Store inspection methods, media, equipment and procedures products

Abstract

A store inspection method, medium, equipment, and program product receive inspection result data for a specified inspection task, determine sub-areas with rectification items based on the inspection result data, generate rectification tasks, and distribute them to the store. Furthermore, the store area to be inspected and its sub-areas are pre-configured for the specified inspection task. After generating the rectification task, the target store area associated with the target sub-area configured in the specified inspection task is determined as the shooting area corresponding to the rectification task through an inheritance mechanism. The shooting parameters corresponding to the rectification task are determined and distributed to the cameras covering the shooting area. This process eliminates the need to wait for store staff to upload rectified images, improving the timeliness of image capture. During the rectification process, images are captured based on pre-configured shooting parameters, eliminating inconsistencies in image perspective, range, and parameters caused by manual shooting by store staff, thereby improving the standardization and reliability of the determined rectification results.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a store inspection method, medium, device and program product. Background Technology

[0002] To ensure the operational quality of offline stores, the industry currently widely adopts video-based intelligent inspection for store management. By collecting and analyzing in-store videos, issues such as non-standard employee operations and unclean environments can be effectively identified. After a problem is discovered, store employees typically need to manually take images as evidence of "problem improvement" after making rectifications themselves, and then upload them. Auditors then manually review and judge the uploaded evidence to determine whether the rectification is complete. However, this method of relying on manual evidence collection and uploading has the following technical problems: First, the real-time nature of evidence collection is poor; employees may delay uploading or even miss uploading, causing the system to be unable to obtain the rectified image data in a timely manner. Second, due to the inconsistent angles, ranges, and parameters of manually taken photos, it is difficult to guarantee the standardization and comparability of image data. In summary, existing technologies lack a technical solution that can automatically and standardizedly acquire rectification verification images, resulting in low automation and insufficient reliability in the evaluation of rectification effectiveness. Summary of the Invention

[0003] In a first aspect, embodiments of this application provide a store inspection method, the method comprising: Receive inspection result data for a specified inspection task, wherein the specified inspection task is configured with at least one in-store area to be inspected in the store, a sub-area associated with the at least one in-store area, and inspection items for the sub-area. Based on the inspection results data, target sub-areas with rectification items are identified; rectification tasks are generated for the rectification items and distributed to the stores; Determine the shooting area and shooting parameters corresponding to the rectification task; wherein, the shooting area is inherited from the target store area associated with the target sub-area configured in the inspection task; The shooting parameters are sent to the cameras covering the shooting area to update the shooting parameters of the corresponding cameras; Acquire images captured by the camera covering the shooting area under updated shooting parameters; the images are used to determine the effectiveness of the rectification task.

[0004] Secondly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the methods described in any embodiment of this application.

[0005] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in any embodiment of this application.

[0006] Fourthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the methods described in any embodiment of this application.

[0007] In this embodiment, inspection result data for a specified inspection task is first received. Then, based on the inspection result data, sub-areas with rectification items are determined, rectification tasks are generated, and distributed to the store. Furthermore, for a specified inspection task, at least one in-store area to be inspected, sub-areas associated with the at least one in-store area, and inspection items for the sub-areas are pre-configured. After generating rectification tasks for rectification items within the target sub-area, the target in-store area associated with the target sub-area configured in the specified inspection task is determined as the shooting area corresponding to the rectification task through an inheritance mechanism. After determining the shooting parameters corresponding to the rectification task, the shooting parameters are distributed to the camera covering the shooting area. The camera covering the shooting area can capture images under the corresponding shooting parameters to determine the rectification effect. On the one hand, the above process can generate rectification tasks, set shooting parameters, and acquire images after identifying the sub-areas with rectification items, without waiting for store staff to upload the rectified images, thus improving the timeliness of image shooting. On the other hand, the image shooting is based on pre-configured shooting parameters during the rectification process, eliminating the problem of inconsistent image perspective, range, and parameters caused by manual shooting by store staff, thereby improving the standardization and reliability of the determined rectification effect.

[0008] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this application. Attached Figure Description

[0009] The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments consistent with this application and, together with the description, serve to explain the technical solutions of this application.

[0010] Figure 1 This is a flowchart of a store inspection method according to an embodiment of this application.

