Money bundle detection method, storage method and related device
By using detection network models in the vault to automatically identify the location and category of money bundles, the time-consuming and labor-intensive and error problems caused by manual inventory are solved, and efficient money bundles are implemented and out-of-warehouse verification is achieved.
Patent Information
- Application Number
- CN202111604033.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-24
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2041-12-24
AI Technical Summary
In the prior art, the process of vacancies' money bundling and out-of-warehouse verification relies on manual inventory, which leads to time-consuming and labor-intensive and prone to inventory errors.
A money bundle detection method is adopted. By obtaining the money bundle picture and inputting the trained detection network model, the money bundle placement position and coin category are automatically identified, and the coin storage data is generated, thereby realizing automatic inventory. The method includes obtaining the revised money bundle picture, identifying the money bundle position and category through the trained detection network model, and converting the recognition results into a raster diagram to determine the stored data.
The speed of the money bundling in and out-of-warehouse verification process has been improved, the missing points or errors caused by manual inventory have been reduced, and the automated money bundling inventory has been realized.
Smart Images

Figure CN114445340B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a money bundle detection method, a warehousing method and related devices. Background Art
[0002] For current vault scenarios, most of the scenarios are operated manually, although there are machines such as sorting machines to help sort the money bundles in the vault. However, for the most critical money bundle entry and exit verification process, the current method of fully manually counting the grid turnover boxes to determine the money bundle information in the grid turnover box is usually used. This is time-consuming and labor-intensive, and may also result in counting errors caused by human errors.
[0003] Therefore the prior art still needs to be improved and enhanced. Summary of the invention
[0004] The technical problem to be solved by the present application is to provide a money bundle detection method, a warehousing method and related devices in view of the deficiencies of the prior art.
[0005] In order to solve the above technical problems, the first aspect of the embodiments of the present application provides a money bundle detection method, wherein the method further includes:
[0006] Obtaining a picture of a bundle of money, and inputting the picture of the bundle of money into a trained detection network model;
[0007] Detecting the money bundle placement position and the coin category corresponding to the money bundle image by using the detection network model;
[0008] The coin storage data corresponding to the money bundle image is determined based on the detected placement position of the money bundle and the type of the coins.
[0009] The money bundle detection method, wherein the step of obtaining a money bundle image and inputting the money bundle image into a trained detection network model specifically comprises:
[0010] Acquire a money bundle image, and correct the money bundle image so that the height of the money bundle in the corrected money bundle image is perpendicular to the boundary of the money bundle image;
[0011] Input the corrected money bundle image into the trained detection network model.
[0012] The money bundle detection method, wherein the training process of the detection network model specifically includes:
[0013] Acquire a training sample set, wherein the training sample set includes a plurality of training money bundle images;
[0014] The preset network model is trained based on the training sample set to obtain the detection network model.
[0015] The money bundle detection method, wherein the step of obtaining a training sample set specifically comprises:
[0016] Acquire a training money bundle picture set, and select some training money bundle pictures from the training money bundle picture set for annotation, so as to obtain a target money bundle picture with annotation information;
[0017] Training the preset network model based on the obtained target money bundle images to obtain a candidate network model;
[0018] The candidate network model is used to label the coin category for each money bundle picture in the training money bundle picture set except the target money bundle picture, so as to obtain a training sample set.
[0019] The money bundle detection method, wherein the step of training a preset network model based on the training sample set to obtain the detection network model specifically includes:
[0020] The candidate network model is used as a preset network model, and the candidate network model is trained based on the training sample set to obtain the detection network model.
[0021] The money bundle detection method, wherein the determining of the money storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category specifically includes:
[0022] Fit the positions of the money bundles to generate a grid map;
[0023] The coin categories corresponding to the money bundle placement positions corresponding to each grid are added to each grid to obtain the coin storage data corresponding to the money bundle image.
[0024] A second aspect of the embodiment of the present application provides a warehousing method, the method comprising:
[0025] Obtaining a money bundle image corresponding to the grid turnover box to be stored, and determining the money storage data corresponding to the money bundle image based on the money bundle detection method as described above;
[0026] comparing the coin storage data with the target data corresponding to the money bundle image;
[0027] When the coin storage data is the same as the target data, the target data is stored and the grid turnover box is put into storage.
