Abnormal order judgment and auditing optimization method and system, medium and program product

Abstract

The invention discloses an abnormal order judgment and auditing optimization method and system, a medium and a program product, and the method comprises the steps: determining a settlement commodity; comparing the settled commodity information with settled commodity information input by the cashier equipment, and judging whether an abnormal order exists or not; if the abnormal order exists, behavior analysis operation is carried out on abnormal order period monitoring, and a preliminary judgment responsible party is obtained; binding the abnormal order, the abnormal order time period monitoring and the preliminary judgment responsible party to obtain to-be-audited information, and pushing the to-be-audited information to a to-be-audited queue; under the condition that an audit result corresponding to the to-be-audited information is received, the audit result is analyzed, the audit result is that the preliminary judgment responsibility party is correct or wrong, and the preliminary judgment responsibility party error indicates that the final judgment responsibility party is opposite to the preliminary judgment responsibility party. According to the invention, the accuracy of missed commodity identification can be improved, the difficulty of responsibility attribution judgment is reduced, the auditing process is simplified, and the auditing efficiency is improved.

Description

Technical Field

[0001] The present application belongs to the field of abnormal order determination and review, and in particular, relates to an abnormal order determination and review optimization method, system, medium and program product. Background Art

[0002] With the rapid development of the retail industry and the increasing demands of consumers for shopping experience, loss prevention management in supermarkets and other retail venues is facing increasingly severe challenges. Traditional loss prevention methods mainly rely on manual monitoring, customer awareness, and post-event investigation and handling. These methods are not only inefficient, but also difficult to detect and handle problems such as missed goods in real time and accurately, resulting in economic losses to supermarkets and reduced customer satisfaction.

[0003] In order to solve these problems, intelligent loss prevention systems have gradually attracted attention and applications in recent years. These systems use advanced video surveillance technology, artificial intelligence algorithms, big data analysis and other technical means to achieve real-time monitoring of the shopping process and intelligent identification of abnormal behaviors. However, the existing intelligent loss prevention systems still have some shortcomings in the determination and review of abnormal orders, such as low accuracy in identifying missed goods, difficulty in determining responsibility, and cumbersome and inefficient review processes. Summary of the invention

[0004] The present application provides a method, system, medium and program product for optimizing abnormal order determination and audit, which are used to improve the accuracy of identifying missed goods, reduce the difficulty of determining responsibility, simplify the audit process and improve the efficiency of the audit.

[0005] In the first aspect, the present application provides an abnormal order determination and review optimization method, which identifies the surveillance video of the cashier area according to a preset algorithm, determines the settlement commodity, and the settlement commodity includes the corresponding settlement commodity information; Compare the settlement product information with the checked-out product information entered into the cash register device to determine whether there is an abnormal order, where the abnormal order is a product order corresponding to the settlement product information that is not included in the checked-out product information; If there is an abnormal order, a behavior analysis operation is performed on the abnormal order time period monitoring to obtain a preliminary determination of the responsible party. The abnormal order time period monitoring is obtained by intercepting the monitoring video for a preset time. The processing time point of the abnormal order is within the abnormal order time period monitoring. The responsible party is preliminarily determined to be the customer's responsibility or the staff's responsibility; Bind abnormal orders, abnormal order time period monitoring and preliminary determination of responsible parties, obtain the pending review information, and push the pending review information to the pending review queue; When the audit result corresponding to the information to be audited is received, the audit result is parsed, and the audit result is that the preliminary determination of the responsible party is correct or the preliminary determination of the responsible party is wrong. The preliminary determination of the responsible party being wrong means that the final determination of the responsible party is opposite to the preliminary determination of the responsible party.

[0006] By adopting the above technical solution, the preset algorithm is used to analyze the surveillance video, accurately identify the settlement product information, and compare it with the settled product information entered by the cash register device. This step greatly improves the recognition accuracy of abnormal orders. Secondly, the behavior analysis of abnormal orders is carried out. By intercepting the surveillance video of a specific period of time, the possible responsible party is accurately located, which effectively reduces the workload of manual review. The abnormal orders, surveillance videos and preliminary judgment results are bound as information to be reviewed and pushed to the queue to be reviewed. This method significantly improves the review efficiency, allowing auditors to quickly obtain all relevant information and make more accurate judgments. Finally, by parsing the audit results, the system can continuously optimize the judgment accuracy and improve the intelligence level of the entire process. The recognition and processing efficiency of abnormal orders is improved, human errors are effectively reduced, and supermarket losses are reduced. At the same time, by analyzing the behavior of customers and staff, it can help merchants discover potential management loopholes or training needs, thereby improving the overall operation quality. The accuracy of identifying missed goods is improved, the difficulty of determining responsibility is reduced, the audit process is simplified, and the efficiency of audit is improved.

[0007] In conjunction with some embodiments of the first aspect, in some embodiments, a behavior analysis operation is performed on abnormal order period monitoring to obtain a preliminary determination of the responsible party, specifically including: Obtaining order information corresponding to the abnormal order, including the time when the abnormal order was entered into the cash register device; Locate the time point in the surveillance video, and intercept the surveillance video of the preset length with the time point as the midpoint to obtain the abnormal order period monitoring; Identify customers and staff members under monitoring during the abnormal order period; Conduct behavioral analysis on customers and staff respectively to determine the last person who has been in contact with the abnormal products corresponding to the abnormal orders; When it is determined that the last handled target is a customer, determine whether the customer has any abnormal settlement operation, where the abnormal settlement operation is an operation in which the abnormal commodity is not handed over to the staff and is placed in the settled area; If there is any abnormal settlement operation, the customer will be initially determined to be responsible; If there is no abnormal calculation operation, the preliminary determination is that the responsible party is the staff member; When it is determined that the last person handled by the employee is a staff member, determine whether the staff member has any abnormal cashier operation. The abnormal cashier operation is an operation in which the abnormal product is not placed in the scanning area but in the settled area. If there is any abnormal cashier operation, the staff will be initially identified as the responsible party; If there is no abnormal cashier operation, it is preliminarily determined that the customer is responsible.

[0008] By adopting the above technical solution, the pertinence and accuracy of the analysis are ensured by accurately locating the order entry time point and intercepting the surveillance video of the corresponding period. This method greatly reduces the length of the video to be analyzed, improves the processing efficiency, and also reduces the amount of data stored and transmitted. Secondly, the behavior analysis of customers and staff is carried out separately to determine the target who finally touched the abnormal goods. This method can clearly restore the process of the incident and effectively avoid the ambiguity of responsibility identification. By judging whether there are abnormal settlement operations or abnormal cashier operations, the system can more objectively and fairly determine the preliminary responsible party, reducing the deviation that may be caused by human judgment. This meticulous behavior analysis method can not only improve the accuracy of abnormal order processing, but also help merchants discover potential management loopholes. For example, if it is found that certain types of goods often appear in abnormal orders, merchants can take corresponding preventive measures, such as adjusting the placement of goods or strengthening employee training. At the same time, this method can also provide valuable data support for subsequent employee training and management optimization. It effectively reduces disputes between customers and employees due to unclear responsibility attribution and improves the efficiency of problem solving.

