Live broadcast product attribute identification method, live broadcast machine, live broadcast system and medium

Abstract

The invention provides a live broadcast product attribute identification method, a live broadcast machine, a live broadcast system and a medium, and relates to the technical field of product identification, the method comprises the following steps: a controller obtains a target identification model; in the live broadcast process, the spectrum camera obtains an actual spectrum image of the live broadcast product and sends the actual spectrum image to the controller; the controller inputs the actual spectral image into a target recognition model to obtain category information; acquiring a corresponding standard spectrum from the cloud equipment according to the category information; and determining the current attribute of the live broadcast product according to the standard spectrum and the actual spectrum image, and sending the current attribute to the user terminal. According to the invention, the current attribute of the live product is visually displayed, and the purchase convenience and purchase efficiency of the user are improved.

Description

Technical Field

[0001] The present invention relates to the field of product identification technology, and in particular to a method for identifying the attributes of a live broadcast product, a live broadcast machine, a live broadcast system and a medium. Background Art

[0002] At present, more and more users tend to shop through online live streaming. For some agricultural products, users cannot intuitively understand the properties of the products such as maturity or freshness, which brings great inconvenience to users' purchases. Summary of the invention

[0003] In view of this, the purpose of the present invention is to overcome the deficiencies in the prior art and provide a method for identifying the attributes of a live broadcast product, a live broadcast machine, a live broadcast system and a medium. The present invention provides the following technical solutions:

[0004] In a first aspect, the present invention provides a method for identifying attributes of a live broadcast product, which is applied to a live broadcast machine, wherein the live broadcast machine includes a spectral camera and a controller, and the method includes: the controller obtains a target recognition model; during the live broadcast process, the spectral camera obtains an actual spectral image of the live broadcast product, and sends the actual spectral image to the controller; the controller inputs the actual spectral image into the target recognition model to obtain category information; obtains a corresponding standard spectrum from a cloud device according to the category information; determines the current attributes of the live broadcast product according to the standard spectrum and the actual spectral image, and sends the current attributes to a user terminal.

[0005] In one embodiment, the standard spectrum includes: a plurality of different spectral colors and a product characteristic score corresponding to each of the spectral colors, and determining the current attribute of the live product based on the standard spectrum and the actual spectral image includes: determining the actual spectral color corresponding to the live product based on the actual spectral image; traversing the standard spectrum based on the actual spectral color, and determining the corresponding product characteristic score as the target score; and determining the current attribute based on the target score.

[0006] In one embodiment, the live broadcast product includes a fresh product, the current attribute includes a current freshness and / or a current maturity, and determining the current attribute based on the target score includes: determining a preset score interval corresponding to the target score based on the target score; and determining a preset attribute corresponding to the preset score interval as the current attribute.

[0007] In one embodiment, obtaining a target recognition model includes: obtaining a product data set, the product data set including actual spectral images of a plurality of different products and category information corresponding to each of the actual spectral images; dividing the product data set into a training set, a test set, and a validation set according to a preset ratio; training a preset recognition model based on the training set, the test set, and the validation set to obtain the target recognition model.

[0008] In one embodiment, before the controller inputs the actual spectrum image into the target recognition model, the controller further includes: performing denoising processing on the actual spectrum image.

[0009] In one embodiment, the method also includes: the controller obtains an initial display image, the initial display image includes an RGB image of the live broadcast product; obtains camera identification information of the spectral camera, determines a target display image based on the initial display image and the camera identification information, and sends the target display image to the user terminal.

[0010] In one embodiment, determining the target live broadcast screen according to the initial display image and the camera identification information includes: adding the camera identification information within a preset area range of the initial display image to obtain the target display image.

[0011] In a second aspect, the present invention provides a live broadcast machine, comprising: a spectral camera and a controller; the controller is used to obtain a target recognition model; during the live broadcast process, the spectral camera is used to obtain an actual spectral image of the live broadcast product, and send the actual spectral image to the controller; the controller is also used to input the actual spectral image into the target recognition model to obtain category information; obtain a corresponding standard spectrum from a cloud device based on the category information; determine the current attributes of the live broadcast product based on the standard spectrum and the actual spectral image, and send the current attributes to a user terminal.

[0012] In a third aspect, the present invention provides a live broadcast system, the system comprising: a cloud device, a user terminal and the live broadcast machine described in the second aspect, the cloud device and the user terminal are respectively communicatively connected to the live broadcast machine.