[0028] The warehousing method, wherein the method further comprises:
[0029] When the coin storage data is different from the target data, a distinguishing grid whose coin storage data is different from the target data is selected, and a prompt is given that the coin placement in the distinguishing grid is wrong.
[0030] A third aspect of an embodiment of the present application provides a computer-readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the steps in the money bundle detection method as described above, and / or to implement the steps in the warehousing method as described above.
[0031] A fourth aspect of the embodiments of the present application provides a terminal device, comprising: a processor, a memory, and a communication bus; the memory stores a computer-readable program that can be executed by the processor;
[0032] The communication bus realizes the connection and communication between the processor and the memory;
[0033] When the processor executes the computer-readable program, the processor implements the steps in the money bundle detection method described above, and / or implements the steps in the warehousing method described above.
[0034] Beneficial effects: Compared with the prior art, the present application provides a money bundle detection method, a warehousing method and related devices, the method comprising obtaining a money bundle image, and inputting the money bundle image into a trained detection network model; detecting the money bundle placement position and the coin category corresponding to the money bundle image through the detection network model; and determining the coin storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category. The present application recognizes the money bundle image through a trained detection network model, and can automatically obtain the money bundle placement position and the coin category in the money bundle image, so as to automatically count the grid turnover box, which can increase the speed of money bundle counting in the money bundle entry and exit verification process on the one hand, and can eliminate the phenomenon of missing points or wrong points caused by manual counting of the panel on the other hand. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without inventive work.
[0036] Figure 1 This is a flow chart of the money bundle detection method provided in this application.
[0037] Figure 2 This is a structural principle diagram of the money bundle detection device provided in this application.
[0038] Figure 3 This is a schematic diagram of the structure of the terminal device provided in this application. DETAILED DESCRIPTION
[0039] The present application provides a money bundle detection method, a storage method and related devices. In order to make the purpose, technical solution and effect of the present application clearer and more specific, the present application is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.
[0040] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.
[0041] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with those in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.
[0042] It should be understood that the sequence numbers and sizes of the steps in this embodiment do not mean the order of execution. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.
[0043] After research, the inventor found that most of the current vault scenes are operated manually, although there are machines such as sorting machines to help sort the money bundles in the vault. However, for the most critical money bundle entry and exit verification process, the current method of completely manually counting the grid turnover boxes to determine the money bundle information in the grid turnover box is usually adopted. This is time-consuming and laborious, and on the other hand, it may also cause counting errors due to human operation errors.
[0044] In order to solve the above problems, in an embodiment of the present application, a picture of a money bundle is obtained, and the picture of the money bundle is input into a trained detection network model; the money bundle placement position and the coin category corresponding to the money bundle picture are detected by the detection network model; and the coin storage data corresponding to the money bundle picture is determined based on the detected money bundle placement position and coin category. The present application recognizes the money bundle picture through a trained detection network model, and can automatically obtain the money bundle placement position and coin category in the money bundle picture, so that the grid turnover box can be automatically counted. This can increase the speed of money bundle counting in the money bundle entry and exit verification process on the one hand, and can eliminate the phenomenon of missed points or wrong points caused by manual counting of the panel on the other hand.
[0045] The application content is further explained below through the description of embodiments in conjunction with the accompanying drawings.
[0046] This embodiment provides a money bundle detection method, such as Figure 1 As shown, the method includes:
[0047] S10, obtaining a picture of a bundle of money, and inputting the picture of the bundle of money into a trained detection network model.
[0048] Specifically, the money bundle image carries several money bundle areas, each of the several money bundle areas corresponds to a money bundle, wherein the types of coins of the money bundles corresponding to the money bundles of the money bundle areas can be all the same, can be partially the same and partially different, or can be all different. For example, the money bundle image includes money bundle area A and money bundle area B, the type of coins of the money bundle corresponding to money bundle area A is 10 yuan, and the type of coins of the money bundle corresponding to money bundle area B is 20 yuan. In one implementation, the money bundle image can be obtained by photographing the grid turnover box with a photographing device, that is, when the grid turnover box needs to be counted, the grid turnover box can be photographed with a photographing device to obtain the money bundle image, and then the money bundle storage data of the money bundle carried by the grid turnover box can be determined based on the money bundle detection method provided in this embodiment.
