Commodity picture generation method and terminal

Abstract

The invention discloses a commodity picture generation method and a terminal, and the method comprises the steps: obtaining package information which comprises commodity identifiers and the number of commodities corresponding to each commodity identifier; constructing a single commodity picture according to the commodity identifier and the commodity quantity; obtaining a splicing template according to the identification number of the commodity identification, wherein the splicing template comprises splicing positions and priority layer information of each splicing position; and randomly obtaining a preset background image, randomly splicing each single commodity image on the splicing position according to the priority layer information, and then superposing the single commodity images on the background image to obtain a package image. The template comprises the position where the single commodity picture can be placed and the priority level between the positions, the priority level is the shielding relation, the real shooting environment can be better simulated, the picture is more natural, the whole splicing process can be automatically completed except for the set package information, and the splicing efficiency is improved. And manual participation is not needed.

Description

Technical Field

[0001] The present invention relates to the field of commodity management, and in particular to a commodity image generation method and terminal. Background Art

[0002] In the current sales method, in addition to directly selling single items, merchants usually sell goods in the form of package combinations. After the goods are combined, users can buy goods in batches in a single order, the consumption amount is larger, and users can also enjoy the comprehensive discounts provided by the merchant; the purpose of guiding user consumption, increasing product sales and increasing total profits is achieved.

[0003] In the process of setting up a package, in order to more intuitively display the products in the package, it is usually necessary to cooperate with picture display. However, if it is necessary to display product pictures, staff are required to make drawings, which takes a lot of manpower and time, and the results achieved are uneven due to the operations of different people. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and a terminal for generating product images, so as to realize automatic generation of product images according to a set meal.

[0005] In order to solve the above technical problems, a technical solution adopted by the present invention is:

[0006] A method for generating a product image comprises the steps of:

[0007] Obtaining package information, the package information including product identifiers and the quantity of products corresponding to each product identifier;

[0008] Constructing a single product image according to the product identifier and the product quantity;

[0009] Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each of the splicing positions;

[0010] A preset background image is randomly obtained, and each of the single product images is randomly spliced ​​at the splicing position according to the priority layer information, and then superimposed on the background image to obtain a package image.

[0011] In order to solve the above technical problems, another technical solution adopted by the present invention is:

[0012] A product image generation terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0013] Obtaining package information, the package information including product identifiers and the quantity of products corresponding to each product identifier;

[0014] Constructing a single product image according to the product identifier and the product quantity;

[0015] Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each of the splicing positions;

[0016] A preset background image is randomly obtained, and each of the single product images is randomly spliced ​​at the splicing position according to the priority layer information, and then superimposed on the background image to obtain a package image.

[0017] The beneficial effects of the present invention are: the product identification uniquely identifies the product, a single product image is constructed according to the number of products identified by the product identification in the package information, and then a preset stitching template and background are obtained. The template includes the positions where the single product image can be placed and the priority levels between the positions. The priority level, i.e., the occlusion relationship, can better simulate the real shooting environment, making the picture more natural. In addition to the need to set the package information, the entire stitching process can be completed automatically without the need for human participation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a method for generating a product image according to an embodiment of the present invention;

[0019] Figure 2 This is a structural schematic diagram of a commodity image generating terminal according to an embodiment of the present invention;

[0020] Description of labels:

[0021] 1. A product image generation terminal; 2. A processor; 3. A memory. DETAILED DESCRIPTION

[0022] In order to explain the technical content, achieved objectives and effects of the present invention in detail, the following is an explanation in combination with the implementation modes and the accompanying drawings.

[0023] Please refer to Figure 1 , a method for generating a product image, comprising the steps of:

[0024] Obtaining package information, the package information including product identifiers and the quantity of products corresponding to each product identifier;

[0025] Constructing a single product image according to the product identifier and the product quantity;

[0026] Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each of the splicing positions;

[0027] A preset background image is randomly obtained, and each of the single product images is randomly spliced ​​at the splicing position according to the priority layer information, and then superimposed on the background image to obtain a package image.

