Banknote sorting method and system, storage medium and program product
Through the combination of acoustic excitation and image acquisition, the pre-trained model is used to identify and sort banknote types, which solves the secondary damage and low accuracy problems caused by contact detection in the prior art, and achieves contactless and efficient and accurate banknote clearing.
Patent Information
- Application Number
- CN202510720214.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-30
- Publication Date
- 2025-08-08
AI Technical Summary
Existing banknote clearing methods usually require contact with banknotes, which may cause secondary damage and insufficient ability to determine local minor defects, resulting in low detection accuracy.
The transmitted sound wave and reflected sound wave are obtained through the acoustic wave excitation detection device, and the surface image is acquired in combination with the image acquisition device, and input it to the pre-trained banknote classification model for banknote type identification, and finally the contactless sorting is realized through the sorting execution device.
The contactless banknote detection is realized, which improves the accuracy and efficiency of banknote clearance and avoids secondary damage to banknotes.
Smart Images
Figure CN120452095A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a banknote sorting method, system, storage medium and program product. Background Art
[0002] During the circulation process, banknotes are prone to stains, damage and other problems. Therefore, when sorting banknotes, they need to be inspected and classified to determine whether there are any problems with them.
[0003] At present, the existing banknote sorting methods usually use physical parameter sensors such as weight, thickness, and transmittance to detect banknotes. Although this method can obtain some physical properties of banknotes, it often requires contact with the banknotes, which may cause secondary damage, and the ability to judge local minor defects is insufficient, which easily leads to low detection accuracy. Summary of the Invention
[0004] The present invention provides a banknote sorting method, system, storage medium and program product, which can realize contactless detection of banknotes and improve the accuracy of banknote sorting.
[0005] According to one aspect of the present invention, there is provided a banknote sorting method, comprising:
[0006] Acquire the transmitted and reflected sound waves corresponding to the banknote to be inspected through the sound wave excitation detection device, and acquire the surface image corresponding to the banknote to be inspected through the image acquisition device;
[0007] Inputting the transmitted sound wave, the reflected sound wave, and the surface image into a pre-trained banknote classification model, and obtaining the banknote type corresponding to the banknote to be inspected output by the banknote classification model;
[0008] The banknotes to be inspected are sorted according to the types of the banknotes through a sorting execution device.
[0009] According to another aspect of the present invention, there is provided a banknote sorting device, comprising:
[0010] The acoustic wave acquisition module is used to acquire the transmitted acoustic wave and the reflected acoustic wave corresponding to the banknote to be inspected through the acoustic wave excitation detection device, and to acquire the surface image corresponding to the banknote to be inspected through the image acquisition device;
[0011] a banknote type acquisition module, configured to input the transmitted sound wave, the reflected sound wave, and the surface image into a pre-trained banknote classification model, and acquire the banknote type corresponding to the banknote to be inspected output by the banknote classification model;
[0012] The banknote sorting module is used to sort the banknotes to be inspected according to the types of the banknotes through a sorting execution device.
[0013] According to another aspect of the present invention, a banknote sorting system is provided, comprising an acoustic wave excitation detection device, an image acquisition device, a sorting execution device and a processor; wherein,
[0014] The acoustic wave excitation detection device is in communication with the processor, and is used to obtain the transmitted acoustic wave and the reflected acoustic wave corresponding to the banknote to be detected, and send the transmitted acoustic wave and the reflected acoustic wave to the processor;
[0015] The image acquisition device is in communication with the processor, and is used to acquire a surface image corresponding to the banknote to be inspected, and send the surface image to the processor;
[0016] The processor is in communication with the sorting execution device, and is configured to execute the banknote sorting method according to any embodiment of the present invention, and send a sorting control instruction to the sorting execution device;
[0017] The sorting execution device is used to implement sorting of the banknotes to be inspected according to the sorting control instructions.
[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and the computer program is used to enable a processor to implement the banknote sorting method described in any embodiment of the present invention when executed.
[0019] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the banknote sorting method according to any embodiment of the present invention.
[0020] The technical solution of the embodiment of the present invention is to obtain the transmitted sound waves and reflected sound waves corresponding to the banknotes to be inspected through an acoustic wave excitation detection device, and obtain the surface image corresponding to the banknotes to be inspected through an image acquisition device; input the transmitted sound waves, reflected sound waves and surface images into a pre-trained banknote classification model, and obtain the banknote type corresponding to the banknotes to be inspected output by the banknote classification model; through a sorting execution device, sorting of the banknotes to be inspected is realized according to the banknote type; by comprehensively evaluating the banknote type through the transmitted sound waves, reflected sound waves and surface images, contactless detection of banknotes can be realized, and the accuracy of banknote sorting can be improved.
