Artificial intelligence-based agricultural product packaging individualized design generation method and system

CN122472005BActive Publication Date: 2026-08-28CHANGCHUN INST OF TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610952671.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-06-30
Publication Date
2026-08-28
Estimated Expiration
2046-06-30

AI Technical Summary

Technical Problem

[0005]本申请提供一种基于人工智能的农产品包装个性化设计生成方法和系统,用以解决现有农产品包装溯源体系中包装印刷内容与种植地块保护记录之间缺乏结构化绑定与动态核验手段的问题

Benefits of technology

由于现有技术中追溯码与包装印刷内容分属独立载体且不存在结构化的绑定关系,本申请通过将种植地块保护记录按生产环节顺序组织为保护农时链,并基于保护农时链构建复合校验结构,在保护事项的时序顺序与包装印刷内容之间建立了关联约束,使得包装内容与种植地块保护记录之间形成了结构化绑定关系,从而使核验环节能够检测到包装内容与保护记录之间的不一致;

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122472005B_ABST
    Figure CN122472005B_ABST
Patent Text Reader

Abstract

The application provides an agricultural product packaging personalized design generation method and system based on artificial intelligence, and relates to the technical field of intelligent packaging, wherein the method comprises the following steps: determining protection matters from the planting plot protection record according to the planting batch and generating a protection agricultural time chain in the order of production links; generating plot history content according to the protection agricultural time chain, and generating an initial packaging design scheme through a layout generation model; extracting regional verification elements of each packaging area and establishing an association relationship in the order of the protection agricultural time chain to form a composite verification structure; splitting the composite verification structure into a global verification layer and a regional verification layer, and writing the global verification layer into a traceability code and the regional verification layer into a packaging area; comparing verification mark numbers to determine a verification mode when scanning and verifying, and comparing protection matters and packaging expression levels with current protection records to determine a verification result. The application improves the accuracy of packaging traceability verification.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of intelligent packaging technology, and in particular to a method and system for generating personalized designs for agricultural product packaging based on artificial intelligence. Background Technology

[0002] In recent years, protection of planting sites has been widely printed on agricultural product packaging as an important endorsement of agricultural product quality. This includes the visual presentation of protection measures such as black soil protection, straw return to the field, organic fertilizer substitution for chemical fertilizer, and crop rotation and fallow. This type of information has become one of the key factors influencing consumers' willingness to buy and the market premium of products. Therefore, how to accurately and standardizedly transform protection measures into packaging content during the packaging design stage and continuously verify the authenticity of the packaging content during the packaging and distribution stage is the core issue facing agricultural product packaging technology today.

[0003] Currently, protective information on agricultural product packaging is usually manually compiled by packaging designers based on written materials provided by the place of origin. Once the packaging is printed, the content of the protective information is fixed on the physical carrier. In the traceability process, existing technologies mainly write basic data such as place of origin information and production date into QR codes or barcodes. Consumers can obtain product origin information by scanning the code. However, the consistency between the protective information printed on the packaging and the actual protection records of the planting site is not included in the scope of traceability verification. The traceability code and the content printed on the packaging are two independent information carriers, and there is no structured binding relationship between the two.

[0004] The aforementioned existing technologies have the following problems in practical applications: The printed content on the packaging remains static after leaving the factory. When the protection records of the planting plots are adjusted due to periodic updates or changes in qualifications, discrepancies may arise between the protection information on the already circulated packaging and the latest records. Existing traceability methods lack the ability to detect such discrepancies. Furthermore, since the traceability code and the printed content on the packaging are carried independently by different carriers, the integrity of the traceability code itself is not affected when the protection information on the packaging is partially replaced or reprinted. Such partial tampering is not easily detected in the existing verification process. In addition, during the verification stage, existing solutions cannot further distinguish the types of abnormalities in the packaging content, and cannot determine whether the abnormality stems from information lag after the natural update of the protection records or from discrepancies caused by human adjustments to the printed content on the packaging. Summary of the Invention

[0005] This application provides a method and system for generating personalized designs for agricultural product packaging based on artificial intelligence, in order to solve the problem that there is a lack of structured binding and dynamic verification methods between the printed content of the packaging and the protection records of the planting plot in the existing traceability system for agricultural product packaging.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for generating personalized agricultural product packaging designs based on artificial intelligence, comprising: Based on the planting batch of the agricultural products to be packaged, various protection items are determined from the protection records of the planting plots, and a protection agricultural time chain is generated according to the production process. Based on the protection of the agricultural time chain, the land plot history is generated and then distributed to each packaging area according to the packaging usage scenario. An initial packaging design scheme is generated through the layout generation model. The protection measures and corresponding packaging expression levels adopted in each packaging area are extracted from the initial packaging design scheme as regional verification elements. The correlation between the regional verification elements is established according to the production link sequence of the agricultural protection chain to form a composite verification structure. The composite verification structure is split into a global verification layer and several regional verification layers. The global verification layer is written with a traceability code, and each regional verification layer is embedded in the corresponding packaging area in the form of a verification identifier, thus forming a verification-enabled agricultural product packaging design scheme. When the agricultural product packaging corresponding to the verified agricultural product packaging design scheme is scanned for verification, the global verification layer is read from the traceability code. The verification method is determined based on the comparison result between the total number of packaging areas recorded in the global verification layer and the actual number of verification marks identified, and the area verification layer corresponding to each verification mark is read. The protection items and packaging expression levels read from each area verification layer are compared with the current planting plot protection records to determine the verification results of agricultural product packaging.

[0007] Optionally, relationships can be established between verification elements in each region according to the production sequence of the agricultural timeline to form a composite verification structure, including: According to the production process sequence corresponding to each protection item in the agricultural protection chain and the current packaging batch, all regional verification elements are arranged; The corresponding history fingerprint is generated for all the regional verification elements after the arrangement in an ordered hash chain manner. Each node in the ordered hash chain is linked sequentially according to the production link of the agricultural time chain. The hash input of each node includes the hash output of the previous node and the regional verification element corresponding to the current node. All regional verification elements, the relationships between regional verification elements, and historical fingerprints are used together as a composite verification structure. Specifically, when the content and arrangement order of all area verification elements are the same, the generated history fingerprints are the same; when the content of any area verification element changes or the arrangement order is adjusted, the output of all subsequent nodes in the ordered hash chain from the changed node changes, and the generated history fingerprints are different.

[0008] Optionally, the composite verification structure can be split into a global verification layer and several regional verification layers, including: Extract the correlation between the verification elements of each region, the total number of packaging regions, and the location information of each packaging region from the composite verification structure. Extract some information from the history fingerprint and write the correlation, the total number of packaging regions, the location information of each packaging region, and the partial information of the history fingerprint into the global verification layer. Extract the content identifier and packaging expression level of the protected items in each packaging area from the composite verification structure, and extract the remaining information from the history fingerprint. Write the content identifier, packaging expression level and the remaining information of the history fingerprint into the area verification layer of the corresponding packaging area. Among them, some information of the resume fingerprint together with the rest of the resume fingerprint information constitutes the complete information of the resume fingerprint. Neither the global verification layer nor the all-area verification layer can independently restore the composite verification structure.

[0009] Optionally, the verification method is determined based on the comparison between the total number of packaging areas recorded in the global verification layer and the actual number of verification identifiers identified, including: When the total number of packaging areas matches the actual number of verification identifiers identified, the corresponding area verification layer in all verification identifiers is read synchronously in the overall verification method, and the composite verification structure is reconstructed in a unified manner in combination with the global verification layer. When the total number of packaging areas is inconsistent with the number of verification marks actually identified, a zone-by-zone verification method is used. The search path is based on the production sequence in the agricultural time chain. The corresponding area verification layer in each verification mark is searched and read one by one according to the location information of each packaging area recorded in the global verification layer, and partial verification is performed based on the area verification elements that have been read.

