Supermarket loss prevention early warning traceability method based on bar code traceability
By using barcode-based traceability methods combined with AI and monitoring technologies, the entire process of supermarket goods can be monitored, solving the problems of abandoned purchases and theft. It provides an efficient early warning and traceability mechanism, forming a closed-loop management system, reducing losses and improving loss prevention efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-25
- Publication Date
- 2026-03-27
AI Technical Summary
The inability to promptly detect abandoned or stolen goods after purchase at supermarket stores, the inability to quickly map the movement of goods and personnel, and the inability to automatically incorporate them into the monitoring system for tracing and early warning, leads to irreparable losses.
By using barcode-based traceability methods, combined with AI technology, monitoring systems, and facial recognition, the entire process of a product being picked up by a customer and collected at the checkout can be monitored. Rules for detecting uncollected items can be set, movement trajectories can be generated and alerts can be pushed out, and facial recognition can be compared in real time to establish a blacklist for prevention.
It enables monitoring and management of goods throughout the entire process, improves loss prevention efficiency and accuracy, provides a clear traceability evidence chain, forms closed-loop management, reduces losses and improves operational efficiency.
Smart Images
Figure CN121745958A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of retail technology, specifically to a supermarket loss prevention and early warning traceability method based on barcode traceability. Background Technology
[0002] Theft prevention in supermarkets is an important issue, and the history of theft prevention is as follows: (1) Using anti-theft devices, such as RIFD cards, to attach anti-theft strips to each item. If a customer takes the item out of the store, an alarm will be triggered. However, this solution has the following drawbacks: First, the cost of the anti-theft strips is high. Second, customers may tear off the anti-theft strips.
[0003] (2) Another method is through video surveillance, as mentioned in the patent above. It uses a camera to capture the process of a person taking a product and identifies what the product is. Then, it identifies the person again at the checkout or exit area. Simultaneously, it determines whether the product has been paid for by detecting whether the barcode scanner has scanned the product identified by the camera. The drawback of this method is that it requires identifying what the product is, which is difficult to achieve.
[0004] The two methods mentioned above only determine whether goods have been taken out of the store. However, a complete method has been invented that automatically alerts customers to theft, provides evidence for tracing the source, and issues a warning for the next theft. (3) Monitoring algorithm, facial and body recognition algorithm to confirm the identity of a person.
[0005] (4) The unique identifier of a product is a barcode, which is usually obtained by monitoring the input of the barcode scanner.
[0006] The existing technology has the following technical problems: (1) It is not possible to know in a timely manner if a supermarket store purchases goods (including standard and non-standard weighed goods) are abandoned.
[0007] (2) Supermarket store goods (including standard and non-standard weighed goods) are purchased and taken away from the store without payment.
[0008] (3) After the abandonment or theft described in (1) and (2) occurs, the supermarket is unable to quickly map the goods and personnel and the movement trajectory of the personnel, so as to provide evidence for tracing the source.
[0009] (4) After the theft described in (1) and (2) occurs, the supermarket cannot automatically include the thief in the monitoring system to realize the reminder of the thief entering the store. Summary of the Invention
[0010] This invention primarily addresses the shortcomings of existing technologies by providing a supermarket loss prevention and early warning traceability method based on barcode traceability. This method boasts high accuracy in judging and issuing early warnings regarding the sold status of goods, and can also provide a clear trajectory, helping to quickly identify problems and mitigate losses. Through the integration of multiple technologies, it achieves full-process monitoring and management of goods from when they are picked up by customers, moved within the store, to the final checkout or when deemed as abnormal behavior. It solves the problem of goods loss caused by theft and customer abandonment in traditional supermarkets, and provides an efficient early warning, traceability, and handling mechanism.
[0011] The above-mentioned technical problems of the present invention are mainly solved by the following technical solutions: A supermarket loss prevention and early warning traceability method based on barcode traceability includes the following steps: Step 1: Obtain the product's sold status.
[0012] Step 2: Monitoring the sales status, which involves monitoring serial port data and connecting to the ERP database to obtain standard product barcode information.
[0013] Step 3: Users set the unpaid checkout detection time and rules for products, and will be notified if the timeout period expires.
[0014] Step 4: Send a notification to your mobile phone that you have not paid.
[0015] Step 5: Generate movement trajectory.
[0016] Step 6: The phone can be retrieved for verification.
[0017] Step 7: Mark as theft and add to the blacklist.
[0018] Step 8: Real-time facial recognition and intrusion alert.
