features on an affine graph with the size of M × N obtained after affine transformation, namely, the frequency domain features obtained by discrete cosine transformation, where the frequency domain transformation is used for compressing data or an image, a signal in a space domain can be converted onto a frequency domain, and the performance of good decorrelation is achieved, and a general expression of the discrete cosine transformation is as follows:
after frequency domain transformation, for a processing object with a large correlation, such as an image, the coefficients are largely concentrated in the upper left corner, the lower right corner is almost all 0, the upper left corner is a low-frequency component, the lower right corner is a high-frequency component, the high-frequency coefficients represent detail information of a target shape, and the low-frequency coefficients represent target contour and gray distribution characteristics in the image, that is, the energy of the image is mainly concentrated in the low-frequency component, that is, key information required when performing denomination recognition by using a small image after affine transformation. Therefore, the invention intercepts the upper left corner area with the total size of T (the length is p and the width is q) from the image after the frequency domain change, and obtains the characteristic array of the image of a certain side of the banknote.
It can be understood that the length of the front face feature array X is T after the front face image of the banknote is obtained and subjected to frequency domain transformation1，X2，X3…XTThen, the same processing steps are carried out on the green channel image of the back image to obtain a back characteristic array Y with the length T1，Y2，Y3…YTConnecting the front feature array and the back feature array to obtain an array X with the size of 2X T1，X2，X3…XT，Y1，Y2，Y3…YTThe array of sizes 2 × T is as follows:
here, considering the case that the two-dimensional frequency domain transform is directed to a generally square matrix, if the data length is abnormal after the transform, the actual image may be filled to meet the requirement of the square matrix form.
In the invention, any visible light channel image on the front surface of the banknote and any visible light channel image on the back surface of the banknote are the same type of visible light channel, such as a green channel image, otherwise if different visible light channel images are extracted, the obtained data characteristics corresponding to the front surface characteristic array and the back surface characteristic array are different, so that the characteristic arrays with the length of 2T cannot be formed in a connected mode.
According to the method, the frequency domain characteristics of the affine image of the visible light channel are extracted, so that the accuracy and stability of subsequent banknote identification are improved conveniently.
Based on any one of the above embodiments, the banknote classification feature obtaining operator includes:
acquiring a banknote classification sample image, and extracting frequency domain characteristics of the banknote classification sample image;
and performing multi-classification training on the frequency domain characteristics of the banknote classification sample images based on a Support Vector Machine (SVM) to obtain the banknote classification characteristic operator.
And the length of the banknote classification characteristic operator is the frequency domain characteristic length of the image to be identified multiplied by the banknote category number.
Specifically, in the upper computer part, a large number of classified banknote samples are obtained in advance, the processing method for extracting the frequency domain features of the banknote samples is consistent with the frequency domain feature processing method for extracting the banknotes to be recognized in the previous embodiment, and the upper computer part needs to perform multi-classification training and learning on the frequency domain features extracted from the large number of classified samples in an SVM mode, so that the feature operator is obtained.
If there are a total of nClass classes, the total feature operator size is 2 xT nClass, taking into account that the length of each frequency domain feature is 2 xT.
According to the invention, a large number of banknote sample images are obtained at the upper computer part for training to obtain the characteristic operators, and the accuracy and efficiency of sample classification and identification can be improved through the diversity training of a large number of samples
Based on any one of the above embodiments, the performing a preset operation on the frequency domain feature of the banknote to be recognized and the banknote classification feature operator to obtain a banknote category score set includes:
and performing multiply-add operation on the frequency domain characteristics of the bank notes to be identified and the bank note classification characteristic operator to obtain a plurality of bank note category scores, and taking the plurality of bank note category scores as the bank note category score set.
Specifically, after frequency domain features of the bank notes to be recognized are extracted at one side of the financial machine tool, namely the lower computer part, the frequency domain features and a total feature operator with the size of 2 × T × nClass are subjected to multiply-add calculation to obtain scores of each category, and the bank note classification information corresponding to the category with the largest score is used as the recognition result of the bank notes to be recognized.
The invention adopts a machine learning method to identify the denomination of the bank note, has strong universality and high algorithm efficiency, can learn iteration and has strong expandability.
In the following, the banknote denomination recognition system provided by the present invention is described, and the banknote denomination recognition system described below and the banknote denomination recognition method described above can be referred to in correspondence with each other.