[0028] From the above description, it can be seen that the beneficial effects of the present invention are: the product identification uniquely identifies the product, a single product image is constructed according to the number of products identified by the product identification in the package information, and then a preset stitching template and background are obtained. The template includes the positions where the single product image can be placed and the priority levels between the positions. The priority level, i.e., the occlusion relationship, can better simulate the real shooting environment, making the picture more natural, and the entire stitching process can be completed automatically except for the need to set the package information, without the need for human participation in the process.

[0029] Furthermore, the step of obtaining the package information includes:

[0030] Constructing commodity information corresponding to the commodity identifier, wherein the commodity information includes a commodity image;

[0031] The constructing a single product image according to the product identifier and the product quantity includes:

[0032] Obtaining a product image corresponding to the product identifier according to the product identifier;

[0033] Copy the product image according to the number of the products to obtain a number of product images that is the same as the number of the products;

[0034] Multiple product images are stitched together to obtain a single product image.

[0035] From the above description, it can be seen that the product information is constructed by associating the product identification with the product information, and the product image corresponding to the product identification is stored in the product information. Then, in the process of generating the package image, the generated product image can be directly called, thereby speeding up the efficiency of generating the package image. Moreover, before splicing, for the same product identification, if there are multiple product quantities corresponding to the package information, the product images corresponding to the same product identification are spliced ​​in advance, and the splicing process is disassembled to speed up the execution speed of the splicing process.

[0036] Furthermore, the randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain the package image includes:

[0037] Execute a preset number of steps of randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and then superimposing them on the background image to obtain a package image, thereby obtaining a preset number of package images;

[0038] Obtaining a preset evaluation function, and scoring each of the package pictures according to the evaluation function to obtain a first evaluation score for each of the package pictures;

[0039] Get the package picture with the highest first evaluation score and return it.

[0040] From the above description, it can be seen that in the actual operation process, if only one stitching is performed, it cannot be guaranteed to obtain the optimal stitching result. Therefore, a preset number and an evaluation function are introduced. After obtaining a preset number of package pictures, the package pictures are scored according to the evaluation function to obtain evaluation scores. The package picture with the highest evaluation score is obtained and returned, which can improve the quality of the final generated package picture.

[0041] Furthermore, it also includes:

[0042] Get the preset number of candidate images;

[0043] The step of executing a preset number of random acquisitions of a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a package image, wherein obtaining the preset number of package images includes:

[0044] When the current execution quantity is less than the preset execution quantity and greater than the preset number of candidate pictures, the second evaluation score of all the package pictures that have been obtained is calculated according to the evaluation function, and then when each of the single product pictures is randomly spliced ​​at the splicing position according to the priority layer information, a real-time evaluation score is obtained by real-time scoring according to the evaluation function. If the real-time evaluation score is less than the minimum value of the second evaluation score, the generation of the package pictures of this round is stopped.

[0045] From the above description, it can be seen that if the number of candidate pictures is set in advance, for example, in the end only the picture with the highest score needs to be selected from 5 pictures, then during the generation process, if 5 pictures have been generated, real-time scoring is performed during the generation of the 6th picture to obtain the second evaluation score. If during the generation process, the second evaluation score is less than the lowest score among the previous 5 pictures, then even if the generation is completed, it is highly likely that it will not be ranked in the top 5 in the score. The generation process can be directly terminated and the next one can be entered to improve the efficiency of completing the preset number of executions.

[0046] Furthermore, before randomly splicing each of the single product images at the splicing position according to the priority layer information, the method includes:

[0047] Scaling each of the single product images according to a preset width and height range;

[0048] The randomly splicing each of the single product images at the splicing position according to the priority layer information includes:

[0049] Each of the scaled single product images is randomly spliced ​​at the splicing position according to the priority layer information.

[0050] From the above description, it can be seen that before splicing single product images, they are first scaled according to the preset width and height range so that the sizes of single product images corresponding to different product logos can be adapted to avoid the situation where the size difference is too large and inconsistent.