[0019] As a preferred option, for non-standard weighed products, barcodes are obtained through AI electronic scale recognition software; for standard products, product information is obtained through high-speed document scanner AI recognition, thereby obtaining the standard product barcode.
[0020] As a preferred option, the monitoring rules include setting the product category labels to be monitored, setting the price range of the monitored products, setting the time period for monitoring the products, and setting the permissions of the personnel issuing warnings.
[0021] As a preferred approach, behavioral analysis is performed when purchasing goods to determine if theft has occurred. If no theft is detected, an abandonment warning is issued and the goods are retrieved; if theft is detected, appropriate action is taken.
[0022] As a preferred method, the alert records are manually checked to confirm whether theft has occurred. If a false alarm is detected, it is deleted. If theft is confirmed, the person is added to a blacklist.
[0023] As a preferred method, the product sale status detection involves taking a picture of the product with a camera, then performing image preprocessing, using deep neural network forward inference to extract the product's AI characteristics, then comparing it with the product feature library and outputting the product. After the customer confirms the product category, the product barcode is output, then written to the cloud list, and finally a unique barcode encoding field is added.
[0024] As a preferred method, the product status detection uses ERP data fields and writes data to the POS to update the database when the product is checked out. Then, it pushes and obtains the checkout signal and writes the checkout time. If it is determined that the product has not been checked out or the time has expired, it generates an alert message, sends it to the mobile device, and uploads it to the cloud.
[0025] As a preferred method, the generation of activity trajectories, unpaid orders, and blacklists involves the detection personnel pushing the generated activity trajectory videos to the client for identification. If theft is detected, the video is downloaded, and the personnel are screenshotted and added to the blacklist; or if the purchase is abandoned, the goods are scanned and verified after the product is inspected, and the unpaid status is changed to product retrieval.
[0026] As a preferred option, in-store reminders use facial recognition via camera to identify individuals against a blacklist. If an individual is on the blacklist, a reminder is sent to their mobile phone.
[0027] The present invention can achieve the following effects: This invention provides a supermarket loss prevention and early warning traceability method based on barcode traceability. Compared with existing technologies, it integrates multiple technologies to achieve full-process monitoring and management of goods from when they are picked up by customers, moved within the store, to the final checkout or when they are identified as abnormal behavior. It solves the problem of goods loss caused by theft and customers abandoning their purchases in traditional supermarkets, and provides an efficient early warning, traceability, and handling mechanism.
[0028] It surpasses the traditional EAS access control alarm model, shifting the loss prevention focus from post-event interception to in-event monitoring and pre-event prevention. Through closed-loop management of "identification-monitoring-early warning-tracing-handling-prevention," it can not only effectively reduce product losses but also improve operational efficiency.
[0029] End-to-end monitoring: Enables full-process tracking of goods from the shelf to the checkout counter, with no blind spots in monitoring.
[0030] Intelligentization and automation: The extensive use of AI technologies (product recognition, behavior analysis, facial recognition) has greatly improved loss prevention efficiency and accuracy, and reduced reliance on human labor.
[0031] Precise traceability: The integrated data chain of "product-barcode-person-trajectory" provides a clear and traceable chain of evidence for every abnormal event.
[0032] Closed-loop management: From early warning to handling (retrieving goods or adding to the blacklist), and then to secondary prevention, a complete management closed loop is formed to continuously optimize the loss prevention effect.
[0033] Flexible and configurable: The rich rule setting function enables the system to adapt to the personalized loss prevention needs of supermarkets of different sizes and types. Attached Figure Description
[0034] Figure 1 This is the business process diagram of the present invention.
[0035] Figure 2 This is the technical logic diagram of the present invention. Detailed Implementation
[0036] The technical solution of the invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.
[0037] Example: Figure 1 and Figure 2 As shown, a supermarket loss prevention and early warning traceability method based on barcode traceability includes the following steps: Step 1: Obtaining Product Sales Status. For non-standard weighed products, the barcode is obtained through AI electronic scale recognition software. For standard products, product information is obtained through high-speed document scanner AI recognition, thereby obtaining the standard product barcode.
[0038] Step 2: Monitoring the sales status, which involves monitoring serial port data and connecting to the ERP database to obtain standard product barcode information.
[0039] The product sale status detection is achieved by taking pictures of the product with a camera, followed by image preprocessing, using deep neural network forward inference to extract the product's AI characteristics, then comparing them with the product feature library and outputting the product. After the customer confirms the product category, the product barcode is output, then written to the cloud list, and finally a unique barcode encoding field is added.
