Method and device for detecting and counting goods on shelf and electronic equipment

Through automated shelf product identification and statistics methods, image segmentation and vector search technology are used to solve the problem of cumbersome human recognition and analysis process, data collection and processing efficiency is improved, and more accurate product analysis is achieved.

CN119006786APending Publication Date: 2024-11-22YUNNAN BAIYAO GRP MEDICINE E-COMMERCE CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411089781.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-08
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In the health care products and pharmaceutical distribution industry, due to the diversity of goods placed by distributors and shelves, the process of identifying and analyzing shelves of product data is cumbersome and time-consuming, resulting in a long cycle of data collection and optimization of marketing solutions.

Method used

By obtaining shelf images, performing image segmentation processing and entity detection, it is converted into product entity vectors, searching for pre-stored product vector identification, determining products and performing statistical processing, and achieving automated product identification and data collection.

Benefits of technology

It improves the standardization accuracy and efficiency of product analysis, greatly improves the efficiency of data collection and processing, and reduces human intervention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006786A_ABST
    Figure CN119006786A_ABST
Patent Text Reader

Abstract

The invention relates to a method and device for detecting and counting goods shelf commodities and electronic equipment, and the method comprises the steps: obtaining a goods shelf image, carrying out the image segmentation processing of the goods shelf image, obtaining a goods entity image, converting the goods entity image into a goods entity vector, and carrying out the detection and statistics of the goods shelf commodities. And searching a pre-stored commodity vector identifier according to the commodity entity vector, determining a commodity corresponding to the commodity entity image according to the searched commodity vector identifier, and performing statistical processing on the commodity. Through the above method, the commodity shelf image is analyzed to determine the contained commodity, statistics is performed according to the commodity information, original manual inspection and analysis of the commodity proportion and category are converted into use of the model to solve problems, the standardization accuracy is improved, and the commodity analysis efficiency is greatly improved. The identification and detection of the commodity images in the whole process are automatically processed, so that the data collection and processing efficiency is greatly improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of image recognition and processing, and in particular to a method, device and electronic device for detecting and counting goods on shelves. Background Art

[0002] In the health care products and pharmaceutical distribution industry, pharmaceutical manufacturers provide distributors with data on product marketing, shelf placement, data analysis, etc., usually through human identification and analysis. However, due to the large number of distributors, the diversity of product placement formats in different regions and shelves, and the diversity of product specifications provided to distributors, the overall cycle of human data collection, analysis, and marketing optimization for existing distributors is one month. Human analysis and identification of the number of products of the company and the number of products of competing companies on the shelves, as well as the collection of the above product specification data and analysis of the data results have become cumbersome and time-consuming. Summary of the invention

[0003] In view of this, the present application proposes a method for detecting and counting goods on shelves to solve the problems embodied in the above-mentioned background technology.

[0004] According to one aspect of the present application, a method for detecting and counting shelf commodities is provided, comprising:

[0005] Get product shelf images;

[0006] Performing image segmentation processing on the commodity shelf image to obtain a commodity entity image;

[0007] Converting the commodity entity image into a commodity entity vector, and searching for a pre-stored commodity vector identifier according to the commodity entity vector;

[0008] The commodity corresponding to the commodity entity image is determined according to the searched commodity vector identifier, and statistical processing is performed on the commodities.

[0009] As an optional implementation scheme of the present application, the commodity shelf image is subjected to image segmentation processing to obtain a commodity entity image, including:

[0010] Performing image segmentation processing on the commodity shelf image to obtain image segmentation area blocks;

[0011] The commodity entity is detected on the image segmented area block, and the commodity entity image is output when the commodity entity is detected.

[0012] As an optional implementation scheme of the present application, before searching for a pre-stored product vector identifier according to the product entity vector, the method further includes:

[0013] Preset the minimum stock keeping unit of the product;

[0014] Acquire commodity images collected according to the minimum inventory unit of the commodity;

[0015] The product image is converted into a product vector identifier and pre-stored in a vector database.

[0016] As an optional implementation scheme of the present application, determining the commodity corresponding to the commodity entity image according to the searched commodity vector identifier includes:

[0017] When multiple product vector identifiers are searched, similarity calculation is performed between the product entity vector and each product vector identifier to obtain a similarity result;

[0018] The product vector identifier closest to the product entity vector is determined according to the similarity result, and the product corresponding to the product entity image is determined.