[0051] Please refer to Figure 2 , a commodity image generating terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0052] Obtaining package information, the package information including product identifiers and the quantity of products corresponding to each product identifier;

[0053] Constructing a single product image according to the product identifier and the product quantity;

[0054] Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each of the splicing positions;

[0055] A preset background image is randomly obtained, and each of the single product images is randomly spliced ​​at the splicing position according to the priority layer information, and then superimposed on the background image to obtain a package image.

[0056] The beneficial effects of the present invention are: the product identification uniquely identifies the product, a single product image is constructed according to the number of products identified by the product identification in the package information, and then a preset stitching template and background are obtained. The template includes the positions where the single product image can be placed and the priority levels between the positions. The priority level, i.e., the occlusion relationship, can better simulate the real shooting environment, making the picture more natural. In addition to the need to set the package information, the entire stitching process can be completed automatically without the need for human participation.

[0057] Furthermore, the step of obtaining the package information includes:

[0058] Constructing commodity information corresponding to the commodity identifier, wherein the commodity information includes a commodity image;

[0059] The constructing a single product image according to the product identifier and the product quantity includes:

[0060] Obtaining a product image corresponding to the product identifier according to the product identifier;

[0061] Copy the product image according to the number of the products to obtain a number of product images that is the same as the number of the products;

[0062] Multiple product images are stitched together to obtain a single product image.

[0063] From the above description, it can be seen that the product information is constructed by associating the product identification with the product information, and the product image corresponding to the product identification is stored in the product information. Then, in the process of generating the package image, the generated product image can be directly called, thereby speeding up the efficiency of generating the package image. Moreover, before splicing, for the same product identification, if there are multiple product quantities corresponding to the package information, the product images corresponding to the same product identification are spliced ​​in advance, and the splicing process is disassembled to speed up the execution speed of the splicing process.

[0064] Furthermore, the randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain the package image includes:

[0065] Execute a preset number of steps of randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and then superimposing them on the background image to obtain a package image, thereby obtaining a preset number of package images;

[0066] Obtaining a preset evaluation function, and scoring each of the package pictures according to the evaluation function to obtain a first evaluation score for each of the package pictures;

[0067] Get the package picture with the highest first evaluation score and return it.

[0068] From the above description, it can be seen that in the actual operation process, if only one stitching is performed, it cannot be guaranteed to obtain the optimal stitching result. Therefore, a preset number and an evaluation function are introduced. After obtaining a preset number of package pictures, the package pictures are scored according to the evaluation function to obtain evaluation scores. The package picture with the highest evaluation score is obtained and returned, which can improve the quality of the final generated package picture.

[0069] Furthermore, it also includes:

[0070] Get the preset number of candidate images;

[0071] The step of executing a preset number of random acquisitions of a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a package image, wherein obtaining the preset number of package images includes:

[0072] When the current execution quantity is less than the preset execution quantity and greater than the preset number of candidate pictures, the second evaluation score of all the package pictures that have been obtained is calculated according to the evaluation function, and then when each of the single product pictures is randomly spliced ​​at the splicing position according to the priority layer information, a real-time evaluation score is obtained by real-time scoring according to the evaluation function. If the real-time evaluation score is less than the minimum value of the second evaluation score, the generation of the package pictures of this round is stopped.

[0073] From the above description, it can be seen that if the number of candidate pictures is set in advance, for example, in the end only the picture with the highest score needs to be selected from 5 pictures, then during the generation process, if 5 pictures have been generated, real-time scoring is performed during the generation of the 6th picture to obtain the second evaluation score. If during the generation process, the second evaluation score is less than the lowest score among the previous 5 pictures, then even if the generation is completed this time, it is highly likely that it will not be ranked in the top 5 scoring values. The generation process can be directly terminated and the next one can be entered to improve the efficiency of completing the preset number of executions.

[0074] Furthermore, before randomly splicing each of the single product images at the splicing position according to the priority layer information, the method includes:

[0075] Scaling each of the single product images according to a preset width and height range;

[0076] The randomly splicing each of the single product images at the splicing position according to the priority layer information includes:

[0077] Each of the scaled single product images is randomly spliced ​​at the splicing position according to the priority layer information.