[0009] In combination with some embodiments of the first aspect, in some embodiments, abnormal orders, abnormal order time period monitoring, and preliminary determination of responsible parties are bound to obtain information to be reviewed, and the information to be reviewed is pushed to a queue to be reviewed, specifically including: Determine the commodity value corresponding to the abnormal commodity in the abnormal order; Generate a viewing control, and play the abnormal order period monitoring when the viewing control is triggered; The product value, viewing controls and the preliminary determination of the responsible party are bound to obtain the information to be reviewed, and the information to be reviewed is pushed to the queue to be reviewed.

[0010] By adopting the above technical solution, the system can conduct a preliminary risk assessment of abnormal orders by determining the value of abnormal goods, which helps auditors quickly identify high-value, high-risk abnormal orders and give priority to cases that may cause greater economic losses. Secondly, generating viewing controls and binding them to abnormal order information greatly improves the convenience and efficiency of the audit process. Auditors only need to trigger the viewing control to directly play the relevant surveillance video without switching between multiple systems or files, which not only saves time but also reduces the possible errors in the information search process. The commodity value, viewing control and preliminary determination of the responsible party are bound together as the information to be audited, so that each piece of information to be audited contains complete data required for decision-making. This integration of information greatly improves the efficiency and accuracy of the audit, and auditors can quickly obtain all relevant information and make more accurate judgments.

[0011] In conjunction with some embodiments of the first aspect, in some embodiments, after binding the commodity value, the viewing control, and the initially determined responsible party, obtaining the information to be reviewed, and pushing the information to be reviewed to the queue to be reviewed, the method further includes: Determine the range of commodity values ​​in the preset table and obtain the corresponding priority; Determine the priority level of each pending information in the pending review queue; Adjust the order of the queue to be reviewed according to the corresponding priority.

[0012] By adopting the above technical solution, the system can automatically classify the risk level of abnormal orders by determining the value range of goods in the preset table and the corresponding priority. This method ensures that abnormal orders related to high-value goods can be handled first, effectively reducing the risk of economic losses that merchants may face. Secondly, this priority sorting method greatly improves the efficiency and resource utilization of the audit process. Auditors can focus on handling high-priority abnormal orders first to ensure that the most important and urgent issues are resolved in a timely manner. This can not only minimize potential economic losses, but also improve the response speed of the entire audit process and enhance the merchant's control over abnormal situations. In addition, this value-based priority sorting method also provides merchants with a flexible management tool. By adjusting the value range and corresponding priority in the preset table, merchants can customize the audit process according to their own business strategies and risk tolerance. For example, during special promotions, the priority of certain product categories can be increased to cope with the possible increased risk of abnormal orders.

[0013] In conjunction with some embodiments of the first aspect, in some embodiments, determining the interval of commodity value in a preset table to obtain the corresponding priority specifically includes: Determine the value interval of the commodity in the preset table and obtain the value priority corresponding to the interval; Determine the flow of people in the cashier area; Determine the priority of the number of people according to the flow of people. When the flow of people is greater than the preset flow threshold, the priority of the number of people is the first priority. When the flow of people is not greater than the preset flow threshold, the priority of the number of people is the second priority. Calculate the weighted average of value priority and number priority to get the corresponding priority.

[0014] By adopting the above technical solution, by determining the value priority corresponding to the commodity value range in the preset table, the system can make preliminary priority division according to the actual value of the commodities in the abnormal order. This method can not only ensure that abnormal orders related to high-value commodities are processed first, but also flexibly adjust the value range according to the actual situation and needs of the merchants, and adapt to retail environments of different types and sizes. Secondly, the introduction of the flow of people in the cashier area as the second dimension of priority determination greatly improves the intelligence and practicality of the system. In the case of large flow of people, the priority is raised to the first level. This approach effectively copes with the problem of abnormal order backlog that may occur during peak hours. By processing abnormal orders in a timely manner, customer waiting time can be reduced and customer satisfaction can be improved. At the same time, additional losses or disputes that may be caused by delayed processing can also be reduced. The value priority and the number of people priority are weighted averaged to obtain the final priority. This method realizes the comprehensive consideration of multiple factors. It can not only balance the impact of commodity value and current operating conditions, but also provide greater flexibility for the system. Merchants can adjust the weights of value priority and number of people priority according to their own business strategies and risk management needs, so as to achieve a more personalized and accurate abnormal order processing process.

[0015] In conjunction with some embodiments of the first aspect, in some embodiments, generating a viewing control, and playing abnormal order period monitoring when the viewing control is triggered, specifically includes: In response to the operation of clicking the view control, the abnormal order period monitoring is played; The abnormal order period monitoring starts from the preset time length from the time point; Mark abnormal products during abnormal order period monitoring.

[0016] By adopting the above technical solution, the convenience and efficiency of the audit process are greatly improved by generating a viewing control and binding it to the abnormal order period monitoring. Auditors can directly play the relevant monitoring video by simply clicking the viewing control, without switching between multiple systems or files. This design not only saves time, but also reduces the errors that may occur in the information search process and improves the accuracy of the audit. Secondly, the abnormal order period monitoring starts playing from the progress of the preset time length from the time point. This functional design greatly improves the audit efficiency. Auditors do not need to watch the entire video, but can directly start viewing from the most relevant time point, which not only saves a lot of time, but also allows auditors to focus on key information more quickly. Especially in the case of handling a large number of abnormal orders, this design can significantly improve the overall audit efficiency, allowing the system to respond to and handle abnormal situations more quickly. The function of marking abnormal products in the abnormal order period monitoring further enhances the readability of video information and the accuracy of the audit. Through intuitive visual markings, auditors can quickly locate and identify the specific products involved in the abnormality, reducing the possibility of misjudgment and omission. This clear visual guidance not only helps to speed up the audit, but also improves the accuracy of the audit results, especially when dealing with complex scenarios or when multiple products are abnormal at the same time.

[0017] In conjunction with some embodiments of the first aspect, in some embodiments, when receiving the audit result corresponding to the information to be audited, after parsing the audit result, the method further includes: Real-time monitoring of whether abnormal orders are entered into the cash register device; If yes, the abnormal order processing completion information is sent to the review terminal; If not, an alarm message is sent to the cash register device, and the alarm message includes the abnormal order.