[0019] As an optional implementation scheme of the present application, determining the commodity corresponding to the commodity entity image includes:

[0020] Determining a corresponding product image according to the product vector identifier;

[0021] The commodity corresponding to the commodity physical image is determined according to the commodity image.

[0022] As an optional implementation scheme of the present application, the statistical processing of commodities includes:

[0023] After determining the commodity corresponding to the commodity physical image, statistics of the commodity are performed according to the commodity information.

[0024] The present application also provides a device for detecting and counting shelf commodities, comprising:

[0025] An image acquisition module is used to acquire commodity shelf images;

[0026] An image processing module is used to perform image segmentation processing on the commodity shelf image to obtain a commodity entity image;

[0027] A vectorization processing module, used for converting the commodity entity image into a commodity entity vector, and searching for a pre-stored commodity vector identifier according to the commodity entity vector;

[0028] The search result processing module is used to determine the commodity corresponding to the commodity entity image according to the searched commodity vector identifier, and perform statistical processing on the commodity.

[0029] As an optional implementation scheme of the present application, the image processing module includes:

[0030] An image segmentation module is used to perform image segmentation processing on the commodity shelf image to obtain image segmentation area blocks;

[0031] The entity detection module is used to perform commodity entity detection on the image segmentation area block and output a commodity entity image when a commodity entity is detected.

[0032] As an optional implementation scheme of the present application, the search result processing module includes:

[0033] A similarity calculation module is used to calculate the similarity between the commodity entity vector and each of the commodity vector identifiers when multiple commodity vector identifiers are searched to obtain a similarity result;

[0034] The module for determining a product vector identification is used to determine a product vector identification closest to the product entity vector according to the similarity result, and to determine the product corresponding to the product entity image.

[0035] The present application also provides an electronic device, including:

[0036] processor;

[0037] a memory for storing processor-executable instructions;

[0038] Wherein, the processor is configured to implement the above-mentioned method of detecting and counting shelf commodities when executing the executable instructions.

[0039] Beneficial effects of this application:

[0040] The present invention obtains a commodity shelf image, performs image segmentation processing on the commodity shelf image, obtains a commodity entity image, converts the commodity entity image into a commodity entity vector, searches for a pre-stored commodity vector identifier according to the commodity entity vector, determines the commodity corresponding to the commodity entity image according to the searched commodity vector identifier, and performs statistical processing on the commodity. The commodity shelf image is analyzed to determine the commodities contained therein, and statistics are performed based on the commodity information, converting the original manual inspection and analysis of commodity proportions and categories into solving problems using models, thereby improving the accuracy of standardization and greatly improving the efficiency of commodity analysis. The automated processing of the recognition and detection of commodity images in the entire process greatly improves the efficiency of data collection and processing.

[0041] Other features and aspects of the present application will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the present application and, together with the description, serve to explain the principles of the present application.

[0043] Figure 1A flow chart showing a method for detecting and counting shelf commodities according to an embodiment of the present application;

[0044] Figure 2 A block diagram of a device for detecting and counting shelf merchandise according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0045] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0046] Among them, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application or simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0047] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0048] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0049] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0050] SKU (Stock Keeping Unit) is the basic unit of inventory measurement, which can be in units of pieces, boxes, pallets, etc. It has now been extended to the abbreviation of product unified numbering, and each product has a unique SKU number.

[0051] Example 1

[0052] Figure 1 A flow chart of a method for detecting and counting shelf commodities according to an embodiment of the present application is shown. Figure 1 As shown, this method can be used to identify and count the goods on the shelf. The flow chart includes:

[0053] S100, acquiring a commodity shelf image;

[0054] In this embodiment, a device with a photographing or video recording function is used to photograph the commodity shelf where the commodities to be counted are to obtain a commodity shelf image. In this embodiment, a shelf where toothpaste is placed is photographed to obtain its corresponding toothpaste commodity shelf image.

[0055] S200, performing image segmentation processing on the commodity shelf image to obtain a commodity entity image;

[0056] In this embodiment, an image segmentation and entity detection model is used to analyze the toothpaste commodity shelf image. The model is based on deep learning, convolutional neural network, target detection, instance segmentation, vector search, machine learning and other algorithms. In computer vision, it can detect the target object in the image and accurately segment the target object. In this embodiment, the detection image is a commodity shelf image, and the target object is the commodity in the commodity shelf image, such as toothpaste in the toothpaste commodity shelf image.

