A product intelligent identification method, system, electronic device and storage medium

Abstract

The present application provides a product intelligent identification method, system, electronic device and storage medium, wherein the product intelligent identification method includes: obtaining an original image including an internal accessory product of a telecommunication switching cabinet, and preprocessing the original image to obtain a preprocessed image; identifying the preprocessed image through a trained recognition model based on a deep neural network, and judging whether there is a set type of internal accessory product of a telecommunication switching cabinet based on feature matching; if so, counting the set type of internal accessory products of the telecommunication switching cabinet until the preprocessed image is traversed to obtain the number of the set type of internal accessory products of the telecommunication switching cabinet. The present application can greatly improve the efficiency of product identification, reduce manual workload, human errors and arbitrariness, and comprehensively improve the intelligence and accuracy of product identification.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a product intelligent identification method, system, electronic device and storage medium. Background Art

[0002] The type and quantity of internal accessories of the telecommunications switching cabinet are important parameters for testing the quality of the telecommunications switching cabinet, which determines whether the product meets its basic working requirements. However, we cannot quickly and accurately judge the type and number of internal accessories in front of the switching cabinet in a purely manual way. Product type and quantity detection is an important part of product testing. The traditional testing process usually uses manual identification and statistics of the number of each product one by one, and then records and saves the obtained data. This is not only inefficient but also difficult to guarantee measurement accuracy. With the rapid development of modern information technology and the gradual development of traditional industrial transformation, automated and intelligent detection and identification of products have become possible and have a broad application market. The traditional method of detecting product types and quantities has low detection efficiency, large errors, and high labor costs, which seriously restricts the technological development in related fields. Summary of the invention

[0003] The present application provides a product intelligent identification method, system, electronic device and storage medium to solve the problems of low detection efficiency, large errors and high labor costs in traditional methods for detecting product types and quantities.

[0004] In a first aspect, the present application provides a product intelligent identification method, comprising: Acquire an original image including the internal accessory products of the telecommunication switching cabinet, and preprocess the original image to obtain a preprocessed image; The pre-processed image is recognized by a trained recognition model based on a deep neural network, and it is determined whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching; If so, the set type of telecommunication switching cabinet internal accessory products are counted until the pre-processed image is traversed to obtain the number of the set type of telecommunication switching cabinet internal accessory products.

[0005] Furthermore, the preprocessing of the original image to obtain a preprocessed image includes: Grayscale conversion, filtering, noise reduction and edge detection are performed on the original image to obtain a preprocessed image.

[0006] Furthermore, the filtering and noise reduction adopts Gaussian filtering, and the edge detection adopts sub-pixel edge extraction.

[0007] Furthermore, before the pre-processed image is recognized by the trained recognition model based on the deep neural network and whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching, the method further includes: Acquire training images of different types of internal accessories of telecommunication switching cabinets, and mark features of the different types of internal accessories of telecommunication switching cabinets; The recognition model based on the deep neural network is trained using the training images to obtain a trained recognition model based on the deep neural network.

[0008] Furthermore, the pre-processed image is recognized by using a trained recognition model based on a deep neural network, and judging whether there is a set type of internal accessory product of a telecommunication switching cabinet based on feature matching includes: Performing rasterization processing on the preprocessed image to divide the preprocessed image into a plurality of raster sub-images; The target in the grid sub-image is detected by using a trained recognition model based on a deep neural network, and it is determined based on feature matching whether the target is a set type of internal accessory product of a telecommunication switching cabinet.

[0009] Furthermore, the target in the grid sub-image is detected by using a trained recognition model based on a deep neural network, and whether the target is a set type of internal accessory product of a telecommunication switching cabinet is determined based on feature matching, including: Setting a bounding box for the target in the grid sub-image through a trained recognition model based on a deep neural network, and extracting target features within the bounding box; It is determined whether the target feature matches the feature of a set type of internal accessory product of a telecommunication switching cabinet. If they match, it is determined that the target is a set type of internal accessory product of a telecommunication switching cabinet.

[0010] Furthermore, if the telecom switching cabinet internal accessory products of the set type exist, the number of the telecom switching cabinet internal accessory products of the set type is counted until the pre-processed image is traversed to obtain the number of the telecom switching cabinet internal accessory products of the set type, and the method further includes: The quantity of the internal accessory products of the set type of the telecommunication switching cabinet is transmitted to the mobile terminal for real-time display and recording.