[0013] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the attribute identification method of the live broadcast product described in the first aspect.

[0014] The present application provides a method for identifying the attributes of a live broadcast product, a live broadcast machine, a live broadcast system and a medium, the method comprising: the controller acquires a target recognition model; during the live broadcast, the spectral camera acquires an actual spectral image of the live broadcast product and sends the actual spectral image to the controller; the controller inputs the actual spectral image into the target recognition model to obtain category information; obtains a corresponding standard spectrum from a cloud device according to the category information; determines the current attributes of the live broadcast product according to the standard spectrum and the actual spectral image, and sends the current attributes to a user terminal. The present invention realizes an intuitive display of the current attributes of the live broadcast product, improving the user's purchase convenience and efficiency.

[0015] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A schematic diagram of a process of a method for identifying attributes of a live broadcast product provided in an embodiment of the present application is shown;

[0018] Figure 2 A schematic diagram of the structure of a live broadcast machine provided in an embodiment of the present application is shown;

[0019] Figure 3 Another schematic diagram of the process of identifying the attributes of a live broadcast product provided by an embodiment of the present application is shown;

[0020] Figure 4 Another schematic diagram of the process of the attribute identification method of the live broadcast product provided in the embodiment of the present application is shown;

[0021] Figure 5 A structural schematic diagram of a live broadcast system provided in an embodiment of the present application is shown.

[0022] Description of main component symbols:

[0023] 200-live broadcast machine; 210-spectral camera; 220-controller; 500-live broadcast system; 510-cloud device; 520-user terminal. DETAILED DESCRIPTION

[0024] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0025] 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. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of the template are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0027] Example 1

[0028] At present, online live shopping has become one of the mainstream shopping methods. The usual form of online live shopping is: using ordinary cameras to obtain product images and send them to user terminals, and introducing the products in combination with explanations. This shopping method has the following disadvantages: for fresh products, such as agricultural products, users cannot intuitively understand the properties of the products such as maturity or freshness. For this, please refer to Figure 1 The present application embodiment provides a method for identifying the attributes of a live broadcast product, which is applied to Figure 2 The live broadcast machine 200 shown includes: a spectral camera 210 and a controller 220, and the method includes steps S110 to S140.

[0029] In step S110, the controller obtains a target recognition model.

[0030] In this embodiment, the target recognition model is used to determine the category information of the live broadcast product based on the actual spectral image of the live broadcast product, and then determine the standard spectrum corresponding to the live broadcast product based on the category information corresponding to the live broadcast product, so as to further determine the attributes of the live broadcast product. It should be noted that the target recognition model is obtained by pre-training a preset recognition model. It can be understood that in another embodiment, the controller can also directly obtain the category information corresponding to the live broadcast product through the human-computer interaction device.

[0031] In one embodiment, see Figure 3The acquisition of the target recognition model described in step S110 includes: steps S111 to S113.

[0032] Step S111 : acquiring a product data set, wherein the product data set includes actual images of a plurality of different products and category information corresponding to each of the actual spectral images.

[0033] It should be noted that the various products include various fresh products, such as corn, cabbage, grapes, etc. The preset recognition model is trained according to the actual spectral images corresponding to the different products and the category information corresponding to each actual spectral image, so as to obtain a target recognition model for identifying the category information of the live broadcast products.

[0034] Step S112, dividing the product data set into a training set, a test set and a validation set according to a preset ratio.

[0035] In this embodiment, the product data set is divided into a training set, a test set, and a validation set according to a preset ratio of 8:1:1. Specifically, the preset division ratio can be determined according to actual conditions and is not limited here.

[0036] Step S113, training the preset recognition model based on the training set, the test set and the verification set to obtain the target recognition model.

[0037] It should be noted that the main function of the training set is to allow the preset recognition model to learn and identify the patterns and rules of the data in the product dataset. The test set is used to evaluate the model performance, and the validation set is used to help select the best module to obtain the target recognition model.

[0038] Step S120: During the live broadcast process, the spectral camera obtains an actual spectral image of the live broadcast product and sends the actual spectral image to the controller.

[0039] During the live broadcast process, the controller obtains the live broadcast screen and also obtains the actual spectral image of the live broadcast product through the spectral camera, so as to determine the current properties of the live broadcast product in real time according to the actual spectral image during the live broadcast process.