[0011] Figure 2A This is a schematic diagram illustrating the process of setting the shooting area and shooting parameters according to an embodiment of this application.

[0012] Figure 2B This is a schematic diagram illustrating the process of viewing the inspection report according to an embodiment of this application.

[0013] Figure 3 This is a business process diagram of an embodiment of this application.

[0014] Figure 4 This is a schematic diagram of the system architecture of an embodiment of this application.

[0015] Figure 5 This is a schematic diagram of a computer device according to an embodiment of this application. Detailed Implementation

[0016] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0017] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items. Additionally, the term “at least one” herein means any combination of at least two of any one or more of a plurality.

[0018] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0019] To enable those skilled in the art to better understand the technical solutions in the embodiments of this application, and to make the above-mentioned objectives, features and advantages of the embodiments of this application more apparent and understandable, the technical solutions in the embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0020] To maintain consistent operational and service standards across offline stores, the industry commonly employs intelligent video inspection for store management. Specifically, intelligent video inspection analyzes store footage to automatically identify various deviations from standard practices. For instance, in the catering industry, it can accurately identify issues such as kitchen staff not wearing gloves or masks while processing food, slippery kitchen floors with oil and debris, and items not being placed according to regulations.

[0021] When such situations are detected, the traditional process typically requires store employees to first rectify the issue on-site, then take photos of the rectified area or behavior, and finally upload the photos to the management platform. Managers then need to manually review these uploads, relying on their personal experience to determine whether the problem has been resolved.

[0022] However, this method increases the workload of store staff and makes it difficult to guarantee the quality of evidence. For example, employees may delay taking photos for several hours due to busy operations during peak hours, or they may only take photos from a deliberately adjusted angle, failing to fully and truthfully reflect the rectification situation, making it difficult to accurately assess the store's rectification results.

[0023] Based on this, this application proposes a store inspection method. The method first acquires inspection result data for a specified inspection task, then determines the target sub-area with rectification items based on the inspection result data, generates a rectification task, and distributes it to the store. Furthermore, after generating the rectification task, the target store area associated with the target sub-area configured in the specified inspection task is determined as the shooting area corresponding to the rectification task. After determining the shooting parameters corresponding to the rectification task, the shooting parameters are distributed to the camera covering the shooting area, so that the camera covering the shooting area can capture images under the corresponding shooting parameters, and the rectification effect is determined using these images. In the above process, the camera acquires images based on the determined configuration information, rather than manually acquiring images. This not only improves the timeliness of image acquisition but also eliminates the problem of inconsistent image perspective, range, and parameters caused by manual shooting by store staff. This allows the system to judge the rectification results under a unified standard of perspective, range, and parameters, thereby improving the standardization and reliability of the determined rectification results. The specific implementation of the embodiments of this application is illustrated below with reference to the accompanying drawings.

[0024] Figure 1 This application illustrates a store inspection method according to an embodiment of the present application, such as... Figure 1 As shown, the method includes: Step S11: Receive inspection result data for a specified inspection task, wherein the specified inspection task is configured with at least one in-store area to be inspected in the store, a sub-area associated with the at least one in-store area, and inspection items for the sub-area. Step S12: Based on the inspection result data, determine the target sub-area with rectification items; generate rectification tasks for the rectification items, and issue the rectification tasks to the store; Step S13: Determine the shooting area and shooting parameters corresponding to the rectification task; wherein, the shooting area is inherited from the target store area associated with the target sub-area configured in the inspection task; Step S14: Send the shooting parameters to the cameras covering the shooting area to update the shooting parameters of the corresponding cameras; Step S15: Obtain images captured by the camera covering the shooting area under the updated shooting parameters; the images are used to determine the execution effect of the rectification task.

[0025] The method described in this application embodiment can be executed by a management platform. This management platform can be an intelligent inspection platform deployed in the cloud, such as a SaaS platform, a local platform deployed privately within an enterprise, an inspection platform based on a private cloud, a hybrid cloud architecture system, or a PaaS service platform. The management platform can connect to one or more stores. A store refers to a physical commercial unit with a fixed business location that provides goods or services to end consumers, such as a retail store, restaurant, bank branch, or brand counter. Various terminal devices can be deployed within the store for image acquisition, such as cameras for capturing pictures or videos, sensors for monitoring temperature, humidity, or pedestrian flow, and smart screens for interaction and display. Control devices for controlling these terminal devices can also be deployed. The control devices can communicate with each terminal device, and the terminal devices can communicate with each other via a local area network. The control devices can communicate with the management platform via Ethernet, and each terminal device can also communicate directly with the management platform via Ethernet. The following example, using one of the stores connected to the management platform, illustrates the solution of this application embodiment.