[0049] In an implementation of this embodiment, since problems such as image content tilt may occur during the shooting of the money bundle image, the money bundle image may be pre-processed to correct the image content in the money bundle image before inputting the money bundle image into the trained detection network model, which can reduce the difficulty of identifying the money bundle image, thereby improving the recognition speed and accuracy of the detection network. Based on this, the steps of obtaining the money bundle image and inputting the money bundle image into the trained detection network model specifically include:
[0050] Acquire a money bundle image, and correct the money bundle image so that the height of the money bundle in the corrected money bundle image is perpendicular to the boundary of the money bundle image;
[0051] Input the corrected money bundle image into the trained detection network model.
[0052] Specifically, when correcting the money bundle picture, the angle and position of the picture content in the money bundle picture can be corrected so that the height of the money bundle in the corrected money bundle picture is perpendicular to the boundary of the money bundle picture. In one implementation, a horizontal correction algorithm (e.g., Hough transform, etc.) can be used to correct the money bundle picture, and the angle and position of the picture content in the money bundle picture are corrected to a preset fixed angle and fixed position, wherein when the picture content of the money bundle picture is at a fixed angle and fixed position, the height of the money bundle in the money bundle picture is perpendicular to the boundary of the money bundle picture. In addition, in order to increase the detection speed of the money bundle picture, after the money bundle picture is obtained, the money bundle picture can be compressed to reduce the number of money bundle pictures, thereby improving the operation speed of the detection network model. Of course, in actual applications, the picture of the money bundle can be obtained by photographing the entire grid turnover box, or by photographing part of the grid turnover box according to actual conditions. That is to say, the grid turnover box can be divided into several box areas according to actual conditions, and then each box area is photographed separately to obtain the money bundle picture corresponding to each box area, and then the money bundle picture is detected separately to obtain the coin storage data corresponding to the entire grid turnover box. In this way, the method of dividing and detecting in different areas can avoid the problem of unclear overall picture due to the large grid turnover box, which affects the detection accuracy.
[0053] The detection network model is a trained deep learning network model, for example, the detection network model adopts an SSD network model, etc. In one implementation, the training process of obtaining the detection network model based on a preset training sample set specifically includes:
[0054] Obtain a training sample set;
[0055] The preset network model is trained based on the training sample set to obtain the detection network model.
[0056] Specifically, the training sample set includes a number of training money bundle images, each of which carries a money bundle area, and each money bundle area corresponds to a money bundle. The money bundle area set formed by each training money bundle image carrying a money bundle area can be divided into a number of money bundle area classes according to the type of currency, each of which corresponds to a type of currency, and the types of wallets corresponding to each money bundle area class are different. In one implementation, the number of money bundle area classes at least includes 100 yuan, 50 yuan, 20 yuan, 10 yuan, 5 yuan, 1 yuan of circulating notes on the market and 100 yuan, 50 yuan, 20 yuan, 10 yuan, 5 yuan, 1 yuan of sealed notes transferred from the People's Bank of China. In addition, in practical applications, in order to improve the training speed of the detection network model, each training money bundle picture in the training sample set carries a money bundle area, that is, each training money bundle picture carries a money bundle. This can reduce the difficulty of identifying the training money bundle pictures, thereby improving the learning speed of the preset network model, and then improving the training speed of the detection network model.
[0057] In an implementation of this embodiment, obtaining a training sample set specifically includes:
[0058] Acquire a training money bundle picture set, and select some training money bundle pictures from the training money bundle picture set for annotation, so as to obtain a target money bundle picture with annotation information;
[0059] Training the preset network model based on the obtained target money bundle images to obtain a candidate network model;
[0060] The candidate network model is used to label the coin category for each money bundle picture in the training money bundle picture set except the target money bundle picture, so as to obtain a training sample set.
