A physical network fusion transaction method and system based on intelligent electronic scales

Abstract

This invention relates to a physical network fusion transaction method and system based on intelligent electronic scales, belonging to the field of intelligent metering equipment transaction management technology. The method includes: querying product information corresponding to a product to be traded; selecting the product information corresponding to the product to be traded; constructing a product matching and recognition model; extracting information features from the product image; associating and recognizing the product information based on the transaction location information; and outputting traded product data; assessing the product inventory status; when the product inventory status meets a preset transaction inventory adjustment threshold, adjusting the product's listing and delisting status and transaction price in real time on the transaction cloud platform using a distributed consistency mechanism; uploading the adjusted product information data and corresponding transaction information to the cloud platform data hub; coordinating online order processing and offline weighing, transaction, and transportation through a message queue; and publishing product information to the transaction cloud platform.

Description

Technical Field

[0001] This invention belongs to the field of intelligent metering equipment transaction management technology, specifically relating to a physical network fusion transaction method and system based on intelligent electronic scales. Background Technology

[0002] With the rapid development of IoT, AI, and cloud computing technologies, physical retail scenarios are gradually evolving towards intelligence, networking, and datafication. In traditional physical stores, the process of commodity transactions typically relies on manual labor to complete operations such as commodity identification, weighing and pricing, inventory registration, and price adjustments. This method is not only inefficient but also susceptible to differences in human experience and operational errors, making it difficult to meet the needs of high-frequency transactions and refined management.

[0003] While some existing smart scales can perform basic weighing and price calculation functions, their product recognition capabilities are limited. They typically require manual selection of product categories or matching through pre-set coding methods, making them ill-suited for real-world trading environments with complex product types and high visual similarity. Furthermore, most smart scales operate as standalone devices with low levels of data integration with online trading platforms, inventory management systems, and delivery systems, easily leading to inconsistencies between online and offline data, outdated inventory status, and untimely price updates.

[0004] The assessment and adjustment of product inventory status largely rely on manual intervention in the backend or timed batch processing, lacking a dynamic assessment mechanism based on real-time transaction data. When multiple transaction nodes operate concurrently, issues such as overselling of inventory, conflicting product statuses, or inconsistent pricing can easily arise, affecting transaction reliability and system stability. Especially in a distributed transaction environment, existing technologies lack sufficient control over the consistency of inventory data and product status, making it difficult to guarantee the synchronization and accuracy of transaction data across different system nodes.

[0005] On the other hand, the existing online trading platforms and offline weighing and trading equipment have limited coordination capabilities. Transaction data, commodity information and logistics information are often processed through multiple systems, resulting in complex data flow links, high system maintenance costs, and difficulties in real-time monitoring and unified management of the transaction process.

[0006] Therefore, there is an urgent need for a technical solution that can integrate smart electronic scales, intelligent product recognition, dynamic inventory assessment, and cloud platform data collaboration to achieve efficient integration of physical and online transactions, improve the intelligence level of the transaction process, and enhance the overall operational efficiency of the system. Summary of the Invention

[0007] To address the aforementioned problems in the existing technology, this invention provides a physical network fusion transaction method based on a smart electronic scale. The objective of this invention can be achieved through the following technical solutions: S1: Query the product information corresponding to the product to be traded, match and filter the product information based on the transaction information database, select the product information corresponding to the product to be traded according to the transaction requirements, and generate the order information for the traded product; S2: Construct a product matching and recognition model based on the order information of the traded products, extract the information features of the product image, convert them into a preset data field format and input them into the product matching and recognition model, and perform association recognition of product information based on the transaction location information, and output the traded product data; S3: Evaluate the inventory status of the goods based on the transaction goods data. When the inventory status of the goods meets the preset transaction inventory adjustment threshold, adjust the listing status and transaction price of the goods in real time in the transaction cloud platform in combination with the distributed consistency mechanism. S4: Upload the adjusted product information data and corresponding transaction information to the cloud platform data hub, coordinate online order processing and offline weighing, transaction and transportation through message queues, and publish product information to the transaction cloud platform.