[0018] By adopting the above technical solution, the system establishes a closed-loop abnormal order processing process by real-time monitoring whether the abnormal order is entered into the cash register. This real-time monitoring mechanism ensures that each identified abnormal order is properly processed, greatly reducing the risk of abnormal orders being missed or delayed. When the abnormal order is successfully entered into the cash register, the system will send abnormal order processing completion information to the audit terminal. This function effectively improves the transparency and traceability of the entire abnormal order processing process. Auditors can timely understand the processing status of abnormal orders without manually following up each case, which greatly improves work efficiency. At the same time, this automated status update mechanism also provides management with real-time business insights, enabling them to better monitor and evaluate the efficiency and effectiveness of abnormal order processing. On the other hand, if the abnormal order is not entered into the cash register, the system will send an alarm message containing abnormal order information to the cash register. It can promptly remind cashiers to pay attention to abnormal orders that have not been processed to prevent omissions and delays. Secondly, this instant feedback mechanism helps cashiers quickly locate and handle problems and improve work efficiency. In addition, by sending specific abnormal order information rather than just simple alerts, the system provides cashiers with directly actionable information, reducing communication costs and processing time.

[0019] In the second aspect, an embodiment of the present application provides an abnormal order determination and review optimization system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0020] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a system, causes the system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0021] In a fourth aspect, an embodiment of the present application provides a computer program product, characterized in that when the computer program product runs on a system, the system executes the method described in any possible implementation manner in the first aspect.

[0022] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. This application provides an abnormal order determination and audit optimization method, which uses a preset algorithm to analyze the surveillance video, accurately identify the settlement product information, and compare it with the settled product information entered by the cash register device. This step greatly improves the recognition accuracy of abnormal orders. Secondly, the behavior analysis of abnormal orders is carried out, and the possible responsible party is accurately located by intercepting the surveillance video of a specific period of time, which effectively reduces the workload of manual review. The abnormal orders, surveillance videos and preliminary judgment results are bound as information to be reviewed and pushed to the queue to be reviewed. This method significantly improves the review efficiency, allowing auditors to quickly obtain all relevant information and make more accurate judgments. Finally, by parsing the audit results, the system can continuously optimize the judgment accuracy and improve the intelligence level of the entire process. The recognition and processing efficiency of abnormal orders is improved, human errors are effectively reduced, and supermarket losses are reduced. At the same time, by analyzing the behavior of customers and staff, it can help merchants discover potential management loopholes or training needs, thereby improving the overall operation quality. The accuracy of identifying missed goods is improved, the difficulty of determining responsibility attribution is reduced, the audit process is simplified, and the efficiency of audit is improved.

[0023] 2. This application provides an optimization method for abnormal order determination and review, which ensures the pertinence and accuracy of the analysis by accurately locating the order entry time point and intercepting the surveillance video of the corresponding period. This method greatly reduces the length of the video to be analyzed, improves processing efficiency, and also reduces the amount of data stored and transmitted. Secondly, the behavior of customers and staff is analyzed separately to determine the target of the last contact with the abnormal goods. This method can clearly restore the process of the incident and effectively avoid the ambiguity of responsibility identification. By judging whether there are abnormal settlement operations or abnormal cashier operations, the system can determine the preliminary responsible party more objectively and fairly, reducing the deviation that may be caused by human judgment. This meticulous behavior analysis method can not only improve the accuracy of abnormal order processing, but also help merchants discover potential management loopholes. For example, if it is found that certain types of goods often appear in abnormal orders, merchants can take corresponding preventive measures, such as adjusting the placement of goods or strengthening employee training. At the same time, this method can also provide valuable data support for subsequent employee training and management optimization. Effectively reduce disputes between customers and employees due to unclear responsibility attribution and improve the efficiency of problem solving.

[0024] 3. This application provides an abnormal order determination and audit optimization method. By determining the value priority corresponding to the commodity value range in a preset table, the system can make preliminary priority division according to the actual value of the commodity in the abnormal order. This method can not only ensure that abnormal orders related to high-value commodities are given priority, but also flexibly adjust the value range according to the actual situation and needs of the merchant, and adapt to retail environments of different types and sizes. Secondly, the introduction of the flow of people in the cashier area as the second dimension of priority determination greatly improves the intelligence and practicality of the system. In the case of large flow of people, the priority is raised to the first level. This approach effectively copes with the problem of abnormal order backlog that may occur during peak hours. By processing abnormal orders in a timely manner, customer waiting time can be reduced, customer satisfaction can be improved, and additional losses or disputes that may be caused by delayed processing can also be reduced. The value priority and the number of people priority are weighted averaged to obtain the final priority. This method realizes the comprehensive consideration of multiple factors. It can not only balance the impact of commodity value and current operating conditions, but also provide greater flexibility for the system. Merchants can adjust the weights of value priority and number priority according to their own business strategies and risk management needs, so as to achieve a more personalized and accurate abnormal order processing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flow chart of an abnormal order determination and review optimization method in an embodiment of the present application.

[0026] Figure 2 It is another flow chart of an abnormal order determination and review optimization method in an embodiment of the present application.

[0027] Figure 3 It is a schematic diagram of the physical device structure of an abnormal order determination and review optimization system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0030] An application scenario of an embodiment of the present application is described below: In a busy supermarket, thousands of customers shop and check out every day. The supermarket has introduced a relatively advanced intelligent loss prevention system, including high-definition surveillance cameras and intelligent settlement systems. However, with the continuous increase in customer flow and the increasing variety of goods, traditional loss prevention measures and the first generation of intelligent systems have been unable to cope with the increasingly complex loss prevention needs.

[0031] In daily operations, supermarkets often face various abnormal order situations. Sometimes customers accidentally leave items at the bottom of the shopping cart and fail to check out, and sometimes cashiers accidentally miss scanning certain items during busy hours. These problems not only cause direct economic losses, but also affect the shopping experience of normal customers and the overall operational efficiency of supermarkets.

[0032] Supermarket management found that although the existing intelligent loss prevention system can capture some abnormal situations, it still has obvious deficiencies in accurately determining abnormal orders and determining the responsibility. The system often generates false alarms, causing innocent customers to be mistakenly intercepted, resulting in unnecessary disputes and complaints. At the same time, real abnormal situations are sometimes missed by the system and cannot be discovered and handled in time.

[0033] In addition, for abnormal orders marked by the system, the review process is often time-consuming and requires manual review of surveillance videos one by one to verify product information and determine who is responsible. This not only greatly increases the workload of managers, but also prolongs the processing time of abnormal orders, affecting the overall operational efficiency of the supermarket.

[0034] In terms of processing priority, existing systems often adopt a simple first-in-first-out principle and cannot flexibly adjust the review order according to the value of the goods and the current operating conditions. This results in some high-value or urgently needed abnormal orders not receiving timely attention and processing, increasing the potential risk of loss. In order to solve the above technical problems, this application provides an abnormal order determination and review optimization method to improve the accuracy of missed goods identification, reduce the difficulty of determining responsibility, simplify the review process and improve the efficiency of the review.

