Pricing method and device, equipment and storage medium

Abstract

The invention relates to the technical field of Internet of Things, and discloses a pricing method, device and equipment and a storage medium, and the method comprises the steps: capturing the biological characteristics of a user through a camera of disk binding equipment, or scanning the identity identification code of the user, obtaining the identity information of the user according to the biological characteristics or the identity identification code of the user, and transmitting the information to the disk binding equipment; the method comprises the following steps: acquiring carrier information through a carrier identification code, establishing a binding relationship between a carrier and a user according to the carrier information and user identity information, monitoring weight change of the carrier based on the binding relationship, identifying an article image on the carrier, and generating a pricing order according to at least one of the weight change and the article image. According to the method, the carrier information is acquired through the carrier identification code and is bound with the user identity information, so that accurate association of the carrier and the user is realized, the pricing order is generated based on the weight change and the image feature, different scenes can be flexibly coped, the pricing accuracy and efficiency are improved, and possible errors caused by manual pricing are effectively avoided.

Description

Technical Field

[0001] This application relates to the field of Internet of Things (IoT) technology, and in particular to a pricing method, apparatus, device, and storage medium. Background Technology

[0002] Against the backdrop of enterprise digital transformation, pricing and payment processes in various scenarios are facing the dual challenges of efficiency and accuracy. Traditional pricing methods rely heavily on manual operation or simple card-swiping systems, such as in cafeteria dining, logistics and warehousing billing, or retail checkout. Payment methods are primarily card swiping, which suffers from issues like easily lost cards, cumbersome replacement processes, high card production costs for enterprises, and complex management. In the pricing process, manual price calculation is inefficient and prone to omissions and errors, especially in scenarios where items need to be retrieved multiple times. These problems are further exacerbated, leading to longer waiting times, decreased customer satisfaction, and reduced revenue. Summary of the Invention

[0003] The main purpose of this application is to provide a pricing method that addresses the technical problems of low accuracy and efficiency in existing pricing methods.

[0004] To achieve the above objectives, this application proposes a pricing method, the method comprising:

[0005] The device captures the user's biometric features through its camera or scans the user's identification code, and obtains the user's identity information based on the biometric features or the identification code.

[0006] Carrier information is obtained through carrier identification encoding. Based on the carrier information and the user identity information, a binding relationship between the carrier and the user is established, and the weight change of the carrier is monitored based on the binding relationship.

[0007] Identify images of objects on a carrier;

[0008] A pricing order is generated based on at least one of the weight change and the item image.

[0009] In one embodiment, the step of generating a pricing order based on at least one of the weight change and the item image includes:

[0010] The image of the item is segmented to obtain the item region;

[0011] Image features of the object region are extracted using a convolutional neural network;

[0012] The image features are compared with the pre-stored feature vectors in the pre-stored item library to obtain a similarity matrix. The item corresponding to the pre-stored feature vector with the highest similarity in the similarity matrix is ​​taken as the target matching result. The pre-stored feature vector is the feature vector extracted from each item in the pre-stored item library.

[0013] When the target matching result is not empty, the corresponding item information is associated with it, and a pricing order is generated based on the corresponding item information.

[0014] When the target matching result is empty, the image is added to the pre-stored item library, and a pricing order is generated according to the preset configuration information. A reminder notification is sent to the staff to check the recognition status.

[0015] In one embodiment, the step of segmenting the item image to obtain item regions includes:

[0016] The highly reflective areas of the object image are processed by a bilateral filter, and the low-contrast areas of the object image are nonlinearly enhanced to obtain the processed image.

[0017] The processed image is segmented using an image segmentation model to obtain the item region. The image segmentation model includes a YOLOv8 segmentation network obtained by extracting the segmentation capability of the SAM pre-trained model through a feature distillation algorithm and fine-tuning it on a preset dataset.

[0018] In one embodiment, the step of segmenting the processed image using an image segmentation model to obtain the item region includes:

[0019] The processed image is input into the encoder of the segmentation network to obtain multi-scale feature maps;

[0020] Channel attention fusion is performed on the multi-scale feature map to generate a modulated feature map;

[0021] The modulated feature map is input into the decoder of the segmentation network to obtain the decoded features;

[0022] Spatial attention is calculated on the multi-scale feature map to obtain the spatial attention calculation result, and the spatial attention calculation result is multiplied by the Hadamard product with the decoded features to obtain the optimized features;

[0023] The optimized features are convolved and normalized to generate an item region probability map;

[0024] The segmentation boundaries are adjusted based on the confidence level of the item region probability map to obtain the final item region.

[0025] In one embodiment, the steps of obtaining carrier information through carrier identification encoding, establishing a binding relationship between the carrier and the user based on the carrier information and the user identity information, and monitoring changes in the carrier weight based on the binding relationship include:

[0026] Scan the pre-printed carrier identification code on the carrier to obtain carrier information;

[0027] The historical orders and the number of currently bound carriers in the user identity information are verified to obtain the verification result;

[0028] When the verification result indicates that there are no unpaid historical orders and the number of currently bound carriers does not exceed a preset binding threshold, a binding relationship between the carrier and the user is generated by associating and storing the carrier information and the user identity information.

[0029] The weight change of the carrier is monitored based on the binding relationship.

[0030] In one embodiment, the step of generating a pricing order based on at least one of the weight change and the item image includes:

[0031] When the weighing device detects that the current carrier is in a bound state, it generates a pricing order with associated corresponding user identity information based on the weight change and the item information on the current carrier. The corresponding user identity information is the user identity information marked in the weighing device according to the binding relationship.