[0010] Optionally, the protection items and packaging expression levels read from each area verification layer are compared with the current planting plot protection records to determine the verification results of agricultural product packaging, including: When the history fingerprint corresponds to the verification elements of each region in the composite verification structure jointly reconstructed by the global verification layer and all regional verification layers, the reconstructed regional verification elements are used as the approved verification items. When the resume fingerprint does not correspond to the verification elements of each region, it is determined that the agricultural product packaging has been tampered with. Extract actual verification items from the various protection measures actually presented on agricultural product packaging, as well as the corresponding packaging areas and packaging expression levels; The approved verification items are compared with the printed verification items, and the validity of each protection item in the printed verification items is determined based on the current protection records of the planted plots, so as to determine the verification results.

[0011] Optionally, the verification results of agricultural product packaging may be determined, including: When the actual printed verification items are consistent with the approved verification items, and all protection items are valid protection items, the agricultural product packaging is deemed to have passed the verification. The actual printed verification items are compared with the approved verification items one by one. When the display intensity of the protected items in the packaging area or packaging expression level in the actual printed verification items is stronger than that in the packaging area or packaging expression level in the approved verification items, the agricultural product packaging is judged to be reprinted packaging beyond the approved level. When the approved verification items are consistent with the actual printed verification items, and there are protective items in the actual printed verification items that are no longer valid in the current planting plot protection record, the agricultural product packaging is determined to be invalid packaging. When the actual printed verification items have missing protection items, added protection items, or inconsistent packaging areas compared to the approved verification items, and do not belong to over-level reprinted packaging or expired packaging, the agricultural product packaging is judged to be inconsistent packaging. Among them, a valid protected item is a protected item that exists in the current protected record of the plantation plot, and the source of the protected item and the number of consecutive records support the packaging expression level corresponding to the protected item.

[0012] Optionally, the step of locating and reading the corresponding area verification layer in each verification identifier one by one in a zone-by-zone verification manner further includes: If there are unreadable verification identifiers, the verification elements of each area that have been read are compared with the associations recorded in the global verification layer; if the protection item corresponding to the unreadable verification identifier can be uniquely determined based on the association, the protection item corresponding to the unreadable verification identifier is inferred; if it cannot be uniquely determined, the corresponding packaging area is determined to be the area to be reviewed. If all verification marks cannot be read, the composite verification structure of the agricultural product packaging is deemed invalid.

[0013] Optionally, after determining that the agricultural product packaging is reprinted packaging beyond its grade, the following further steps may be taken: The protective items in the actual print verification items that exceed the approved verification items are identified as packaging contents to be corrected; After downgrading, rewriting, or deleting the packaging content to be corrected, the corresponding packaging design content for the packaging area is regenerated. The composite verification structure, corresponding global verification layer, regional verification layers, and verification identifiers are regenerated based on the regenerated packaging design. The specific rules for downgrading are as follows: When the current packaging expression level of the content to be corrected is Level 1, it is downgraded to Level 2, meaning the presentation style of icon plus title plus detailed description is changed to icon plus title, and the detailed description text is deleted; when the current packaging expression level is Level 2, it is downgraded to Level 3, meaning the presentation style of icon plus title is changed to a text list only, and the icon element is deleted. The downgraded packaging content is then regenerated by the layout generation model based on the adjusted site history content and packaging size parameters. If the packaging content to be corrected needs to be deleted rather than downgraded, the corresponding area will not be left empty after deletion, and the layout generation model will reallocate the packaging area content based on the remaining protection items.

[0014] Optionally, after determining that the agricultural product packaging is expired packaging, the following steps are also included: In the verification results, identify the expired protection items and determine the packaging contents corresponding to the expired protection items as packaging contents that will no longer be displayed; Update the packaging expression level of the remaining valid protection items in the physical verification items based on the packaging expression level supported by the current plantation protection record.

[0015] Optionally, the protective measures and corresponding packaging expression levels adopted for each packaging area can be extracted from the initial packaging design as area verification elements, including: Identify the actual protective measures adopted from each packaging area of ​​the initial packaging design, and extract the content identifiers of each protective measure and the packaging area identifiers where the protective measures are located; The corresponding packaging expression level is determined based on how the protective measures are presented in the initial packaging design for each packaging area; Content identification, packaging area identification, and packaging expression level will be used as the regional verification elements for the packaging area. The area verification elements do not include the layout, font, and color scheme information of the protected items within the packaging area.

[0016] Optionally, various protection items can be identified from the planting plot protection records, and a protection agricultural time chain can be generated according to the production process, including: Extract the record sources and consecutive record counts for each protection item corresponding to the planting batch from the protection records of the planting plots; The various protection items are arranged according to the time sequence of the corresponding production links to form a protection agricultural time chain with each protection item as a node. Each node in the protection agricultural time chain records the source of the corresponding protection item and the number of consecutive records.

[0017] Secondly, this application provides an artificial intelligence-based personalized design generation system for agricultural product packaging, comprising: The determination module is used to determine various protection items from the planting plot protection record based on the planting batch of the agricultural products to be packaged, and generate a protection agricultural time chain according to the production process sequence; The generation module is used to generate land parcel history content based on the agricultural time chain protection, allocate the land parcel history content to each packaging area according to the packaging usage scenario, and generate the initial packaging design scheme through the layout generation model. The extraction module is used to extract the protection measures adopted by each packaging area and the corresponding packaging expression level from the initial packaging design scheme as regional verification elements. The module establishes the correlation between the regional verification elements according to the production link sequence of the agricultural protection chain to form a composite verification structure. The splitting module is used to split the composite verification structure into a global verification layer and several regional verification layers. The global verification layer is written with a traceability code, and each regional verification layer is embedded in the corresponding packaging area in the form of a verification identifier, forming a verification-enabled agricultural product packaging design scheme. The verification module is used to read the global verification layer from the traceability code when the agricultural product packaging corresponding to the verification agricultural product packaging design scheme is scanned for verification. It determines the verification method based on the comparison result between the total number of packaging areas recorded in the global verification layer and the actual number of verification marks identified, and reads the area verification layer corresponding to each verification mark. The comparison module is used to compare the protection items and packaging expression levels read from each area verification layer with the current planting plot protection records to determine the verification results of agricultural product packaging.

[0018] The technical solution provided in this application has the following beneficial effects: Since the traceability code and the packaging printing content are independent carriers in the existing technology and there is no structured binding relationship, this application organizes the protection records of planting plots into a protection agricultural time chain according to the production process, and constructs a composite verification structure based on the protection agricultural time chain. It establishes an association constraint between the temporal sequence of the protection items and the packaging printing content, so that a structured binding relationship is formed between the packaging content and the protection records of planting plots, thereby enabling the verification process to detect inconsistencies between the packaging content and the protection records. Furthermore, since there is a risk that the entire verification information may be tampered with without being detected when a single carrier carries all the verification information, this application splits the composite verification structure into a global verification layer and several regional verification layers and embeds them into the traceability code and the verification identifier of each packaging area, so that the verification information is distributed and stored on multiple physical carriers. When any carrier is tampered with individually, the verification data in that carrier will be inconsistent with the verification data in other carriers, so that it can be detected in the verification stage. Furthermore, since existing verification methods are not capable of distinguishing between different types of packaging content anomalies, this application compares the approved verification items, the actual printed verification items, and the current planting plot protection records in a multi-dimensional manner during the verification stage. This allows for the differentiation of different types of anomalies, such as carrier tampering, over-level reprinted packaging, expired record packaging, and inconsistent content packaging, based on the specific patterns of inconsistencies in the comparison results. This enables the verification results to provide a targeted basis for subsequent packaging disposal. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the overall scene interaction of a method for generating personalized agricultural product packaging design based on artificial intelligence disclosed in this application.

[0021] Figure 2 This is an overall flowchart of a method for generating personalized agricultural product packaging design based on artificial intelligence, as disclosed in this application.

[0022] Figure 3 This is a schematic diagram of an ordered hash chain node connection structure disclosed in this application.