[0057] S300, converting the commodity entity image into a commodity entity vector, and searching for a pre-stored commodity vector identifier according to the commodity entity vector;

[0058] In actual application, after the image segmentation is performed on the commodity shelf image, the commodity entity image obtained cannot yet analyze the specific corresponding commodity. For example, the staff imports the commodity image of the toothpaste commodity shelf into the model. The model can detect that the image contains commodities, but cannot analyze that the detected commodity is toothpaste. Therefore, after obtaining the commodity entity image, the commodity entity image needs to be further processed. In this embodiment, the commodity entity image is analyzed from a vector perspective, converted into a commodity entity vector, and a vector search engine is used to search the pre-stored commodity vector identifier in the vector database.

[0059] S400: Determine the commodity corresponding to the commodity entity image according to the searched commodity vector identifier, and perform statistical processing on the commodities.

[0060] The vector search engine is used to search the vector database for the product vector identifier that is closest to the product entity vector. It should be noted that there is a one-to-one correspondence between the product vector identifier and the sku entity. Therefore, the sku entity corresponding to the product entity image can be determined based on the closest product vector identifier. Finally, based on the correspondence between the product entity image and the sku entity, the product in the product entity image is determined, and the products in the product shelf image are counted.

[0061] As an optional implementation scheme of the present application, the commodity shelf image is subjected to image segmentation processing to obtain a commodity entity image, including: performing image segmentation processing on the commodity shelf image to obtain an image segmentation area block; performing commodity entity detection on the image segmentation area block, and outputting the commodity entity image when the commodity entity is detected.

[0062] Use image segmentation and entity detection models for specific scenarios of commodity shelves to perform image segmentation processing on commodity shelf images, segment multiple image segmentation area blocks, and then perform commodity entity detection on these segmentation area blocks separately. This has better detection results than the open source target detection & segmentation model.

[0063] As an optional implementation scheme of the present application, before searching for the pre-stored product vector identifier based on the product entity vector, it also includes: presetting the minimum inventory unit of the product; presetting the minimum inventory unit of the product; obtaining the product image collected according to the minimum inventory unit of the product; converting the product image into a product vector identifier, and pre-storing it in a vector database.

[0064] In this embodiment, the minimum inventory unit of the commodity is preset based on the sku entity, the existing sku commodity is photographed to obtain the commodity image, the minimum inventory unit of the commodity corresponding to the commodity image is vectorized, the obtained vector is used as the commodity vector identifier, and is saved in the vector database for the similarity comparison of the commodity entity vector. It should be noted that, in order to facilitate commodity management, the minimum inventory unit of the commodity is set according to the actual commodity type, that is, each commodity has its corresponding minimum inventory unit. For example, for commodity entity A and commodity entity B, commodity entity A and commodity entity B are both a sku entity. The staff sets the minimum inventory unit of commodity entity A to three pieces, the minimum inventory unit of commodity B to two boxes, and so on. When pre-storing the commodity vector identifier, the commodity image of commodity A is vectorized according to the minimum inventory unit of three pieces. If the commodity image of commodity A has five pieces, the commodity image is processed based on the minimum inventory unit, and then the vectorization is still performed according to the minimum inventory unit of three pieces. The purpose is to ensure that there is a commodity vector identifier of commodity A with a minimum inventory unit of three pieces.

[0065] As an optional implementation scheme of the present application, determining the product corresponding to the product entity image based on the searched product vector identifier includes: when multiple product vector identifiers are searched, calculating the similarity between the product entity vector and each of the product vector identifiers to obtain a similarity result; determining the product vector identifier closest to the product entity vector based on the similarity result, and determining the product corresponding to the product entity image.

[0066] In this step, the purpose is to search for the product vector identifier that is closest to the product entity vector and then determine the specific product. The number of products in the product entity image is not necessarily its minimum inventory unit, so a similarity calculation method can be used. For example, if similar product vector identifier 1, product vector identifier 2, and product vector identifier 3 are searched, the similarity of the product entity vector is calculated with product vector identifier 1, product vector identifier 2, and product vector identifier 3. For example, according to the similarity result, it is determined that product vector identifier 1 is closest to the product entity vector, and the product entity vector and product vector identifier 1 are processed accordingly.

[0067] As an optional implementation scheme of the present application, determining the commodity corresponding to the commodity physical image includes: determining the corresponding commodity image according to the commodity vector identifier; and determining the commodity corresponding to the commodity physical image according to the commodity image.