[0032] In one embodiment, after the step of generating a pricing order based on at least one of the weight change and the item image, the method further includes:

[0033] While connected to the network, the system calls the backend user account corresponding to the user's identity information to deduct the payment.

[0034] In the event of a network outage, the user's biometrics are compared with local characteristics, and an offline transaction record is created if the comparison is successful.

[0035] Network connectivity is detected at fixed time intervals. When the network connectivity meets the requirement of network connection recovery and the network load is lower than a preset threshold, the offline transaction record is marked as pending determination.

[0036] After successful payment, the pending status will be updated to completed status.

[0037] When the pending status is detected to have not been updated for more than a preset time, the corresponding user will be added to the disk binding restriction list.

[0038] Furthermore, to achieve the above objectives, this application also proposes a pricing device, the pricing device comprising:

[0039] The user information acquisition module is used to capture the user's biometric features through the camera of the binding device, or scan the user's identity identification code, and obtain the user's identity information based on the user's biometric features or the identity identification code;

[0040] The weight detection module is used to obtain carrier information through carrier identification code, establish a binding relationship between the carrier and the user based on the carrier information and the user identity information, and monitor the weight change of the carrier based on the binding relationship.

[0041] The image recognition module is used to recognize images of objects on a carrier.

[0042] The pricing module is used to generate a pricing order based on at least one of the weight change and the item image.

[0043] In addition, to achieve the above objectives, this application also proposes a pricing device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the pricing method as described above.

[0044] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and which, when executed by a processor, implements the steps of the pricing method described above.

[0045] The technical solution proposed in this application captures user biometrics or scans user identification codes using a camera on a data binding device. Based on the user's biometrics or identification code, user identity information is obtained. Carrier information is obtained through a carrier identification code. A binding relationship is established between the carrier and the user based on the carrier information and the user's identity information. Weight changes on the carrier are monitored based on this binding relationship, and images of items on the carrier are identified. A pricing order is generated based on at least one of the weight changes and the item images. This application achieves precise association between the carrier and the user by obtaining carrier information through carrier identification codes and binding it to user identity information. The generation of pricing orders based on weight changes and image features allows for flexible adaptation to different scenarios, improving the accuracy and efficiency of pricing and effectively avoiding errors that may occur with manual pricing. Attached Figure Description

[0046] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0047] 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, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating the pricing method of this application in Embodiment 1;

[0049] Figure 2 This is a flowchart illustrating the pricing method of this application in Embodiment 2.

[0050] Figure 3 This is a flowchart illustrating the pricing method of this application in Embodiment 3;

[0051] Figure 4 A schematic diagram of the pricing process for a single-vegetable-weighing equipment;

[0052] Figure 5 A schematic diagram of the intelligent identification and pricing process for dishes;

[0053] Figure 6 This is a schematic diagram of the feature recognition process;

[0054] Figure 7 This is a schematic diagram of the network detection and verification process;

[0055] Figure 8 This is a schematic diagram of the module structure of the pricing device in an embodiment of this application.

[0056] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0057] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0058] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0059] Existing pricing methods rely on manual operation or simple card-swiping systems, which have technical problems such as easy card loss, cumbersome replacement process, low efficiency and error-proneness of manual pricing.

[0060] Therefore, in order to overcome the above-mentioned defects, this application provides a solution that obtains carrier information through carrier identification encoding and binds it with user identity information, thereby realizing the precise association between carrier and user. It generates pricing orders based on weight changes and image features, which can flexibly cope with different scenarios, improve the accuracy and efficiency of pricing, and effectively avoid errors that may occur in manual pricing.

[0061] It should be noted that the executing entity of each embodiment of this application can be a computing service system with data processing, network communication, and program execution functions, such as an electronic system or a pricing system capable of implementing the above functions. The following description uses a pricing system (hereinafter referred to as "the system") as an example to illustrate the following embodiments.

[0062] Based on this, the embodiments of this application provide a pricing method, referring to Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the pricing method of this application.

[0063] In this embodiment, the pricing method includes steps S10 to S40:

[0064] Step S10: Capture the user's biometric features through the camera of the binding device, or scan the user's identity code, and obtain the user's identity information based on the user's biometric features or the identity code.

[0065] It should be understood that this step involves two methods of obtaining user identity information: First, the binding device is equipped with a camera to capture the user's biometric features, such as facial images or other biometric information. These biometric features are unique and can accurately identify the user. Second, the system can also obtain user identity information by scanning the user's identification code. This identification code can be a barcode, QR code, or Radio Frequency Identification (RFID) tag, or other form of encoding. These codes store the user's identity information, and scanning them allows for quick retrieval of the user's identity information. Based on the captured user biometric features or the scanned identification code, the system will further retrieve and obtain complete user identity information from the database.

[0066] Step S20: Obtain carrier information through carrier identification encoding, establish a binding relationship between carrier and user based on carrier information and user identity information, and monitor carrier weight changes based on the binding relationship.

[0067] It should be noted that the carrier identification code (such as barcode, RFID tag, etc.) is pre-printed or embedded on the carrier and is used to uniquely identify the carrier.

[0068] It should be understood that the system uses carrier identification codes to obtain relevant information about the carrier (such as plates, trays, etc.). The system reads the carrier identification code through a scanning device to obtain the carrier's unique identification information (carrier information). At the same time, the system collects user identity information, which can be a user identification code (such as an employee number or membership card number) or user biometrics (such as facial information). By matching and storing the carrier's identification information with the user's identity information, a binding relationship between the carrier and the user is established, and then the weight changes of the carrier are monitored based on this binding relationship.

[0069] Step S30: Identify the image of the item on the carrier.