[0023] Figure 4 This is a schematic diagram of the spatial layout of the packaging area and the verification mark as disclosed in this application.

[0024] Figure 5 This is a flowchart of the verification method determination and verification result judgment disclosed in this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] In the field of agricultural product packaging, planting site protection records are a collection of information describing the various protective measures implemented during the planting process of agricultural products. These records are stored in a database, with each record corresponding to the implementation of protective measures for a single planting site in a single planting batch. Each protective measure is a specific protection item included in the planting site protection record, describing a particular protective action performed on the planting site during production, such as black soil protection, straw return to the field, organic fertilizer substitution for chemical fertilizer, and crop rotation / fallow. The planting batch is the specific planting cycle identifier for the agricultural product to be packaged. Planting batches are used to distinguish production activities of the same plot in different planting cycles. For example, the planting batch identifier for a batch of rice planted on a plot in the spring of 2024 can be represented as 2024S-A.

[0027] See Figure 1 As shown, the method proposed in this application involves the following interactive entities: a planting plot protection record database 101 for storing protection information for each batch of each plot; a packaging design generation server 102 obtains protection information from the planting plot protection record database 101 and executes steps S100 to S400 for packaging design generation and verification structure construction; a printing terminal 103 receives the verified agricultural product packaging design scheme output by the packaging design generation server 102 and completes the physical packaging printing to form an agricultural product packaging entity 104; a barcode verification terminal 105 scans the traceability code and verification mark on the agricultural product packaging entity 104 in the circulation process, executes steps S500 to S600 for verification, and the barcode verification terminal 105 obtains the latest planting plot protection record through the planting plot protection record query interface 106 for comparison.

[0028] See Figure 2 As shown, based on the above scenario, this application discloses a method for generating personalized agricultural product packaging designs based on artificial intelligence. The overall process of this method includes six steps, from S100 to S600. Taking the spring 2024 batch of rice planted on a certain plot A as an example, the following details each step in the complete business cycle of "packaging design generation, physical printing, distribution and shelf placement, initial consumer verification, annual update of protection records, and secondary verification".

[0029] S100. Based on the planting batch of the agricultural products to be packaged, determine various protection items from the protection record of the planting plot, and generate a protection agricultural time chain according to the production process.

[0030] The purpose of this step is to organize the various protection items in the planting plot protection record according to the time sequence of the production process, forming a protection agricultural time chain with a temporal structure.

[0031] The agricultural protection timeline is an ordered chain structure with various protection items as nodes and the production process sequence as the connection relationship. Each node in the agricultural protection timeline corresponds to a protection item in the planting process. The order of the nodes reflects the sequential relationship of the production process corresponding to the protection item in the agricultural production cycle. Each node in the agricultural protection timeline also records the source of the corresponding protection item and the number of consecutive records.

[0032] The source of records refers to the entity providing or collecting information on protected matters. Different sources of records reflect the methods of information collection and the level of confirmation. The number of consecutive records refers to the number of batches in which the same protected matter appears consecutively in the protection records of a planting plot. The number of consecutive records reflects the continuity of the protection of that matter in that plot. The production stage refers to the time-based phases in the agricultural production cycle, such as pre-sowing preparation, sowing period, growing season, and harvest period.

[0033] The specific implementation process of this step is as follows: First, based on the planting batch of the agricultural product to be packaged, this step retrieves all protection items corresponding to the planting batch from the planting plot protection record database 101. The protection item information includes the protection item name, execution time, record source, and number of consecutive records. Then, this step maps the protection item to the corresponding production stage based on the execution time field of each protection item. The mapping relationship is determined by a preset mapping table between protection items and production stages. This mapping table is pre-stored in the packaging design generation server 102. Table 1 is the mapping table between protection items and production stages, and its specific contents are shown in the table below.

[0034] Table 1. Mapping of Protection Items to Production Processes

[0035] For protection items corresponding to the entire production cycle, since the protection items for the entire cycle run through the entire planting process and do not correspond to a single point in time, this step arranges the protection items for the entire cycle at the end of the protection time chain when constructing the protection time chain, so as to distinguish between the protection items for the entire cycle and the protection items for each stage.

[0036] The sources of records are divided into three levels of authority based on the method of information collection and the level of confirmation, as shown in Table 2.

[0037] Table 2. Classification of Source Authority Levels

[0038] The reason for classifying the sources of records into three levels of authority is that the credibility of protection information from different sources varies: verification data from local agricultural authorities, confirmed through on-site government verification, has the highest credibility and is therefore classified as Level L1; certification results from third-party testing institutions, with certification reports issued by independent organizations, have the next highest credibility and are classified as Level L2; records self-reported by farmers, although verified at the grassroots level, are based on information provided by the farmers themselves and have relatively low credibility, thus being classified as Level L3. The authority level is set at three levels because the existing channels for collecting information on agricultural product protection are mainly divided into these three categories in actual production. A two-level system would fail to distinguish between third-party certification and farmer reporting, while a four-level or higher system would increase management complexity, and the actual collection channels are insufficient to support a more granular classification.

[0039] In practical engineering applications, assume that plot A implemented the following protection measures in the spring planting batch of 2024: straw return to the field corresponds to the pre-sowing preparation stage, with an authoritative record source level of L1 and five consecutive records; organic fertilizer substitution for chemical fertilizer corresponds to the sowing period, with an authoritative record source level of L2 and three consecutive records; black soil protection corresponds to the entire cycle, with an authoritative record source level of L1 and six consecutive records; and crop rotation and fallow correspond to the planting cycle transition period, with an authoritative record source level of L3 and two consecutive records. Arranged chronologically according to the production stages, the non-full-cycle protection measures are, in chronological order, crop rotation and fallow, straw return to the field, and organic fertilizer substitution for chemical fertilizer. The full-cycle protection measure, black soil protection, is placed at the end. Therefore, the node order of the agricultural protection timeline is: the first node corresponds to crop rotation and fallow, the second node to straw return to the field, the third node to organic fertilizer substitution for chemical fertilizer, and the fourth node to black soil protection.

[0040] This application maps scattered protection items to corresponding production stages according to a mapping table and organizes them into a protection agricultural time chain in chronological order. This enables subsequent steps to allocate packaging content and build verification associations based on the ordered structure of the protection agricultural time chain, thereby providing a chronological basis for the structured binding between packaging content and planting plot protection records.

[0041] S200: Generate land parcel history content based on the agricultural timeline protection, allocate the land parcel history content to each packaging area according to the packaging usage scenario, and generate the initial packaging design scheme through the layout generation model.

[0042] In step S100, the various protection items have been organized into a protection agricultural season chain according to the production process. Since each node in the protection agricultural season chain only records the identification information and record attributes of the protection items, the packaging design scheme needs to transform this information into textual descriptions and visual presentation content that consumers can understand. Therefore, the purpose of this step is to transform the protection items in the protection agricultural season chain into land history content suitable for packaging display, and to arrange the land history content into a specific packaging design scheme through a layout generation model.

[0043] The land plot history is a collection of textual and icon-based descriptions generated based on protection information for each node in the agricultural protection chain. This history is used on the packaging to visually present to consumers the various protection measures implemented during the production process of the planting plot. The packaging usage scenario refers to the way agricultural product packaging is presented in actual use and the main viewing surface for consumers. The packaging area is an independent display surface within the packaging usage scenario, divided by function and location; each packaging area corresponds to a physical area on the packaging entity.

[0044] Packaging usage scenarios are categorized based on the physical structure of the packaging and its retail display method. Common usage scenarios include retail gift boxes, bulk bagged packaging, and shipping boxed packaging. Retail gift boxes are divided into three areas: a main front display area, a side information area, and a bottom instruction area. The main front display area presents the most important protective information to attract consumer attention, the side information area presents supplementary protective information, and the bottom instruction area presents basic information and a traceability code. Bulk bagged packaging is divided into two areas: a front display area and a back information area. Shipping boxed packaging is divided into two areas: a front labeling area and a side detail area. The number of packaging areas is determined by the usage scenario, and the number of protective information items each area can accommodate is determined based on the area size and the output of the layout generation model.