[0068] Continuing with the above example, when it is determined that the product vector identifier 1 is closest to the product entity vector, since the product image and the product vector identifier are in a one-to-one correspondence, the correspondence between the product entity vector and the product image can be indirectly determined. The specific product corresponding to the product entity image can be known based on the product image.

[0069] As an optional implementation scheme of the present application, the statistical processing of the commodities includes: after determining the commodities corresponding to the commodity physical images, statistical processing of the commodities according to the commodity information.

[0070] By continuously inputting the commodity shelf image into the image segmentation and entity detection model, the commodity entity image is continuously obtained, and the commodity contained in the commodity shelf image is finally determined according to the one-to-one correspondence between the commodity image of the sku commodity and the commodity vector identifier in the vector database. In this embodiment, statistics are performed according to the category and quantity of the commodity. For example, it is determined that the commodities contained in the commodity shelf image are commodity A, commodity B, and commodity C, and the above commodities all belong to the toothpaste category, then the statistical information can be: commodity category name: toothpaste; commodity category number: toothpaste 3; where toothpaste 3 is a commodity that is identified as 3 belonging to toothpaste. Then the analysis result for a certain commodity shelf image is: the category contained in the commodity shelf image is toothpaste, and the number of commodities identified as the toothpaste category is 3. Users can also perform statistics based on commodity information according to other needs, such as counting the specific quantity of each commodity, etc. The statistical results are returned to the backend in json data format for staff to view.

[0071] As an optional implementation scheme of the present application, a data update mode is preset; after the product image is acquired again, the product vector identifier corresponding to the product image is updated in the vector database according to the data update mode.

[0072] In this embodiment, the data update mode adopts the t+1 mode, where t+1 means that the vector database is updated or processed at the first time point after date t. The first time point can be the first working day, the first rest day, etc. The staff sets the meaning corresponding to "1" according to the actual situation.

[0073] The product images acquired again may be stock product images or incremental product images, and their corresponding product vector identifiers are all new to the vector database. The product vector identifiers in the vector database are updated through the vector search engine according to the set t+1.

[0074] Through the above method, the product shelf images are analyzed to determine the products they contain, and statistics are performed based on the product information. The original manual inspection and analysis of product proportions and categories is transformed into using models to solve problems, which improves the accuracy of standardization and greatly improves the efficiency of product analysis. The automated processing of product image recognition and detection in the entire process greatly improves the efficiency of data collection and processing.

[0075] Example 2

[0076] Based on the same principle as the above method, a device for detecting and counting shelf commodities is also proposed, see Figure 2 , a device 100 for detecting and counting shelf commodities according to an embodiment of the present disclosure includes:

[0077] An image acquisition module 110 is used to acquire a commodity shelf image;

[0078] An image processing module 120 is used to perform image segmentation processing on the commodity shelf image to obtain a commodity entity image;

[0079] A vectorization processing module 130, configured to convert the commodity entity image into a commodity entity vector, and search for a pre-stored commodity vector identifier according to the commodity entity vector;

[0080] The search result processing module 140 is used to determine the commodity corresponding to the commodity entity image according to the searched commodity vector identifier, and perform statistical processing on the commodities.

[0081] As an optional implementation scheme of the present application, the image processing module 120 includes:

[0082] An image segmentation module 121 is used to perform image segmentation processing on the commodity shelf image to obtain image segmentation area blocks;

[0083] The entity detection module 122 is used to perform commodity entity detection on the segmented area block of the image, and output a commodity entity image when a commodity entity is detected.

[0084] As an optional implementation of the present application, the search result processing module 140 includes:

[0085] A similarity calculation module 141 is used to calculate the similarity between the commodity entity vector and each of the commodity vector identifiers when multiple commodity vector identifiers are searched, so as to obtain a similarity result;

[0086] The product vector identification determination module 142 is used to determine the product vector identification closest to the product entity vector according to the similarity result, and determine the product corresponding to the product entity image.

[0087] Obviously, those skilled in the art should understand that all or part of the processes in the above-mentioned embodiment method can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned control methods. The modules or steps of the present invention can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. Optionally, they can be implemented by program codes executable by computing devices, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules respectively, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the present invention is not limited to any specific combination of hardware and software.

[0088] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment method can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned control methods. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (Flash Memory), a hard disk (Hard Disk Drive, abbreviated as: HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above-mentioned types of memory.