[0070] In practice, images of objects on a carrier can be captured using a camera or other image acquisition devices, and then visual analysis of the objects placed on the carrier can be performed using image recognition technology.

[0071] Step S40: Generate a pricing order based on at least one of the weight change and the item image.

[0072] In this application, after the binding relationship between the carrier and the user is established, a pricing order can be generated based on the weight change of the carrier or the image features of the items on the carrier (or a combination of both). Specifically, the system monitors the weight change of the carrier through a weighing device, and calculates the cost of the corresponding items by combining preset item unit prices and weight information; simultaneously, the system can also identify the image features of the items on the carrier, match the corresponding item information, and generate a pricing order based on the item's unit price and quantity. This process can choose to use weight change, item image features, or a combination of both to generate orders according to the actual application scenario, thereby achieving flexible and accurate pricing functionality.

[0073] It should be noted that this application includes a weighing-based pricing mode and a dish recognition-based pricing mode. The weighing-based pricing mode is based on weight monitoring equipment, which generates a pricing order by capturing weight changes. The dish recognition-based pricing mode relies on image recognition equipment, which uses image features to identify dishes and generate a pricing order. In practical applications, the system intelligently selects the pricing signal source based on the hardware status. When the weight monitoring equipment is functioning normally, the weighing-based pricing mode is used first. When the image recognition equipment is functioning normally, the system switches to the dish recognition-based pricing mode. If one device malfunctions, the other device is automatically activated for pricing. When both devices are functioning normally, the system flexibly selects the appropriate signal source to generate a pricing order based on user needs.

[0074] As one implementation, step S40 in this embodiment may include: when the weighing device verifies that the current carrier is in a bound state, generating a pricing order associated with corresponding user identity information based on the weight change and the item information on the current carrier, wherein the corresponding user identity information is user identity information marked in the weighing device according to the binding relationship.

[0075] In practice, when the weighing equipment detects that the carrier is in a bound state—that is, the carrier has been successfully bound to a user—the system calculates the fee based on the carrier's weight change and detailed information about the items on the carrier (such as item type, unit price, etc.). Simultaneously, the system associates the generated pricing order with the corresponding user's detailed identity information based on the previously established binding relationship. This corresponding user identity information refers to the user information marked in the weighing equipment through the binding relationship, ensuring that the pricing order accurately belongs to the correct user.

[0076] It should be noted that the pricing method of this application has good scalability and can be applied to multiple real-world scenarios, such as cafeteria dining, logistics warehousing billing, or retail POS.

[0077] For example, in a logistics warehousing environment, goods or pallets are embedded or printed with carrier identification codes. When goods are handled, stored, or shipped out, staff scan the carrier identification codes and the operator's identification (such as employee number) using a scanning device. The system dynamically calculates costs (such as handling fees, storage fees, etc.) based on the scanned information, combined with changes in the weight of the goods (monitored by weighing equipment) and image recognition results (identifying the appearance features of the goods). In retail stores, products are usually labeled with barcodes or QR codes containing product information. In response to a checkout request at the cashier, the system scans the barcodes or QR codes on the products with a barcode scanner. At the same time, the customer's identity information (such as verification via bank card, mobile payment, or membership card) is captured. A pricing order is generated based on changes in the weight of the goods (monitored by weighing equipment) and image recognition results (identifying the appearance features of the goods).

[0078] This embodiment obtains carrier information through carrier identification encoding and binds it to user identity information, achieving a precise association between the carrier and the user. It generates pricing orders based on weight changes and image features, flexibly adapting to different scenarios, improving pricing accuracy and efficiency, and effectively avoiding errors that may occur with manual pricing. Furthermore, by marking the user identity information corresponding to the carrier in the weighing equipment, monitoring weight changes, and generating pricing orders associated with the user identity information when the carrier is in a bound state, it achieves automated and precise pricing, avoiding interference from human factors, improving pricing efficiency and accuracy, and ensuring the accurate association between pricing orders and user identities, facilitating subsequent management and settlement.

[0079] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. On this basis, refer to Figure 2 Step S40 may include steps S401 to S405:

[0080] Step S401: Segment the item image to obtain the item region.

[0081] It should be understood that an image of an object on a carrier can be divided into multiple object regions based on the shape, color, or other characteristics of the object, with each region corresponding to one object.

[0082] In one implementation, step S401 in this embodiment may include: processing the highly reflective areas of the object image using a bilateral filter and performing nonlinear enhancement on the low-contrast areas of the object image to obtain a processed image; segmenting the processed image using an image segmentation model to obtain the object region, wherein the image segmentation model includes a YOLOv8 segmentation network obtained by extracting the segmentation capability of the SAM pre-trained model using a feature distillation algorithm and fine-tuning it on a preset dataset.

[0083] This step describes the preprocessing of the object image to improve image quality, enabling more accurate subsequent image segmentation and feature extraction. Specifically, highly reflective areas in the image are processed using a bilateral filter. Bilateral filters can remove noise while maintaining edge sharpness, effectively reducing the interference of highly reflective areas on image segmentation. For low-contrast areas, a nonlinear enhancement method is used to improve image contrast and enhance image detail. Nonlinear enhancement can improve contrast by adjusting the image's histogram distribution or by utilizing neural networks.

[0084] It should be noted that image segmentation models are deep learning-based models capable of automatically identifying different regions in an image. Feature distillation is a technique used to optimize model performance. Its principle is to transfer knowledge from a large pre-trained model (usually a powerful and complex model, such as a residual network or a VGG (Visual Geometry Group Network)) to a smaller model. This smaller model (i.e., the segmentation network) learns the feature representations and decision-making processes of the larger model, thus maintaining high performance while reducing computational resource consumption. In this application, the segmentation capabilities of the SAM open-source network are distilled into the YOLOv8-SEG model used by the device using a feature distillation algorithm, and fine-tuned on a proprietary dataset. This allows the segmentation network to better understand and recognize the features of objects, thereby more accurately segmenting object regions in the image.