[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0023] Figure 1 This is a flow chart of a banknote sorting method provided according to the first embodiment of the present invention;
[0024] Figure 2 This is a flow chart of a banknote sorting method provided according to the second embodiment of the present invention;
[0025] Figure 3 This is a flow chart of another banknote sorting method provided according to the second embodiment of the present invention;
[0026] Figure 4 This is a schematic structural diagram of a banknote sorting device provided according to a third embodiment of the present invention;
[0027] Figure 5 It is a structural diagram of a banknote sorting system provided according to the fourth embodiment of the present invention. DETAILED DESCRIPTION
[0028] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.
[0029] It should be noted that the terms "first," "second," "target," etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products, or apparatus.
[0030] Example 1
[0031] Figure 1A flow chart of a banknote sorting method is provided for the first embodiment of the present invention. This embodiment is applicable to the situation where banknotes are evaluated and classified for defects. The method can be executed by a banknote sorting device, which can be implemented in the form of hardware and / or software. Typically, the banknote sorting device can be configured in the banknote sorting system provided in the fourth embodiment of the present invention. Figure 1 As shown, the method includes:
[0032] S110 , obtaining the transmitted sound wave and the reflected sound wave corresponding to the banknote to be inspected through the sound wave excitation detection device, and obtaining the surface image corresponding to the banknote to be inspected through the image acquisition device.
[0033] Among them, the acoustic wave excitation detection device may include a transmitting unit and a receiving unit. The transmitting unit includes a signal generator, a power amplifier and a focused ultrasonic transducer. The signal generator is used to generate a specific waveform, such as a pulse or a frequency-modulated continuous wave. The focused ultrasonic transducer can be a piezoelectric ceramic transducer with a center frequency in the range of 1-10 MHz, which can be optimized according to the material of the banknote and the type of damage to be detected. The transducer transmits a focused acoustic beam to the banknote in a non-contact manner (for example, 5-20 mm from the surface of the banknote, etc.). Typically, a single or array transducer can be used to scan or cover key areas of the banknote.
[0034] The receiving unit contains one or more highly sensitive ultrasonic transducers, paired with or independently of the transmitting unit, to synchronously or sequentially capture sound waves transmitted through the banknote (transmitted sound waves) and echoes reflected from the banknote's surface or internal layers (reflected sound waves). Transmitted sound waves primarily reveal the banknote's overall density and uniformity, as well as the presence of penetrating defects (such as holes). Reflected sound waves (especially when collected at multiple points or in a scanning manner) can more precisely reveal subtle structural anomalies near the banknote's surface and interior, such as internal tears, delamination, material fatigue, and damage from attachments like tape.
[0035] The image acquisition device can be a high-resolution linear or area array charge coupled device (CCD) / complementary metal oxide semiconductor (CMOS) camera equipped with a uniform, stable light source, such as a light-emitting diode (LED) bar light source, a ring light source, or a coaxial light source. For example, the image acquisition device can be installed in a closed darkroom and can capture images of the front and back of a banknote separately or simultaneously.
[0036] In a specific example, a banknote transport device can stably transport the banknotes to be inspected at a predetermined speed (e.g., 5-15 banknotes / second) into the inspection area and to the sorting port. For example, the banknote transport device can be a high-speed conveyor belt or roller system. Within the inspection area, an acoustic wave excitation detection device can capture transmitted and reflected acoustic waves corresponding to the banknotes to be inspected, and an image capture device can capture a surface image of the banknotes to be inspected.
[0037] S120: Input the transmitted sound wave, the reflected sound wave, and the surface image into a pre-trained banknote classification model, and obtain the banknote type corresponding to the banknote to be inspected output by the banknote classification model.
[0038] The input of the banknote classification model can be the transmitted sound wave, reflected sound wave, and surface image corresponding to the banknote to be inspected, and the output can be the banknote type and the corresponding confidence level. In this embodiment, an initial banknote classification model can be established based on a convolutional neural network algorithm or a recurrent neural network algorithm and predefined initial model parameters. Then, training samples can be screened from a pre-established damaged banknote feature library. The training samples can then be used to train, optimize, and verify the initial banknote classification model, adjusting the model parameters of the initial banknote classification model until a banknote classification model that meets preset training termination conditions (e.g., a loss function value less than a preset threshold, the number of iterations reaching a preset threshold, etc.) is obtained.