[0008] Specifically, the method for querying product information is as follows: indexing and querying the transaction information database based on product identification features, which include product code features and product image features; and matching the product identification features with the pre-stored product information in the transaction information database to obtain product information corresponding to the product to be traded.

[0009] Specifically, the method for filtering the information of goods to be traded is as follows: the information of goods to be traded is compared based on preset filtering parameters, the filtering parameters including the tradable status parameter, the inventory status parameter and the pricing validity parameter; when the information of goods to be traded meets the conditions corresponding to the filtering parameters, it is written into the candidate goods data set and the information of goods to be traded is marked as tradable.

[0010] Specifically, the method for generating the order information of the traded goods is as follows: retrieve the pricing parameters, determine the pricing method based on the demand for the traded goods, the pricing method includes: weight pricing and quantity pricing, obtain the transaction amount information, assign the corresponding identifier index to the transaction amount information according to the product identifier for association and binding, and encapsulate the associated and bound data into a traded goods order data object in combination with the identifier index to generate the traded goods order information.

[0011] Specifically, the method for constructing the product matching and recognition model is as follows: Based on the order information of the traded goods, basic product feature data is extracted. The basic product feature data includes product identification information, product category information, and pricing parameter information. The product identification information and product category information are then converted into numerical feature vectors through feature encoding to obtain a product identification feature input set. Based on the product identification feature input set, the model parameters are initialized, the mapping relationship between product features and product identifiers is established, the model weight parameters and discrimination threshold parameters are fixed and stored, and the corresponding product identifier results are output based on the input product identification features to construct a product matching and identification model.

[0012] Specifically, the method for extracting information features from the product image is as follows: obtain the same input image data scale as the product image to be identified, input the unified product image into the image feature extraction network of the product matching and recognition model, the image feature extraction network is based on multi-layer feature extraction, including low-level visual features and high-level semantic features, the low-level visual features include color distribution features and edge contour features; the high-level semantic features include texture structure features and shape structure features, to generate information features of the product image.

[0013] Specifically, the process of outputting the transaction commodity data is as follows: inputting the commodity identification feature input set corresponding to the commodity to be traded into the commodity matching and identification model, determining the confidence level of the commodity identification result, and when the identification confidence level meets the preset threshold condition, performing consistency verification on the associated commodity information, and encapsulating the commodity identifier and commodity weight according to the preset data structure, and outputting the transaction commodity data.

[0014] Specifically, the method for assessing the inventory status of goods is as follows: Real-time inventory data of the corresponding goods is queried based on the transaction goods data and compared with preset inventory benchmark parameters, including a safety stock threshold, a minimum tradable inventory threshold, and a replenishment warning threshold; when the available inventory is higher than the safety stock threshold, the inventory status of the goods is determined to be in a normal tradable state; when the available inventory is lower than the safety stock threshold but higher than the minimum tradable inventory threshold, the inventory status of the goods is determined to be in a warning trading state; when the available inventory is lower than the minimum tradable inventory threshold, the inventory status of the goods is determined to be in a non-tradable state; and the inventory status of the goods is determined based on the comparison results.

[0015] Specifically, the distributed consistency mechanism coordinates and controls data update operations among multiple nodes of the transaction cloud platform. Before executing the operation of adjusting the status of a product or the transaction price, the target product data is locked. After the product status or transaction price adjustment is completed, the distributed lock is released, and the update result is synchronized to each distributed node, triggering a consistency rollback to re-coordinate.

[0016] Specifically, the method for uploading the product information data to the cloud platform data hub is as follows: the adjusted product information data and corresponding transaction information are encapsulated, and before the data is uploaded, the data message is sent to the cloud platform data access interface based on a secure communication protocol, and a data update notification is issued to the associated transaction processing layer.

[0017] Specifically, the product matching and recognition model also includes a product feature-assisted matching algorithm: to perform auxiliary verification on the product recognition results output by the product matching and recognition model, calculate the feature similarity between the current product image features and the historical feature samples of the target product, and confirm the validity of the product recognition result when the feature similarity meets the preset similarity threshold condition; when the feature similarity does not meet the preset similarity threshold condition, trigger candidate product re-sorting or secondary recognition processing to correct the product recognition result.