[0085] As one implementation method, the step of segmenting the processed image using an image segmentation model to obtain the item region in this embodiment may further include: inputting the processed image into the encoder of the segmentation network to obtain a multi-scale feature map; performing channel attention fusion on the multi-scale feature map to generate a modulated feature map; inputting the modulated feature map into the decoder of the segmentation network to obtain decoded features; performing spatial attention calculation on the multi-scale feature map to obtain a spatial attention calculation result, and performing a Hadamard product between the spatial attention calculation result and the decoded features to obtain optimized features; performing convolution and normalization processing on the optimized features to generate an item region probability map; and adjusting the segmentation boundary according to the confidence level of the item region probability map and the weight change of the carrier to obtain the final item region.

[0086] In its implementation, the system inputs the processed image into the encoder part of the segmentation network to extract the image's feature information and generate multi-scale feature maps. Multi-scale feature maps refer to image features extracted at different scales, capable of capturing both details and global information within the image. The encoder extracts image features step-by-step through operations such as convolutional layers and pooling layers, generating feature maps at different scales.

[0087] Next, channel attention fusion is performed on these multi-scale feature maps. Channel attention mechanisms (such as SENet (Squeeze-and-Excitation Networks)) dynamically adjust the importance of each channel to generate modulated feature maps, thereby enhancing the ability to express key features. The modulated feature maps are then input into the decoder part of the segmentation network. The decoder gradually recovers the image's detailed information through operations such as deconvolutional layers and upsampling layers, ultimately obtaining the decoded feature maps. Simultaneously, spatial attention is calculated on the multi-scale feature maps. Spatial attention mechanisms (such as Convolutional Block Attention Module (CBAM)) can highlight important spatial regions in the image.

[0088] Furthermore, the spatial attention calculation result is combined with the decoded feature map using a Hadamard product (i.e., element-wise multiplication) to further optimize the feature map and enhance its ability to focus on the target region. The optimized feature map undergoes convolution and normalization to generate an item region probability map, which represents the probability that each pixel belongs to an item region. Finally, the segmentation boundaries are adjusted based on the confidence level (probability value) of the item region probability map. Thresholding or edge optimization algorithms are used to accurately determine the boundaries of the item regions, thus obtaining the final item regions.

[0089] Step S402: Extract image features of the object region using a convolutional neural network.

[0090] It should be understood that after obtaining the object region, a convolutional neural network (CNN) is used to extract image features for each object region, which may include color distribution, texture features, shape features, etc. A CNN is a deep learning architecture that progressively extracts local and global features of an image through operations such as convolutional layers, activation layers, and pooling layers. In the convolutional layers, the convolutional kernel slides across the image and performs convolution operations, extracting local features such as edges and textures; the activation layers introduce non-linearity to enhance the model's expressive power; and the pooling layers reduce the dimensionality of the feature maps through downsampling operations, preserving important information. Through multiple layers of convolution and pooling operations, the CNN can extract discriminative feature vectors from the image of the object region. These feature vectors can effectively represent information such as the shape, color, and texture of the object (i.e., image features).

[0091] Step S403: Compare the image features with the pre-stored feature vectors in the pre-stored item library to obtain a similarity matrix, and take the item corresponding to the pre-stored feature vector with the highest similarity in the similarity matrix as the target matching result, wherein the pre-stored feature vector is the feature vector extracted from each item in the pre-stored item library.

[0092] In its implementation, the system first extracts pre-stored feature vectors related to the carrier identification code from a pre-stored item database. These pre-stored feature vectors are used to describe the characteristic information of known items. Then, the image features are compared with these pre-stored feature vectors to calculate their similarity. The similarity matrix is ​​a two-dimensional array that records the similarity value between the image feature vector and each pre-stored feature vector. By analyzing the similarity matrix, the pre-stored feature vector with the highest similarity is found, and the item corresponding to this vector is the target matching result. This process utilizes distance or similarity measurement methods between feature vectors (such as Euclidean distance, cosine similarity, etc.) to achieve fast and accurate identification of items in the image.

[0093] Step S404: When the target matching result is not empty, associate it with the corresponding item information and generate a pricing order based on the corresponding item information.

[0094] When the system finds the most similar pre-stored feature vector through image feature comparison and determines the target matching result, it means that the item in the image has been successfully identified. At this point, the target matching result is associated with the corresponding item information in the pre-stored item database. This item information may include detailed data such as the item's name, price, and specifications. Subsequently, a pricing order is generated based on the item information.

[0095] Step S405: When the target matching result is empty, add the image to the pre-stored item library, generate a pricing order according to the preset configuration information, and send a reminder notification to the staff to check the recognition status.

[0096] If the target matching result is empty, meaning the system failed to identify the item, the system adds the image to the pre-stored item library to expand its contents. Simultaneously, the system generates a pricing order based on preset configuration information, which can be a default price or pricing rules, to handle cases involving unknown items.

[0097] This embodiment obtains item regions through image segmentation and matches them with a pre-stored item database, enabling fast and accurate item identification. Upon successful matching, item information is associated with the item to generate an order; otherwise, the image is added to the item database and an order is generated. This improves the accuracy and speed of item recognition and allows for continuous enrichment of the pre-stored item database to accommodate the pricing needs of more items. Furthermore, polarization compensation and nonlinear enhancement are applied to the item images to effectively improve image quality. A series of operations, including channel attention fusion, decoding, spatial attention calculation, and Hadamard product, are performed on the processed images to further enhance the precision and accuracy of image segmentation.