[0078] From the above description, it can be seen that before splicing single product images, they are first scaled according to the preset width and height range so that the sizes of single product images corresponding to different product logos can be adapted to avoid the situation where the size difference is too large and inconsistent.

[0079] The above-mentioned product image generation method and terminal of the present invention can be applied to the scenario where it is necessary to generate package images for a package consisting of products, especially to generate package images in scenarios such as KTV where there are no standard images of products, which is explained below through a specific implementation method.

[0080] Please refer to Figure 1 , Embodiment 1 of the present invention is:

[0081] A method and terminal for generating a product image, comprising the steps of:

[0082] S0. Constructing commodity information corresponding to the commodity identifier, wherein the commodity information includes a commodity picture;

[0083] In an optional implementation, product information is constructed according to the product SKU (Stock Keeping Unit), and product images are stored therein, where the product images are processed to retain only the outline of the product;

[0084] For supermarkets and other scenes, packaged goods for sale usually have standard SKUs that can be used (by scanning product barcodes). However, in special scenes such as KTVs, except for alcoholic beverages, snacks, nuts, snacks, fruit plates, etc., do not have standard SKUs, and it is also difficult to obtain product images. Therefore, product images are taken in advance and processed, and only the outline of the product is cut out, so that it can be better integrated with the background image during later use;

[0085] S1. Obtain package information, where the package information includes a product identifier and the quantity of the product corresponding to each product identifier;

[0086] S2. constructing a single product image according to the product identifier and the product quantity, including:

[0087] S21, obtaining a product image corresponding to the product identifier according to the product identifier;

[0088] S22, copying the product images according to the number of the products, to obtain a number of product images that is the same as the number of the products;

[0089] S23, stitching together the multiple product images to obtain a single product image;

[0090] S3. Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each splicing position;

[0091] In an optional implementation, the splicing template further includes a resolution, the splicing position is identified by a splicing coordinate, and the priority layer information identifies the covering method of images between different positions. The higher the priority of the splicing position, the higher the top layer of the single product image placed. For example, there are three single product images, and the priority layer information is 1, 2, and 3 respectively; the single product image at the splicing position of priority 1 can cover the product images of priorities 2 and 3, and the single product image of priority 2 can cover the product image of priority 3. In the process of splicing the single product images, the single product images are placed in each splicing position in turn, and it is determined whether the covering method needs to be adjusted according to the priority layer information.

[0092] S4. Scaling each of the single product images according to a preset width and height range to obtain a scaled single product image for subsequent generation of a package image. The scaling includes:

[0093] Scale the obtained single product image so that the sizes of the several product images will not be greatly different; the width or height can be scaled to a certain ratio of the background image size, for example, if the width ratio of the product image is less than 0.1, it can be scaled to 0.1, if it is greater than 0.5, it can be scaled to 0.5, or if the height ratio is less than 0.4, it can be scaled to 0.4, and if it is greater than 0.9, it can be scaled to 0.9, etc.; retain the original width-to-height ratio of the single product image to prevent the single product image from displaying abnormally;

[0094] S5, randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a set meal image;

[0095] In an optional implementation, before step S5, the method includes: generating a background image library, which can be drawn / photographed by professionals, or generating a background image of a specified type through tools such as AIGC (Artificial Intelligence Generated Content);

[0096] An evaluation function is established. The evaluation function can be established based on the difference between the color of the product and the background, the proportion of the product covered by other products, the neatness and order of the product sorting, etc. The function value is used to evaluate the quality of the generated package image. The higher the value, the better the quality. For example:

[0097] The product coverage area accounts for the proportion of the entire background image -0.3|;

[0098] In an optional implementation, a preset number of steps S5 are executed to obtain a preset number of package pictures, and then the step S6 is further included: obtaining a preset evaluation function, scoring each of the package pictures according to the evaluation function to obtain a first evaluation score for each of the package pictures, obtaining the package picture with the highest first evaluation score and returning it;