[0011] In a second aspect, the present application provides a product intelligent identification system, comprising: An image acquisition module is used to acquire an original image including the internal accessory products of the telecommunication switching cabinet, and preprocess the original image to obtain a preprocessed image; A product recognition module is used to recognize the pre-processed image through a trained recognition model based on a deep neural network, and determine whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching; The product counting module is used to count the internal accessory products of the telecommunication switching cabinet of the set type, if any, until the pre-processed image is traversed to obtain the number of the internal accessory products of the telecommunication switching cabinet of the set type.

[0012] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the product intelligent identification method as described above when executing the computer program.

[0013] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the product intelligent identification method as described above is implemented.

[0014] The above technical solution of the present application has the following advantages: The first aspect of the present application provides a method for intelligent product identification, which obtains an original image including an internal accessory product of a telecommunication switching cabinet and preprocesses the original image to obtain a preprocessed image, and then identifies the preprocessed image through a trained recognition model based on a deep neural network, and determines whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching. If so, the set type of internal accessory product of the telecommunication switching cabinet is counted until the preprocessed image is traversed, and finally the number of the set type of internal accessory product of the telecommunication switching cabinet is obtained. High-performance machine vision and deep neural network technology are used to automatically identify the type and quantity of the internal accessory product of the telecommunication switching cabinet. The operation is simple, and the efficiency of identifying the internal accessory product of the telecommunication switching cabinet can be greatly improved, the manual workload, human errors and arbitrariness can be reduced, and the intelligence and accuracy of identifying the internal accessory product of the telecommunication switching cabinet can be comprehensively improved.

[0015] It can be understood that the beneficial effects of the second, third and fourth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flowchart of the product intelligent identification method provided in this application; Figure 2 A schematic diagram of the structure of the product intelligent identification system provided for this application; Figure 3 A schematic diagram of the structure of the electronic device provided in this application. DETAILED DESCRIPTION

[0018] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0019] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0020] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0021] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways. "Multiple" means "two or more".

[0022] The specific implementation methods of the present application are further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0023] like Figure 1As shown, an embodiment of the present application provides a product intelligent identification method, which specifically includes the following steps: acquiring an original image including internal accessory products of a telecommunication switching cabinet, and preprocessing the original image to obtain a preprocessed image; identifying the preprocessed image through a trained recognition model based on a deep neural network, and judging whether there are internal accessory products of a set type of telecommunication switching cabinet based on feature matching; if so, counting the internal accessory products of the telecommunication switching cabinet of the set type until the preprocessed image is traversed to obtain the number of internal accessory products of the telecommunication switching cabinet of the set type.

[0024] The detection of the number of accessories inside the telecommunications switching cabinet is an important part of the detection of the telecommunications switching cabinet. The traditional detection process usually uses manual identification and counting of the number of accessories one by one, and then records and saves the obtained data. This is not only inefficient but also difficult to ensure measurement accuracy. The goal of the embodiment of the present application is to quickly and conveniently identify the type and quantity of internal accessories of the telecommunications switching cabinet, and quickly analyze and process the detection results. The business mainly includes accessory type identification, accessory quantity statistics, analysis and processing of detection results, etc.

[0025] Computer vision can intuitively and accurately identify the type and quantity of objects. Computer vision technology refers to taking photos of the appearance of products with mobile phones, computers and other related equipment, and processing the photos at the pixel level. It can filter out irrelevant noise information in the image and retain the required image features. The retained image features are then analyzed and matched with the built-in model to obtain the accurate product type and count the number of each product through the number of matches.

[0026] Compared with traditional detection methods, the embodiment of the present application uses high-performance machine vision and deep neural network technology to automatically identify the type and quantity of accessories inside the telecommunications switching cabinet. Through real-time detection of the collected images, data analysis can be used to obtain the type and quantity of the products. The data can be directly transmitted to the mobile terminal through the cloud for real-time display and recording, thereby simplifying the manual operation process.

[0027] In some embodiments, preprocessing the original image to obtain a preprocessed image includes: performing grayscale conversion, filtering and noise reduction, and edge detection on the original image to obtain a preprocessed image.

[0028] In some embodiments, the filtering and noise reduction uses Gaussian filtering, and the edge detection uses sub-pixel edge extraction.

[0029] In actual production and processing, many parts are not so easy to distinguish due to the influence of complex environmental factors. Therefore, in order to quickly and accurately identify them, they must be converted to grayscale. The color of each pixel of the RGB image corresponds to a point in three-dimensional space, while the color of the grayscale image pixel can be represented by a straight line. Therefore, it is easy to conclude that the conversion of color images to grayscale images is essentially to seek a mapping in three-dimensional space.