[0098] Based on the first embodiment of this application, in the third embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 Step S20 may include steps S201 to S204:

[0099] Step S201: Scan the pre-printed carrier identification code on the carrier to obtain carrier information.

[0100] Understandably, the system can obtain relevant information about the carrier by scanning the pre-printed carrier identification code on the carrier using a scanning device. The scanning device can be a barcode scanner, QR code scanner, or RFID reader, etc., used to read the carrier identification code and convert it into recognizable digital information.

[0101] Step S202: Verify the number of historical orders and currently bound carriers in the user identity information to obtain the verification result.

[0102] To further maintain the smooth operation of the system, the system verifies the user's historical orders and the number of currently bound carriers in the user's identity information to check whether the user has any unpaid historical orders and whether the number of currently bound carriers exceeds the preset binding threshold.

[0103] Step S203: When the verification result indicates that there are no unpaid historical orders and the number of currently bound carriers does not exceed a preset binding threshold, a binding relationship between the carrier and the user is generated by associating and storing the carrier information and the user identity information.

[0104] If the verification result shows that the user has no unpaid historical orders and the number of currently bound carriers does not exceed the preset binding threshold, then the carrier information and user identity information will be associated and stored together. That is, the carrier's identification information will be bound together with the user's identity information to generate a binding relationship between the carrier and the user, which will be stored in the system's database.

[0105] Step S204: Monitor the weight change of the carrier based on the binding relationship.

[0106] After establishing a binding relationship between the carrier and the user, the system continuously monitors changes in the carrier's weight using weighing equipment. For example, when a user adds or removes items from the carrier, the weighing equipment detects these weight changes. The system uses this weight change information, combined with the previously established binding relationship between the carrier and the user, to further determine the user's identity information associated with that carrier. In this way, the system can link weight changes to specific users, providing an accurate weight data basis for subsequent pricing operations, thereby achieving precise pricing based on weight changes.

[0107] This embodiment obtains carrier information by scanning carrier identification codes, captures user biometrics or scans identity identification codes to obtain user identity information, and verifies user identity information. A binding relationship is only generated when there are no unpaid historical orders and the number of bound carriers does not exceed the threshold, effectively avoiding illegal binding and duplicate binding, ensuring the security of the pricing system, protecting the legitimate rights and interests of users, and maintaining the normal operation of the system.

[0108] In one implementation, step S40 may further include: in a network-connected state, deducting funds from the backend user account corresponding to the user's identity information; in a network-interrupted state, comparing the user's biometric features with local features, and creating an offline transaction record when the comparison passes; detecting network connectivity at fixed time intervals, and marking the offline transaction record as pending confirmation when the network connectivity meets the requirement of network connection recovery and the network load is below a preset threshold; updating the pending confirmation status to completed status after successful deduction; and adding the corresponding user to the binding restriction list when the pending confirmation status is not updated for a preset time.

[0109] In practice, when the network connection is normal, the system calls the backend user account corresponding to the user's identity information to perform the deduction operation. The backend user account is a database storing user payment information. The system communicates with the backend server via network requests to complete the deduction process. If the network connection is interrupted, the system compares the user's biometric features with locally stored features. If the comparison is successful, the system creates an offline transaction record and synchronizes it after the network is restored.

[0110] In addition, the system checks network connectivity at fixed time intervals. When the network connection is restored and the network load is below a preset threshold, the system marks offline transaction records as pending confirmation, indicating that the transaction record needs further confirmation of its payment status. If the deduction is successful, the pending confirmation transaction record is updated to completed, indicating that the transaction has been successfully completed. If the pending confirmation transaction record is not updated within a preset time, the corresponding user is added to the restricted list, restricting certain user operations.

[0111] This embodiment, when connected to the network, can directly access the user's backend account for deductions, ensuring the immediacy and convenience of transactions. In the event of a network outage, offline transaction records are created through biometric comparison to guarantee transaction continuity and prevent transaction failures due to network issues. Furthermore, by detecting network connectivity and load conditions, the processing of offline transaction records is rationally arranged to ensure the finality of transactions. Restrictions are implemented for pending statuses that have not been updated for extended periods, effectively preventing potential risks and ensuring system security.

[0112] For ease of understanding, please refer to Figure 4 , Figure 5 , Figure 6 , Figure 7 ( Figure 4 A schematic diagram of the pricing process for a single-vegetable-weighing equipment; Figure 5 A schematic diagram of the intelligent identification and pricing process for dishes; Figure 6 This is a schematic diagram of the feature recognition process;

[0113] Figure 7 This is an illustration of the network detection and verification process (see diagram), but it is not intended to limit the pricing method of this application. Applying this pricing method to a closed dining scenario, as a comprehensive digital dining equipment pricing and payment solution, the overall system can include two stages: the dining identity binding stage and the pricing and payment stage. In the pricing and payment stage, it provides food ordering pricing, weighing pricing, and Artificial Intelligence (AI) food recognition pricing. Through the combination of hardware and software, accurate price calculation is achieved. During the recognition process, facial recognition, QR code display, and other methods are used to identify the payer, further enabling account deduction payment. Specifically, it includes:

[0114] 1. One weighing device per dish (weighing-based pricing mode)

[0115] First, barcodes (i.e., carrier identification codes) need to be generated in advance and printed on the plates. Plate-binding and weighing equipment are then used. Before meals, the cafeteria manager configures the menu via a personal computer (PC) and then configures the associated dishes on the equipment. Diners can then take their own plates, complete the plate binding and food retrieval, and the pricing and settlement are automatically completed. No cafeteria staff assistance is required during the entire meal process, significantly reducing manpower and improving dining efficiency. Diners can take smaller portions more frequently based on their individual dietary needs, reducing food waste and making it more economical. The specific steps are as follows:

[0116] (1) The canteen administrator configures the menu information on the canteen management terminal, including entering the dish name, picture, price and other information, and associates the dish information with the daily menu (i.e. the item information on the carrier).