[0099] In an optional embodiment, it further includes obtaining a preset number of candidate images. In the process of looping step S5, if the current execution number is less than the preset execution number and greater than the number of pre-sale candidate images, the second evaluation score of all the package images that have been obtained is calculated according to the evaluation function, and when executing step S5 of the current round, when splicing each of the single product images at the splicing position randomly according to the priority layer information, a real-time evaluation score is obtained by real-time scoring according to the evaluation function. If the real-time evaluation score is less than the minimum value of the second evaluation score, the generation of the package images of this round is stopped, and step S5 is restarted until the preset number of package images are obtained;

[0100] In an optional implementation, S6 further includes: receiving user selection information, and displaying a package picture corresponding to the selection information; that is, providing a preset number of package picture sets for the user to choose, and obtaining the final display result by combining the automatic evaluation points and the manual selection, thereby further ensuring the final display result while reducing the pressure on the staff;

[0101] For example, the preset number of candidate images is 5, and the preset number is 20 times. After executing step S5 5 times in a loop, the second evaluation scores of the 5 package images are calculated. When executing step S5 for the 6th time, when each single product image is randomly spliced ​​at the splicing position according to the priority layer information, a real-time evaluation score is obtained according to the evaluation function. If the real-time evaluation score is already less than the lowest value of all the second evaluation scores, the 6th execution of step S5 is stopped, and the package image is no longer generated, and the 7th step S5 is re-executed. Because the template and background of the package image are randomly selected, the score of the next image cannot be predicted. Therefore, a larger preset number is set to obtain more images for selection, thereby increasing the probability of obtaining a high-evaluation-score image. If the score is less than the previous preset number of candidate images during the process of generating the image, it means that there is a high probability that the generation of this image cannot obtain a higher-scoring image. Then the generation of the package image this time is directly stopped, and the generation process of the next package image is started, which can speed up the generation efficiency of the final preset number of images provided to the user for selection.

[0102] If the real-time evaluation score is greater than the minimum value of the second evaluation score, it can be determined whether the next cycle needs to calculate the real-time evaluation score based on the remaining number of cycles. If the execution time of the remaining number of cycles is greater than the calculation time of the real-time evaluation score, the real-time evaluation score calculation continues. Otherwise, the remaining number of cycles is directly executed to ensure the overall efficiency of generating images.

[0103] Please refer to Figure 2 , Embodiment 2 of the present invention is:

[0104] A product image generation terminal 1 includes a processor 2, a memory 3, and a computer program stored in the memory 3 and executable on the processor 2. When the processor 2 executes the computer program, each step in the first embodiment is implemented.

[0105] In summary, the present invention provides a product image generation method and terminal, which stores processed product images when setting product SKUs, and takes out product images according to corresponding product identifiers when updating packages, and then pre-splices the product images according to the number of products corresponding to the product identifiers in the package information to obtain a single product image corresponding to each product identifier, and then obtains a template with the same number of splicing positions as the number of product identifiers, places the single product image at the splicing position in the template position, and adjusts the covering relationship between the product images according to the priority layer information configured at the splicing position, and finally combines the template with the spliced ​​single product image and the background image to obtain the package image, and the entire process does not require human participation, thus achieving automation.

[0106] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for generating product images, It is characterized in that Includes steps: Obtaining package information, the package information including product identifiers and the quantity of products corresponding to each product identifier; Constructing a single product image according to the product identifier and the product quantity; Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each of the splicing positions; A preset background image is randomly obtained, and each of the single product images is randomly spliced ​​at the splicing position according to the priority layer information, and then superimposed on the background image to obtain a package image.

2. A method for generating a product image according to claim 1, It is characterized in that The step of obtaining package information includes: Constructing commodity information corresponding to the commodity identifier, wherein the commodity information includes a commodity image; The constructing a single product image according to the product identifier and the product quantity includes: Obtaining a product image corresponding to the product identifier according to the product identifier; Copy the product image according to the number of the products to obtain a number of product images that is the same as the number of the products; Multiple product images are stitched together to obtain a single product image.