[0045] The implementation process of this step is as follows: First, based on the name of the protected item, the source of the record, and the number of consecutive records for each node in the agricultural protection timeline, this step generates the corresponding plot history text. For example, if the record source for the straw return node is L1 and the number of consecutive records is five, then the corresponding plot history text generated will be "The planting plot of this product has implemented straw return for five consecutive batches, as verified and confirmed by the local agricultural authorities." Then, this step determines the division method of the packaging area based on the packaging usage scenario. Further, this step allocates the plot history content to each packaging area according to the importance of the protected items. The method for synthesizing the importance is a lexicographical order: first grouped by authority level, then arranged in descending order of the number of consecutive records within each group. All protected items in group L1 are placed before group L2, group L2 is placed before group L3, and within the same authority level group, they are arranged in descending order of the number of consecutive records. Protected items ranked higher in the arrangement are given priority for allocation to the main display area.

[0046] The layout generation model is a deep learning-based graphic layout model used to automatically generate layout schemes for packaging based on input land parcel history content and packaging size parameters. The construction process of the layout generation model is as follows: First, existing compliant agricultural product packaging samples are collected, and the position coordinates and text content of each packaging area are labeled for each sample. Then, the labeled samples are classified according to packaging usage scenarios, and layout sub-models for each scenario are constructed. The layout generation model in this application can select a specific model architecture according to actual needs. For example, this application adopts an architecture based on a combination of variational autoencoder and conditional generative adversarial network. The variational autoencoder is responsible for mapping the packaging layout to feature vectors in the latent space. The conditional generative adversarial network uses the text length of the land parcel history content, the number of icons, and the physical size of the packaging as conditional inputs to generate layout feature vectors in the latent space. Then, the decoder decodes the layout feature vectors into the position coordinates and size parameters of text and icons in each packaging area.

[0047] In practice, the latent space dimension of the variational autoencoder is set to 128 dimensions, the encoder consists of three fully connected layers, and the decoder structure is symmetrical to the encoder. The generator of the conditional generative adversarial network consists of four fully connected layers, and the discriminator consists of three fully connected layers. At least two thousand compliant packaging samples are used during training, with each sample labeled with the top-left corner coordinates, width, height, and start and end positions of the text content within each packaging area. The Adam optimizer is used during training, with an initial learning rate of 0.0001, a batch size of 32, and at least two hundred training epochs. Training stops when the layout aesthetics loss on the validation set no longer decreases after twenty consecutive epochs. Layout sub-models are constructed according to packaging usage scenarios, and layout sub-models for different scenarios are trained independently. For example, the layout sub-model for the retail gift box scenario is trained using retail gift box packaging samples, and the layout sub-model for the bulk bag scenario is trained using bag packaging samples.

[0048] During the training phase, the layout generation model learned the layout rules under different combinations of text length and packaging size from compliant packaging samples. The training objective function includes two items: layout aesthetics loss and text readability loss. The text readability loss measures whether the text area size meets the preset minimum font size threshold. The minimum font size threshold ranges from 6 to 9, and can be set according to the packaging size and the consumer's reading distance. In this embodiment, it is set to 8, which is based on the minimum font size that consumers can recognize with their naked eyes at common retail shelf distances.

[0049] The layout aesthetics loss is used to measure whether the generated layout conforms to visual aesthetic standards. The layout aesthetics loss consists of a weighted sum of three sub-items: the first sub-item is an element overlap penalty, which measures the proportion of overlap between text and icon areas in each packaging area. A penalty is applied when the overlap exceeds 5%. The second sub-item is a white space ratio constraint, which measures the deviation between the proportion of blank area in each packaging area and the preset target white space ratio. The target white space ratio ranges from 20% to 40%, and is set to 30% in this embodiment. The third sub-item is an alignment deviation measure, which measures the variance between the horizontal starting coordinates of multiple text elements within the same packaging area. A smaller variance indicates a higher degree of alignment. The weight coefficients for the three sub-items are 0.4, 0.3, and 0.3, respectively. This weighting is based on the fact that element overlap has the greatest impact on the consumer's reading experience, hence it is given the highest weight; white space and alignment have similar impacts, hence they are given the same weight.

[0050] Following the aforementioned embodiment, assuming the agricultural product to be packaged is a retail gift box, the packaging area is divided into a front main display area A1, a side information area A2, and a bottom explanatory area A3. Arranged by importance, black soil protection is group L1 with six consecutive records, ranking first; straw return to the field is group L1 with five consecutive records, ranking second; organic fertilizer substitution for chemical fertilizer is group L2, ranking third; and crop rotation and fallow are group L3, ranking fourth. The front main display area A1 can accommodate two protection items; therefore, black soil protection and straw return to the field are assigned to A1, organic fertilizer substitution for chemical fertilizer is assigned to A2, and crop rotation and fallow are assigned to A3. This step inputs the allocation results into the layout generation model to obtain the initial packaging design scheme.

[0051] This application uses a lexicographical order to combine the authority level of the record source and the number of consecutive records to rank importance, giving the distribution of packaging design content a clear priority rule, and automatically arranging the layout through a layout generation model, thus avoiding the subjective arbitrariness of manual layout.

[0052] S300. Extract the protection measures and corresponding packaging expression levels adopted for each packaging area from the initial packaging design scheme as regional verification elements. Establish the correlation between the regional verification elements according to the production link sequence of the agricultural protection chain to form a composite verification structure.

[0053] In step S200, an initial packaging design scheme has been generated through the layout generation model. Since the initial packaging design scheme itself is only layout information and does not have anti-tampering and verification capabilities, this step needs to extract key information related to verification from the initial packaging design scheme to construct a verification structure.

[0054] The specific implementation process of this step is as follows: First, this step identifies the actual protected items adopted from each packaging area of ​​the initial packaging design scheme, and extracts the content identifier and the packaging area identifier of each protected item. The content identifier is the coded identifier of the protected item in the planting plot protection record system. The content identifier is used to identify and distinguish the category of protected item. The content identifier is generated by performing a hash operation on the string of the protected item name and then extracting the first eight hexadecimal characters. The reason for extracting the first eight characters is that eight hexadecimal characters can provide approximately 4.3 billion different coding combinations, which has high distinguishability for the number of agricultural product protected item categories. At the same time, eight characters occupy less storage space in the verification identifier. When the content identifiers generated by different protected items conflict, the number of truncation bits is increased or the content identifier is regenerated by combining with the preset protected item coding table. The packaging area identifier is the coded identifier of each packaging area in the initial packaging design scheme. For example, the packaging area identifier of the main display area on the front is A1.

[0055] Next, this step determines the corresponding packaging expression level based on how the protected information is presented in the initial packaging design for each packaging area. The packaging expression level is a graded indicator based on the intensity of the protected information's presentation in the packaging area. It quantifies the strength of the display of a particular protected information on the packaging; a lower value indicates a stronger display. Table 3 shows the rules for determining the packaging expression level, as detailed below: Table 3 Rules for Determining Packaging Expression Level

[0056] The area validation element is a triple of information extracted from a packaging area in the initial packaging design scheme for subsequent verification. The area validation element includes three fields: content identifier, packaging area identifier, and packaging expression level. The area validation element does not include layout position, font, and color information. Since these visual parameters can have multiple reasonable variations under the same packaging expression level, excluding these parameters can prevent design fine-tuning from triggering validation failure.

[0057] Furthermore, this step establishes relationships between verification elements in each region according to the production chain sequence for protecting the agricultural season, thus forming a composite verification structure. The composite verification structure is a complete verification data body containing three parts: all regional verification elements, the relationships between regional verification elements, and historical fingerprints.