[0039] The damaged coin feature library can be a structured database containing a large number of representative damaged coin samples. These samples include high-resolution image data, corresponding acoustic response data (raw signals or extracted features), manual or semi-automatic annotation of damage type (such as holes, missing corners, tears, tape repairs, stains, wrinkles, internal damage, etc.) and severity, and the final classification label (such as circulated, uncirculated and destroyed, requiring manual review, etc.). This damaged coin feature library can be used to train, validate, and optimize banknote classification models.
[0040] Optionally, in this embodiment, model compression and distillation technology can be used to process the trained banknote classification model to obtain a lightweight model suitable for edge devices, and the lightweight model can be deployed to high-performance edge computing devices to achieve real-time reasoning.
[0041] Optionally, the banknote type may include circulated, pending review, and / or recommended for destruction. In one optional example, first, a banknote classification model can be used to determine the damage type (e.g., surface stains, penetrating defects, etc.) and damage degree corresponding to the banknote to be inspected based on the transmitted sound wave, reflected sound wave, and surface image. Then, the banknote type corresponding to the banknote to be inspected can be determined based on the current damage type and damage degree, as well as a preset mapping relationship between damage type, damage degree range, and banknote type.
[0042] In this embodiment, by setting multi-dimensional banknote types, it is possible to achieve fine classification of banknotes and improve the precision of banknote sorting.
[0043] S130 , using a sorting execution device, sorting the banknotes to be inspected according to the banknote types.
[0044] The sorting execution device can be a fast-response pneumatic or electric paddle, channel switch, or other mechanical device. Specifically, after determining the banknote type of the banknote to be inspected, a corresponding control instruction can be generated and sent to the sorting execution device to control the sorting execution device to sort and collect the banknotes according to their type.
[0045] Optionally, sorting the banknotes to be inspected according to the banknote types by a sorting execution device may include:
[0046] According to the banknote type and the preset mapping relationship between the banknote type and the collection slot, obtaining the collection slot corresponding to the banknote to be inspected;
[0047] The banknotes to be inspected are sorted into corresponding collecting slots by a sorting execution device.
[0048] In one optional example, a unique collection slot can be pre-assigned for each banknote type. For example, if the banknote type is circulated, the corresponding collection slot is the circulation slot; if the banknote type is pending review, the corresponding collection slot is the review slot; and if the banknote type is recommended for destruction, the corresponding collection slot is the destruction slot. Thus, when sorting banknotes, the collection slot corresponding to the current banknote type can be first determined. Based on this collection slot, a corresponding control instruction can be generated and sent to the sorting execution device. Upon receiving this control instruction, the sorting execution device can execute the control instruction and sort the current banknote to be inspected into the specified collection slot.
[0049] In this embodiment, by automatically sorting the banknotes to be inspected into corresponding collection slots according to the banknote types, automatic classification and sorting of the banknotes can be achieved, which can improve the efficiency of banknote sorting.
[0050] The technical solution of the embodiment of the present invention is to obtain the transmitted sound waves and reflected sound waves corresponding to the banknotes to be inspected through an acoustic wave excitation detection device, and obtain the surface image corresponding to the banknotes to be inspected through an image acquisition device; input the transmitted sound waves, reflected sound waves and surface images into a pre-trained banknote classification model, and obtain the banknote type corresponding to the banknotes to be inspected output by the banknote classification model; through a sorting execution device, sorting of the banknotes to be inspected is realized according to the banknote type; by comprehensively evaluating the banknote type through the transmitted sound waves, reflected sound waves and surface images, contactless detection of banknotes can be realized, and the accuracy of banknote sorting can be improved.
[0051] Example 2
[0052] Figure 2 This is a flow chart of a banknote sorting method provided in Example 2 of the present invention. This embodiment is a further refinement of the above technical solution. The technical solution in this embodiment can be combined with one or more of the above implementations. Figure 2 As shown, the method includes:
[0053] S210 , obtaining the transmitted sound wave and the reflected sound wave corresponding to the banknote to be inspected through the sound wave excitation detection device, and obtaining the surface image corresponding to the banknote to be inspected through the image acquisition device.