[0040] The product status detection uses ERP data fields and writes data to the POS to update the database when the product is checked out. Then, it pushes and obtains the checkout signal and writes the checkout time. If it is determined that the product has not been checked out or the timeout has expired, it generates an alert message, sends it to the mobile device, and uploads it to the cloud.
[0041] Step 3: Users set the unpaid checkout detection time and rules for products. If the timeout period expires, a notification will be displayed indicating that the product has not been paid for. The monitoring rules include setting the product category tags to be monitored, setting the price range of the monitored products, setting the time period for monitoring products, and setting the permissions of the personnel issuing alerts.
[0042] When purchasing goods, behavioral analysis is performed to determine if theft has occurred. If no theft is found, an abandonment warning is issued and the goods are retrieved; if theft is detected, the relevant actions are taken.
[0043] Manually check the warning records to confirm whether theft has occurred. If it is a false alarm, delete it. If the theft is confirmed, add it to the blacklist.
[0044] Step 4: Send a notification to your mobile phone that you have not paid.
[0045] Step 5: Generate movement trajectory.
[0046] Step 6: The phone can be retrieved for verification.
[0047] Step 7: Mark as theft and add to the blacklist. The tracking, unpaid orders, and blacklist generation are achieved by having personnel push the generated activity tracking video to the client for verification. If theft is detected, the video is downloaded, and a screenshot of the personnel is added to the blacklist; or if the purchase was abandoned, the item is scanned and verified after the transaction is completed, and the unpaid status is changed to item recovery.
[0048] Step 8: Real-time facial recognition and intrusion alert. The in-store alert uses a camera to perform facial recognition and compares the person against a blacklist. If the person is on the blacklist, an alert is sent to their mobile phone.
[0049] Working principle: Phase 1: Monitoring Start-up and Status Tracking.
[0050] (a) Obtaining product information: Standard products (pre-packaged goods): When a customer picks up a product, the AI recognition system deployed near the shelf is activated. It captures images of the product using a camera and performs real-time recognition using a deep learning model to obtain its unique barcode information.
[0051] Non-standard products (bulk weighed products): When customers weigh their products on the AI electronic scale, the system automatically identifies the product category through image recognition software and generates or associates a temporary barcode.
[0052] Cloud registration: Once a product is identified, its barcode information is immediately written to a cloud monitoring list, along with a unique code and a start timestamp, marking that the product has entered the "picked up, ready to be sold" monitoring state.
[0053] (ii) Real-time monitoring of sales status: Data source: The system connects in real time with the supermarket's ERP / POS system database to monitor cash register transaction data.
[0054] Status Update: When a customer successfully scans a product and pays at the checkout, the POS system updates its database. Upon receiving this checkout signal, the system updates the corresponding product's barcode status to "sold" in the cloud list and records the checkout time.
[0055] Phase Two: Early Warning Triggering and Manual Intervention.
[0056] (I) Intelligent Early Warning Rule Engine: Rule Configuration: Administrators can flexibly configure monitoring rules according to the actual needs of the supermarket, including: Product categories: High-loss categories will be the focus of monitoring (such as cosmetics and electronic products).
[0057] Price range: Focus on monitoring high-value goods.
[0058] Time period: Strengthen monitoring during peak hours or night shifts and other high-risk periods.
[0059] Detection time: Set a reasonable time threshold for "from picking up to checkout" for each type of product.
[0060] Warning Triggered: The system continuously compares cloud-based lists and POS data. Once a product's "pending sale" status exceeds the preset detection time, the system immediately determines it as "overdue for payment".
[0061] (II) Closed-loop system for early warning response and handling: Information push: Warning information (including product name, pickup time, etc.) is pushed to the loss prevention officer's mobile app in real time.
[0062] Tracing the source of the incident: The system automatically links with the store's surveillance cameras to generate a video of the movement trajectory of the person carrying the item, and pushes it to the loss prevention officer to provide intuitive on-site evidence.
[0063] Manual analysis and handling: Loss prevention personnel analyze video footage to assess behavior and take appropriate measures. If an item is deemed abandoned: Loss prevention staff locate the abandoned item based on the tracking information and scan it using a mobile app for verification. The system then updates the item's status from "not paid for" to "recovered," completing the closed loop.
[0064] If the theft is determined: Loss prevention personnel download the video footage of the incident as evidence, take screenshots of the individuals involved, and add them to the system's blacklist.
[0065] False alarm handling: If the system makes a mistake, the loss prevention officer can delete the warning record directly on the App.
[0066] Phase Three: Blacklist Management and Secondary Prevention (a) Blacklist establishment: All screenshots and information of thieves confirmed by manual verification are centrally stored in the system's blacklist database.