[0058] See Figure 3 As shown, this step establishes the association using an ordered hash chain. An ordered hash chain is a data integrity verification structure based on chained hash operations. Nodes in the ordered hash chain are linked sequentially according to the production stage order of the agricultural time protection chain. The calculation process of the ordered hash chain is as follows: First, all the area verification elements after arrangement are numbered from E1 to Em according to the agricultural time protection chain order, where m is the total number of area verification elements; as shown... Figure 3 As shown, the packaging batch identifier enters the first hash node 31 from the input terminal 30. The first hash node 31 concatenates the packaging batch identifier with E1 and performs a hash operation to obtain the output value of the first hash node 31. The second hash node 32 concatenates the output value of the first hash node 31 with E2 and performs a hash operation to obtain the output value of the second hash node 32. And so on, each hash node concatenates the output value of the previous hash node with the area verification element corresponding to the current node and performs a hash operation. Finally, the output value of the hash node 34 is output from the history fingerprint output terminal 35, and this output value is the history fingerprint.

[0059] The historical fingerprint is a fixed-length hexadecimal string used as a unique digest to identify the content and order of all verification elements in the region. The hash algorithm used in this application can be selected according to actual needs; for example, the SHA-256 algorithm can be used, whose output is a fixed 64-digit hexadecimal string. Since hash algorithms are commonly used, the specific implementation of hash algorithms will not be elaborated upon in this application.

[0060] Due to the chain-like computational characteristics of ordered hash chains, when the content and arrangement order of all area verification elements are the same, the inputs of each hash node are the same, and therefore the generated history fingerprints are the same; when the content of any area verification element changes or the arrangement order is adjusted, the inputs of all subsequent hash nodes starting from the changed node change, and therefore the generated history fingerprints are different.

[0061] Following the aforementioned embodiment, the regional verification elements corresponding to the four protection items arranged in the order of the agricultural protection time chain are as follows: E1 corresponds to crop rotation and fallow at level A3 and level 3; E2 corresponds to straw return to the field at level A1 and level 1; E3 corresponds to organic fertilizer replacing chemical fertilizer at level A2 and level 2; and E4 corresponds to black soil protection at level A1 and level 1. Assuming the packaging batch identifier is 2024S-A, after each hash node calculates in sequence, the final historical fingerprint is a 64-bit hexadecimal string.

[0062] This application uses an ordered hash chain to associate the verification elements of each region into a complete verification chain according to the order of the protection agricultural time chain. This ensures that any modification to the content or order of the regional verification elements will cause the history fingerprint to change, thereby giving the composite verification structure anti-tampering capability.

[0063] S400: The composite verification structure is split into a global verification layer and several regional verification layers. The global verification layer is written with a traceability code, and each regional verification layer is embedded in the corresponding packaging area in the form of a verification mark, forming a verification-enabled agricultural product packaging design scheme.

[0064] In step S300, a composite verification structure has been constructed. Since the composite verification structure is a complete data collection, if the composite verification structure is written as a whole into a single carrier, the entire verification information will become invalid if the single carrier is tampered with. Therefore, this step splits the composite verification structure into a global verification layer and several regional verification layers and embeds them into different carriers respectively.

[0065] The global verification layer is an information layer in the composite verification structure that describes the overall relationship and the overall characteristics of the packaging structure. The global verification layer includes the total number of packaging areas, the relationship between the verification elements of each area, the location information of each packaging area, and some fragments of the history fingerprint.

[0066] The location information is the correspondence between the packaging area identifier and the verification identifier on the packaging entity in the preset coordinate range. The location information is used to guide the scanning device to locate each verification identifier in the area-by-area verification method in step S500.

[0067] The regional verification layer is an information layer describing the verification information of protective items within a single packaging area. It includes the content identifiers of each protective item within that area, the packaging's level of protection, and fragments of its historical fingerprint. The traceability code is a printed identifier that carries the global verification layer data in a machine-readable encoding format such as a QR code. The verification mark is a miniature machine-readable identifier generated by encoding the regional verification layer information; the verification mark can take the form of a miniature QR code, a data matrix code, or an invisible watermark, etc.

[0068] The specific implementation process of this step is as follows: First, this step extracts global information from the composite verification structure and writes it into the global verification layer. Then, this step extracts local information of each packaging area from the composite verification structure and writes it into each area verification layer. For the splitting of the history fingerprint, this step adopts an alternating bit splitting method: the complete 64-bit hexadecimal string history fingerprint is split according to the parity of the character positions. The 32-character substring composed of odd-numbered characters is written into the global verification layer; the 32-character substring composed of even-numbered characters is divided into equal-length segments according to the number of packaging areas. Each area verification layer obtains an even-numbered character segment of equal length. If the segment is not divisible, the last area verification layer obtains all remaining characters. Since neither the global verification layer nor the area verification layer can reconstruct the complete history fingerprint on its own, only by combining the two can the complete composite verification structure be reconstructed.

[0069] Furthermore, this step involves writing the global verification layer into the traceability code, and embedding the verification layers for each region into the corresponding packaging area as verification identifiers. See [link / reference]. Figure 4 As shown, following the aforementioned embodiment, the agricultural product packaging 50 includes a front main display area 51, a side information area 52, and a bottom description area 53. A traceability code 54 is located in the lower left corner of the bottom description area 53. A first verification mark 55 is embedded in the front main display area 51, a second verification mark 56 is embedded in the side information area 52, and a third verification mark 57 is embedded in the bottom description area 53. The printing terminal 103 completes the printing of the physical packaging according to the verified agricultural product packaging design scheme, forming the agricultural product packaging entity 104; where solid lines indicate the packaging area boundaries, and dashed lines indicate the embedding positions of the verification marks.

[0070] This application splits the composite verification structure and embeds it into multiple physical carriers, so that any tampering with any carrier will cause the verification data in that carrier to be inconsistent with the verification data in other carriers, thus enabling it to be detected during the verification stage.

[0071] After steps S100 to S400 are completed, the printing terminal 103 completes the printing of the physical packaging of the rice product according to the verified agricultural product packaging design scheme, and the packaging then enters the circulation process.

[0072] Following the aforementioned embodiment, assuming that this batch of rice was packaged, printed, and put on the shelves for sale in June 2024, the main display area 51 on the front presents the two protection measures of black soil protection and straw return to the field in the form of icons, titles, and detailed descriptions. The side information area 52 presents the substitution of chemical fertilizers with organic fertilizers in the form of icons and titles. The bottom description area 53 presents crop rotation and fallow in the form of a text list. The traceability code 54 and three verification marks 55, 56, and 57 are all printed in their corresponding positions. During the distribution process, this packaging will undergo a consumer scanning and verification process. Steps S500 and S600 describe the specific implementation of the verification process.

[0073] S500: When the agricultural product packaging corresponding to the verified agricultural product packaging design scheme is scanned for verification, the global verification layer is read from the traceability code. The verification method is determined based on the comparison result between the total number of packaging areas recorded in the global verification layer and the actual number of verification marks identified. The area verification layer corresponding to each verification mark is read.

[0074] During the distribution process, agricultural product packaging entity 104 may experience damage to the packaging area or blurred verification marks, which may result in the number of verification marks that can actually be identified being inconsistent with the total number of packaging areas recorded in the global verification layer. Therefore, this step requires selecting an appropriate verification method based on the comparison results.

[0075] See Figure 5 As shown, the specific implementation process of this step includes the following sub-steps: S501. Scan the traceability code and read the global verification layer. Specifically, the scanning and verification terminal 105 scans the traceability code 54 on the agricultural product packaging entity 104, reads the data of the global verification layer from the traceability code 54, and extracts the total number of packaging areas from the global verification layer.

[0076] S502. Scan each packaging area and count the number of verification marks. Specifically, the barcode scanning and verification terminal 105 scans each packaging area of ​​the agricultural product packaging entity 104, identifies and counts the number of actual and readable verification marks.