[0054] S220 , obtaining a transmitted sound wave feature vector corresponding to the transmitted sound wave and a reflected sound wave feature vector corresponding to the reflected sound wave, and obtaining a composite sound wave feature vector based on the transmitted sound wave feature vector and the reflected sound wave feature vector.
[0055] Specifically, first, the transmitted and reflected sound waves can be preprocessed to obtain preprocessed transmitted and reflected sound waves; for example, the preprocessing method can include filtering, amplification, analog-to-digital conversion, etc. Then, corresponding characteristic parameters can be extracted from the preprocessed transmitted and reflected sound waves, and the extracted characteristic parameters can be combined to obtain a transmitted and reflected sound wave feature vector. Finally, the transmitted and reflected sound wave feature vectors can be combined to obtain a composite sound wave feature vector.
[0056] Optionally, different characteristic parameters can be set for the transmitted sound waves and the reflected sound waves. For example, the characteristic parameters corresponding to the transmitted sound waves may include attenuation coefficient, sound speed change, etc., and the characteristic parameters corresponding to the reflected sound waves may include energy, phase, amplitude of specific frequency components, echo delay time, etc.
[0057] Optionally, obtaining a composite sound wave feature vector according to the transmitted sound wave feature vector and the reflected sound wave feature vector may include:
[0058] The transmitted sound wave feature vector and the reflected sound wave feature vector are integrated and weighted to obtain a composite sound wave feature vector.
[0059] In an optional example, when combining the transmitted sound wave eigenvector and the reflected sound wave eigenvector, first, the transmitted sound wave eigenvector and the reflected sound wave eigenvector can be concatenated to obtain an initial composite sound wave eigenvector; then, the weight value corresponding to each eigenvalue can be determined based on the characteristic parameters corresponding to each eigenvalue in the initial composite sound wave eigenvector, and the mapping relationship between the preset characteristic parameters and the weight value; finally, each eigenvalue can be multiplied by the corresponding weight value to obtain the final composite sound wave eigenvector. The weight value corresponding to each characteristic parameter can be equal or unequal.
[0060] In this embodiment, by integrating and weighting the characteristic parameters from the transmission and reflection modes, a more complete and rich composite acoustic wave characteristic vector is formed, which can more comprehensively characterize the physical state of the banknote and significantly improve the detection rate and recognition accuracy of complex, hidden and diverse defects.
[0061] S230: Perform acoustic deep feature extraction on the composite sound wave feature vector to obtain an acoustic feature vector, and perform visual deep feature extraction on the surface image to obtain a visual feature vector.
[0062] Specifically, a specialized acoustic feature extraction network can be used to learn a deep representation of the composite sound wave feature vector to obtain an acoustic feature vector. For example, the acoustic feature extraction network can be a variant of a one-dimensional convolutional neural network or a recurrent neural network, pre-trained based on an acoustic feature library. Furthermore, a pre-trained convolutional neural network, or a lightweight convolutional neural network architecture customized for banknote image characteristics, can be used to extract visual features from the surface image to obtain a visual feature vector. Visual features can include texture, color, shape, defects, and so on.
[0063] S240 , using a pre-trained banknote classification model, and according to the acoustic feature vector and the visual feature vector, obtaining the banknote type corresponding to the banknote to be inspected.
[0064] In an alternative embodiment, the input to the banknote classification model may also be acoustic and visual feature vectors, and the output may be the banknote type. In this embodiment, the training data for the banknote classification model can be adjusted to acoustic and visual feature vectors labeled with the banknote type to obtain the current banknote classification model. Thus, after obtaining the current acoustic and visual feature vectors, they can be input into the current banknote classification model to obtain the banknote type corresponding to the banknote to be inspected.
[0065] Optionally, obtaining the banknote type corresponding to the banknote to be inspected according to the acoustic feature vector and the visual feature vector using a pre-trained banknote classification model may include:
[0066] The acoustic feature vector and the visual feature vector are subjected to feature fusion to obtain a multimodal feature vector, and a pre-trained banknote classification model is used to obtain the banknote type corresponding to the banknote to be inspected according to the multimodal feature vector.
[0067] In an optional embodiment, the input to the banknote classification model may also be a multimodal feature vector obtained by fusing acoustic and visual feature vectors. When training the banknote classification model, acoustic and visual feature vectors corresponding to representative banknotes may be obtained. These feature vectors are then fused to obtain a multimodal feature vector. This multimodal feature vector is then labeled with the corresponding banknote type to generate training data. This training data is then used for model training to obtain the final banknote classification model.