3. A method for generating a product image according to claim 1, It is characterized in that The randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a package image includes: Execute a preset number of steps of randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and then superimposing them on the background image to obtain a package image, thereby obtaining a preset number of package images; Obtaining a preset evaluation function, and scoring each of the package pictures according to the evaluation function to obtain a first evaluation score for each of the package pictures; Get the package picture with the highest first evaluation score and return it.

4. A method for generating a product image according to claim 3, It is characterized in that Also includes: Get the preset number of candidate images; The step of executing a preset number of random acquisitions of a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a package image, wherein obtaining the preset number of package images includes: When the current execution quantity is less than the preset execution quantity and greater than the preset number of candidate pictures, the second evaluation score of all the package pictures that have been obtained is calculated according to the evaluation function, and then when each of the single product pictures is randomly spliced ​​at the splicing position according to the priority layer information, a real-time evaluation score is obtained by real-time scoring according to the evaluation function. If the real-time evaluation score is less than the minimum value of the second evaluation score, the generation of the package pictures of this round is stopped.

5. A method for generating a product image according to claim 1, It is characterized in that Before randomly splicing each of the single product images at the splicing position according to the priority layer information, the method includes: Scaling each of the single product images according to a preset width and height range; The randomly splicing each of the single product images at the splicing position according to the priority layer information includes: Each of the scaled single product images is randomly spliced ​​at the splicing position according to the priority layer information.

6. A product image generation terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the following steps are implemented: Obtaining package information, the package information including product identifiers and the quantity of products corresponding to each product identifier; Constructing a single product image according to the product identifier and the product quantity; Acquire a splicing template according to the number of identifications of the commodity identification, wherein the splicing template includes splicing positions and priority layer information of each of the splicing positions; A preset background image is randomly obtained, and each of the single product images is randomly spliced ​​at the splicing position according to the priority layer information, and then superimposed on the background image to obtain a package image.

7. A commodity image generating terminal according to claim 6, It is characterized in that The step of obtaining package information includes: Constructing commodity information corresponding to the commodity identifier, wherein the commodity information includes a commodity image; The constructing a single product image according to the product identifier and the product quantity includes: Obtaining a product image corresponding to the product identifier according to the product identifier; Copy the product image according to the number of the products to obtain a number of product images that is the same as the number of the products; Multiple product images are stitched together to obtain a single product image.

8. A commodity image generating terminal according to claim 6, It is characterized in that The randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a package image includes: Execute a preset number of steps of randomly obtaining a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and then superimposing them on the background image to obtain a package image, thereby obtaining a preset number of package images; Obtaining a preset evaluation function, and scoring each of the package pictures according to the evaluation function to obtain a first evaluation score for each of the package pictures; Get the package picture with the highest first evaluation score and return it.

9. A commodity image generating terminal according to claim 8, It is characterized in that Also includes: Get the preset number of candidate images; The step of executing a preset number of random acquisitions of a preset background image, randomly splicing each of the single product images at the splicing position according to the priority layer information, and superimposing them on the background image to obtain a package image, wherein obtaining the preset number of package images includes: When the current execution quantity is less than the preset execution quantity and greater than the preset number of candidate pictures, the second evaluation score of all the package pictures that have been obtained is calculated according to the evaluation function, and then when each of the single product pictures is randomly spliced ​​at the splicing position according to the priority layer information, a real-time evaluation score is obtained by real-time scoring according to the evaluation function. If the real-time evaluation score is less than the minimum value of the second evaluation score, the generation of the package pictures of this round is stopped.

10. A commodity image generating terminal according to claim 6, It is characterized in that Before randomly splicing each of the single product images at the splicing position according to the priority layer information, the method includes: Scaling each of the single product images according to a preset width and height range; The randomly splicing each of the single product images at the splicing position according to the priority layer information includes: Each of the scaled single product images is randomly spliced ​​at the splicing position according to the priority layer information.

Images