[0117] (2) Generate barcodes (i.e. carrier identification codes) for the plates in batches, provide the barcodes to the manufacturer, and print them on the plates.

[0118] (3) The canteen manager associates the dishes of the current meal with the weighing equipment and displays the current unit price and weight of the dishes (i.e., the information of the items on the carrier).

[0119] (4) After receiving their trays in the cafeteria, diners can scan the QR code (i.e., carrier identification code) on their trays using a mobile app to complete the tray binding process; or they can use a tray binding device to place the tray in a designated location. The device will scan the barcode on the tray and collect the user's facial information or the QR code information from the app (i.e., user identity information) to complete employee identity verification and check whether the current person has any unpaid orders or has bound multiple trays. If the requirements are met, the binding between the diner and the tray is completed (i.e., establishing a binding relationship between the carrier and the user). If there are any unpaid orders that have exceeded the time limit, the user will be prompted to complete the order payment on their mobile phone before binding the tray. If the number of bound trays exceeds the limit, the binding time and the amount already consumed will be displayed. The user can choose to unbind the trays they want to unbind, and the corresponding order will be paid immediately after unbinding.

[0120] (5) After the meal tray is bound, the diner places the tray in the designated area of ​​the meal weighing equipment. The equipment will automatically identify the user information bound to the meal tray (i.e., the binding relationship) and calculate the maximum weight to be taken based on the price of the dishes. After the diner takes the food from the dish tray of the weighing equipment and puts it into the meal tray, the system will automatically capture the weight change (i.e., monitor the weight change of the carrier). If the maximum weight to be taken is exceeded, an alarm will be issued to remind the user. After the user takes the food and removes the tray, the price is calculated based on the difference in the weight taken to generate this order (i.e., generate a priced order based on the weight change and item information).

[0121] (6) After a user has taken multiple dishes, the system will automatically scan for orders that have not generated new transactions within 15 minutes, automatically initiate payment, and unbind the plates.

[0122] 2. AI-powered menu item recognition and pricing (menu recognition and pricing mode)

[0123] This solution integrates AI-powered food recognition equipment. Before meals, cafeteria managers configure the menu via PC and then photograph the dishes on the device, linking the dishes to the images. During the meal, diners simply place their trays in the recognition area. The device quickly displays the price of each dish based on the photographed image and calculates the entire order. Cafeteria managers then verify that all dishes have been placed and click the payment button to complete the payment. Payment can also be completed offline if the device is disconnected from the internet. The specific steps are as follows:

[0124] (1) The canteen administrator configures the menu information on the canteen management terminal, including entering the dish name, picture, price and other information, and associates the dish information with the daily menu (i.e. the item information on the carrier).

[0125] (2) When the device is turned on every day, it will automatically download the menu information and facial feature values ​​to the device. If the device has no network, it will automatically retrieve the menu data synchronized in the previous few days. Then the canteen manager will take photos of the trays and the dishes that have not been recorded. Multiple photos can be taken from multiple angles. The device application will segment and save the images according to the edges of the dishes, and extract the feature values ​​through the dish feature extraction algorithm. The images and feature values ​​will be associated with the dishes and saved (i.e., the pre-stored item library).

[0126] (3) Before the meal service starts, the canteen manager clicks the cashier function, which will automatically check whether the dishes for the meal have been photographed and whether there are enough photos. If the requirements are met, the cashier can be identified and paid for; otherwise, the manager will be prompted to take photos.

[0127] (4) After the user picks up the food, he puts the tray into the device’s recognition area. The device will first automatically detect whether the tray has been completely placed in the recognition area. If the tray is detected but not completely placed, the user will be prompted to adjust the tray.

[0128] (5) After the tray is fully placed, the identified image is segmented using an algorithm. Before segmentation, polarization compensation and nonlinear enhancement are performed on the image. The segmented item regions are obtained through the image segmentation model. The feature values ​​of each segmented image (i.e., item region) are compared (i.e., image feature matching). If the comparison is successful, the price and quantity of each dish are given. For any unidentified anomalies, the image can be directly associated with the dish by clicking on the interface for real-time learning. The learned feature values ​​are added to the dish library (i.e., updating the pre-stored item library). After three overall comparisons with all prices being the same, the overall comparison time is controlled within 1 second, and the payment button is displayed.

[0129] (6) When the network is not interrupted, the user clicks the payment button, the device captures a facial photo, performs online facial comparison to confirm the user's information, and calls the user's account in the background to deduct the payment (i.e., deduction under network connection). If the user is disconnected from the network when making payment, the facial photo captured by the device is compared with the device's local feature values. If the user is identified, the device records the order information such as the identified image, price, and person on the device. At the same time, the device will start a scheduled task to perform a network probe every 5 minutes. If the probe is successful and it is not during the peak transaction period, all offline orders will be synchronized to the background. If the payment is successful immediately, the device will return a success message; otherwise, it will return a synchronization success message with an unknown status. The device will periodically synchronize orders with unknown statuses. Orders that have not been successfully paid will generate pending payment orders in the background and remind the user to make payment. The background will also deduct the payment periodically every two hours. Users can also actively make payments via their mobile phones. If payment is not made within 2 days, the user will be blacklisted and will not be able to make any payments until the repayment is completed.