Fig. 3 is a schematic structural view of a banknote denomination recognition system provided by the present invention, as shown in fig. 3, including: an extraction module 31, a calculation module 32 and a recognition module 33, wherein:
the extraction module 31 is used for acquiring the bank notes to be identified and extracting the frequency domain characteristics of the bank notes to be identified; the calculation module 32 is configured to obtain a banknote classification feature operator, perform preset operation on the frequency domain feature of the banknote to be identified and the banknote classification feature operator, and obtain a banknote classification score set; the recognition module 33 is configured to determine that the banknote corresponding to the category with the largest score in the banknote category score set is classified as the recognition result of the banknote to be recognized.
The invention identifies the denomination of the bank note by adopting a machine learning method, has the characteristics of strong universality and high execution efficiency, and particularly has better identification effect on the bank note which is deformed and stained.
Based on the above embodiment, the extraction module 31 includes: a determination sub-module 311, a transformation sub-module 312, an extraction sub-module 313, and a connection sub-module 314, wherein:
the determining submodule 311 is configured to determine any front preset visible light channel image and any back preset visible light channel image of the banknote to be identified; the transformation submodule 312 is configured to perform linear planar affine transformation on each of the front-side preset visible light channel images and the back-side preset visible light channel images to obtain front-side affine images and back-side affine images having the same preset breadth; the extraction submodule 313 is configured to extract discrete cosine transform features of the front affine image and the back affine image respectively, and intercept a feature array of a preset size area range from the discrete cosine transform features to obtain a front feature array and a back feature array; the connection sub-module 314 is configured to connect the front side feature array and the back side feature array to obtain the frequency domain features.
Based on any of the above embodiments, the extraction module 31 further includes a padding sub-module 315, where the padding sub-module 315 is configured to perform feature data padding on the frequency domain features based on a square matrix data form if the frequency domain features have data length abnormality.
Based on any of the above embodiments, the any front-side preset visible light channel image and the any back-side preset visible light channel image in the determination sub-module 311 are the same visible light channel image.
Based on any of the above embodiments, the calculation module 32 includes an obtaining sub-module 321 and a calculation sub-module 322, where:
the obtaining submodule 321 is configured to obtain a banknote classification sample image, and extract frequency domain features of the banknote classification sample image; and performing multi-classification training on the frequency domain characteristics of the banknote classification sample images based on a Support Vector Machine (SVM) to obtain the banknote classification characteristic operator.
And the length of the banknote classification characteristic operator is the frequency domain characteristic length of the image to be identified multiplied by the banknote category number.
The calculating submodule 322 is configured to perform multiply-add operation on the frequency domain feature of the banknote to be identified and the banknote classification feature operator to obtain a plurality of banknote category scores, and use the plurality of banknote category scores as the banknote category score set.
The invention adopts a machine learning method to identify the denomination of the bank note, has strong universality and high algorithm efficiency, can learn iteration and has strong expandability.
Fig. 4 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 4: a processor (processor)410, a communication Interface 420, a memory (memory)430 and a communication bus 440, wherein the processor 410, the communication Interface 420 and the memory 430 are communicated with each other via the communication bus 440. The processor 410 may invoke logic instructions in the memory 430 to perform a method of banknote denomination recognition, the method comprising: acquiring a bank note to be recognized, and extracting frequency domain characteristics of the bank note to be recognized; acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set; and determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
In addition, the logic instructions in the memory 430 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program storable on a non-transitory computer readable storage medium, the computer program, when executed by a processor, being capable of executing the banknote denomination recognition method provided by the above-mentioned methods, the method comprising: acquiring a bank note to be recognized, and extracting frequency domain characteristics of the bank note to be recognized; acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set; and determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method of banknote denomination identification provided by the above methods, the method comprising: acquiring a bank note to be recognized, and extracting frequency domain characteristics of the bank note to be recognized; acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set; and determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A banknote denomination identification method, comprising:
acquiring a bank note to be recognized, and extracting frequency domain characteristics of the bank note to be recognized;
acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be recognized and the banknote classification characteristic operator to obtain a banknote classification score set;
and determining the banknote corresponding to the category with the highest score in the banknote category score set as the recognition result of the banknote to be recognized.