[0077] S503. In this step, the total number of packaging areas is compared with the actual number of verification marks identified. If the two are consistent, step S504 is executed; if the two are inconsistent, step S505 is executed.

[0078] S504. This step reads the corresponding area verification layer in all verification identifiers in a holistic verification manner, and reconstructs the composite verification structure in a unified manner with the global verification layer. The holistic verification method can completely verify all the information of the composite verification structure.

[0079] S505. This step performs verification in a zone-by-zone manner. The zone-by-zone verification method uses the production sequence within the agricultural supply chain as the search path. Based on the location information of each packaging area recorded in the global verification layer, it searches for and reads the corresponding area verification layer from each verification identifier, and performs partial verification based on the read area verification elements. During the zone-by-zone verification process, if any verification identifier cannot be read, this step compares the read area verification elements with the correlation recorded in the global verification layer, and infers the protection items corresponding to the unreadable verification identifier based on the correlation. If all verification identifiers cannot be read, this step determines that the composite verification structure of the agricultural product packaging is invalid.

[0080] Taking the aforementioned embodiment as an example, the inference process is explained as follows: Assuming that the second verification identifier 56 among the three verification identifiers cannot be read, the scanning verification terminal 105 has read the area verification layer corresponding to the main display area 51 on the front of the first verification identifier 55. The area verification elements recorded in this area verification layer include black soil protection and straw return to the field. At the same time, it reads the area verification layer corresponding to the bottom description area 53 of the third verification identifier 57. The area verification elements recorded in this area verification layer include crop rotation and fallow. The correlation recorded in the global verification layer shows that the node order of the agricultural time chain protection is crop rotation and fallow, straw return to the field, organic fertilizer substitution for chemical fertilizer, and black soil protection. Moreover, the position information in the global verification layer shows that the second verification identifier 56 corresponds to the side information area 52. Therefore, when the protection item missing from the read area verification elements and located in the side information area 52 can be uniquely identified as organic fertilizer substitution for chemical fertilizer, it is inferred that the protection item corresponding to the second verification identifier 56 is organic fertilizer substitution for chemical fertilizer. If it cannot be uniquely identified based on the correlation and position information, then the side information area 52 corresponding to the second verification identifier 56 is determined as the area to be reviewed.

[0081] This application automatically selects the verification method based on the comparison results between the total number of packaging areas and the number of verification marks, enabling the verification process to adapt to various situations with different packaging integrity levels.

[0082] S600: Compare the protection items and packaging expression levels read from each area verification layer with the current planting plot protection records to determine the verification results of agricultural product packaging.

[0083] In step S500, the data of the global verification layer and each area verification layer have been read. Since the read verification layer data reflects the status of the planting plot protection record when the packaging is designed, but the planting plot protection record may be updated after the packaging is printed, this step needs to compare the protection items and packaging expression level in the verification layer with the latest planting plot protection record.

[0084] The specific implementation process of this step includes the following sub-steps: S601. Reconstruct the composite verification structure and verify the historical fingerprint. Specifically, after the global verification layer and all regional verification layers jointly reconstruct the composite verification structure, this step verifies the correspondence between the historical fingerprint and the verification elements of each region. Specifically, this step first combines the odd-numbered character fragments in the global verification layer with the even-numbered character fragments in each regional verification layer according to their positions to restore the complete original historical fingerprint. Then, according to the arrangement order of the reconstructed regional verification elements, the ordered hash chain calculation is re-executed to obtain the recalculated historical fingerprint. Finally, this step compares the recalculated historical fingerprint with the restored original historical fingerprint bit by bit. Approved verification items are the set of regional verification elements confirmed to be untampered after the integrity verification of the composite verification structure. Carrier tampering refers to the abnormal state where the data in the verification mark or traceability code on the packaging is artificially modified, causing the composite verification structure to fail the integrity verification. When the two are consistent, this step uses the reconstructed regional verification elements as approved verification items and executes step S602; when the two are inconsistent, this step determines that the agricultural product packaging has been tampered with.

[0085] S602. This step extracts the actual printed verification items from the various protective items actually presented on the agricultural product packaging entity 104, along with their corresponding packaging areas and packaging expression levels. The actual printed verification items are a set of protective item information directly extracted from the actual printed content of the agricultural product packaging. The data structure of the actual printed verification items is the same as that of the area verification elements, including three fields: content identifier, packaging area identifier, and packaging expression level. This step uses the optical character recognition module on the barcode verification terminal 105 to identify the actual printed protective item text and icons in each packaging area from the packaging image, and determines the corresponding packaging expression level according to the rules in Table 3 based on the identified presentation method, thereby extracting the actual printed verification items.

[0086] The workflow of the optical character recognition module is as follows: First, images of each side of the agricultural product packaging are captured by the camera of the barcode verification terminal 105. Then, the images are preprocessed, including tilt correction and noise reduction. Next, a pre-trained text detection model is used to locate the text regions in the image. Finally, character recognition is performed on the detected text regions to obtain the text result. The mapping process from the OCR recognition result to the content identifier is as follows: The recognized protection item name text is subjected to the same hash operation as in step S300, and the first eight hexadecimal characters are extracted. The result is the content identifier of the protection item. The consistency of the protection item can be confirmed by comparing the content identifier with the content identifier stored in the verification identifier. The logic for determining the packaging expression level is as follows: If the recognized packaging area contains icon elements, title text, and detailed description text, it is determined to be Level 1 Emphasis Level; if only icon elements and title text exist without detailed description, it is determined to be Level 2 Standard Level; if only a text list exists without icon elements, it is determined to be Level 3 Brief Level.

[0087] S603. This step compares the approved verification items with the actual printed verification items, and obtains the latest planting plot protection record through the planting plot protection record query interface 106. Based on the current planting plot protection record, the validity of each protection item in the actual printed verification items is determined. Valid protection items are those that exist in the current planting plot protection record, and whose record source and consecutive record count support the packaging expression level corresponding to the protection item. Table 4 shows the validity determination rules, the specific content of which is shown in the table below: Table 4 Validity Judgment Rules

[0088] The threshold values ​​for consecutive recording times range from one to five. The basis for setting the first-level threshold to three and the second-level threshold to two in this application is as follows: The first-level packaging level is presented with an icon, title, and detailed description, providing the strongest visual impact and having the greatest influence on consumer purchasing decisions. Therefore, it requires high continuity of implementation. Setting the threshold to three means that the protected item must be recorded in at least three consecutive planting batches. In actual agricultural production, a planting batch typically lasts six months to a year; therefore, three consecutive batches correspond to one and a half to three years of continuous implementation records, excluding occasional implementation. The second-level packaging level is presented with an icon and title, providing a moderate visual impact. Setting the threshold to two means that the protected item must be recorded in at least two consecutive planting batches, confirming that the protected item is not only implemented in a single batch. The third-level packaging level is presented only as a text list, providing the weakest visual impact. Therefore, no consecutive recording time threshold is set; the existence of the protected item in the current planting plot's protection record is sufficient to meet the third-level requirement.

[0089] The specific rules for determining the verification results are as follows: In the first scenario, when all fields of the printed verification items are consistent with those of the approved verification items, and all protection items meet the above validity judgment rules, this step determines that the agricultural product packaging has passed the verification.

[0090] Following the aforementioned embodiment, suppose that in August 2024, a consumer purchases this batch of rice products at a retail store and uses their mobile phone to scan the traceability code 54 for verification. After completing steps S501 to S602, the scanning verification terminal 105 extracts the approved verification items and the printed verification items. At this time, black soil protection exists in the current planting plot protection record and has been recorded six times consecutively with an authority level of L1. The number of consecutive records is not less than three and the authority level meets the first-level requirement. Therefore, black soil protection is a valid protection item and supports the first-level packaging expression level. Organic fertilizer replacing chemical fertilizer exists in the current planting plot protection record and has been recorded three times consecutively with an authority level of L2, meeting the second-level requirement. Straw returning to the field has been recorded five times consecutively, meeting the second-level requirement. Crop rotation and fallow exist in the current record, meeting the third-level requirement. Since the printed verification items are consistent with the approved verification items and all four protection items are valid protection items, the agricultural product packaging passes the verification. The scanning verification terminal 105 displays a prompt to the consumer: "Verification passed, packaging information is consistent with the planting plot protection record."