[0018] Specifically, a physical network integrated transaction system based on intelligent electronic scales is characterized by comprising: Product information query and order placement module: Query the product information corresponding to the product to be traded, match and filter the product information to be traded based on the transaction information database, select the product information corresponding to the product to be traded according to the transaction requirements, and generate the order information for the traded product; Transaction matching and recognition module: Constructs a product matching and recognition model based on the order information of the transaction products, extracts information features of the product images, converts them into a preset data field format and inputs them into the product matching and recognition model, and performs association recognition of product information based on the transaction location information, and outputs transaction product data; Inventory Status Assessment Module: Assess the inventory status of goods based on the transaction goods data. When the inventory status of goods meets the preset transaction inventory adjustment threshold conditions, adjust the listing status and transaction price of goods in real time in the transaction cloud platform in combination with the distributed consistency mechanism. Platform data synchronization module: Uploads the adjusted product information data and corresponding transaction information to the cloud platform data hub, coordinates online order processing and offline weighing, transaction and transportation through message queues, and publishes product information to the transaction cloud platform.

[0019] The beneficial effects of this invention are as follows: This invention constructs a product information query and order generation mechanism to automatically match and filter product information for transactions, avoiding operational errors caused by manual product selection and improving the accuracy and efficiency of the transaction initiation stage. By introducing a product matching and recognition model, feature extraction is performed on product image information and associated recognition is performed by combining it with transaction location information, enabling automated product identification and confirmation, effectively reducing the need for manual intervention, and improving the accuracy of product identification and the level of intelligence in the transaction process.

[0020] Based on transaction product data, the system performs real-time assessments of product inventory status and dynamically adjusts product listing / delisting status and transaction prices when preset inventory adjustment thresholds are met, incorporating a distributed consistency mechanism. This effectively avoids inventory conflicts and price inconsistencies in multi-node concurrent transaction scenarios, improving the stability and reliability of the transaction system in a distributed environment. Simultaneously, the distributed consistency control mechanism ensures the consistency of inventory data, product status data, and price data within the transaction cloud platform, enhancing the system's data security and controllability.

[0021] The adjusted product and transaction information is uniformly uploaded to the cloud platform's data hub. A message queue mechanism is used to achieve coordinated operation between online order processing, offline weighing and transactions, and logistics transportation, simplifying the data flow and improving the overall system response speed and scalability. This deep integration of physical and online transactions enables real-time release of product information and closed-loop management of the transaction process, helping to improve transaction efficiency, reduce system maintenance costs, and provide a reliable data foundation for subsequent data analysis and business expansion. Attached Figure Description

[0022] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0023] Figure 1 This is a schematic diagram of the framework of a physical network fusion transaction method and system based on an intelligent electronic scale according to the present invention.

[0024] Figure 2 This is a schematic diagram of the business model of a physical network fusion transaction method and system based on an intelligent electronic scale according to the present invention. Detailed Implementation

[0025] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0026] Please see Figure 1 A physical network fusion transaction method based on smart electronic scales: S1: Query the product information corresponding to the product to be traded, match and filter the product information based on the transaction information database, select the product information corresponding to the product to be traded according to the transaction requirements, and generate the order information for the traded product; S2: Construct a product matching and recognition model based on the order information of the traded products, extract the information features of the product image, convert them into a preset data field format and input them into the product matching and recognition model, and perform association recognition of product information based on the transaction location information, and output the traded product data; S3: Evaluate the inventory status of the goods based on the transaction goods data. When the inventory status of the goods meets the preset transaction inventory adjustment threshold, adjust the listing status and transaction price of the goods in real time in the transaction cloud platform in combination with the distributed consistency mechanism. S4: Upload the adjusted product information data and corresponding transaction information to the cloud platform data hub, coordinate online order processing and offline weighing, transaction and transportation through message queues, and publish product information to the transaction cloud platform.