[0030] During the image acquisition process, due to the different complexity of product structures, noise in the image is inevitable, and noise will affect the system's recognition and judgment of the detection area. Therefore, noise reduction filtering plays an irreplaceable role in the entire detection system. There are linear filtering methods and nonlinear filtering methods for noise processing. For example, mean filtering is a linear method, which uses a mean algorithm to average the grayscale values ​​of the image to achieve the effect of noise reduction and smoothing the image. Median filtering is a nonlinear method. However, for the detection of parts with relatively high precision requirements, these two filtering methods cannot achieve the expected results. Therefore, the embodiment of the present application adopts another reliable filtering method-Gaussian filtering. Using a Gaussian filter, high-precision detection tasks can be completed.

[0031] The sub-pixel precision contour represents the boundary between two areas in the image, where the grayscale value of one area is greater than the grayscale threshold, while the grayscale value of the other area is less than the grayscale threshold. In order to obtain this boundary, the discreteness of the image needs to be converted into a continuous function, and this conversion can be completed by the method of bilinear interpolation. In the industrial production of product inspection, the images of parts collected by industrial cameras are often pixel-precise. In the inspection of parts, it is necessary to achieve a higher accuracy than the image pixel resolution. Therefore, extracting sub-pixel accuracy from the image is the only effective way to achieve high-precision requirements. The filter canny can be used to extract the sub-pixel edge of the target region of the part, and the edge composed of pixels can be directly returned with sub-pixel accuracy.

[0032] The processed images are marked according to factors such as image size, pixel size, aspect ratio, and clarity quantitative indicators, and the marked images are matched with the recognition model to determine the optimal model detection parameters to ensure the accuracy and efficiency of model detection.

[0033] In some embodiments, the pre-processed image is recognized by using a trained deep neural network-based recognition model, and before determining whether there is a set type of internal accessories product of the telecommunications switching cabinet based on feature matching, it also includes: obtaining training images of different types of internal accessories products of the telecommunications switching cabinet, and marking the features of the different types of internal accessories products of the telecommunications switching cabinet; training the deep neural network-based recognition model through the training images to obtain a trained deep neural network-based recognition model.

[0034] In some embodiments, the pre-processed image is recognized by a trained recognition model based on a deep neural network, and whether there is a set type of internal accessories product for a telecommunications switching cabinet based on feature matching is determined, including: rasterizing the pre-processed image to divide the pre-processed image into multiple grid sub-images; detecting targets in the grid sub-images by a trained recognition model based on a deep neural network, and determining whether the targets are the set type of internal accessories product for a telecommunications switching cabinet based on feature matching.

[0035] In some embodiments, the target in the grid sub-image is detected by using a trained deep neural network-based recognition model, and whether the target is a set type of internal accessory product for a telecommunications switching cabinet is determined based on feature matching, including: setting a bounding box for the target in the grid sub-image by using a trained deep neural network-based recognition model, and extracting target features within the bounding box; determining whether the target features match features of a set type of internal accessory product for a telecommunications switching cabinet, and if so, determining that the target is the set type of internal accessory product for a telecommunications switching cabinet.

[0036] In some embodiments, if present, the set type of internal accessory products of the telecommunications switching cabinet are counted until the preprocessed image is traversed and the number of internal accessory products of the telecommunications switching cabinet of the set type is obtained, and it also includes: transmitting the number of internal accessory products of the telecommunications switching cabinet of the set type to the mobile terminal for real-time display and recording.

[0037] The intelligent recognition system accurately obtains the product type by processing and analyzing the product photos. First, the key frame image is grayed to remove some unnecessary color information, and then it is further processed by Gaussian filtering for noise reduction. If the shooting environment is too dark, the system will automatically correct the brightness of the photo to obtain more accurate detection results.

[0038] The recognition model can be a YOLOv5 model. In the overall detection process of the system, the product image is pre-processed to determine whether it meets the template requirements, and the images that do not meet the requirements of the model are upgraded. The images that meet the requirements are put into the improved YOLOv5 model for target detection. The model is mainly divided into three modules: Backbone module, Neck module, and Head module. The Backbone mainly includes: Conv convolution, BN layer, activation function, and the v6.2 version uses the Silu activation function. The purpose of Conv is to learn target features, the purpose of the BN layer is to prevent overfitting and facilitate model training, and the activation function is to introduce nonlinear factors to facilitate the model to better fit the target object.