[0091] In the second scenario, this step compares the actual printed verification items with the approved verification items item by item. If the level of display of the protected items in the packaging area or corresponding packaging expression level of the actual printed verification items is stronger than that in the packaging area or packaging expression level of the approved verification items, this step determines that the agricultural product packaging is an over-grade reprinted packaging. Over-grade reprinted packaging refers to an abnormal packaging state in which the level of display of the protected items actually printed on the packaging exceeds the level of display originally authorized in the packaging design scheme.

[0092] Following the aforementioned embodiment, suppose that during the distribution process, a third party reprinted the packaging of this batch of rice, changing the presentation of crop rotation and fallow from the third-level presentation in the bottom description area A3 to the first-level presentation in the main display area A1 on the front to improve the product's appearance. At this time, after the consumer scans the code for verification, the scanning verification terminal 105 finds that the packaging area of ​​crop rotation and fallow in the actual printed verification item is marked as A1 and the packaging expression level is first-level, but the packaging area of ​​crop rotation and fallow in the approved verification item is marked as A3 and the packaging expression level is third-level. The display strength in the actual printed verification item is higher than that in the approved verification item. Therefore, it is determined that the packaging of this agricultural product is a reprinted packaging that exceeds the level. After determining that the packaging is a reprinted product of an inappropriate level, this step identifies crop rotation and fallow as the packaging content to be corrected. The packaging content to be corrected refers to the set of protected items in the reprinted product of an inappropriate level that exceed the scope of authorization. This step downgrades or deletes the packaging content to be corrected, regenerates the packaging design content of the corresponding packaging area, and re-executes steps S300 and S400 to regenerate the composite verification structure and the corresponding global verification layer, verification layers of each area, and verification identifier.

[0093] In the third scenario, when the approved verification items match the printed verification items, and the printed verification items contain protection items that are no longer valid protection items in the current planting plot protection record, this step determines the agricultural product packaging as record-invalidated packaging. Record-invalidated packaging refers to an abnormal packaging state where the protection items printed on the packaging are consistent with the verification data at the time of packaging design, but due to updates to the planting plot protection record, some protection items no longer meet the validity determination rules. Continuing with the aforementioned example, suppose that part of the inventory of this batch of rice products remained unsold in the warehouse until March 2025. In January 2025, the local agricultural authority conducted an annual review of plot A. The review results showed that the plot did not continue to implement organic fertilizer substitution for chemical fertilizer from autumn 2024 to spring 2025, causing the consecutive record number of organic fertilizer substitution for chemical fertilizer to drop from three to zero, no longer meeting the support conditions for any packaging expression level.

[0094] At this point, after the consumer scans the code for verification, the verification terminal 105 obtains the current planting plot protection record through the planting plot protection record query interface 106. The record shows that organic fertilizer substitution for chemical fertilizer no longer exists or the consecutive recording count is zero. Therefore, organic fertilizer substitution for chemical fertilizer is no longer a valid protection item, and this step determines that the agricultural product packaging is a record-invalidated package. After determining it to be a record-invalidated package, this step marks organic fertilizer substitution for chemical fertilizer as an invalidated protection item in the verification result and determines the corresponding packaging content as packaging content that will no longer be displayed. This step, based on the record source authority level and consecutive recording count of the other three valid protection items in the current planting plot protection record, re-determines the packaging expression level supported by the remaining valid protection items according to the validity determination rules in Table 4.

[0095] In the fourth scenario, when the actual printed verification items differ from the approved verification items in terms of missing protection items, added protection items, or inconsistent packaging areas, but this inconsistency does not constitute reprinted packaging beyond its approved level or expired packaging records, this step determines the agricultural product packaging as inconsistent in content. Inconsistent packaging is used to characterize an abnormal state where there is a general inconsistency between the actual content presented on the packaging and the approved content confirmed by the composite verification structure. For example, protection items may be omitted, additional unapproved protection items may be added, or protection items may be printed in unauthorized packaging areas.

[0096] In another specific implementation, assuming that by the time of the review in March 2025, not only will the continuous record of organic fertilizer substitution for chemical fertilizer in plot A be zero, but the continuous record of crop rotation and fallow will also be zero due to the continuous planting of plots without fallow. In this case, two protection items in the packaging will be judged to be invalid records. In this step, the verification results will simultaneously mark organic fertilizer substitution for chemical fertilizer and crop rotation and fallow as invalid protection items. The barcode verification terminal 105 will show consumers the prompt "The records of some protection items in the packaging have been updated. Please refer to the latest records" and list the specific names of the invalid protection items.

[0097] This application compares the approved verification items, the actual printed verification items, and the current planting plot protection records from multiple dimensions. It can distinguish four different types of packaging anomalies based on the specific patterns of inconsistencies in the comparison results: carrier tampering, reprinting beyond the designated level, record invalidation, and content inconsistency. This allows the verification results to provide a targeted basis for subsequent packaging disposal.

[0098] This application also provides an artificial intelligence-based personalized design generation system for agricultural product packaging, the system comprising: The determination module is used to identify various protection items from the planting plot protection record based on the planting batch of the agricultural products to be packaged, and generate a protection agricultural time chain according to the production process.

[0099] The generation module is used to generate land parcel history content based on the agricultural time chain, allocate the land parcel history content to each packaging area according to the packaging usage scenario, and generate the initial packaging design scheme through the layout generation model.

[0100] The extraction module is used to extract the protection measures adopted by each packaging area and the corresponding packaging expression level from the initial packaging design scheme as regional verification elements. The module establishes a correlation between the regional verification elements according to the production link sequence of the agricultural protection chain to form a composite verification structure.

[0101] The splitting module is used to split the composite verification structure into a global verification layer and several regional verification layers. The global verification layer is written with a traceability code, and each regional verification layer is embedded in the corresponding packaging area in the form of a verification identifier, forming a verification-enabled agricultural product packaging design scheme.

[0102] The verification module is used to read the global verification layer from the traceability code when the agricultural product packaging corresponding to the verification agricultural product packaging design scheme is scanned for verification. It determines the verification method based on the comparison result between the total number of packaging areas recorded in the global verification layer and the actual number of verification marks identified, and reads the area verification layer corresponding to each verification mark.

[0103] The comparison module is used to compare the protection items and packaging expression levels read from each area verification layer with the current planting plot protection records to determine the verification results of agricultural product packaging.

[0104] The AI-based personalized design generation system for agricultural product packaging in this application is used to implement the aforementioned AI-based personalized design generation method for agricultural product packaging. Therefore, the specific implementation of the AI-based personalized design generation system for agricultural product packaging can be found in the embodiment section of the AI-based personalized design generation method for agricultural product packaging described above. The specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.