[0027] In this embodiment, the method for querying product information is as follows: indexing and querying the transaction information database based on product identification features, wherein the product identification features include product code features and product image features; and matching the product identification features with the product information pre-stored in the transaction information database to obtain product information corresponding to the product to be traded.

[0028] In this embodiment, the method for filtering the information of goods to be traded is as follows: the information of goods to be traded is compared based on preset filtering parameters, the filtering parameters including the tradable status parameter, the inventory status parameter and the pricing validity parameter; when the information of goods to be traded meets the conditions corresponding to the filtering parameters, it is written into the candidate goods data set and the information of goods to be traded is marked as tradable.

[0029] In this embodiment, the method for generating the order information of the traded goods is as follows: retrieve the pricing parameters, determine the pricing method based on the demand for the traded goods, the pricing method includes: weight pricing and quantity pricing, obtain the transaction amount information, assign the corresponding identifier index to the transaction amount information according to the product identifier for association and binding, and encapsulate the associated and bound data into a traded goods order data object in combination with the identifier index to generate the traded goods order information.

[0030] In this embodiment, a product—a Fuji apple—is placed on the smart electronic scale. The product information pre-stored in the system includes: product identifier: ID_001; product category: fruit, code 01; unit price: 6.00 yuan / kg.

[0031] I. Input of Product Features and Construction of Mapping Relationships During the transaction, the smart electronic scale collects images of the goods to be traded, their weight, and the location information of the transaction. The system first extracts basic characteristic data of the goods based on the order information, including the goods identification number, goods category, and pricing parameters.

[0032] The basic characteristics of the goods extracted from the order information of the traded goods are as follows: , After numerical encoding, it is converted into vector form: .

[0033] II. Construction of Image Feature Extraction Network An image feature extraction network is constructed, which is a multi-layer convolutional neural network, including a low-level visual feature extraction layer and a high-level semantic feature extraction layer.

[0034] (1) Low-level visual feature extraction: Low-level visual features are extracted through the first layer of the convolutional neural network.

[0035] Main color ratio: Red 0.72; Edge density: 0.35; Texture roughness: 0.41; Forming low-level visual feature vectors: f low =[0.72,0.35,0.41].

[0036] (2) High-level semantic feature extraction: High-level networks extract semantic information based on low-level features.

[0037] Confidence level for circular structure: 0.88; Fruit category probability: 0.91; Forming a high-level semantic feature vector: f high =[0.88,0.91].

[0038] (3) Image feature fusion By fusing low-level visual features with high-level semantic features, we obtain the product image feature vector: , III. Mapping of Product Characteristics and Product Labels The image features are concatenated with the basic features to form the model input: , After inputting this feature vector into the product matching and recognition model, the model output is as follows: , The corresponding recognition confidence level is: , The model recognition result is judged based on the preset confidence threshold.

[0039] IV. Product Feature-Assisted Matching Algorithm The system calls the product feature-assisted matching algorithm to match the current image features f img Features of historical samples f img Similarity calculation is performed using (ID_001)=[0.70,0.36,0.40,0.87,0.92]. , Since S≥0.85, the product identification result is confirmed to be valid, and the product is confirmed to be a Fuji apple (ID_001); when S≤0.85, candidate product reordering or secondary identification processing is triggered to correct the identification result.

[0040] V. Generation of Transaction Commodity Data After the product identification is confirmed, the system associates and encapsulates the product identification information, product weight, pricing parameters and transaction location information to generate transaction product data and outputs it to the subsequent inventory assessment and cloud platform synchronization module.

[0041] Assuming the weighing result is: Product weight: w = 0.80 kg, calculate the transaction amount: , The output transaction product data is as follows: Product ID: ID_001; Product weight: 0.80kg; Unit price: 6.00 yuan / kg; Transaction amount: 4.80 yuan; Identification confidence level: 0.96.