[0068] Correspondingly, when identifying the banknote type, multimodal fusion recognition can be used. Specifically, after obtaining the acoustic and visual feature vectors, they can be further fused to obtain a multimodal feature vector that can represent the overall state of the banknote. This multimodal feature vector is then input into the banknote classification model to determine the banknote type corresponding to the banknote under inspection.
[0069] Among them, multimodal fusion recognition refers to combining data features from different sensors (such as acoustic sensors and image sensors) and performing comprehensive analysis through specific fusion algorithms (such as feature layer fusion, decision layer fusion or hybrid fusion) to obtain more accurate recognition results than a single sensor, which is used for damage degree assessment and classification of damaged coins.
[0070] In this embodiment, banknote type identification is performed by using a multimodal feature vector obtained by fusing acoustic feature vectors and visual feature vectors. This can fully utilize the advantages of acoustic wave signals in detecting internal structure and material uniformity, as well as the strengths of image signals in identifying surface stains, printing defects, obvious damage, etc., thereby improving the accuracy and robustness of banknote type identification.
[0071] Optionally, feature fusion methods can include at least one of feature vector concatenation, element-wise weighted sum / product, attention-guided feature fusion, and feature fusion via a pre-trained fusion network. For example, attention-guided feature fusion can be based on a cross-attention mechanism, allowing acoustic features to focus on related visual areas, and visual features to focus on related acoustic areas. Feature fusion via a pre-trained fusion network can learn the complex interactions between features of the two modalities.
[0072] In this embodiment, by adopting a multi-dimensional feature fusion method, a more powerful and robust fusion feature can be formed than a single modality or simple late decision fusion, which can improve the flexibility of feature fusion.
[0073] S250 : Obtain the collection slot corresponding to the banknote to be inspected according to the banknote type and a preset mapping relationship between the banknote type and the collection slot.
[0074] S260 , sorting the banknotes to be inspected into corresponding collection slots through a sorting execution device.
[0075] In a specific implementation of this embodiment, the process of the banknote sorting method can be as follows: Figure 3 As shown. Specifically, first, the banknotes to be inspected are sent into the inspection area by the banknote conveyor. When the banknotes reach the predetermined position, the main control unit triggers the acoustic wave excitation detection device and the image acquisition device to collect acoustic and image signals. After necessary corrections and enhancements, usable acoustic data and image data are formed. Then, the edge artificial intelligence processor performs inference calculations based on the acoustic data and image data, and outputs an assessment of the degree of damage and classification results of the banknotes. Finally, the main control unit receives the classification results of the artificial intelligence processor and instructs the sorting execution device to direct the banknotes to the corresponding physical channels and collection slots. Among them, the edge artificial intelligence processor refers to an artificial intelligence processing unit deployed locally (or proximal), which is responsible for receiving sensor data, running a pre-trained banknote classification model, performing real-time inference calculations, and outputting the classification results of the banknotes. It is characterized by low latency and high efficiency, and can reduce dependence on cloud computing and network bandwidth.
[0076] In this embodiment, acoustic wave excitation technology is introduced to collect the vibration response characteristics of banknotes. Combined with image acquisition, multimodal feature fusion recognition is achieved. At the same time, artificial intelligence models are deployed locally on edge devices to achieve real-time intelligent judgment and sorting execution, providing a non-contact, high-speed, and accurate method for identifying and sorting damaged banknotes. Non-contact means that during the entire detection and classification process, key sensors (acoustic and visual) do not come into physical contact with the banknotes, avoiding secondary damage or wear to the banknotes.
[0077] The technical solution of the embodiment of the present invention is to obtain the transmitted sound wave and the reflected sound wave corresponding to the banknote to be inspected through the sound wave excitation detection device, and obtain the surface image corresponding to the banknote to be inspected through the image acquisition device; obtain the transmitted sound wave feature vector corresponding to the transmitted sound wave, and the reflected sound wave feature vector corresponding to the reflected sound wave, and obtain the composite sound wave feature vector based on the transmitted sound wave feature vector and the reflected sound wave feature vector; perform acoustic deep feature extraction on the composite sound wave feature vector to obtain the acoustic feature vector, and perform visual deep feature extraction on the surface image to obtain the visual feature vector; through pre-training The banknote classification model obtains the banknote type corresponding to the banknote to be inspected based on the acoustic feature vector and the visual feature vector; obtains the collection slot corresponding to the banknote to be inspected based on the banknote type and the preset mapping relationship between the banknote type and the collection slot; sorts the banknotes to be inspected to the corresponding collection slot through the sorting execution device; and through the fusion of the transmitted sound wave feature vector and the reflected sound wave feature vector, and deep feature extraction of the composite sound wave feature vector and the surface image, and then identifies the banknote type based on the acoustic feature vector and the visual feature vector, which can improve the accuracy and robustness of banknote damage identification and sorting.