[0035] Combine the following Figure 1 , a method for optimizing abnormal order determination and review in an embodiment of the present application is described: See also Figure 1 , which is a flow chart of an abnormal order determination and review optimization method in an embodiment of the present application.

[0036] S101, identifying the surveillance video of the cashier area according to a preset algorithm and determining the settlement commodity; The system identifies the surveillance video of the checkout area according to a preset algorithm and determines the settlement goods, which include the corresponding settlement goods information.

[0037] This step uses computer vision and image recognition technology to analyze the surveillance video of the checkout area in real time. The system uses pre-trained deep learning models, such as convolutional neural networks (CNN) or regional convolutional neural networks (R-CNN), to identify products in the video. These algorithms can accurately extract product features from complex backgrounds, including shape, color, size, and other information. The system may also use object tracking algorithms, such as Kalman filters or optical flow methods, to track the movement of products on the checkout counter.

[0038] Through these technologies, the system can detect and identify each item passing through the checkout counter in real time and record its detailed information, such as product category, quantity, etc. This information forms the basis of settlement product information and provides accurate data support for subsequent comparison and anomaly detection.

[0039] For example, at a busy supermarket checkout, the system might work like this: The camera captured a customer checking out with five items in his shopping cart.

[0040] The system's image recognition algorithm quickly identified the five items: 2 bottles of mineral water, 1 bag of bread, 1 box of milk and 1 bag of potato chips.

[0041] Object tracking algorithms follow each item as it moves from the shopping cart to the checkout counter.

[0042] The system records detailed settlement product information, including the name, quantity and timestamp of each item passing through the checkout counter.

[0043] In this way, the system can accurately capture the information of all commodities that actually pass through the checkout counter, laying the foundation for the subsequent abnormal order determination. This automated identification process not only improves efficiency, but also reduces human errors, greatly improving the accuracy of the entire settlement process.

[0044] S102, comparing the settlement commodity information with the settled commodity information entered into the cash register device, and if it is determined that there is an abnormal order, performing a behavior analysis operation on the abnormal order time period monitoring to obtain a preliminary determination of the responsible party; The system compares the settlement commodity information with the checked-out commodity information entered by the cash register device to determine whether there are abnormal orders. Abnormal orders are commodity orders corresponding to the settlement commodity information that is not included in the checked-out commodity information. In this step, the system compares the settlement commodity information identified in S101 with the checked-out commodity information actually entered by the cash register device in real time. This process uses data matching algorithms, such as hash table lookup or binary search tree, to quickly compare two sets of data. The system checks whether each item passing through the checkout counter is correctly scanned and entered, and also checks whether there are any items that have been scanned but not actually passed through the checkout counter.

[0045] This comparison not only considers the type and quantity of goods, but also the time sequence in which the goods pass through the checkout counter and are scanned to ensure an exact match. If any inconsistency is found, the system will mark it as a potential abnormal order. This method can effectively identify various anomalies, including missed scans, repeated scans, and even intentional fraud by cashiers or customers.

[0046] For example: The settlement commodity information identified by the system in S101: 2 bottles of mineral water, 1 bag of bread, 1 box of milk, and 1 bag of potato chips.

[0047] The checkout product information entered into the cash register: 2 bottles of mineral water, 1 bag of bread, 1 box of milk.

[0048] The system performed a comparison and found that a bag of potato chips existed in the checkout product information, but was missing in the checked-out product information.

[0049] The system immediately marks the transaction as an abnormal order and records the missing product information (1 bag of potato chips).

[0050] Through this precise comparison mechanism, the system can detect and mark various types of abnormal orders in real time, including missed items, multiple items, or items that are not consistent with the items actually passed through the checkout counter. This automated anomaly detection greatly improves the efficiency and accuracy of the loss prevention system, and can detect problems as soon as they occur, thereby providing timely warnings for subsequent processing.

[0051] If there is an abnormal order, a behavior analysis operation is performed on the abnormal order time period monitoring to obtain a preliminary determination of the responsible party. The abnormal order time period monitoring is obtained by intercepting the monitoring video for a preset time. The processing time point of the abnormal order is located in the abnormal order time period monitoring. The responsible party is preliminarily determined to be the customer's responsibility or the staff's responsibility. How to implement this step specifically will be described below in conjunction with another embodiment, and will not be repeated here.

[0052] When the system detects an abnormal order, it immediately triggers an in-depth analysis of the surveillance video of the corresponding period. This step involves complex computer vision and artificial intelligence technologies, including behavior recognition algorithms, posture estimation, and action sequence analysis. The system will trace back a preset period of time (for example, 2 minutes) from the time of the abnormal order and extract the surveillance video during this period for analysis.

[0053] The behavior analysis operation focuses on two main entities: customers and cashiers. The system tracks their action sequences and identifies key behaviors, such as customers putting items into shopping carts and cashiers scanning items. By analyzing the timing and characteristics of these behaviors, the system can preliminarily determine the party responsible for abnormal orders.

[0054] For example, in a specific scenario: The system detected that a bag of potato chips had passed the checkout counter without being scanned.

[0055] The system extracts surveillance video 2 minutes before the abnormal order occurs.

[0056] The behavioral analysis algorithm recognizes the customer's action sequence: pick up the potato chips - put them in the bottom of the shopping cart - put other items on top.

[0057] The system also analyzes the cashier's behavior: scanning other items normally - not noticing the potato chips at the bottom of the shopping cart.

[0058] Based on these behavioral analyses, the system initially determined that the responsible party was the customer, and that the omission might have been unintentional.

[0059] This automated behavior analysis can not only quickly identify the initial responsible party, but also provide strong support for subsequent manual review. The system's judgment takes into account the intentionality and visibility of the behavior, and can distinguish between unintentional omissions and deliberate evasion of payment. This greatly improves the efficiency and accuracy of abnormal order processing and provides supermarket management with a more reliable basis for decision-making.

[0060] In addition, if it is determined that there are no abnormal orders, there is no emergency in this checkout operation and the following steps are not required.

[0061] S103, bind the abnormal order, abnormal order time period monitoring and preliminary determination of the responsible party, obtain the pending review information, and push the pending review information to the pending review queue; Bind abnormal orders, abnormal order time period monitoring and preliminary determination of responsible parties, obtain pending review information, and push the pending review information to the pending review queue. Specifically: determine the commodity value corresponding to the abnormal commodity in the abnormal order; Generate a viewing control, and play the abnormal order period monitoring when the viewing control is triggered. Specifically: in response to the operation of clicking the viewing control, play the abnormal order period monitoring; The abnormal order period monitoring starts from the preset time length from the time point; Mark abnormal products in abnormal order period monitoring; Bind the product value, viewing controls and the preliminary responsible party to obtain the information to be reviewed, and push the information to be reviewed to the review queue; Determine the interval of the commodity value in the preset table and obtain the corresponding priority. Specifically: determine the interval of the commodity value in the preset table and obtain the value priority corresponding to the interval; Determine the flow of people in the cashier area; Determine the priority of the number of people according to the flow of people. When the flow of people is greater than the preset flow threshold, the priority of the number of people is the first priority. When the flow of people is not greater than the preset flow threshold, the priority of the number of people is the second priority. Calculate the weighted average of the value priority and the number of people priority to get the corresponding priority; Determine the priority level of each pending information in the pending review queue; Adjust the order of the queue to be reviewed according to the corresponding priority.