[0130] It should be noted that the pricing method of this application is a general technical solution, and its implementing entity can be any device or system with corresponding functions, not limited to a specific scenario or device. In the above example, multiple implementing entities are involved, such as canteen managers, diners, plate-binding equipment, weighing equipment, and AI food recognition equipment. This division of entities is for the purpose of facilitating understanding and explaining the specific implementation process of the pricing method in practical application scenarios. For example, the canteen manager is responsible for configuring the menu and associating dishes, diners complete the binding and food collection operations, while the plate-binding equipment and weighing equipment are responsible for functions such as identity authentication and weight monitoring, respectively. These specific implementing entities are not essential components of the pricing method of this application, but rather are equipment or personnel roles selected according to the actual application scenario. The core of this application lies in the pricing method itself, and the pricing logic implemented through carrier identification coding, user identity information, weight changes, and image recognition technologies.

[0131] In practical applications, the pricing method of this application can be applied to dining scenarios. Specifically, in a buffet checkout scenario, an E-Meal System is implemented using the pricing method of this application. This system includes a smart food dispensing machine with an embedded smart scale. This device has multiple recognition functions to meet the needs of different scenarios. The smart food dispensing machine is equipped with a smart scale that can identify tray information through chip or barcode scanning technology. Its recognition range and height can be adjusted according to actual needs to ensure accurate recognition. The device is also equipped with a tablet computer (PAD) for displaying food information, operation prompts, and user interaction. The fixed position of the PAD can be designed according to the device model and usage habits. In addition, the smart food dispensing machine has a variety of auxiliary functions, such as a saliva mask and an alarm light, to prevent droplet transmission and trigger alarms for unauthorized operations, ensuring the hygiene and safety of the dining environment. The smart scale can monitor the weight changes of the food in real time and automatically calculate the cost according to preset pricing rules. It also supports multiple payment methods, such as mobile phone operation and contactless payment, improving the user's dining experience. Furthermore, the device also has functions such as allergen reminders and energy intake reminders to help users make reasonable food selections.

[0132] The E-Meal System provides services across five major modules: meal allowance distribution, online ordering, one-stop payment, merchant management, and intelligent reporting. It helps businesses reduce management costs, improve employee service experience, and continuously explores the unique characteristics of corporate dining in different industries, focusing on building core competitiveness and forming comprehensive dining solutions tailored to specific industries. Simultaneously, through practical exploration of the B2B2C (Business to Business to Consumer) dining operation model, it assists enterprise clients in their digital transformation with refined management strategies, diverse and convenient payment methods, excellent intelligent experience, complete digital monitoring strategies, and strong self-developed terminal capabilities. It primarily addresses issues in corporate dining scenarios such as long waiting times, low payment efficiency, limited payment methods, and high system maintenance costs.

[0133] The innovations of the E-Meal Pass system implemented using the pricing method of this application are as follows: Deep integration of mode binding and pricing; the weighing mode combines dish association and plate binding for rapid pricing and automatic deduction for multiple pickups; the dish recognition mode allows for cloud-based menu configuration and offline dish association, supporting dish changes at any time, multiple settlements, and offline payment even during network interruptions; image processing employs multiple innovative methods, combined with attention technology to optimize segmentation, and dynamically updates the item database when matching is empty; strict verification of user binding information ensures transaction security and achieves accurate association between pricing and users.

[0134] Compared to existing technologies, the E-Meal payment system is both convenient and secure. It supports facial recognition via applications and mini-programs, utilizing facial recognition algorithms to enable device-based facial recognition and QR code payments, reducing card production costs and the risk of fraudulent transactions for businesses. In the event of a network outage, offline payments are made through local biometric comparison; records are processed after network recovery to ensure continuous and complete payments. Pricing is fast and accurate, leveraging weighing, recognition, and advanced image technology. Automatic deductions reduce manual intervention, support multiple pricing modes, and rigorous verification ensures pricing accuracy and transaction security. Menu management is cloud-based and linked offline, enabling rapid information updates, reducing manual operations, allowing for menu changes and multiple settlements at any time, adapting to dynamic changes and personalized needs, and improving operational efficiency.

[0135] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the pricing method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0136] This application also provides a pricing device; please refer to... Figure 8 The pricing device includes:

[0137] The user information acquisition module 10 is used to capture the user's biometric features through the camera of the binding device, or scan the user's identity identification code, and obtain the user's identity information based on the user's biometric features or the identity identification code.

[0138] The weight detection module 20 is used to obtain carrier information through carrier identification code, establish a binding relationship between the carrier and the user based on the carrier information and the user identity information, and monitor the weight change of the carrier based on the binding relationship.

[0139] Image recognition module 30 is used to recognize images of objects on a carrier;

[0140] The pricing module 40 is used to generate a pricing order based on at least one of the weight change and the item image.

[0141] The pricing device provided in this application, employing the pricing method described in the above embodiments, can solve the technical problem of low accuracy and efficiency in existing pricing methods. Compared with the prior art, the beneficial effects of the pricing device provided in this application are the same as those of the pricing method provided in the above embodiments, and other technical features in the pricing device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0142] This application provides a pricing device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the pricing method in Embodiment 1 above.

[0143] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0144] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0145] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the pricing method in the above embodiments.

[0146] The aforementioned computer-readable storage medium may be included in the pricing device or may exist independently and not assembled into the pricing device.