[0078] Example 3
[0079] Figure 4 This is a schematic diagram of the structure of a banknote sorting device provided in Example 3 of the present invention. Figure 4 As shown, the device includes: an acoustic wave acquisition module 310, a banknote type acquisition module 320 and a banknote sorting module 330; wherein,
[0080] The acoustic wave acquisition module 310 is used to acquire the transmitted acoustic wave and the reflected acoustic wave corresponding to the banknote to be inspected through the acoustic wave excitation detection device, and to acquire the surface image corresponding to the banknote to be inspected through the image acquisition device;
[0081] a banknote type acquisition module 320 for inputting the transmitted sound wave, the reflected sound wave, and the surface image into a pre-trained banknote classification model, and acquiring the banknote type corresponding to the banknote to be inspected output by the banknote classification model;
[0082] The banknote sorting module 330 is used to sort the banknotes to be inspected according to the banknote types through a sorting execution device.
[0083] The technical solution of the embodiment of the present invention is to obtain the transmitted sound waves and reflected sound waves corresponding to the banknotes to be inspected through an acoustic wave excitation detection device, and obtain the surface image corresponding to the banknotes to be inspected through an image acquisition device; input the transmitted sound waves, reflected sound waves and surface images into a pre-trained banknote classification model, and obtain the banknote type corresponding to the banknotes to be inspected output by the banknote classification model; through a sorting execution device, sorting of the banknotes to be inspected is realized according to the banknote type; by comprehensively evaluating the banknote type through the transmitted sound waves, reflected sound waves and surface images, contactless detection of banknotes can be realized, and the accuracy of banknote sorting can be improved.
[0084] Optionally, the banknote type acquisition module 320 includes:
[0085] a feature vector acquisition unit, configured to acquire a transmitted sound wave feature vector corresponding to the transmitted sound wave and a reflected sound wave feature vector corresponding to the reflected sound wave, and to acquire a composite sound wave feature vector based on the transmitted sound wave feature vector and the reflected sound wave feature vector;
[0086] a feature extraction unit, configured to perform acoustic deep feature extraction on the composite sound wave feature vector to obtain an acoustic feature vector, and perform visual deep feature extraction on the surface image to obtain a visual feature vector;
[0087] The banknote type acquisition unit is used to acquire the banknote type corresponding to the banknote to be inspected according to the acoustic feature vector and the visual feature vector using a pre-trained banknote classification model.
[0088] Optionally, the banknote type acquisition unit is specifically used to perform feature fusion on the acoustic feature vector and the visual feature vector to obtain a multimodal feature vector, and obtain the banknote type corresponding to the banknote to be inspected according to the multimodal feature vector through a pre-trained banknote classification model.
[0089] Optionally, the feature fusion method includes at least one of feature vector concatenation, element-level weighted sum / product, feature fusion guided by an attention mechanism, and feature fusion through a pre-trained fusion network.
[0090] Optionally, the feature vector acquisition unit is specifically configured to integrate and weight the transmitted sound wave feature vector and the reflected sound wave feature vector to obtain a composite sound wave feature vector.
[0091] Optionally, the banknote sorting module 330 is specifically configured to obtain the collection slot corresponding to the banknote to be inspected according to the banknote type and a preset mapping relationship between the banknote type and the collection slot;
[0092] The banknotes to be inspected are sorted into corresponding collecting slots by a sorting execution device.
[0093] Optionally, the banknote type includes circulated, pending review, and / or recommended for destruction.
[0094] The banknote sorting device provided in the embodiment of the present invention can execute the banknote sorting method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0095] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.