[0061] Specifically, the annotation information of the target money bundle image can be formed by manual annotation, wherein the annotation information can include the type of currency corresponding to the money bundle area and the position information of the money bundle area in the money bundle image, so that the preset network model can learn the characteristics of the currency type and the money bundle area carried by the target money bundle image, and then predict the placement position of the money bundle and the category of the currency. In addition, the money bundle area set formed by the money bundle area carried by each target money bundle image is also divided into several candidate money bundle area classes according to the currency type, and the several candidate money bundle area classes correspond to the above several money bundle area classes one by one, and the currency types corresponding to the candidate money bundle area classes are the same as the currency types corresponding to the corresponding money bundle area classes. For example, several money bundle area classes include 100 yuan, 50 yuan, 20 yuan, 10 yuan, 5 yuan, and 1 yuan of circulating banknotes on the market and 100 yuan, 50 yuan, 20 yuan, 10 yuan, 5 yuan, and 1 yuan of sealed banknotes allocated by the People's Bank of China, then several candidate money bundle area classes include 100 yuan, 50 yuan, 20 yuan, 10 yuan, 5 yuan, and 1 yuan of circulating banknotes on the market and 100 yuan, 50 yuan, 20 yuan, 10 yuan, 5 yuan, and 1 yuan of sealed banknotes allocated by the People's Bank of China.
[0062] After obtaining the labeled target money bundle picture, the preset network model is pre-trained based on the labeled target money bundle picture to obtain a candidate network model, wherein the candidate network model can learn the image features in the money bundle picture to predict the placement position of the money bundle and the type of coins in the money bundle picture. Then, the candidate network model is used to predict the predicted placement position and coin type of each training money bundle picture except the target money bundle picture in the training picture set, and the predicted placement position and coin type are used as the labeling information of the training money bundle picture to obtain a training sample set. This embodiment automatically labels all the remaining original picture data through the candidate network model, greatly reducing the time of manual labeling.
[0063] In this embodiment, when determining the training sample set, the pre-trained candidate network model is determined, and the labeling information of some training money bundle pictures in the training sample set is determined by the candidate network model, that is, the labeling information of some training money bundle pictures in the training sample set may have the problem of low accuracy. Therefore, in order to ensure the model accuracy of the detection network model, when training the detection network model based on the training sample set, the candidate network model can be used as the preset network model, so that the money bundle picture feature recognition ability learned during the pre-training of the candidate network model can be used, thereby ensuring the model accuracy of the detection network model.
[0064] Based on this, the candidate network model is used as a preset network model, and the candidate network model is trained based on the training sample set to obtain the detection network model.
[0065] Specifically, the model parameters of the candidate network model are model parameters trained based on the target money bundle picture, and the candidate network model can learn the picture features of the training money bundle picture, so that the candidate network model is used as the preset network model, which can improve the training accuracy of the trained detection network model. Wherein, when the candidate network model is trained based on the training sample set, the training money bundle picture is input into the candidate network model, and the predicted money bundle placement position and predicted coin category corresponding to the training money bundle picture are determined by the candidate network model, and then the loss item is determined based on the annotation information carried by the training money bundle picture, and then the candidate network model is trained based on the loss item to obtain the detection network model.
[0066] S20. Detecting the money bundle placement position and the coin category corresponding to the money bundle image through the detection network model.
[0067] Specifically, the money bundle picture may carry one money bundle area or multiple money bundle areas. When carrying one money bundle area, the money bundle placement position and the coin category may be the placement position of the money bundle area and the coin type of the money bundle in the money bundle area; when carrying multiple money bundle areas, the money bundle placement position and the coin category are also multiple, and the multiple money bundle placement positions and coin categories correspond one-to-one to the multiple money bundle areas. The money bundle placement position and the coin category are the position of the corresponding money bundle area in the money bundle picture, and the coin type of the money bundle in the money bundle area.
[0068] S30, determining the coin storage data corresponding to the money bundle image based on the detected placement position of the money bundle and the coin category.