[0040] In step S130 , the controller inputs the actual spectrum image into the target recognition model to obtain category information.

[0041] The target recognition model is used to perform category recognition on the actual spectral image to obtain category information corresponding to the actual spectral image. It can be understood that the category information includes product classification information. For example, the actual spectral image corresponding to corn is input into the target recognition model, and the corresponding category information can be determined to be corn. Since different categories of products have different corresponding standard spectra, before obtaining the standard spectrum corresponding to the live broadcast product, it is necessary to first determine the category information corresponding to the live broadcast product, so as to obtain the standard spectrum corresponding to the live broadcast product from the cloud device according to the category information.

[0042] In one embodiment, before the controller inputs the actual spectrum image into the target recognition model, the controller further includes: performing denoising processing on the actual spectrum image.

[0043] In this embodiment, in order to improve the image quality of the actual spectral image and improve the accuracy of the output result of the target recognition model, the actual spectral image is denoised by methods such as mean filtering, median filtering or Gaussian filtering to remove noise in the image.

[0044] In step S140 , the controller obtains a corresponding standard spectrum from a cloud device according to the category information.

[0045] It can be understood that the cloud device pre-stores a plurality of different category information and standard spectra corresponding to each category information. According to the category information corresponding to the live broadcast product, the data pre-stored in the cloud device is traversed to obtain the corresponding standard spectrum.

[0046] Step S150: The controller determines the current attributes of the live broadcast product according to the standard spectrum and the actual spectrum image, and sends the current attributes to the user terminal.

[0047] In one embodiment, the standard spectrum includes: a plurality of different spectral colors and product characteristic scores corresponding to each of the spectral colors, see Figure 4 , the step S150 of determining the current attribute of the live broadcast product according to the standard spectrum and the actual spectrum image includes: steps S151 to S153.

[0048] Step S151, determining the actual spectral color corresponding to the live broadcast product according to the actual spectral image.

[0049] It can be understood that different categories of live broadcast products have corresponding actual spectral colors that are different, and live broadcast products of the same category have corresponding actual spectral colors that are different due to different current attributes.

[0050] Step S152: traverse the standard spectrum according to the actual spectrum color, and determine the corresponding product characteristic score as the target score.

[0051] The standard spectrum includes a variety of different spectral colors. Different spectral colors correspond to different scores. For example, purple corresponds to a score below 85 points, yellow corresponds to a score between 85 and 90 points, orange corresponds to a score between 90 and 95 points, and green corresponds to a score between 95 and 100 points. Different scores represent different current attributes.

[0052] Step S153: determining the current attribute according to the target score.

[0053] In this embodiment, the standard spectrum is traversed according to the actual spectrum color to obtain the target score of the live broadcast product, and the corresponding current attribute can be determined according to the preset score interval to which the target score belongs.

[0054] In one embodiment, the live broadcast product includes a fresh product, the current attribute includes a current freshness and / or a current maturity, and determining the current attribute based on the target score includes: determining a preset score interval corresponding to the target score based on the target score; and determining a preset attribute corresponding to the preset score interval as the current attribute.

[0055] In this embodiment, for example, for corn, a score below 85 indicates immaturity, a score between 85 and 90 indicates slightly mature, a score between 90 and 95 indicates good maturity, and a score between 95 and 100 indicates overmatureness. In another embodiment, the actual spectrum color and the standard spectrum may be directly sent to the user terminal, and the user may judge the current attribute by comparing the actual spectrum color with the standard spectrum. In another real-time method, the target score may be directly sent to the user terminal, and the user may intuitively understand the current attribute of the live broadcast product based on the target score.

[0056] In one embodiment, the method also includes: the controller obtains an initial display image, the initial display image includes an RGB image of the live broadcast product; obtains camera identification information of the spectral camera, determines a target display image based on the initial display image and the camera identification information, and sends the target display image to the user terminal.

[0057] In this embodiment, the device used to obtain the actual spectral image of the live broadcast product is a spectral camera. In order to improve the credibility of the current attributes, while sending the live broadcast picture to the user, the spectral camera identification information will be added in the preset area of ​​the live broadcast picture. The identification information is the officially authorized certification mark and the product information of the corresponding spectral camera.