[0089] Example 3

[0090] Furthermore, an electronic device is proposed, comprising:

[0091] processor;

[0092] a memory for storing processor-executable instructions;

[0093] The processor is configured to implement the method for detecting and counting shelf commodities described in Example 1 when executing the executable instructions.

[0094] The electronic device of the embodiment of the present disclosure includes a processor and a memory for storing processor executable instructions, wherein the processor is configured to implement any of the above-mentioned methods for detecting and counting shelf commodities when executing the executable instructions.

[0095] It should be noted that the number of processors may be one or more. At the same time, the electronic device of the embodiment of the present disclosure may also include an input device and an output device. The processor, memory, input device, and output device may be connected via a bus or in other ways, which are not specifically limited here.

[0096] The memory is a computer-readable storage medium for the method of detecting and counting shelf commodities, and can be used to store software programs, computer executable programs, and various modules, such as the program or module corresponding to the method of detecting and counting shelf commodities in the embodiment of the present disclosure. The processor executes various functional applications and data processing of the electronic device by running the software programs or modules stored in the memory.

[0097] The input device can be used to receive input numbers or signals. The signal can be a key signal related to user settings and function control of the device / terminal / server. The output device can include a display device such as a display screen.

[0098] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for detecting and counting goods on shelves, characterized in that: include: Get product shelf images; Performing image segmentation processing on the commodity shelf image to obtain a commodity entity image; Converting the commodity entity image into a commodity entity vector, and searching for a pre-stored commodity vector identifier according to the commodity entity vector; The commodity corresponding to the commodity entity image is determined according to the searched commodity vector identifier, and statistical processing is performed on the commodities.

2. The method for detecting and counting shelf commodities according to claim 1, characterized in that: The commodity shelf image is subjected to image segmentation processing to obtain a commodity entity image, including: Performing image segmentation processing on the commodity shelf image to obtain image segmentation area blocks; The commodity entity is detected on the image segmented area block, and the commodity entity image is output when the commodity entity is detected.

3. The method for detecting and counting shelf commodities according to claim 1, characterized in that: Before searching the pre-stored product vector identifier according to the product entity vector, the method further includes: Preset the minimum stock keeping unit of the product; Acquire commodity images collected according to the minimum inventory unit of the commodity; The product image is converted into a product vector identifier and pre-stored in a vector database.

4. The method for detecting and counting shelf commodities according to claim 1, characterized in that: The step of determining the commodity corresponding to the commodity entity image according to the searched commodity vector identifier includes: When multiple product vector identifiers are searched, similarity calculation is performed between the product entity vector and each product vector identifier to obtain a similarity result; The product vector identifier closest to the product entity vector is determined according to the similarity result, and the product corresponding to the product entity image is determined.

5. The method for detecting and counting shelf commodities according to claim 3, characterized in that: The determining the commodity corresponding to the commodity entity image includes: Determining a corresponding product image according to the product vector identifier; The commodity corresponding to the commodity physical image is determined according to the commodity image.

6. The method for detecting and counting shelf commodities according to claim 5, characterized in that: The statistical processing of commodities includes: After determining the commodity corresponding to the commodity physical image, statistics of the commodity are performed according to the commodity information.

7. A device for detecting and counting goods on shelves, characterized in that: include: An image acquisition module is used to acquire commodity shelf images; An image processing module is used to perform image segmentation processing on the commodity shelf image to obtain a commodity entity image; A vectorization processing module, used for converting the commodity entity image into a commodity entity vector, and searching for a pre-stored commodity vector identifier according to the commodity entity vector; The search result processing module is used to determine the commodity corresponding to the commodity entity image according to the searched commodity vector identifier, and perform statistical processing on the commodity.

8. The device for detecting and counting shelf commodities according to claim 7, characterized in that: The image processing module comprises: An image segmentation module is used to perform image segmentation processing on the commodity shelf image to obtain image segmentation area blocks; The entity detection module is used to perform commodity entity detection on the image segmentation area block and output a commodity entity image when a commodity entity is detected.

9. The device for detecting and counting shelf commodities according to claim 7, characterized in that: The search result processing module includes: A similarity calculation module is used to calculate the similarity between the commodity entity vector and each of the commodity vector identifiers when multiple commodity vector identifiers are searched to obtain a similarity result; The module for determining a product vector identification is used to determine a product vector identification closest to the product entity vector according to the similarity result, and to determine the product corresponding to the product entity image.

10. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the method for detecting and counting shelf commodities as described in any one of claims 1 to 6 when executing the executable instructions.