[0042] In this embodiment, the method for constructing the product matching and recognition model is as follows: Based on the order information of the traded goods, basic product feature data is extracted. The basic product feature data includes product identification information, product category information, and pricing parameter information. The product identification information and product category information are then converted into numerical feature vectors through feature encoding to obtain a product identification feature input set. Based on the product identification feature input set, the model parameters are initialized, the mapping relationship between product features and product identifiers is established, the model weight parameters and discrimination threshold parameters are fixed and stored, and the corresponding product identifier results are output based on the input product identification features to construct a product matching and identification model.

[0043] In this embodiment, the method for extracting information features from the product image is as follows: obtaining the same input image data scale for the product image to be identified, and inputting the unified product image into the image feature extraction network of the product matching and recognition model. The image feature extraction network is based on multi-layer feature extraction, including low-level visual features and high-level semantic features. The low-level visual features include color distribution features and edge contour features; the high-level semantic features include texture structure features and shape structure features, thereby generating information features of the product image.

[0044] In this embodiment, as Figure 2As shown, consumers can directly query information about tradable goods on the electronic scale, including product name, price, inventory status, and delivery range. This product information is obtained in real-time from the transaction cloud platform via the electronic scale's communication module and then cached and displayed locally.

[0045] After a consumer selects a desired item using the electronic scale, the item is added to their shopping list. For items requiring physical verification, the scale acquires an image of the item via its image acquisition module and automatically identifies and confirms it based on its built-in item matching and recognition model. Simultaneously, it uses a weighing sensor to obtain the item's weight information, generating corresponding transaction data. The system automatically calculates the transaction amount based on this data and displays the order details to the consumer on the scale's interface.

[0046] After the consumer confirms the purchase, the electronic scale encapsulates the transaction data into online order information and sends it to the transaction cloud platform via a secure communication mechanism. Upon receiving the order information, the transaction cloud platform creates the order and generates a payment request, then sends a payment QR code or payment confirmation information back to the electronic scale. The consumer can then complete the online payment directly on the electronic scale.

[0047] After payment is completed, the electronic scale synchronizes the order status update information to the transaction cloud platform. Based on the order information, the transaction cloud platform automatically triggers the delivery process and pushes the order information to the delivery dispatch system. The delivery system combines the order's delivery address information to generate a delivery task and assigns the corresponding delivery personnel to complete the delivery of the goods, enabling the goods to be delivered directly from the store to the consumer's designated address.

[0048] Through the above technical process, this embodiment realizes an integrated business model that combines physical transactions and online transactions with a single Dahua smart electronic scale as the core. This allows operators to complete product display, transaction processing, payment settlement, and delivery coordination without the need for additional cash register equipment, independent ordering terminals, or third-party systems, significantly improving operational efficiency and consumer transaction experience.

[0049] In this embodiment, the process of outputting the traded commodity data is as follows: inputting the commodity identification feature input set corresponding to the commodity to be traded into the commodity matching and identification model, determining the confidence level of the commodity identification result, and when the identification confidence level meets the preset threshold condition, performing consistency verification on the associated commodity information, and encapsulating the commodity identifier and commodity weight according to the preset data structure, and outputting the traded commodity data.

[0050] In this embodiment, the method for assessing the inventory status of goods is as follows: Real-time inventory data of the corresponding goods is queried based on the transaction goods data and compared with preset inventory benchmark parameters, which include a safety stock threshold, a minimum tradable inventory threshold, and a replenishment warning threshold; when the available inventory is higher than the safety stock threshold, the inventory status of the goods is determined to be in a normal tradable state; when the available inventory is lower than the safety stock threshold but higher than the minimum tradable inventory threshold, the inventory status of the goods is determined to be in a warning trading state; when the available inventory is lower than the minimum tradable inventory threshold, the inventory status of the goods is determined to be in a non-tradable state; and the inventory status of the goods is determined based on the comparison results.

[0051] In this embodiment, the distributed consistency mechanism coordinates and controls data update operations among multiple nodes of the transaction cloud platform. Before executing the operation of adjusting the status of a product or the transaction price, the target product data is locked. After the product status or transaction price adjustment is completed, the distributed lock is released, and the update result is synchronized to each distributed node, triggering a consistency rollback to re-coordinate.