[0096] Example 4
[0097] Figure 5 This is a schematic diagram of the structure of a banknote sorting system provided by the fourth embodiment of the present invention. The banknote sorting system 40 may include an acoustic wave excitation detection device 41, an image acquisition device 42, a sorting execution device 43 and a processor 44; wherein,
[0098] The acoustic wave excitation detection device 41 is communicatively connected to the processor 44, and is used to obtain the transmitted acoustic waves and reflected acoustic waves corresponding to the banknote to be inspected, and send the transmitted acoustic waves and reflected acoustic waves to the processor 44. Acoustic wave excitation refers to the non-contact detection of banknotes using acoustic waves of a specific frequency (typically, ultrasonic frequency band). Acoustic waves are generated by a transmitter (such as a piezoelectric ceramic transducer) to stimulate the banknote to generate mechanical vibrations or acoustic wave propagation, and then the receiver captures the reflected or transmitted acoustic wave signals. Based on the characteristics of the acoustic wave signal, such as flight time, attenuation coefficient, spectrum distribution, phase change, etc., the physical integrity of the banknote is analyzed, and it is particularly good at detecting internal structural damage (such as internal tearing, delamination, material fatigue, etc.) and damage under certain surface coverings (such as specific tapes).
[0099] The image acquisition device 42 is in communication with the processor 44 and is configured to acquire a surface image corresponding to the banknote to be inspected and send the surface image to the processor 44 .
[0100] The processor 44 is in communication with the sorting execution device 43 and is configured to execute the banknote sorting method according to any embodiment of the present invention and send a sorting control instruction to the sorting execution device 43 .
[0101] The sorting execution device 43 is used to sort the banknotes to be inspected according to the sorting control instructions.
[0102] In an optional example, the banknote sorting system 40 may be composed of a banknote sorting device and an edge artificial intelligence processor. The banknote sorting device may include a main control unit, a banknote conveying device, an acoustic wave excitation detection device 41, an image acquisition device 42, and a sorting execution device 43. When sorting banknotes, first, the banknote conveying device sends the banknotes to be inspected into the detection area, and the acoustic wave excitation detection device 41 and the image acquisition device 42 respectively acquire acoustic wave data (including transmitted acoustic waves and reflected acoustic waves) and image data; then, the main control unit sends the acoustic wave data and image data to the edge artificial intelligence processor, and the edge artificial intelligence processor obtains the banknote type corresponding to the banknote to be inspected based on the acoustic wave data and image data through a pre-trained banknote classification model, and generates corresponding sorting control instructions according to the banknote type and sends them to the main control unit; finally, the main control unit controls the sorting execution device 43 based on the sorting control instructions to sort the banknotes to be inspected into the corresponding collection slots.
[0103] The technical solution of the embodiment of the present invention is to obtain the transmitted sound waves and reflected sound waves corresponding to the banknotes to be inspected through an acoustic wave excitation detection device, and obtain the surface image corresponding to the banknotes to be inspected through an image acquisition device; input the transmitted sound waves, reflected sound waves and surface images into a pre-trained banknote classification model through a processor, and obtain the banknote type corresponding to the banknotes to be inspected output by the banknote classification model, and generate a sorting control instruction according to the banknote type; sort the banknotes to be inspected according to the sorting control instruction through a sorting execution device; evaluate the banknote type by comprehensively evaluating the transmitted sound waves, reflected sound waves and surface images, thereby realizing contactless detection of banknotes and improving the accuracy of banknote sorting.
[0104] It should be noted that the components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit implementations of the inventions described and / or claimed herein.
[0105] Optionally, the banknote sorting system 40 may also include a memory connected to the processor 44 in communication, such as a read-only memory (ROM), a random access memory (RAM), etc., wherein the memory stores a computer program that can be executed by at least one processor 44, and the processor 44 can perform various appropriate actions and processes according to the computer program stored in the read-only memory or the computer program loaded from the storage unit into the random access memory. Various programs and data required for the operation of the banknote sorting system 40 can also be stored in the RAM. The processor 44, ROM and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.
[0106] Multiple components in the banknote sorting system 40 are connected to the I / O interface, including: an input unit; an output unit, such as various types of displays, speakers, etc.; a storage unit, such as a magnetic disk, an optical disk, etc.; and a communication unit, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit allows the banknote sorting system 40 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.
[0107] Processor 44 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of processor 44 include, but are not limited to, a central processing unit, a graphics processing unit, various specialized artificial intelligence computing chips, various processors running machine learning model algorithms, a digital signal processor, and any other suitable processor, controller, microcontroller, etc. Processor 44 executes the various methods and processes described above, such as the banknote sorting method.