[0105] The above provides a detailed description of the method and system for generating personalized agricultural product packaging designs based on artificial intelligence, as provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for generating personalized agricultural product packaging designs based on artificial intelligence, characterized in that, include: Based on the planting batch of the agricultural products to be packaged, various protection items are determined from the protection records of the planting plots, and a protection agricultural time chain is generated according to the production process. Based on the protected agricultural time chain, the land plot history content is generated, and the land plot history content is allocated to each packaging area according to the packaging usage scenario. An initial packaging design scheme is generated through the layout generation model. The protection measures and corresponding packaging expression levels adopted in each packaging area are extracted from the initial packaging design scheme as regional verification elements. The correlation between the regional verification elements is established according to the production link sequence of the agricultural protection chain to form a composite verification structure. The composite verification structure is divided into a global verification layer and several regional verification layers. The global verification layer is written with a traceability code, and each regional verification layer is embedded in the corresponding packaging area in the form of a verification identifier, thus forming a verification-enabled agricultural product packaging design scheme. When the agricultural product packaging corresponding to the verified agricultural product packaging design scheme is scanned for verification, the global verification layer is read from the traceability code, and the verification method is determined based on the comparison result between the total number of packaging areas recorded in the global verification layer and the actual number of verification marks identified, and the area verification layer corresponding to each verification mark is read. The protection items and packaging expression levels read from each area verification layer are compared with the current planting plot protection records to determine the verification results of the agricultural product packaging; The process of establishing correlations between verification elements in each region according to the production sequence of the protected agricultural time chain to form a composite verification structure includes: According to the production process sequence and current packaging batch of each protection item in the agricultural protection chain, all regional verification elements are arranged. A corresponding history fingerprint is generated for all the regional verification elements after the arrangement in an ordered hash chain manner. Each node in the ordered hash chain is linked sequentially according to the production link order of the protected agricultural time chain. The hash input of each node includes the hash output of the previous node and the regional verification element corresponding to the current node. The entire area verification elements, the relationships between the area verification elements, and the history fingerprint are collectively used as a composite verification structure. When the content and arrangement order of all regional verification elements are the same, the generated resume fingerprints are the same; when the content of any regional verification element changes or the arrangement order is adjusted, the output of all subsequent nodes of the ordered hash chain from the changed node changes, and the generated resume fingerprints are different. The step of splitting the composite verification structure into a global verification layer and several regional verification layers includes: Extract the correlation between the verification elements of each region, the total number of packaging regions, and the location information of each packaging region from the composite verification structure, and extract some information from the history fingerprint. Write the correlation, the total number of packaging regions, the location information of each packaging region, and the partial information of the history fingerprint into the global verification layer. Extract the content identifier and packaging expression level of the protected items in each packaging area from the composite verification structure, and extract the remaining information from the history fingerprint. Write the content identifier, the packaging expression level and the remaining information of the history fingerprint into the area verification layer of the corresponding packaging area. The partial information of the resume fingerprint, together with the remaining information of the resume fingerprint, constitutes the complete information of the resume fingerprint. Neither the global verification layer nor the all-region verification layer can independently reconstruct the composite verification structure.

2. The method according to claim 1, characterized in that, The method for determining the verification method based on the comparison between the total number of packaging areas recorded in the global verification layer and the actual number of verification identifiers identified includes: When the total number of packaging areas is consistent with the number of actual identified verification identifiers, the corresponding area verification layer in all verification identifiers is read synchronously in the overall verification method, and the composite verification structure is reconstructed in a unified manner in combination with the global verification layer. When the total number of packaging areas is inconsistent with the number of verification identifiers actually identified, a zone-by-zone verification method is used. The production link sequence in the agricultural protection chain is used as the search path. The corresponding area verification layer in each verification identifier is searched and read one by one according to the location information of each packaging area recorded in the global verification layer. Partial verification is performed based on the area verification elements that have been read.

3. The method according to claim 2, characterized in that, The process of comparing the protection items and packaging expression levels read from each regional verification layer with the current planting plot protection record to determine the verification result of the agricultural product packaging includes: When the history fingerprint corresponds to each region's verification element in the composite verification structure jointly reconstructed by the global verification layer and all region verification layers, the reconstructed region verification elements are used as approved verification items. When the historical fingerprint does not correspond to the verification elements of each region, it is determined that the agricultural product packaging has been tampered with. Extract actual verification items from the various protection measures actually presented on agricultural product packaging, as well as the corresponding packaging areas and packaging expression levels; The approved verification items are compared with the actual verification items, and the validity of each protection item in the actual verification items is determined based on the current protection record of the planting plot, so as to determine the verification result.

4. The method according to claim 3, characterized in that, The determination of the verification result of the agricultural product packaging includes: When the actual printed verification items are consistent with the approved verification items, and all protection items are valid protection items, the agricultural product packaging is deemed to have passed the verification. The actual printing verification items are compared with the approved verification items one by one. When the display intensity of the protected item in the packaging area or packaging expression level corresponding to the actual printing verification item is stronger than that in the packaging area or packaging expression level corresponding to the approved verification item, the agricultural product packaging is determined to be over-grade reprinted packaging. When the approved verification item is consistent with the actual printed verification item, and there is a protection item in the actual printed verification item that is no longer a valid protection item in the current planting plot protection record, the agricultural product packaging is determined to be a record invalid packaging. When the actual printed verification item has missing protection items, added protection items, or inconsistent packaging areas compared to the approved verification item, and it does not belong to over-level reprinted packaging or expired packaging, the agricultural product packaging is determined to be inconsistent packaging. Among them, a valid protection item is a protection item that exists in the current planting plot protection record, and the record source and consecutive record count of the protection item support the packaging expression level corresponding to the protection item.

5. The method according to claim 2, characterized in that, The step of locating and reading the corresponding area verification layer in each verification identifier one by one using an area-by-area verification method also includes: If there are unreadable verification identifiers, the verification elements of each area that have been read are compared with the associations recorded in the global verification layer; if the protection item corresponding to the unreadable verification identifier can be uniquely determined based on the association, the protection item corresponding to the unreadable verification identifier is inferred; if it cannot be uniquely determined, the corresponding packaging area is determined to be the area to be reviewed. If all verification identifiers cannot be read, the composite verification structure of the agricultural product packaging is deemed invalid.

6. The method according to claim 4, characterized in that, After determining that the agricultural product packaging is reprinted beyond its permitted level, the following also applies: The protective items in the actual print verification items that are higher than the approved verification items are identified as packaging contents to be corrected; After downgrading or deleting the packaging content to be corrected, the corresponding packaging design content for the packaging area is regenerated. Based on the regenerated packaging design, the composite verification structure and the corresponding global verification layer, regional verification layers, and verification identifiers are regenerated.

7. The method according to claim 4, characterized in that, After determining that the agricultural product packaging is expired, the following steps are also included: In the verification results, identify the expired protection items and determine the packaging contents corresponding to the expired protection items as packaging contents that will no longer be displayed; Based on the packaging expression level supported by the current planting plot protection record, update the packaging expression level of the remaining valid protection items in the printed verification item.

8. An artificial intelligence-based personalized design generation system for agricultural product packaging, used to execute the artificial intelligence-based personalized design generation method for agricultural product packaging as described in any one of claims 1 to 7, characterized in that, include: The determination module is used to determine various protection items from the planting plot protection record based on the planting batch of the agricultural products to be packaged, and generate a protection agricultural time chain according to the production process sequence; The generation module is used to generate land parcel history content based on the agricultural time chain protection, allocate the land parcel history content to each packaging area according to the packaging usage scenario, and generate the initial packaging design scheme through the layout generation model. The extraction module is used to extract the protection measures adopted by each packaging area and the corresponding packaging expression level from the initial packaging design scheme as regional verification elements. The module establishes the correlation between the regional verification elements according to the production link sequence of the agricultural protection chain to form a composite verification structure. The splitting module is used to split the composite verification structure into a global verification layer and several regional verification layers. The global verification layer is written with a traceability code, and each regional verification layer is embedded in the corresponding packaging area in the form of a verification identifier, forming a verification-enabled agricultural product packaging design scheme. The verification module is used to read the global verification layer from the traceability code when the agricultural product packaging corresponding to the verification agricultural product packaging design scheme is scanned for verification. It determines the verification method based on the comparison result between the total number of packaging areas recorded in the global verification layer and the actual number of verification marks identified, and reads the area verification layer corresponding to each verification mark. The comparison module is used to compare the protection items and packaging expression levels read from each area verification layer with the current planting plot protection records to determine the verification results of agricultural product packaging.

Citation Information

Patent Citations

  • Agricultural and sideline product production and processing record information efficient and trusted iteration tracing system and method

    CN107644239A

  • Agricultural product block chain data storage and tracing method and device and storage medium

    CN121860654A