2. The banknote denomination recognition method according to claim 1, wherein the obtaining of the banknote to be recognized and the extraction of the frequency domain features of the banknote to be recognized comprises:
determining any front preset visible light channel image and any back preset visible light channel image of the bank note to be identified;
respectively carrying out linear planar affine transformation on any front preset visible light channel image and any back preset visible light channel image to obtain a front affine image and a back affine image with the same preset breadth size;
respectively extracting discrete cosine transform characteristics of the front affine image and the back affine image, and intercepting a characteristic array of a preset size area range from the discrete cosine transform characteristics to obtain a front characteristic array and a back characteristic array;
and connecting the front face characteristic array and the back face characteristic array to obtain the frequency domain characteristics.
3. The banknote denomination recognition method of claim 2, wherein the front face feature array and the back face feature array are concatenated to obtain the frequency domain features, and thereafter further comprising:
and if the frequency domain features have abnormal data length, performing feature data filling on the frequency domain features based on a square matrix data form.
4. A banknote denomination identification method according to claim 2 or 3, wherein any one of the front side preset visible light channel images and any one of the back side preset visible light channel images are the same visible light channel image.
5. The banknote denomination recognition method of claim 1, wherein the obtaining of the banknote classification characteristic operator comprises:
acquiring a banknote classification sample image, and extracting frequency domain characteristics of the banknote classification sample image;
and performing multi-classification training on the frequency domain characteristics of the banknote classification sample images based on a Support Vector Machine (SVM) to obtain the banknote classification characteristic operator.
6. The banknote denomination recognition method of claim 5, wherein the length of the banknote classification feature operator is the number of banknote categories multiplied by the frequency domain feature length of the image to be recognized.
7. The banknote denomination recognition method according to claim 1, wherein the step of performing a preset operation on the frequency domain features of the banknote to be recognized and the banknote classification feature operator to obtain a banknote category score set comprises:
and performing multiply-add operation on the frequency domain characteristics of the bank notes to be identified and the bank note classification characteristic operator to obtain a plurality of bank note category scores, and taking the plurality of bank note category scores as the bank note category score set.
8. A banknote denomination recognition system, comprising:
the extraction module is used for acquiring the bank notes to be identified and extracting the frequency domain characteristics of the bank notes to be identified;
the calculation module is used for acquiring a banknote classification characteristic operator, and performing preset operation on the frequency domain characteristic of the banknote to be identified and the banknote classification characteristic operator to obtain a banknote classification score set;
and the recognition module is used for determining that the banknote corresponding to the category with the highest score in the banknote category score set is classified as the recognition result of the banknote to be recognized.
9. An electronic device comprising a memory, a processor and a computer program stored on said memory and executable on said processor, characterized in that said processor, when executing said program, carries out the steps of the banknote denomination recognition method according to any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the banknote denomination recognition method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111658954.8A CN114419644A (en) | 2021-12-30 | 2021-12-30 | Banknote denomination recognition method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111658954.8A CN114419644A (en) | 2021-12-30 | 2021-12-30 | Banknote denomination recognition method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114419644A true CN114419644A (en) | 2022-04-29 |
Family
ID=81269852
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111658954.8A Pending CN114419644A (en) | 2021-12-30 | 2021-12-30 | Banknote denomination recognition method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114419644A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050271265A1 (en) * | 2004-06-04 | 2005-12-08 | Xerox Corporation | Method and system for classifying scanned-media |
| US20100027969A1 (en) * | 2002-06-10 | 2010-02-04 | Alattar Adnan M | Identification and protection of video |
| CN105118048A (en) * | 2015-07-17 | 2015-12-02 | 北京旷视科技有限公司 | Method and device for identifying copying certificate image |