[0062] In this step, the system integrates and organizes the information obtained in the previous steps to form a complete information package for review. This process involves data structure design and the application of information management systems. The system will create a data object that binds together the details of the abnormal order, the corresponding surveillance video clips, and the preliminary responsible party.

[0063] At the same time, the system will dynamically calculate the priority of each pending review information based on the value of the abnormal goods and the current flow of people in the cashier area. This involves multi-factor decision-making algorithms, such as weighted average method or hierarchical analysis method (AHP). The system will maintain a dynamic queue for pending review and sort the pending review information according to the calculated priority.

[0064] The specific implementation can be as follows: The system creates an information object to be reviewed, which contains the following: Abnormal order ID: AO20240828001 Unusual product: Potato chips (value: 5 yuan) Surveillance video clip: 2024-08-28 15:30:00 to 15:32:00 Preliminary determination of responsible party: Customer Generated viewing controls: a clickable video play button The system queries the preset table to determine the priority of the product value: 0-10 yuan: low priority (1 point) 11-50 yuan: medium priority (2 points) 50 yuan or more: High priority (3 points) The potato chips are worth 5 yuan, which is a low priority item and gets 1 point.

[0065] The system detects the current flow of people in the cashier area: Current traffic: 20 people / minute Preset traffic threshold: 15 people / minute The current flow of people is greater than the threshold, and the priority of people is the first priority (2 points) The system calculates the final priority: Value priority (weight 0.6): 1 * 0.6 = 0.6 Number of people priority (weight 0.4): 2 * 0.4 = 0.8 Final priority: 0.6 + 0.8 = 1.4 The system inserts this pending review information into the pending review queue with a priority of 1.4 and adjusts the queue order according to the priority.

[0066] In this way, the system can intelligently manage and sort the information to be reviewed, ensuring that important abnormal orders are processed in a timely manner, while also taking into account the current operating conditions, improving overall review efficiency and resource utilization.

[0067] S104, when receiving the audit result corresponding to the information to be audited, analyzing the audit result; When the audit result corresponding to the information to be audited is received, the audit result is parsed, and the audit result is that the preliminary determination of the responsible party is correct or the preliminary determination of the responsible party is wrong. The preliminary determination of the responsible party being wrong means that the final determination of the responsible party is opposite to the preliminary determination of the responsible party.

[0068] When the system receives the results of manual review, it will immediately parse and process them. This step involves natural language processing (NLP) technology and decision tree algorithms to understand and classify the review results. The system will classify the review results as preliminary correct or preliminary wrong and update the relevant database.

[0069] This process is not just a simple record of the results, but also includes in-depth analysis of the audit results to continuously optimize the system's judgment ability. The system will use machine learning algorithms, such as support vector machines (SVM) or random forests, to learn the features extracted from these manual audit results and continuously improve its initial judgment accuracy.

[0070] For example, in practice it may be as follows: The system receives the audit results submitted by the auditor: Abnormal order ID: AO20240828001 Audit result: Preliminary judgment is correct Audit Notes: The customer accidentally omitted the item, but has been reminded and paid the fee System analysis audit results: Confirm that the initial judgment is correct Extract keywords: unintentional omission, reminded, back payment System update database: Mark AO20240828001 as reviewed Record audit results and comments Update relevant statistics, such as correct judgment rate System learning and optimization: Add the case as training data to the machine learning model Adjust relevant parameters, such as the weight of items at the bottom of the shopping cart In this way, the system can not only accurately record and process the audit results of each abnormal order, but also continuously learn and improve to improve the accuracy of future judgments. This self-optimization mechanism ensures that the system can adapt to the ever-changing shopping environment and new abnormal situations, and continuously improve loss prevention efficiency.

[0071] S105, real-time monitoring of whether abnormal orders are entered into the cash register device; In this step, the system continuously monitors the data stream of the cash register device and checks in real time whether any previously marked abnormal orders have been correctly entered. This process involves real-time data stream processing technology and database query optimization. The system uses efficient memory data structures, such as hash tables or B-trees, to quickly match abnormal order information.

[0072] At the same time, the system also uses an event-driven architecture to process real-time data. Whenever a new item is entered into the cash register, the system triggers an event and immediately compares it with the list of unprocessed abnormal orders. This method ensures that abnormal orders are discovered and processed in the shortest possible time.

[0073] An example to illustrate this process: The system maintains a list of unprocessed exception orders, which includes: Abnormal order ID: AO20240828001 Unusual product: Potato chips Amount: 5 yuan After confirming with the customer, the cashier scanned the bag of potato chips again.

[0074] The cash register device sends the product entry information: Product: Potato chips Amount: 5 yuan Timestamp: 2024-08-28 15:35:20 When the system receives this information, it immediately triggers a comparison event: Quickly find a list of unprocessed exception orders Matched to AO20240828001 Confirm that the product and amount are consistent The system marks AO20240828001 as processed: Update abnormal order status Record processing time Remove from unprocessed list Through this real-time monitoring mechanism, the system can capture this information as soon as the abnormal order is properly processed. This not only improves the response speed of the entire loss prevention system, but also ensures that each abnormal order is handled promptly and properly, greatly reducing the possibility of omissions and delays.

[0075] S106, sending abnormal order processing completion information to the review terminal; If so, the abnormal order processing completion information is sent to the review terminal.

[0076] When the system confirms that the abnormal order has been correctly entered into the cash register, it will immediately send a notification of processing completion to the audit terminal. This step involves network communication and message queue technology. The system may use WebSocket or message queue systems (such as RabbitMQ) to achieve real-time communication to ensure that information can be delivered to the auditor quickly and reliably.

[0077] This instant notification mechanism not only allows auditors to understand the processing status of abnormal orders in a timely manner, but also improves the transparency and efficiency of the entire audit process. Auditors do not need to frequently check manually, and the system will actively push the latest information.

[0078] For example, in practice: The system confirms that the abnormal order AO20240828001 has been processed correctly.

[0079] The system generates a processing completion message: Abnormal order ID: AO20240828001 Processing status: Completed Processing time: 2024-08-28 15:35:20 Processing method: Scan and record The system sends this information to the audit terminal via a WebSocket connection.

[0080] After the audit terminal receives the information: Update the order status to processed on the interface Play a short beep Display a pop-up message in the notification area The system also updates the status of the abnormal order in the database.