[0058] In one embodiment, determining the target live broadcast screen according to the initial display image and the camera identification information includes: adding the camera identification information within a preset area range of the initial display image to obtain the target display image.

[0059] The preset area range may be any area on the initial display image that the user can intuitively understand, and may be specifically selected based on actual conditions.

[0060] The attribute identification method of the live broadcast product provided in the embodiment of the present application obtains the target identification model through the controller; during the live broadcast process, the spectral camera obtains the actual spectral image of the live broadcast product, and sends the actual spectral image to the controller; the controller inputs the actual spectral image into the target identification model to obtain category information; obtains the corresponding standard spectrum from the cloud device according to the category information; determines the current attribute of the live broadcast product according to the standard spectrum and the actual spectral image, and sends the current attribute to the user terminal, thereby realizing the intuitive display of the current attribute of the live broadcast product and improving the user's purchase convenience and efficiency.

[0061] Example 2

[0062] Also, see again Figure 2 , the embodiment of the present application also provides a live broadcast machine 200, including: a spectrum camera 210 and a controller 220;

[0063] The controller 220 is used to obtain a target recognition model;

[0064] During the live broadcast process, the spectral camera 210 is used to obtain an actual spectral image of the live broadcast product and send the actual spectral image to the controller 220;

[0065] The controller 220 is also used to input the actual spectrum image into the target recognition model to obtain category information; obtain the corresponding standard spectrum from the cloud device according to the category information; determine the current attributes of the live broadcast product according to the standard spectrum and the actual spectrum image, and send the current attributes to the user terminal.

[0066] In one embodiment, the standard spectrum includes: a plurality of different spectral colors and a product characteristic score corresponding to each of the spectral colors. The controller 220 is further used to determine the actual spectral color corresponding to the live product according to the actual spectral image; traverse the standard spectrum according to the actual spectral color, and determine the corresponding product characteristic score as the target score; and determine the current attribute according to the target score.

[0067] In one embodiment, the live broadcast product includes a fresh product, the current attribute includes a current freshness and / or a current maturity, and the controller 220 is further used to determine a preset score interval corresponding to the target score based on the target score; and determine a preset attribute corresponding to the preset score interval as the current attribute.

[0068] In one embodiment, the controller 220 is also used to obtain a product data set, which includes actual spectral images of a variety of different products and category information corresponding to each of the actual spectral images; divide the product data set into a training set, a test set and a verification set according to a preset ratio; and train a preset recognition model based on the training set, the test set and the verification set to obtain the target recognition model.

[0069] In one implementation, the controller 220 is further configured to perform denoising on the actual spectral image.

[0070] In one embodiment, the controller 220 is also used to obtain an initial display image, which includes an RGB image of the live broadcast product; obtain camera identification information of the spectral camera, determine a target display image based on the initial display image and the camera identification information, and send the target display image to the user terminal.

[0071] The live broadcast machine 200 provided in the embodiment of the present application can execute the attribute identification method of the live broadcast product provided in the above-mentioned method embodiment 1, and will not be repeated here to avoid repetition.

[0072] The live broadcast machine provided in the embodiment of the present application obtains the target recognition model through the controller; during the live broadcast process, the spectral camera obtains the actual spectral image of the live broadcast product, and sends the actual spectral image to the controller; the controller inputs the actual spectral image into the target recognition model to obtain category information; obtains the corresponding standard spectrum from the cloud device according to the category information; determines the current attributes of the live broadcast product according to the standard spectrum and the actual spectral image, and sends the current attributes to the user terminal, thereby intuitively displaying the current attributes of the live broadcast product, improving the user's purchase convenience and purchase efficiency, and intuitively displaying the current attributes of the live broadcast product, improving the user's purchase convenience and purchase efficiency.

[0073] Example 3

[0074] Also, see Figure 5, the embodiment of the present application provides a live broadcast system 500, including: a cloud device 510, a user terminal 520 and a live broadcast machine 200 as described in Example 2, the cloud device 510 and the user terminal 520 are respectively connected to the live broadcast machine 200. Specifically, the live broadcast machine 200 includes: a spectral camera 210 and a controller 220; the controller 220 is used to obtain a target recognition model; during the live broadcast process, the spectral camera 210 is used to obtain an actual spectral image of the live broadcast product, and send the actual spectral image to the controller 220; the controller 220 is also used to input the actual spectral image into the target recognition model to obtain category information; obtain the corresponding standard spectrum from the cloud device according to the category information; determine the current attributes of the live broadcast product according to the standard spectrum and the actual spectral image, and send the current attributes to the user terminal.