[0069] Specifically, the coin storage data is used to reflect the coin storage information in the money bundle storage device corresponding to the money bundle image, wherein the coin storage information includes the money bundle storage position and the coin category corresponding to the money bundle. In one implementation, since money bundles are generally stored in grid turnover boxes, the coin storage data is the money bundle storage information in the grid turnover box, that is, the coin storage data includes the coin category of the money bundle stored in each grid in the grid turnover box, and the storage position of each money bundle in the grid turnover box.
[0070] In one implementation, the determining of the coin storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category specifically includes:
[0071] Fit the positions of the money bundles to generate a grid map;
[0072] The coin categories corresponding to the money bundle placement positions corresponding to each grid are added to each grid to obtain the coin storage data corresponding to the money bundle image.
[0073] Specifically, after obtaining the money bundle placement position of each money bundle, each money bundle placement position can be linearly fitted to obtain a number of horizontally and vertically intersecting fitting lines, and the number of horizontally and vertically intersecting fitting lines can be pixel grid images, and each grid in the grid image corresponds to a money bundle placement position, that is, each grid image corresponds to a money bundle area. In a specific implementation, before detecting the money bundle image, the money bundle image is corrected so that the height of the money bundle in the money bundle image is perpendicular to the boundary of the money bundle image. Thus, the regional edges of each money bundle area are on the X and Y axes, so that when fitting the money bundle placement position to generate a grid map, linear fitting can be used to obtain a number of fitting lines that intersect along the X-axis direction and along the Y-axis direction to obtain a grid map of the grid in the simulated grid turnover box. After the grid map is obtained, the coin category corresponding to the money bundle placement position corresponding to each grid is added to each grid to obtain the coin storage data corresponding to the money bundle image.
[0074] In summary, the present embodiment provides a money bundle detection method, the method comprising obtaining a money bundle image, and inputting the money bundle image into a trained detection network model; detecting the money bundle placement position and the coin category corresponding to the money bundle image through the detection network model; and determining the coin storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category. The present application recognizes the money bundle image through a trained detection network model, and can automatically obtain the money bundle placement position and the coin category in the money bundle image, so that the grid turnover box can be automatically counted. This can increase the speed of money bundle counting in the money bundle entry and exit verification process on the one hand, and can prevent the phenomenon of missed points or wrong points caused by manual counting of the panel on the other hand.
[0075] Based on the above money bundle detection method, this embodiment provides a warehousing method, which includes:
[0076] Obtaining a money bundle image corresponding to the grid turnover box to be stored, and determining the money storage data corresponding to the money bundle image based on the money bundle detection method as described above;
[0077] comparing the coin storage data with the target data corresponding to the money bundle image;
[0078] When the coin storage data is the same as the target data, the target data is stored and the grid turnover box is put into storage.
[0079] In one implementation, the method further includes:
[0080] When the coin storage data is different from the target data, a distinguishing grid whose coin storage data is different from the target data is selected, and a prompt is given that the coin placement in the distinguishing grid is wrong.
[0081] Specifically, the money bundle image is obtained by collecting images of the grid turnover box to be stored. The target data is pre-stored money bundle information for recording the money bundles stored in the grid turnover box. The obtained coin storage data can be compared with the target data to compare the money bundles stored in the grid turnover box with the money bundles recorded in the target data, so as to count the money bundles. In this way, it can be detected whether there are problems such as money bundles being randomly placed, money bundles being missing, and money bundles being replaced in the grid turnover box, thereby realizing automatic storage and counting.
[0082] Based on the above money bundle detection method, this embodiment provides a money bundle detection device, such as Figure 2 As shown, the device comprises:
[0083] The acquisition module 100 is used to acquire a picture of a bundle of money and input the picture of the bundle of money into a trained detection network model;
[0084] A detection module 200, configured to detect the money bundle placement position and the coin category corresponding to the money bundle image through the detection network model;
[0085] The determination module 300 is used to determine the coin storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category.
[0086] Based on the money bundle detection method described above, this embodiment provides a computer-readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the steps in the money bundle detection method described in the above embodiment.