[0081] In this way, auditors can grasp the processing progress of each abnormal order in real time without frequent queries or manual updates. This not only improves work efficiency, but also reduces the possibility of human errors. At the same time, this automated status update also provides management with real-time business insights, enabling them to better monitor and evaluate the efficiency and effectiveness of abnormal order processing.

[0082] S107: Send an alarm message to the cash register device.

[0083] If not, an alarm message is sent to the cash register device, and the alarm message includes the abnormal order.

[0084] If the system detects that an abnormal order has not been processed within a certain period of time, it will send an alarm message to the relevant cash register equipment. This step involves timed task management and multi-channel communication technology. The system will use a timer to monitor unprocessed abnormal orders, and once the preset time threshold is exceeded, the alarm mechanism will be triggered.

[0085] The alert information may be sent through multiple channels, such as screen display of the cash register device, sound reminder, or even push notification to the cashier’s mobile device. This multi-channel alert mechanism can ensure that the information is delivered in time and will not be ignored.

[0086] Through the above embodiments, computer vision and deep learning algorithms are used to automatically identify settlement commodities, and real-time comparison is performed with the information entered in the cash register device, so that abnormal orders can be detected quickly and accurately, greatly improving the detection efficiency and accuracy. The system can automatically analyze the surveillance video of the abnormal order period, and preliminarily determine the responsible party through AI technologies such as behavior recognition, providing an important basis for subsequent manual review, and improving the efficiency of abnormal order processing. By setting a priority mechanism, the order of the queue to be reviewed is dynamically adjusted according to factors such as commodity value and traffic flow, ensuring that important abnormal orders are processed in a timely manner, and optimizing the overall review process. The system can monitor in real time whether abnormal orders are entered into the cash register device, and send processing completion information to the review terminal in a timely manner, or send an alarm to the cash register device, realizing rapid response and closed-loop management. Through machine learning algorithms to learn manual review results, the accuracy of the system's preliminary judgment is continuously improved, so that the system can adapt to the ever-changing shopping environment. The automation and intelligence of the entire system greatly improve the efficiency of discovering and processing abnormal orders, effectively reduce commodity losses, and improve the operating efficiency and profitability of retail enterprises. The system tracks and records the entire process of abnormal orders, enhances the transparency and traceability of the entire process, and is conducive to subsequent analysis and management decisions.

[0087] In the following, another embodiment is combined to describe in detail the above embodiment step S103, if there is an abnormal order, then the abnormal order period monitoring is performed behavior analysis operation, and the preliminary determination of the responsible party is obtained. Figure 2 , another abnormal order determination and review optimization method in the embodiment of the present application is described: See also Figure 2 , is another flow chart of an abnormal order determination and review optimization method in an embodiment of the present application.

[0088] S201. Obtain order information corresponding to the abnormal order; The order information corresponding to the abnormal order is obtained, and the order information includes the time point when the abnormal order is entered into the cash register device.

[0089] This step is the starting point of abnormal order analysis. The system needs to extract detailed information related to abnormal orders from the database or order management system. This process involves database query optimization and data integration technology. The system will use efficient query algorithms, such as index search or hash search, to quickly locate and extract relevant data of abnormal orders.

[0090] Order information usually includes order number, product list, total amount, payment method, etc., but what is particularly important for abnormal order determination is the time when the order is entered into the cash register. This time point will serve as a key reference point for subsequent video analysis. The system may use a time series database or a relational database with timestamp index to store and retrieve this information to ensure efficient time-related queries.

[0091] Specific implementation example: The system received an abnormal order alert, the order number is AO20240828002.

[0092] The system sends a query request to the order database: SQL: SELECT * FROM orders WHERE order_id ='AO20240828002' The database returns the order information: Order number: AO20240828002 Product list: [Milk (1), Bread (1), Apple (3)] Total amount: 35.50 yuan Actual payment: 28.50 yuan Entry time: 2024-08-28 14:23:15 The system parses the information returned, paying special attention to the entry time 2024-08-28 14:23:15.

[0093] The system stores this information in a temporary data structure in memory for quick access in subsequent steps.

[0094] Through this step, the system obtains key information about abnormal orders, especially the entry time, which lays the foundation for subsequent video analysis and behavior judgment. This precise time point positioning can greatly reduce the scope of videos that need to be analyzed and improve the efficiency and accuracy of subsequent processing.

[0095] S202, locating a time point in the monitoring video, and intercepting the monitoring video of a preset length with the time point as the midpoint to obtain abnormal order period monitoring; This step involves video processing and time series analysis technology. The system needs to accurately locate specific time points in a large amount of surveillance video data and extract relevant clips. This process uses video indexing technology and timestamp matching algorithms to quickly locate the target time period.

[0096] The system will first search for the corresponding video file in the video storage system according to the order entry time obtained in S201. Then, using the timestamp information in the video encoding, the system can accurately locate the entry time point. In order to capture the complete transaction process, the system will take this time point as the center, extend forward and backward by a preset time length (for example, 2 minutes), and thus extract a complete video clip.

[0097] Specific implementation example: The system reads the order entry time obtained in S201: 2024-08-28 14:23:15 The system queries the video storage system and finds the video file of the corresponding date and time period: File name: CCTV_Checkout3_20240828_14.mp4 The system uses video processing libraries such as FFmpeg to locate the timestamp of 14:23:15 The system calculates the interception range: Start time = 14:23:15 - 2 minutes = 14:21:15 End time = 14:23:15 + 2 minutes = 14:25:15 The system performs video slicing operations: Command: ffmpeg -i CCTV_Checkout3_20240828_14.mp4 -ss 14:21:15 -to 14:25:15 -c copy AO20240828002_clip.mp4 The system generates a new video clip: AO20240828002_clip.mp4, with a total length of 4 minutes Through this step, the system successfully extracted short clips directly related to abnormal orders from a large amount of surveillance videos. This greatly reduced the amount of data required for subsequent analysis and improved efficiency. At the same time, by including video content before and after order entry, the system ensured that all key behaviors that may lead to abnormalities were captured, providing complete video materials for subsequent behavior analysis.

[0098] S203, determining the customers and staff members under monitoring during the abnormal order period; In this step, the system needs to use computer vision and machine learning techniques to identify and track people in the video. This involves multiple technical fields such as face recognition, human pose estimation, and target tracking. The system uses pre-trained deep learning models such as YOLO (You Only Look Once) or Faster R-CNN to detect and classify people in the video.

[0099] To distinguish between customers and staff, the system may use the staff's unique uniforms or logos as features. At the same time, the system will also consider the position and movement pattern of the characters in the picture. For example, staff usually stand behind the cash register, while customers move in front of the cash register.

[0100] Specific implementation example: The system loads a pre-trained person detection model.