[0075] The live broadcast system 500 provided in the embodiment of the present invention can realize the functions of the live broadcast machine 200 provided in the above-mentioned embodiment 2, and will not be described again here to avoid repetition.

[0076] Example 4

[0077] In addition, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the attribute identification method of the live broadcast product provided in Example 1 is implemented.

[0078] In this embodiment, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0079] The computer-readable storage medium provided in this embodiment can implement the attribute identification method of the live broadcast product provided in Example 1. To avoid repetition, it will not be repeated here.

[0080] In all examples shown and described herein, any specific values ​​should be interpreted as merely exemplary and not as limiting, and thus other examples of the exemplary embodiments may have different values.

[0081] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0082] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for identifying the attributes of a live broadcast product, characterized in that: Applied to a live broadcast machine, the live broadcast machine includes a spectral camera and a controller, and the method includes: The controller acquires a target recognition model; During the live broadcast process, the spectral camera obtains an actual spectral image of the live broadcast product and sends the actual spectral image to the controller; The controller inputs the actual spectrum image into the target recognition model to obtain category information; obtains the corresponding standard spectrum from the cloud device according to the category information; determines the current attributes of the live broadcast product according to the standard spectrum and the actual spectrum image, and sends the current attributes to the user terminal.

2. The attribute identification method of a live broadcast product according to claim 1, characterized in that: The standard spectrum includes: a plurality of different spectrum colors and product characteristic scores corresponding to each of the spectrum colors. The current attribute of the live broadcast product is determined based on the standard spectrum and the actual spectrum image, including: Determine the actual spectrum color corresponding to the live broadcast product according to the actual spectrum image; Traversing the standard spectrum according to the actual spectrum color, and determining the corresponding product characteristic score as a target score; The current attribute is determined according to the target score.

3. The attribute identification method of the live broadcast product according to claim 2 is characterized in that: The live broadcast product includes a fresh product, the current attribute includes current freshness and / or current maturity, and determining the current attribute according to the target score includes: Determine a preset score range corresponding to the target score according to the target score; The preset attribute corresponding to the preset score interval is determined as the current attribute.

4. The attribute identification method of a live broadcast product according to claim 1, characterized in that: The acquiring of the target recognition model comprises: Acquire a product data set, wherein the product data set includes actual spectral images of a plurality of different products and category information corresponding to each of the actual spectral images; Dividing the product data set into a training set, a test set and a validation set according to a preset ratio; The preset recognition model is trained based on the training set, the test set and the verification set to obtain the target recognition model.

5. The attribute identification method of the live broadcast product according to claim 4 is characterized in that: Before the controller inputs the actual spectrum image into the target recognition model, the controller further includes: The actual spectrum image is subjected to denoising processing.

6. The attribute identification method of a live broadcast product according to any one of claims 1 to 5, characterized in that: The method further comprises: The controller acquires an initial display image, wherein the initial display image includes an RGB image of the live broadcast product; The camera identification information of the spectral camera is obtained, a target display image is determined according to an initial display image and the camera identification information, and the target display image is sent to the user terminal.

7. The attribute identification method of a live broadcast product according to claim 6, characterized in that: The determining of the target live broadcast screen according to the initial display image and the camera identification information includes: adding the camera identification information within a preset area range of the initial display image to obtain the target display image.

8. A live broadcast machine, characterized in that: include: Spectral camera and controller; The controller is used to obtain a target recognition model; During the live broadcast process, the spectral camera is used to obtain an actual spectral image of the live broadcast product and send the actual spectral image to the controller; The controller is further configured to input the actual spectrum image into the target recognition model to obtain category information; and obtain a corresponding standard spectrum from a cloud device according to the category information; The current attribute of the live broadcast product is determined according to the standard spectrum and the actual spectrum image, and the current attribute is sent to the user terminal.

9. A live broadcast system, characterized in that: The system includes: a cloud device, a user terminal and the live broadcast machine as described in claim 8, and the cloud device and the user terminal are respectively communicatively connected to the live broadcast machine.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the attribute identification method of the live broadcast product described in any one of claims 1-7 is implemented.

Images