An algorithm for automatically generating product display images based on electronic scales

Abstract

The present invention discloses an algorithm for automatically generating a commodity display image based on an electronic scale, comprising the following steps: S1, placing the commodity at the set position of the electronic scale; S2, identifying the commodity and determining whether there is a corresponding commodity display image, if yes, displaying the corresponding commodity display image, if not, executing step S3; S3, collecting an image of the commodity and performing quality analysis on the image, if the quality is qualified, executing step S4, if the quality is unqualified, ending; S4, generating a commodity display image after computing and processing the image. The present invention can help operators quickly find commodities through commodity display images, with high accuracy and no need for manual matching of images, saving a lot of labor costs, and at the same time, no longer relying on the cloud, no need for AI chips, and no need for edge servers, improving cashier efficiency without increasing any equipment investment.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic scales, and in particular to an algorithm for automatically generating a commodity display image based on an electronic scale. Background Art

[0002] The product display picture is a very important function in the weighing and checkout process. Because some operators are illiterate, it is a very important ability for operators to quickly select products through the product display picture.

[0003] The traditional solution now is that IT operation and maintenance personnel configure pictures for each product in the cloud. Most of these pictures are downloaded from the Internet, which has copyright risks and requires professional maintenance, which is very costly. Summary of the invention

[0004] Based on this, it is necessary to provide an algorithm for automatically generating product display images based on electronic scales with low cost, high accuracy and high checkout efficiency to address the above technical problems.

[0005] An algorithm for automatically generating a product display image based on an electronic scale includes the following steps:

[0006] S1. Place the product at the set position of the electronic scale;

[0007] S2, identifying the product and determining whether there is a corresponding product display picture, if yes, displaying the corresponding product display picture, if no, executing step S3;

[0008] S3, collecting the image of the product and performing quality analysis on the image. If the quality is qualified, executing step S4; if the quality is unqualified, ending;

[0009] S4. Generate a product display image after performing computational processing on the image.

[0010] In one embodiment, the electronic scale in step S1 includes:

[0011] Pedestal;

[0012] A tray is mounted on the base, and a weight sensor is provided at the bottom of the tray, and the weight sensor can detect the weight of the goods in the tray;

[0013] A columnar mounting frame is fixed on the side of the base, a coding machine is installed on the columnar mounting frame, and the coding machine is connected to a display screen;

[0014] A camera, wherein the camera is rotatably mounted on the columnar mounting frame, and a camera head of the camera corresponds to the tray;

[0015] The embedded mainboard is connected with the weight sensor, the coding machine, the display screen and the camera respectively.

[0016] In one embodiment, in step S2, identifying the product and determining whether there is a corresponding product display image includes:

[0017] S21, collecting pictures of the goods in the tray;

[0018] S22, comparing the product image with the product display images in the sample library to determine the similarity;

[0019] S23. When the similarity is greater than a set range, it is determined that the product has a corresponding product display picture.

[0020] In one of the embodiments, in step S22, the comparison of the product image with the product display image in the sample library includes: comparing the shape, color, size and surface texture of the product.

[0021] In one embodiment, in step S3, collecting the image of the commodity and performing quality analysis on the image includes:

[0022] S31, taking photos of the product from multiple angles to obtain images;

[0023] S32: Screen all the pictures to find the pictures with the highest definition.

[0024] In one embodiment, the step S4 comprises:

[0025] S41, extracting the feature values ​​{x of the foreground image and the background image from the screened image i ,y i};

[0026] S42, enlarging the feature values ​​of the foreground image and the background image into {x i ',y i '};

[0027] S43, calculate the mask after a layer of full connection;

[0028] S44, obtaining image I by weighted summation;

[0029] S45, multiplying the image I and the mask to obtain a final generated image I′;

[0030] S46: Use the image I' as a product display image.

[0031] In one embodiment, in step S45, the calculation formula of image I′ is as follows:

[0032] I=α·x i ′+β·y i '

[0033] I′=I·mask

[0034] Among them, α and β represent weighting coefficients.

[0035] In one embodiment, the method for improving the accuracy of the product display picture includes: when the foreground of the picture is simple, the weighting coefficient β increases in value; when the background of the picture is simple, the weighting coefficient α increases in value.

[0036] The above algorithm for automatically generating product display pictures based on electronic scales can not only automatically identify products and determine whether there are corresponding product display pictures, but also collect images of products and generate corresponding product display pictures after processing. It can help operators quickly find products through product display pictures with high accuracy and no longer requires manual matching of pictures, saving a lot of labor costs. At the same time, it no longer relies on the cloud, does not require AI chips, and does not require edge servers, thus improving checkout efficiency without increasing any equipment investment. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 It is a structural schematic diagram of an electronic scale of the present invention;

[0039] Figure 2 The present invention is a flowchart of an algorithm for automatically generating a commodity display image based on an electronic scale. DETAILED DESCRIPTION

[0040] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Preferred embodiments of the present invention are provided in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present invention more thorough and comprehensive.