[0101] The system analyzes the video AO20240828002_clip.mp4 frame by frame: Frame 1 (14:21:15): 2 people detected Frame 100 (14:21:19): 3 people detected ... (intermediate frames omitted) Frame 7200 (14:25:15): 2 people detected The system categorizes detected people: Person 1: Positioned behind the cashier counter, wearing a red uniform -> marked as a staff member Person 2: Moving in front of the cash register, in casual clothes -> marked as Customer A Character 3: Appears at 14:22:30, leaves at 14:24:45, in casual clothes -> marked as Customer B The system generates character tracking data: Staff: In the whole process, ID=Staff001 Customer A: On screen throughout the entire process, ID=Customer001 Customer B: 14:22:30-14:24:45 In the screen, ID=Customer002 The system associates person tracking data with video timestamps and stores them as structured data.

[0102] Through this step, the system successfully identified and distinguished the customers and staff in the video, laying the foundation for subsequent behavioral analysis. This precise character recognition and tracking not only helps the system determine who is the main participant in the transaction, but also provides important clues for analyzing abnormal behavior, such as identifying possible coordinated schemes or misoperations.

[0103] S204, performing behavior analysis operations on customers and staff respectively, and determining the last handling target who has been in contact with the abnormal product corresponding to the abnormal order; This step involves complex behavior recognition and sequence analysis techniques. The system needs to understand and analyze the action sequences of people in the video, especially the behaviors of interacting with the product. This is usually achieved using deep learning models such as spatiotemporal convolutional networks (3D CNN) or long short-term memory networks (LSTM).

[0104] The system focuses on several key behaviors: picking up items, putting down items, handing over items, scanning items, etc. By analyzing the time series of these behaviors, the system can track the whereabouts of each item and determine the last person who touched the abnormal item.

[0105] Specific implementation example: The system loads the pre-trained behavior recognition model.

[0106] Behavior sequence in the system analysis video AO20240828002_clip.mp4: 14:21:20 - Customer001 picks up the milk 14:21:25 - Customer001 puts the milk at the cash register 14:21:30 - Staff001 scans milk 14:21:35 - Customer001 picks up the bread 14:21:40 - Customer001 puts the bread on the cash register 14:21:45 - Staff001 scans bread 14:21:50 - Customer001 picked up the apples (3) 14:21:55 - Customer001 puts the apple on the cash register 14:22:00 - Staff001 scans apples (only 2) 14:22:05 - Staff001 moves the goods to the packing area 14:22:10 - Customer001 starts payment The system analyzes the behavior sequence and finds anomalies: Actual number of apples: 3 Scan quantity: 2 Difference: 1 apple not scanned The system determines the last person to come into contact with the unscanned Apple: Last contact time: 14:22:05 Contact: Staff001 System generated conclusion: The last target handled was Staff001 (staff member) Through this detailed behavioral analysis, the system successfully tracked the entire journey of the abnormal product (unscanned apples) and determined that the last person to touch the product was a staff member. This precise behavioral analysis not only helps determine the responsible party, but also provides valuable information for subsequent training and process optimization. For example, in this case, it may be necessary to remind staff to pay special attention to the accuracy of the quantity when handling multiple identical products.

[0107] S205: When it is determined that the last handled target is the customer, determining whether the customer has any abnormal settlement operation; When it is determined that the last handled target is the customer, determine whether the customer has performed an abnormal settlement operation. An abnormal settlement operation is an operation in which the abnormal product is not handed over to the staff and is placed in the settled area.

[0108] This step focuses on customer behavior, especially abnormal operations during the checkout process. The system needs to use advanced computer vision algorithms to analyze the customer's hand movements and the movement trajectory of the goods. This may involve technologies such as optical flow analysis, motion recognition, and object tracking.

[0109] The system pays special attention to whether the customer hides the product, moves it quickly, or bypasses the scanning area. At the same time, the system also analyzes the customer's body language, such as whether he or she shows nervousness or deliberate avoidance.

[0110] Specific implementation example: The system loads a model specifically designed to analyze settlement behavior.

[0111] The system analyzes the behavior sequence of customer Customer001: 14:21:50 - Pick up apples (3) 14:21:55 - Place the apple at the cash register 14:22:00 - Standing and waiting for scanning 14:22:05 - Observation staff scanning 14:22:10 - Start payment The system checks the following key points: Are all products placed in the scanning area: Yes Is there any action to hide the product: No Is there any behavior of fast moving goods: No Is the customer's posture abnormal: No System judgment result: No abnormal settlement operation was found by the customer The system generates a behavior analysis report: "Customer 001 behaved normally during the checkout process, placing all items in the scanning area without any abnormal or suspicious behavior. No deliberate hiding or rapid movement of items was found." In this example, the system did not detect any abnormal checkout behavior by the customer. This indicates that although there are unscanned items, the customer may not be responsible. This detailed behavior analysis not only helps determine the responsible party, but also helps distinguish between intentional fraud and unintentional negligence. For example, in this case, it is likely that the staff was negligent during the scanning process rather than the customer's intentional behavior. This precise judgment is crucial to fairly handling abnormal orders and maintaining good customer relationships.

[0112] S206. Determine the preliminary determination that the responsible party is the customer; If there are abnormal settlement operations, the customer will be preliminarily determined to be responsible.

[0113] In this step, the system needs to combine the analysis results of the previous steps to make a preliminary judgment on responsibility. This process involves the application of decision tree algorithms or rule engines. The system will determine whether to attribute responsibility to the customer based on the preset judgment criteria and the results of behavioral analysis.

[0114] The key to the judgment is whether the customer has any clear abnormal settlement operation. If so, the system will tend to initially determine that the customer is responsible. However, the system will also consider other factors, such as the value of the goods, the customer's shopping history, etc., to ensure the comprehensiveness of the judgment.

[0115] S207. Determine the preliminary determination that the responsible party is the staff member; If there is no abnormal calculation operation, the preliminary determination is that the responsible party is the staff.

[0116] This step is performed on the basis of the determination in S206 that the customer is not responsible. The system needs to further analyze the behavior of the staff to determine whether the responsibility should be attributed to the staff. This process involves more in-depth behavior analysis and rule determination.

[0117] The system will focus on the staff's scanning operations, attention allocation, and operational norms. At the same time, it will also consider working environment factors, such as passenger flow at the time and whether there are any interferences. This requires a combination of computer vision technology and contextual analysis capabilities.

[0118] S208. When it is determined that the last person handled by the cashier is a staff member, determine whether the staff member has performed any abnormal cashier operation.

[0119] When it is determined that the last person handled is a staff member, it is determined whether the staff member has performed an abnormal cashier operation, which is an operation of placing an abnormal commodity in the settled area instead of in the scanning area. The technical principle of this step is similar to that of step S205, and will not be described in detail here.

[0120] If there is any abnormal cashier operation, the staff will be initially identified as the responsible party; If there is no abnormal cashier operation, it is preliminarily determined that the customer is responsible.