[0087] Based on the above money bundle detection method, the present application also provides a terminal device, such as Figure 3 As shown, it includes at least one processor (processor) 20; display screen 21; and memory (memory) 22, and may also include a communications interface (Communications Interface) 23 and a bus 24. Among them, the processor 20, the display screen 21, the memory 22 and the communication interface 23 can communicate with each other through the bus 24. The display screen 21 is configured to display a preset user guide interface in the initial setting mode. The communication interface 23 can transmit information. The processor 20 can call the logic instructions in the memory 22 to execute the method in the above embodiment.
[0088] In addition, the logic instructions in the memory 22 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.
[0089] The memory 22 is a computer-readable storage medium that can be configured to store software programs, computer executable programs, such as program instructions or modules corresponding to the methods in the embodiments of the present disclosure. The processor 20 executes functional applications and data processing by running the software programs, instructions or modules stored in the memory 22, that is, implementing the methods in the above embodiments.
[0090] The memory 22 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 22 may include a high-speed random access memory and may also include a non-volatile memory. For example, a variety of media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, may also be a transient storage medium.
[0091] In addition, the specific process of loading and executing the multiple instruction processors in the above-mentioned storage medium and the terminal device has been described in detail in the above-mentioned method, and will not be described one by one here.
[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A money bundle detection method, characterized in that: The method further comprises: Obtaining a picture of a bundle of money, and inputting the picture of the bundle of money into a trained detection network model; Detecting the money bundle placement position and the coin category corresponding to the money bundle image by using the detection network model; Determining the coin storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category; obtaining the money bundle image and inputting the money bundle image into the trained detection network model specifically includes: Acquire a money bundle image, and correct the money bundle image so that the height of the money bundle in the corrected money bundle image is perpendicular to the boundary of the money bundle image; Input the corrected money bundle image into the trained detection network model; The method of determining the coin storage data corresponding to the money bundle image based on the detected money bundle placement position and coin category specifically includes: Fit the positions of the money bundles to generate a grid map; Add the coin categories corresponding to the money bundle placement positions corresponding to each grid into each grid to obtain the coin storage data corresponding to the money bundle image; The money storage data is used to reflect the coin type and money bundle placement position of the money bundle corresponding to the money bundle image.
2. The money bundle detection method according to claim 1, characterized in that: The training process of the detection network model specifically includes: Acquire a training sample set, wherein the training sample set includes a plurality of training money bundle images; The preset network model is trained based on the training sample set to obtain the detection network model.
3. The money bundle detection method according to claim 2, characterized in that: The obtaining of the training sample set specifically includes: Acquire a training money bundle picture set, and select some training money bundle pictures from the training money bundle picture set for annotation, so as to obtain a target money bundle picture with annotation information; Training the preset network model based on the obtained target money bundle images to obtain a candidate network model; The candidate network model is used to label the coin category for each money bundle picture in the training money bundle picture set except the target money bundle picture, so as to obtain a training sample set.
4. The money bundle detection method according to claim 3, characterized in that: The step of training the preset network model based on the training sample set to obtain the detection network model specifically includes: The candidate network model is used as a preset network model, and the candidate network model is trained based on the training sample set to obtain the detection network model.
5. A warehousing method, characterized in that: The method comprises: Obtaining a money bundle picture corresponding to the grid turnover box to be stored, and determining the money storage data corresponding to the money bundle picture based on the money bundle detection method according to any one of claims 1 to 4; comparing the coin storage data with the target data corresponding to the money bundle image; When the coin storage data is the same as the target data, the target data is stored and the grid turnover box is put into storage.
6. The warehousing method according to claim 5, characterized in that: The method further comprises: When the coin storage data is different from the target data, a distinguishing grid whose coin storage data is different from the target data is selected, and a prompt is given that the coin placement in the distinguishing grid is wrong.
7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the money bundle detection method according to any one of claims 1 to 4, and / or to implement the steps of the warehousing method according to any one of claims 5 to 6.
8. A terminal device, characterized in that: include: Processor, memory and communication bus; The memory stores a computer-readable program executable by the processor; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, the processor implements the steps of the money bundle detection method according to any one of claims 1 to 4, and / or implements the steps of the warehousing method according to any one of claims 5 to 6.
Citation Information
Patent Citations
Target detection method and device and computer-readable storage medium
CN109325538A
Vault money bundle state detection method and system based on image recognition and medium
CN111339957A