[0108] In some embodiments, the banknote sorting method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit. In some embodiments, part or all of the computer program can be loaded and / or installed on the banknote sorting system 40 via a ROM and / or a communication unit. When the computer program is loaded into the RAM and executed by the processor 44, one or more steps of the banknote sorting method described above can be performed. Alternatively, in other embodiments, the processor 44 can be configured to execute the banknote sorting method by any other appropriate means (e.g., by means of firmware).
[0109] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays, application specific integrated circuits, application specific standard products, system-on-a-chip systems, on-load programmable logic devices, computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0110] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.
[0111] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the foregoing.
[0112] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks, wide area networks, blockchain networks, and the Internet.
[0113] A computing system may include clients and servers. The client and server are generally remote from each other and typically interact through a communication network. The client and server relationship arises through computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server.
[0114] This embodiment may also include a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the banknote sorting method provided by any embodiment of the present invention.
[0115] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.
[0116] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.
Claims
1. A banknote sorting method, characterized in that: include: Acquire the transmitted and reflected sound waves corresponding to the banknote to be inspected through the sound wave excitation detection device, and acquire the surface image corresponding to the banknote to be inspected through the image acquisition device; Inputting the transmitted sound wave, the reflected sound wave, and the surface image into a pre-trained banknote classification model, and obtaining the banknote type corresponding to the banknote to be inspected output by the banknote classification model; The banknotes to be inspected are sorted according to the types of the banknotes through a sorting execution device.
2. The method according to claim 1, characterized in that Inputting the transmitted sound wave, the reflected sound wave, and the surface image into a pre-trained banknote classification model, and obtaining the banknote type corresponding to the banknote to be inspected output by the banknote classification model, comprising: Obtaining a transmitted sound wave feature vector corresponding to the transmitted sound wave and a reflected sound wave feature vector corresponding to the reflected sound wave, and obtaining a composite sound wave feature vector based on the transmitted sound wave feature vector and the reflected sound wave feature vector; Performing acoustic deep feature extraction on the composite sound wave feature vector to obtain an acoustic feature vector, and performing visual deep feature extraction on the surface image to obtain a visual feature vector; The banknote type corresponding to the banknote to be inspected is obtained according to the acoustic feature vector and the visual feature vector through a pre-trained banknote classification model.
3. The method according to claim 2, characterized in that Obtaining the banknote type corresponding to the banknote to be inspected based on the acoustic feature vector and the visual feature vector using a pre-trained banknote classification model, including: The acoustic feature vector and the visual feature vector are subjected to feature fusion to obtain a multimodal feature vector, and a pre-trained banknote classification model is used to obtain the banknote type corresponding to the banknote to be inspected according to the multimodal feature vector.
4. The method according to claim 3, characterized in that The feature fusion method includes at least one of feature vector concatenation, element-level weighted sum / product, feature fusion guided by an attention mechanism, and feature fusion through a pre-trained fusion network.
5. The method according to claim 2, characterized in that Obtaining a composite sound wave feature vector according to the transmitted sound wave feature vector and the reflected sound wave feature vector, including: The transmitted sound wave feature vector and the reflected sound wave feature vector are integrated and weighted to obtain a composite sound wave feature vector.
6. The method according to claim 1, characterized in that Sorting the banknotes to be inspected according to the banknote types is performed by a sorting execution device, including: According to the banknote type and the preset mapping relationship between the banknote type and the collection slot, obtaining the collection slot corresponding to the banknote to be inspected; The banknotes to be inspected are sorted into corresponding collecting slots by a sorting execution device.
7. The method according to any one of claims 1 to 6, characterized in that Note types include Circulate, Awaiting Review, and / or Recommended for Destruction.
8. A banknote sorting system, characterized in that: It includes an acoustic wave excitation detection device, an image acquisition device, a sorting execution device and a processor; wherein, The acoustic wave excitation detection device is in communication with the processor, and is used to obtain the transmitted acoustic wave and the reflected acoustic wave corresponding to the banknote to be detected, and send the transmitted acoustic wave and the reflected acoustic wave to the processor; The image acquisition device is in communication with the processor, and is used to acquire a surface image corresponding to the banknote to be inspected, and send the surface image to the processor; The processor is in communication with the sorting execution device, and is configured to execute the banknote sorting method according to any one of claims 1 to 7, and send a sorting control instruction to the sorting execution device; The sorting execution device is used to implement sorting of the banknotes to be inspected according to the sorting control instructions.
9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and the computer program is used to enable a processor to implement the banknote sorting method according to any one of claims 1 to 7 when executed.
10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the banknote sorting method according to any one of claims 1 to 7.