| CN105184954A (en) * | 2015-08-14 | 2015-12-23 | 深圳怡化电脑股份有限公司 | Banknote detecting method and banknote detecting device |
| CN105427446A (en) * | 2015-11-06 | 2016-03-23 | 东方通信股份有限公司 | Authentic identification method for paper money based on magnetic signal |
| US20170206671A1 (en) * | 2014-07-22 | 2017-07-20 | Hewlett-Packard Development Company, L.P. | Rapid Image Registration |
| CN109727362A (en) * | 2018-11-15 | 2019-05-07 | 恒银金融科技股份有限公司 | Banknote magnetic signal recognition methods based on discrete Fourier transform |
| CN111738079A (en) * | 2020-05-19 | 2020-10-02 | 武汉卓目科技有限公司 | Banknote denomination recognition method and device |
| US20210150768A1 (en) * | 2019-11-14 | 2021-05-20 | Alibaba Group Holding Limited | Techniques to dynamically gate encoded image components for artificial intelligence tasks |
-
2021
- 2021-12-30 CN CN202111658954.8A patent/CN114419644A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100027969A1 (en) * | 2002-06-10 | 2010-02-04 | Alattar Adnan M | Identification and protection of video |
| US20050271265A1 (en) * | 2004-06-04 | 2005-12-08 | Xerox Corporation | Method and system for classifying scanned-media |
| US20170206671A1 (en) * | 2014-07-22 | 2017-07-20 | Hewlett-Packard Development Company, L.P. | Rapid Image Registration |
| CN105118048A (en) * | 2015-07-17 | 2015-12-02 | 北京旷视科技有限公司 | Method and device for identifying copying certificate image |
| CN105184954A (en) * | 2015-08-14 | 2015-12-23 | 深圳怡化电脑股份有限公司 | Banknote detecting method and banknote detecting device |
| CN105427446A (en) * | 2015-11-06 | 2016-03-23 | 东方通信股份有限公司 | Authentic identification method for paper money based on magnetic signal |
| CN109727362A (en) * | 2018-11-15 | 2019-05-07 | 恒银金融科技股份有限公司 | Banknote magnetic signal recognition methods based on discrete Fourier transform |
| US20210150768A1 (en) * | 2019-11-14 | 2021-05-20 | Alibaba Group Holding Limited | Techniques to dynamically gate encoded image components for artificial intelligence tasks |
| CN111738079A (en) * | 2020-05-19 | 2020-10-02 | 武汉卓目科技有限公司 | Banknote denomination recognition method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | A reliable method for paper currency recognition based on LBP | |
| AU2011288069B2 (en) | Valuable file identification method and identification system, device thereof | |
| CN105894656A (en) | Banknote image recognition method | |
| CN106952393B (en) | Paper money identification method and device, electronic equipment and storage medium | |
| CN107103683B (en) | Paper money identification method and device, electronic equipment and storage medium | |
| CN104969268A (en) | Authentication of security documents and mobile device to carry out the authentication | |
| CN102722708A (en) | Method and device for classifying sheet media | |
| CN103886309A (en) | Method for identifying dollar denominations through facial recognition | |
| CN111738079A (en) | Banknote denomination recognition method and device | |
| CN108665603B (en) | Method and device for identifying currency type of paper money and electronic equipment | |
| Sawant et al. | Currency recognition using image processing and minimum distance classifier technique | |
| Gautam | Indian currency detection using image recognition technique | |
| CN106898078B (en) | Port currency version identification method and device | |
| CN108460775A (en) | A kind of forge or true or paper money recognition methods and device | |
| CN201465141U (en) | Device for distinguishing and recording serial member of RMB (Renminbi) | |
| CN114419644A (en) | Banknote denomination recognition method and system | |
| Rajan et al. | An extensive study on currency recognition system using image processing | |
| CN112396638A (en) | Image processing method, terminal and computer readable storage medium | |
| Sooruth et al. | Automatic South African Coin Recognition Through Visual Template Matching | |
| CN111275879B (en) | Currency identification method and device and mobile terminal | |
| Muhamad et al. | Image-Based Processing of Paper Currency Recognition and Fake Identification: A Review | |
| CN107958531B (en) | Paper money identification method and device and equipment | |
| CN106652168B (en) | Method and device for identifying currency value of paper money | |
| Vishnu et al. | Currency detection using similarity indices method | |
| CN108537945B (en) | Bill watermark detection method and system and self-service equipment |
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 | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: 430073 No. 006, 20th floor, business project (China Pharmaceutical Technology Trading Market), No. 1, xiyaojian Road, north of Gaoxin Avenue and Heying Road, East Lake New Technology Development Zone, Wuhan City, Hubei Province Applicant after: Wuhan Zhuomu Technology Co.,Ltd. Address before: 430073 No. 006, 20th floor, business project (China Pharmaceutical Technology Trading Market), No. 1, xiyaojian Road, north of Gaoxin Avenue and Heying Road, East Lake New Technology Development Zone, Wuhan City, Hubei Province Applicant before: WUHAN ZMVISION TECHNOLOGY Co.,Ltd. Country or region before: China |