[0052] In this embodiment, the method for uploading the product information data to the cloud platform data hub is as follows: the adjusted product information data and the corresponding transaction information are encapsulated, and before the data is uploaded, the data message is sent to the cloud platform data access interface based on the secure communication protocol, and a data update notification is issued to the associated transaction processing layer.

[0053] In this embodiment, the product matching and recognition model further includes a product feature-assisted matching algorithm: to perform auxiliary verification on the product recognition result output by the product matching and recognition model, calculate the feature similarity between the current product image features and the historical feature samples of the target product, and confirm the validity of the product recognition result when the feature similarity meets the preset similarity threshold condition; when the feature similarity does not meet the preset similarity threshold condition, trigger candidate product reordering or secondary recognition processing to correct the product recognition result.

[0054] This invention also provides a physical network fusion transaction system based on a smart electronic scale, specifically including: Product information query and order placement module: Query the product information corresponding to the product to be traded, match and filter the product information to be traded based on the transaction information database, select the product information corresponding to the product to be traded according to the transaction requirements, and generate the order information for the traded product; Transaction matching and recognition module: Constructs a product matching and recognition model based on the order information of the transaction products, extracts information features of the product images, converts them into a preset data field format and inputs them into the product matching and recognition model, and performs association recognition of product information based on the transaction location information, and outputs transaction product data; Inventory Status Assessment Module: Assess the inventory status of goods based on the transaction goods data. When the inventory status of goods meets the preset transaction inventory adjustment threshold conditions, adjust the listing status and transaction price of goods in real time in the transaction cloud platform in combination with the distributed consistency mechanism. Platform data synchronization module: Uploads the adjusted product information data and corresponding transaction information to the cloud platform data hub, coordinates online order processing and offline weighing, transaction and transportation through message queues, and publishes product information to the transaction cloud platform.

[0055] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A physical network fusion transaction method based on intelligent electronic scales, characterized in that, include: S1: Query the product information corresponding to the product to be traded, match and filter the product information based on the transaction information database, select the product information corresponding to the product to be traded according to the transaction requirements, and generate the order information for the traded product; S2: Construct a product matching and recognition model based on the order information of the traded products, extract the information features of the product image, convert them into a preset data field format and input them into the product matching and recognition model, and perform association recognition of product information based on the transaction location information, and output the traded product data; S3: Evaluate the inventory status of the goods based on the transaction goods data. When the inventory status of the goods meets the preset transaction inventory adjustment threshold, adjust the listing status and transaction price of the goods in real time in the transaction cloud platform in combination with the distributed consistency mechanism. S4: Upload the adjusted product information data and corresponding transaction information to the cloud platform data hub, coordinate online order processing and offline weighing, transaction and transportation through message queues, and publish product information to the transaction cloud platform.

2. The method according to claim 1, characterized in that, The method for querying product information is as follows: indexing the transaction information database based on product identification features, which include product code features and product image features; and matching the product identification features with the pre-stored product information in the transaction information database to obtain product information corresponding to the product to be traded.

3. The method according to claim 1, characterized in that, The method for filtering the information of goods to be traded is as follows: the information of goods to be traded is compared based on preset filtering parameters, the filtering parameters include the tradability status parameter, the inventory status parameter and the pricing validity parameter; when the information of goods to be traded meets the corresponding conditions of the filtering parameters, it is written into the candidate goods data set and the information of goods to be traded is marked as tradable.

4. The method according to claim 1, characterized in that, The method for generating the order information of the traded goods is as follows: retrieve the pricing parameters, determine the pricing method based on the demand for the traded goods, the pricing method includes: weight pricing and quantity pricing, obtain the transaction amount information, assign the corresponding identifier index to the transaction amount information according to the product identifier for association and binding, and encapsulate the associated and bound data into a traded goods order data object in combination with the identifier index to generate the traded goods order information.

5. The method according to claim 2, characterized in that, The method for constructing the product matching and recognition model is as follows: Based on the order information of the traded goods, basic product feature data is extracted. The basic product feature data includes product identification information, product category information, and pricing parameter information. The product identification information and product category information are then converted into numerical feature vectors through feature encoding to obtain a product identification feature input set. Based on the product identification feature input set, the model parameters are initialized, the mapping relationship between product features and product identifiers is established, the model weight parameters and discrimination threshold parameters are fixed and stored, and the corresponding product identifier results are output based on the input product identification features to construct a product matching and identification model.