[0039] Through the above operations, the image is divided into different feature sub-graphs and is responsible for extracting features from the image. The Neck module is placed between the Backbone module and the Head module. It fuses and outputs features through the feature pyramid network in order to better utilize the features extracted by the Backbone module. The Head module is the processing module after feature extraction. The function of this part is to classify and locate, and finally output information statistics.

[0040] In addition, including the loss function module, the loss function used by YOLOv5 is divided into three parts: BoundingBox loss function (box_loss), the improved YOLOv5 model uses the EIOU loss function optimized by the normal distribution function as the regression loss of the target box, box_loss is used to indicate the size and precise position of the target box; classification loss (cls_loss), calculates whether the predicted box and the corresponding category are correct; confidence loss (obj_loss): used to calculate the confidence of the target. The role of confidence: used to indicate the credibility of the predicted box, the value range is 0-1, the larger the value, the more likely the target is in the predicted box.

[0041] Each time the training is performed, the detection model is fed with processed image data of a specific size. The loss is accumulated through multiple iterations of forward propagation training, and this loss is used to perform a back propagation to update the gradient and continuously optimize the model parameters.

[0042] In the process of object detection, the input feature map is divided into a certain number of grids (N*N), and each grid detects the object that falls into it, and predicts the bounding box of the objects contained in all grids, the location confidence, and the probability of all categories at one time. Feature extraction and matching are performed through the backbone convolutional network, and finally non-maximum suppression processing is performed, and then objects can be detected in the image.

[0043] The embodiment of this application uses YOLOv5s as the core algorithm model. After targeted multiple improvements, optimizations and enhanced training, it can efficiently and accurately realize the recognition, calibration and counting of products. Its detection accuracy reaches 98.5%, which can meet the actual detection needs in production scenarios. The YOLOv5 algorithm model has reached a high level in terms of timeliness, accuracy, stability and other aspects.

[0044] According to the characteristics of high target concentration, high familiarity of single target features, and strong background interference in the switch cabinet detection, the backbone network, feature pyramid network, convolution layer, pooling layer, and loss function of the algorithm model are improved and optimized. In the training stage, highly consistent anchor frame calibration is used for targets with high concentration, global features with different performance characteristics are added for enhanced training, and the binary method is used to iteratively determine the optimal value of NMS parameters.

[0045] Adopting an adaptive preprocessing mechanism, in order to address the defects of poor algorithm model adaptability and the input of multi-size images in real scenes, an adaptive preprocessing mechanism is added to identify and pre-adjust input images of different sizes, and dynamically set the detection algorithm parameters according to the adjusted size. Taking into account the negative impact of the algorithm model detection defects on the system, the counting mechanism is optimized, and the same anchor frames are compared and verified based on the quick sorting algorithm, and the intersection threshold is set to reduce the error rate caused by false detection, missed detection and repeated counting.

[0046] The product intelligent identification method provided in the embodiment of the present application obtains an original image including an internal accessory product of a telecommunication switching cabinet and preprocesses the original image to obtain a preprocessed image, and then identifies the preprocessed image through a trained recognition model based on a deep neural network, and determines whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching. If so, the set type of internal accessory product of the telecommunication switching cabinet is counted until the preprocessed image is traversed, and finally the number of the set type of internal accessory product of the telecommunication switching cabinet is obtained. High-performance machine vision and deep neural network technology are used to automatically identify the type and quantity of the internal accessory product of the telecommunication switching cabinet. The operation is simple, and the efficiency of identifying the internal accessory product of the telecommunication switching cabinet can be greatly improved, the manual workload, human errors and arbitrariness can be reduced, and the intelligence and accuracy of identifying the internal accessory product of the telecommunication switching cabinet can be comprehensively improved.

[0047] Corresponding to the product intelligent identification method described in the above embodiment, Figure 2 As shown, the embodiment of the present application also provides a product intelligent identification system, and the product intelligent identification system 200 includes: The image acquisition module 201 is used to acquire the original image including the internal accessory products of the telecommunication switching cabinet, and pre-process the original image to obtain a pre-processed image; The product recognition module 202 is used to recognize the pre-processed image through a trained recognition model based on a deep neural network, and determine whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching; The product counting module 203 is used to count the internal accessory products of the telecommunication switching cabinet of the set type, if any, until the pre-processed images are traversed to obtain the number of the internal accessory products of the telecommunication switching cabinet of the set type.

[0048] It should be noted that the information interaction, execution process, etc. between the above-mentioned modules / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0049] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0050] The present application embodiment also provides an electronic device 300, such as Figure 3 As shown, it includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and executable on the processor 302. When the processor 302 executes the computer program 303, the steps of the product intelligent identification method provided in the first aspect are implemented.