[0041] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only and do not represent the only implementation method.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention 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 related listed items.

[0043] See also Figure 1-2 As shown, an embodiment of the present invention provides an algorithm for automatically generating a commodity display image based on an electronic scale, comprising the following steps:

[0044] S1. Place the product at the set position of the electronic scale. This facilitates the subsequent image capture of the product and improves the accuracy of image capture of the product. In this embodiment, the product can be placed on the tray of the electronic scale and correspond to the camera position of the camera 6.

[0045] S2. Identify the product and determine whether there is a corresponding product display picture. If yes, display the corresponding product display picture. If not, execute step S3. In this embodiment, displaying the corresponding product display picture is to facilitate the operator's re-checking and confirmation.

[0046] S3, collecting the image of the product and performing quality analysis on the image. If the quality is qualified, executing step S4; if the quality is unqualified, ending;

[0047] S4. Generate a product display image after performing computational processing on the image.

[0048] The above algorithm for automatically generating product display pictures based on electronic scales can not only automatically identify products and determine whether there are corresponding product display pictures, but also collect images of products and generate corresponding product display pictures after processing. It can help operators quickly find products through product display pictures with high accuracy and no longer requires manual matching of pictures, saving a lot of labor costs. At the same time, it no longer relies on the cloud, does not require AI chips, and does not require edge servers, thus improving checkout efficiency without increasing any equipment investment.

[0049] In one embodiment of the present invention, the electronic scale in step S1 includes:

[0050] Base 2;

[0051] A tray 1 is mounted on the base, and a weight sensor is provided at the bottom of the tray, and the weight sensor can detect the weight of the goods in the tray 1;

[0052] A columnar mounting frame 3 is fixed on the side of the base 2. A coding machine 4 is installed on the columnar mounting frame 3. The coding machine 4 is connected to a display screen 5. The display screen 5 is mainly used to display a product display picture of the current product.

[0053] Camera 6, the camera 6 is rotatably mounted and connected to the columnar mounting frame 3, and the camera head of the camera 6 corresponds to the tray 1; in this embodiment, the camera 6 is mainly used to capture images of commodities, and by setting the camera 6 to be rotatable, images of commodities can be captured from multiple angles, which is convenient for subsequent picture screening.

[0054] The embedded mainboard is connected to the weight sensor, the coding machine 4, the display screen 5 and the camera 6 respectively.

[0055] In this embodiment, the weight sensor can detect the weight data of the goods and send it to the embedded mainboard. After receiving the weight data, the embedded mainboard processes it to generate sales data, and the embedded mainboard can control the coding machine 4 to print out the sales data.

[0056] In an embodiment of the present invention, in step S2, identifying the product and determining whether there is a corresponding product display image includes:

[0057] S21, collecting pictures of the goods in the tray 1;

[0058] S22, comparing the product image with the product display images in the sample library to determine the similarity;

[0059] S23, when the similarity is greater than the set range, it is determined that the product has a corresponding product display picture. For example, when the similarity is greater than 95%, it can be directly determined that the product has a corresponding product display picture.

[0060] In this embodiment, when identifying goods, the requirements for image collection are not high. The camera 6 only needs to capture the image of the goods in the tray 1 directly below. This not only reduces the waste of memory and processing resources in the embedded motherboard, but also facilitates improving the efficiency of comparison with the product display pictures in the sample library.

[0061] In one embodiment of the present invention, in step S22, the comparison of the product image with the product display image in the sample library includes: comparing the shape, color, size and surface texture of the product. In this embodiment, the surface texture of the product is first identified to determine whether it is a product of the same type, and then, based on the size, color and size of the product, it is determined whether it is a product of the same specification (the prices of different specifications may be different, which will affect the final consumption amount), thereby improving the accuracy of product recognition.

[0062] In an embodiment of the present invention, in step S3, collecting the image of the commodity and performing quality analysis on the image includes:

[0063] S31, taking photos of the product at multiple angles to obtain images; for example, the camera head of the camera 6 takes photos at multiple angles such as directly below, forward (30-60°), and backward (30-60°).

[0064] S32: Screen all the pictures to find the pictures with the highest definition.

[0065] In this embodiment, the purpose of taking pictures of the product at multiple angles is to select multiple angles for presenting products of different shapes to obtain the best picture display effect. In addition, the interference of external environment such as light can be reduced, thereby improving the accuracy of product picture collection.