6. The method according to claim 5, characterized in that, The method for extracting information features from product images is as follows: obtain the same input image data scale for the product images to be identified, input the unified product images into the image feature extraction network of the product matching and recognition model, the image feature extraction network is based on multi-layer feature extraction, including low-level visual features and high-level semantic features, the low-level visual features include color distribution features and edge contour features; the high-level semantic features include texture structure features and shape structure features, to generate information features of the product images.

7. The method according to claim 4, characterized in that, The process of outputting the traded commodity data is as follows: inputting the commodity identification feature input set corresponding to the commodity to be traded into the commodity matching and identification model, determining the confidence level of the commodity identification result, and when the identification confidence level meets the preset threshold condition, performing consistency verification on the associated commodity information, and encapsulating the commodity identifier and commodity weight according to the preset data structure, and outputting the traded commodity data.

8. The method according to claim 2, characterized in that, The method for assessing the inventory status of the goods is as follows: query the real-time inventory data of the corresponding goods based on the transaction goods data, and compare it with the preset inventory benchmark parameters, which include the safety stock threshold, the minimum tradable inventory threshold, and the replenishment warning threshold. When the available inventory is higher than the safety stock threshold, the inventory status of the product is determined to be in a normal tradable state. When the available inventory is below the safety stock threshold but above the minimum tradable inventory threshold, the inventory status of the product is determined to be a warning transaction status. When the available inventory is below the minimum tradable inventory threshold, the inventory status of the product is determined to be non-tradable. And determine the inventory status of goods based on the comparison results.

9. The method according to claim 4, characterized in that, The distributed consistency mechanism coordinates and controls data update operations across multiple nodes on the transaction cloud platform. Before executing operations to adjust the status of a product or the transaction price, the target product data is locked. After the product status or transaction price adjustment is completed, the distributed lock is released, and the update result is synchronized to each distributed node, triggering a consistency rollback for re-coordination.

10. The method according to claim 2, characterized in that, The method for uploading the product information data to the cloud platform data hub is as follows: the adjusted product information data and corresponding transaction information are encapsulated, and before the data is uploaded, the data message is sent to the cloud platform data access interface based on a secure communication protocol, and a data update notification is issued to the associated transaction processing layer.

11. The method according to claim 7, characterized in that, The product matching and recognition model also includes a product feature-assisted matching algorithm: to perform auxiliary verification on the product recognition results output by the product matching and recognition model, calculate the feature similarity between the current product image features and the historical feature samples of the target product, and confirm the validity of the product recognition result when the feature similarity meets the preset similarity threshold condition; when the feature similarity does not meet the preset similarity threshold condition, trigger candidate product reordering or secondary recognition processing to correct the product recognition result.

12. A physical network fusion transaction system based on a smart electronic scale, used to perform the method as described in any one of claims 1-11, characterized in that, include: Product information query and order placement module: Query the product information corresponding to the product to be traded, match and filter the product information to be traded based on the transaction information database, select the product information corresponding to the product to be traded according to the transaction requirements, and generate the order information for the traded product; Transaction matching and recognition module: Constructs a product matching and recognition model based on the order information of the transaction products, extracts information features of the product images, converts them into a preset data field format and inputs them into the product matching and recognition model, and performs association recognition of product information based on the transaction location information, and outputs transaction product data; Inventory Status Assessment Module: Assess the inventory status of goods based on the transaction goods data. When the inventory status of goods meets the preset transaction inventory adjustment threshold conditions, adjust the listing status and transaction price of goods in real time in the transaction cloud platform in combination with the distributed consistency mechanism. Platform data synchronization module: Uploads the adjusted product information data and corresponding transaction information to the cloud platform data hub, coordinates online order processing and offline weighing, transaction and transportation through message queues, and publishes product information to the transaction cloud platform.

Images