[0051] In applications, the electronic device may include, but is not limited to, a processor and a memory. Figure 3 This is only an example of an electronic device and does not constitute a limitation on the electronic device. It may include more or fewer components than shown in the figure, or combine certain components, or different components, such as input and output devices, network access devices, etc. Input and output devices may include cameras, audio acquisition / playback devices, display screens, etc. Network access devices may include a network module for wirelessly communicating with external devices.

[0052] In applications, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0053] In applications, the memory may be an internal storage unit of an electronic device in some embodiments, such as a hard disk or memory of the electronic device. In other embodiments, the memory may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device. The memory may also include both an internal storage unit and an external storage device of the electronic device. The memory is used to store operating systems, applications, boot loaders, data, and other programs, such as program codes of computer programs. The memory may also be used to temporarily store data that has been output or is to be output.

[0054] The embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0055] The present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electrical carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0056] Those of ordinary skill in the art will appreciate that the devices and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0057] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices, which can be electrical, mechanical or other forms.

[0058] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such 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, and should all be included in the protection scope of the present application.

Claims

1. A product intelligent identification method, characterized in that: include: Acquire an original image including the internal accessory products of the telecommunication switching cabinet, and preprocess the original image to obtain a preprocessed image; The pre-processed image is recognized by a trained recognition model based on a deep neural network, and it is determined whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching; If so, the set type of telecommunication switching cabinet internal accessory products are counted until the pre-processed image is traversed to obtain the number of the set type of telecommunication switching cabinet internal accessory products.

2. The product intelligent identification method according to claim 1, characterized in that: The preprocessing of the original image to obtain a preprocessed image includes: Grayscale conversion, filtering, noise reduction and edge detection are performed on the original image to obtain a preprocessed image.

3. The product intelligent identification method according to claim 2, characterized in that: The filtering and noise reduction adopts Gaussian filtering, and the edge detection adopts sub-pixel edge extraction.

4. The product intelligent identification method according to claim 1, characterized in that: Before the pre-processed image is recognized by the trained recognition model based on the deep neural network and whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching, the method further includes: Acquire training images of different types of internal accessories of telecommunication switching cabinets, and mark features of the different types of internal accessories of telecommunication switching cabinets; The recognition model based on the deep neural network is trained using the training images to obtain a trained recognition model based on the deep neural network.

5. The product intelligent identification method according to claim 1, characterized in that: The pre-processed image is recognized by the trained recognition model based on the deep neural network, and whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching, including: Performing rasterization processing on the preprocessed image to divide the preprocessed image into a plurality of raster sub-images; The target in the grid sub-image is detected by using a trained recognition model based on a deep neural network, and it is determined based on feature matching whether the target is a set type of internal accessory product of a telecommunication switching cabinet.

6. The product intelligent identification method according to claim 5, characterized in that: The target in the grid sub-image is detected by using a trained recognition model based on a deep neural network, and whether the target is a set type of internal accessory product of a telecommunication switching cabinet is determined based on feature matching, including: Setting a bounding box for the target in the grid sub-image through a trained recognition model based on a deep neural network, and extracting target features within the bounding box; It is determined whether the target feature matches the feature of a set type of internal accessory product of a telecommunication switching cabinet. If they match, it is determined that the target is a set type of internal accessory product of a telecommunication switching cabinet.

7. The product intelligent identification method according to claim 1, characterized in that: If so, the set type of telecommunication switch cabinet internal accessory products are counted until the pre-processed image is traversed to obtain the number of the set type of telecommunication switch cabinet internal accessory products, and further includes: The quantity of the internal accessory products of the set type of the telecommunication switching cabinet is transmitted to the mobile terminal for real-time display and recording.

8. A product intelligent identification system, characterized in that: include: An image acquisition module is used to acquire an original image including the internal accessory products of the telecommunication switching cabinet, and preprocess the original image to obtain a preprocessed image; A product recognition module is used to recognize the pre-processed image through a trained recognition model based on a deep neural network, and determine whether there is a set type of internal accessory product of the telecommunication switching cabinet based on feature matching; The product counting module is used to count the internal accessory products of the telecommunication switching cabinet of the set type, if any, until the pre-processed image is traversed to obtain the number of the internal accessory products of the telecommunication switching cabinet of the set type.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the product intelligent identification method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the product intelligent identification method according to any one of claims 1 to 7 is implemented.

Images