[0067] (II) Proactive Prevention: Facial recognition deployment: Cameras deployed at the supermarket entrance will perform facial recognition in real time.
[0068] Real-time comparison and alerts: Captured facial information is compared at high speed with a blacklist database. Once a match is found, the system immediately sends a "blacklisted person intrusion alert" to administrators via mobile device, enabling them to pay attention and intervene at the first moment, achieving proactive prevention.
[0069] In summary, this supermarket loss prevention and early warning traceability method based on barcode traceability integrates multiple technologies to achieve full-process monitoring and management of goods from the moment they are picked up by a customer, moved within the store, to the final checkout or when deemed as abnormal behavior. It solves the problem of product loss in traditional supermarkets caused by theft and customer abandonment of purchases, and provides an efficient early warning, traceability, and handling mechanism.
[0070] A closed-loop intelligent loss prevention system has been built, covering the entire chain from product identification to final transaction completion or processing. No longer relying on traditional exit access control alarms, it integrates loss prevention management into the entire customer shopping process through AI visual recognition, real-time data comparison, and behavioral analysis, achieving an organic combination of in-process monitoring, post-event traceability, and pre-event prevention.
[0071] The above description is only a specific embodiment of the present invention, but the structural features of the present invention are not limited thereto. Any changes or modifications made by those skilled in the art within the scope of the present invention are covered by the patent scope of the present invention.
Claims
1. A supermarket loss prevention early warning and traceability method based on barcode traceability, characterized in that... The following steps are included: Step 1: Obtain the product's sold status; Step 2: Sales status monitoring, which involves monitoring serial port data and ERP database to obtain standard product barcode information; Step 3: Users set the unpaid checkout detection time and rules for products, and prompt that the product has not been checked out if the timeout is exceeded. Step 4: Send a notification to the mobile phone indicating that payment has not been made. Step 5: Generate movement trajectory; Step 6: The phone can be retrieved and used for verification; Step 7: Mark as theft and add to the blacklist; Step 8: Real-time facial recognition and intrusion alert.
2. The supermarket loss prevention early warning traceability method based on barcode traceability according to claim 1, characterized in that: For non-standard weighed products, barcodes are obtained through AI electronic scale recognition software; for standard products, product information is obtained through high-speed document scanner AI recognition, thereby obtaining the standard product barcode.
3. The supermarket loss prevention and early warning traceability method based on barcode traceability according to claim 1, characterized in that: The monitoring rules include setting the product category labels to be monitored, setting the price range of the monitored products, setting the time period for monitoring products, and setting the permissions of the personnel issuing warnings.
4. The supermarket loss prevention early warning and traceability method based on barcode traceability according to claim 3, characterized in that: When purchasing goods, behavioral analysis is performed to determine if theft has occurred. If no theft is found, an abandonment warning is issued and the goods are retrieved; if theft is detected, the relevant actions are taken.
5. The supermarket loss prevention early warning and traceability method based on barcode traceability according to claim 4, characterized in that: Manually check the warning records to confirm whether theft has occurred. If it is a false alarm, delete it. If the theft is confirmed, add it to the blacklist.
6. The supermarket loss prevention early warning and traceability method based on barcode traceability according to claim 1, characterized in that: The product sale status detection is achieved by taking pictures of the product with a camera, followed by image preprocessing, using deep neural network forward inference to extract the product's AI characteristics, then comparing them with the product feature library and outputting the product. After the customer confirms the product category, the product barcode is output, then written to the cloud list, and finally a unique barcode encoding field is added.
7. The supermarket loss prevention and early warning traceability method based on barcode traceability according to claim 6, characterized in that: The product status detection uses ERP data fields and writes data to the POS to update the database when the product is checked out. Then, it pushes and obtains the checkout signal and writes the checkout time. If it is determined that the product has not been checked out or the timeout has expired, it generates an alert message, sends it to the mobile device, and uploads it to the cloud.
8. The supermarket loss prevention and early warning traceability method based on barcode traceability according to claim 7, characterized in that: The system generates activity tracks, unpaid orders, and blacklists by having inspectors push videos of the generated activity tracks to the client for verification. If theft is detected, the video is downloaded, and the person is added to the blacklist. Alternatively, if a purchase is abandoned, the product is scanned and verified after the product is inspected, and the unpaid status is changed to product retrieval.
9. The supermarket loss prevention and early warning traceability method based on barcode traceability according to claim 8, characterized in that: The system uses facial recognition via camera to identify individuals who are on a blacklist. If an individual is on the blacklist, a notification is sent to their mobile phone.