[0121] The above embodiments have the following beneficial effects: By accurately locating the order entry time and intercepting the surveillance video of the corresponding period, the pertinence and accuracy of the analysis are ensured. This method greatly reduces the length of the video to be analyzed, improves processing efficiency, and also reduces the amount of data stored and transmitted. Secondly, the behavior of customers and staff is analyzed separately to determine the last person who has contacted the abnormal goods. This method can clearly restore the process of the incident and effectively avoid the ambiguity of responsibility identification. By judging whether there are abnormal settlement operations or abnormal cashier operations, the system can more objectively and fairly determine the preliminary responsible party, reducing the deviation that may be caused by human judgment. This meticulous behavior analysis method can not only improve the accuracy of abnormal order processing, but also help merchants discover potential management loopholes. For example, if it is found that certain types of goods often appear in abnormal orders, merchants can take corresponding preventive measures, such as adjusting the location of goods or strengthening employee training. At the same time, this method can also provide valuable data support for subsequent employee training and management optimization. It effectively reduces disputes between customers and employees due to unclear responsibility and improves the efficiency of problem solving.

[0122] The following describes the system in the embodiment of the present invention from the perspective of hardware processing. Figure 3, which is a schematic diagram of the physical device structure of an abnormal order determination and review optimization system provided in an embodiment of the present application.

[0123] It should be noted that Figure 3 The structure of the system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0124] like Figure 3 As shown, the system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0125] The following components are connected to the I / O interface 305: an input section 306 including a camera, an infrared sensor, etc.; an output section 307 including a liquid crystal display (LCD) and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0126] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0127] It should be noted that the computer-readable medium shown in the embodiment of the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.

[0128] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Among them, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0129] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the system described in the above embodiment; or may exist independently without being assembled into the system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of a system, the system implements the method provided in the above embodiment.

[0130] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0131] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0132] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0133] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A method for optimizing abnormal order determination and review, characterized in that: include: Identify the surveillance video of the cashier area according to a preset algorithm to determine the settlement commodity, wherein the settlement commodity includes corresponding settlement commodity information; Compare the settlement commodity information with the checked-out commodity information entered by the cash register device to determine whether there is an abnormal order, wherein the abnormal order is a commodity order corresponding to the settlement commodity information not included in the checked-out commodity information; If the abnormal order exists, a behavior analysis operation is performed on the abnormal order time period monitoring to obtain a preliminary determination of the responsible party, wherein the abnormal order time period monitoring is obtained by intercepting the monitoring video for a preset time period, and the processing time point of the abnormal order is within the abnormal order time period monitoring. The preliminary determination of the responsible party is the customer's responsibility or the staff's responsibility; Binding the abnormal order, the abnormal order time period monitoring and the initially determined responsible party to obtain the pending review information, and pushing the pending review information to the pending review queue; When the audit result corresponding to the information to be audited is received, the audit result is parsed, and the audit result is that the preliminary determination of the responsible party is correct or the preliminary determination of the responsible party is wrong. The preliminary determination of the responsible party being wrong means that the final determination of the responsible party is opposite to the preliminary determination of the responsible party.

2. The method according to claim 1, characterized in that The above-mentioned behavior analysis operation of monitoring abnormal order periods to obtain a preliminary determination of the responsible party specifically includes: Acquire order information corresponding to the abnormal order, wherein the order information includes the time point when the abnormal order is entered into the cash register device; Locating the time point in the monitoring video, and intercepting the monitoring video of the preset time length with the time point as the midpoint to obtain the abnormal order period monitoring; Identify the customers and staff members who are monitored during the abnormal order period; Performing behavior analysis operations on the customer and the staff respectively to determine the last handling target who has been in contact with the abnormal product corresponding to the abnormal order; When it is determined that the last handled target is the customer, determining whether the customer has an abnormal settlement operation, wherein the abnormal settlement operation is an operation of not handing the abnormal commodity to the staff and placing it in the settled area; If the abnormal settlement operation mentioned above exists, it is preliminarily determined that the responsible party is the customer; If the abnormal computing operation does not exist, determining that the party responsible for the preliminary determination is the staff's responsibility; When it is determined that the last handled target is the staff member, determining whether the staff member has performed an abnormal cashier operation, wherein the abnormal cashier operation is an operation of not placing the abnormal commodity in the scanning area but placing it in the settled area; If the abnormal cashier operation exists, the staff member is determined to be the party responsible in the preliminary determination; If the abnormal cashier operation does not exist, it is determined that the party responsible is initially determined to be the customer's responsibility.

3. The method according to claim 1 or 2, characterized in that: The step of binding the abnormal order, the abnormal order time period monitoring and the preliminary determination responsible party to obtain the pending review information, and pushing the pending review information to the pending review queue specifically includes: Determine the commodity value corresponding to the abnormal commodity in the abnormal order; Generate a viewing control, and play the abnormal order period monitoring when the viewing control is triggered; The commodity value, the viewing control and the initially determined responsible party are bound to obtain the information to be reviewed, and the information to be reviewed is pushed to a queue to be reviewed.

4. The method according to claim 3, characterized in that After binding the commodity value, the viewing control and the initially determined responsible party to obtain the information to be reviewed, and pushing the information to be reviewed to the queue to be reviewed, the method further includes: Determine the value interval of the commodity in a preset table and obtain the corresponding priority; Determine the priority of each piece of information to be reviewed in the review queue; The arrangement order of the queue to be reviewed is adjusted according to the corresponding priority.

5. The method according to claim 4, characterized in that Determining the interval of the commodity value in the preset table and obtaining the corresponding priority specifically includes: Determine the value interval of the commodity in a preset table, and obtain the value priority corresponding to the interval; Determining the flow of people in the checkout area; Determine the priority of the number of people according to the flow of people, when the flow of people is greater than a preset flow threshold, the priority of the number of people is the first priority, when the flow of people is not greater than the preset flow threshold, the priority of the number of people is the second priority; The value priority and the number priority are weighted averaged to obtain a corresponding priority.

6. The method according to claim 3, characterized in that: The generating of the viewing control and playing the abnormal order period monitoring when the viewing control is triggered specifically includes: In response to the operation of clicking the viewing control, playing the abnormal order period monitoring; The abnormal order period monitoring starts playing from the progress of the preset time length from the time point; The abnormal commodity is marked in the abnormal order period monitoring.

7. The method according to claim 1, characterized in that In the case where the audit result corresponding to the information to be audited is received, after parsing the audit result, the method further includes: Real-time monitoring of whether the abnormal order is entered into the cash register device; If yes, the abnormal order processing completion information is sent to the review terminal; If not, an alarm message is sent to the cash register device, wherein the alarm message includes the abnormal order.

8. An abnormal order determination and review optimization system, characterized in that: The system comprises: One or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the system to execute the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a system, the system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a system, the system is caused to execute the method according to any one of claims 1 to 7.

Images