[0147] The aforementioned computer-readable storage medium carries one or more programs. When the aforementioned one or more programs are executed by the pricing device, the pricing device causes the pricing device to: obtain carrier information through carrier identification encoding, and establish a binding relationship between the carrier and the user based on the carrier information and the collected user identity information; monitor the weight change of the carrier based on the binding relationship, and generate a pricing order based on the weight change and the item information on the carrier; and / or identify the image features of the items on the carrier, and generate a pricing order based on the image features and the binding relationship.

[0148] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0149] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described pricing method, which can solve the technical problem of low accuracy and efficiency of existing pricing methods. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the pricing method provided in the above embodiments, and will not be repeated here.

[0150] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification and drawings, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A pricing method, characterized in that, The method includes the following steps: The device captures the user's biometric features through its camera or scans the user's identification code, and obtains the user's identity information based on the biometric features or the identification code. Carrier information is obtained through carrier identification encoding. Based on the carrier information and the user identity information, a binding relationship between the carrier and the user is established, and the weight change of the carrier is monitored based on the binding relationship. Identify images of objects on a carrier; A pricing order is generated based on at least one of the weight change and the item image.

2. The pricing method as described in claim 1, characterized in that, The step of generating a pricing order based on at least one of the weight change and the item image includes: The image of the item is segmented to obtain the item region; Image features of the object region are extracted using a convolutional neural network; The image features are compared with the pre-stored feature vectors in the pre-stored item library to obtain a similarity matrix. The item corresponding to the pre-stored feature vector with the highest similarity in the similarity matrix is ​​taken as the target matching result. The pre-stored feature vector is the feature vector extracted from each item in the pre-stored item library. When the target matching result is not empty, the corresponding item information is associated with it, and a pricing order is generated based on the corresponding item information. When the target matching result is empty, the image is added to the pre-stored item library, and a pricing order is generated according to the preset configuration information. A reminder notification is sent to the staff to check the recognition status.

3. The pricing method as described in claim 2, characterized in that, The step of segmenting the object image to obtain the object region includes: The highly reflective areas of the object image are processed by a bilateral filter, and the low-contrast areas of the object image are nonlinearly enhanced to obtain the processed image. The processed image is segmented using an image segmentation model to obtain the item region. The image segmentation model includes a YOLOv8 segmentation network obtained by extracting the segmentation capability of the SAM pre-trained model through a feature distillation algorithm and fine-tuning it on a preset dataset.

4. The pricing method as described in claim 3, characterized in that, The step of segmenting the processed image using an image segmentation model to obtain the item region includes: The processed image is input into the encoder of the segmentation network to obtain multi-scale feature maps; Channel attention fusion is performed on the multi-scale feature map to generate a modulated feature map; The modulated feature map is input into the decoder of the segmentation network to obtain the decoded features; Spatial attention is calculated on the multi-scale feature map to obtain the spatial attention calculation result, and the spatial attention calculation result is multiplied by the Hadamard product with the decoded features to obtain the optimized features; The optimized features are convolved and normalized to generate an item region probability map; The segmentation boundaries are adjusted based on the confidence level of the item region probability map to obtain the final item region.

5. The pricing method as described in any one of claims 1 to 4, characterized in that, The steps of obtaining carrier information through carrier identification encoding, establishing a binding relationship between the carrier and the user based on the carrier information and the user identity information, and monitoring changes in the carrier's weight based on the binding relationship include: Scan the pre-printed carrier identification code on the carrier to obtain carrier information; The historical orders and the number of currently bound carriers in the user identity information are verified to obtain the verification result; When the verification result indicates that there are no unpaid historical orders and the number of currently bound carriers does not exceed a preset binding threshold, a binding relationship between the carrier and the user is generated by associating and storing the carrier information and the user identity information. The weight change of the carrier is monitored based on the binding relationship.

6. The pricing method according to any one of claims 1 to 4, characterized in that, The step of generating a pricing order based on at least one of the weight change and the item image includes: When the weighing device detects that the current carrier is in a bound state, it generates a pricing order with associated corresponding user identity information based on the weight change and the item information on the current carrier. The corresponding user identity information is the user identity information marked in the weighing device according to the binding relationship.

7. The pricing method according to any one of claims 1 to 4, characterized in that, After the step of generating a pricing order based on at least one of the weight change and the item image, the method further includes: While connected to the network, the system calls the backend user account corresponding to the user's identity information to deduct the payment. In the event of a network outage, the user's biometrics are compared with local characteristics, and an offline transaction record is created if the comparison is successful. Network connectivity is detected at fixed time intervals. When the network connectivity meets the requirement of network connection recovery and the network load is lower than a preset threshold, the offline transaction record is marked as pending determination. After successful payment, the pending status will be updated to completed status. When the pending status is detected to have not been updated for more than a preset time, the corresponding user will be added to the disk binding restriction list.

8. A pricing device, characterized in that, The pricing device includes: The user information acquisition module is used to capture the user's biometric features through the camera of the binding device, or scan the user's identity identification code, and obtain the user's identity information based on the user's biometric features or the identity identification code; The weight detection module is used to obtain carrier information through carrier identification code, establish a binding relationship between the carrier and the user based on the carrier information and the user identity information, and monitor the weight change of the carrier based on the binding relationship. The image recognition module is used to recognize images of objects on a carrier. The pricing module is used to generate a pricing order based on at least one of the weight change and the item image.

9. A pricing device, characterized in that, The pricing device includes: a memory, a processor, and a pricing program stored in the memory and executable on the processor, wherein the pricing program, when executed by the processor, implements the pricing method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a pricing program, which, when executed by a processor, implements the pricing method as described in any one of claims 1 to 7.

Images