[0066] In one embodiment of the present invention, step S4 includes:

[0067] S41, extracting the feature values ​​{x of the foreground image and the background image from the screened image i ,y i};

[0068] S42, enlarging the feature values ​​of the foreground image and the background image into {x i ',y i '};

[0069] S43, calculate the mask through a layer of full connection;

[0070] S44, obtaining image I by weighted summation;

[0071] S45, multiplying the image I and the mask to obtain a final generated image I′;

[0072] S46: Use the image I' as a product display image.

[0073] In this embodiment, the selected images are passed through the CNN network EfficientNet to extract the feature values ​​of the foreground image and the background image {x i ,y i}, and enlarge the two feature maps into {x i ',y i '}, then, execute steps S43-S45 in sequence to obtain the final image I', so that the generated image is smoother and more beautiful, so it is convenient to use in actual situations.

[0074] In one embodiment of the present invention, in step S45, the calculation formula of the image I′ is as follows:

[0075] I=α·x i ′+β·y i '

[0076] I′=I·mask

[0077] Among them, α and β represent weighting coefficients.

[0078] In one embodiment of the present invention, the method for improving the accuracy of the product display picture includes: when the foreground of the picture is simple, the weighting coefficient β increases in value; when the background of the picture is simple, the weighting coefficient α increases in value.

[0079] In summary, the present invention has the following advantages:

[0080] 1) No AI chip, chip module or AI server is required. It is completely based on the motherboard of the scale itself, which has obvious cost advantages.

[0081] 2) No human supervision is required, the algorithm automatically generates display pictures that belong to the product through video streams.

[0082] 3) This algorithm is based on the deep learning CNN network EfficientNet, and further tailored to the algorithm so that it can operate efficiently on ordinary X86 chips.

[0083] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The above-described embodiments only express several implementation methods of the present application, but they cannot be understood as limiting the scope of the present application. 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 application, which all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the attached claims.

Claims

1. An algorithm for automatically generating product display images based on electronic scales, It is characterized in that The following steps are involved: S1. Place the product at the set position of the electronic scale; S2, identifying the product and determining whether there is a corresponding product display picture, if yes, displaying the corresponding product display picture, if no, executing step S3; S3, collecting the image of the product and performing quality analysis on the image. If the quality is qualified, executing step S4; if the quality is unqualified, ending; S4, generating a product display image after computing and processing the image; The step S4 comprises: S41, extracting the feature values ​​{x of the foreground image and the background image from the screened image i ,y i }; S42, enlarging the feature values ​​of the foreground image and the background image into {x i ',y i '}; S43, calculate the mask after a layer of full connection; S44, obtaining image I by weighted summation; S45, multiplying the image I and the mask to obtain a final generated image I′; S46: Use the image I' as a product display image.

2. The algorithm for automatically generating a product display image based on an electronic scale as claimed in claim 1, It is characterized in that The electronic scale in step S1 comprises: Pedestal; A tray is mounted on the base, and a weight sensor is provided at the bottom of the tray, and the weight sensor can detect the weight of the goods in the tray; A columnar mounting frame is fixed on the side of the base, a coding machine is installed on the columnar mounting frame, and the coding machine is connected to a display screen; A camera, wherein the camera is rotatably mounted on the columnar mounting frame, and a camera head of the camera corresponds to the tray; The embedded mainboard is connected with the weight sensor, the coding machine, the display screen and the camera respectively.

3. The algorithm for automatically generating a product display image based on an electronic scale as claimed in claim 2, It is characterized in that In step S2, identifying the product and determining whether there is a corresponding product display image includes: S21, collecting pictures of the goods in the tray; S22, comparing the product image with the product display images in the sample library to determine the similarity; S23. When the similarity is greater than a set range, it is determined that the product has a corresponding product display picture.

4. The algorithm for automatically generating a product display image based on an electronic scale as claimed in claim 3, It is characterized in that In the step S22, the comparison of the product image with the product display image in the sample library includes: comparing the shape, color, size and surface texture of the product.

5. The algorithm for automatically generating a product display image based on an electronic scale as claimed in claim 1, It is characterized in that In step S3, collecting the image of the commodity and performing quality analysis on the image includes: S31, taking photos of the product from multiple angles to obtain images; S32: Screen all the pictures to find the pictures with the highest definition.

6. The algorithm for automatically generating a product display image based on an electronic scale as claimed in claim 5, It is characterized in that In step S45, the calculation formula of image I' is as follows: I=α·x i ′+β·y i ′ I′=I·mask Among them, α and β represent weighting coefficients.

7. The algorithm for automatically generating a product display image based on an electronic scale as claimed in claim 6, It is characterized in that The method for improving the accuracy of the commodity display picture includes: when the foreground of the picture is simple, the weighting coefficient β increases in value; when the background of the picture is simple, the weighting coefficient α increases in value.

Images