[0026] In step S11, inspection tasks can be generated on the management platform. These tasks can be automatically generated by the management platform or manually created by inspection personnel. For example, an inspection cycle can be configured on the management platform, which can periodically retrieve inspection tasks from the inspection task pool according to this cycle as the store's inspection tasks. Each retrieved inspection task may include some or all of the tasks in the inspection task pool. In some embodiments, multiple candidate inspection tasks can be obtained, wherein the task information of the candidate inspection tasks is pre-configured with the trigger conditions for the candidate inspection tasks. For example, multiple inspection tasks in the inspection task pool can be used as candidate inspection tasks. If a store meets the trigger conditions of any candidate inspection task, then that candidate inspection task is determined as the store's designated inspection task. When the number of stores is greater than one, the trigger conditions met by different stores may be the trigger conditions of different candidate inspection tasks. In this way, inspection tasks can be customized for stores, thereby achieving dynamic and precise inspections. For example, taking a restaurant as an example, candidate inspection tasks could include service quality inspections of store staff, inspections of table and chair arrangement neatness, and floor cleanliness inspections. For the table and chair arrangement neatness inspection task, a preset threshold number of improperly arranged tables and chairs can be used as the trigger condition for this inspection task. Only when the store meets this trigger condition will the table and chair arrangement neatness inspection task be considered as an inspection task for that store.

[0027] In some embodiments, a store is physically divided into multiple functional areas, referred to as in-store areas. Taking a restaurant as an example, a restaurant may include, but is not limited to, at least some of the following in-store areas: a kitchen area, a customer area, and a storage area. Each in-store area may include several sub-areas; for example, a customer area may include a dining area and / or a cashier area. One or more cameras may be deployed in each sub-area for video recording. The inspection task can configure the in-store areas to be inspected, the sub-areas associated with the in-store areas, and inspection items for the sub-areas. Inspection items are used to check specified tasks, such as the service quality of store staff, the neatness of table and chair arrangement, and the cleanliness of the floor, as described in the previous embodiments. Inspection items can be pre-configured on the management platform, and the configured inspection items can be associated with the corresponding in-store areas and sub-areas. Inspection tasks can be generated based on the inspection items and their associated in-store areas and sub-areas.

[0028] After generating an inspection task, the management platform can automatically assign the task to the inspectors, or the task administrator can manually assign it. Furthermore, based on the store areas and sub-areas configured in the inspection task, images of the corresponding sub-areas can be acquired (hereinafter referred to as inspection images) and sent to the inspectors. Inspection images can be single-frame images or videos composed of multiple consecutive frames. They can be automatically captured by cameras deployed within the sub-area, or captured by store staff or inspectors using devices with video capture capabilities such as mobile phones, tablets, and cameras. Inspectors can generate inspection result data based on the image content and upload it to the management platform. It should be noted that all image acquisition activities have obtained informed consent from the relevant area management and the personnel being filmed. The scope of acquisition is limited to the workplace and complies with the company's internal management system. Simultaneously, the collection and processing of video data strictly comply with relevant laws and regulations and are used only for explicitly authorized business purposes such as inspection, training, and compliance audits, ensuring the legality and legitimacy of image acquisition.

[0029] In embodiments where the inspection images are videos, the large data volume of videos makes it inefficient to determine inspection results by having inspectors review the videos. Therefore, video frames containing violations detected by the detection model from the original video can be used as violation frames. These violation frames are then sent to the inspectors, who generate and upload the inspection results based on them. The original video is the video captured during the execution of a specified inspection task. This way, inspectors only need to review the violation frames to determine the inspection results, effectively improving the efficiency of video inspections. Specifically, business personnel can pre-configure a violation event database according to management needs. This database contains at least one violation event entry, each associated with a violation event name, a corresponding detection target, a detection attribute, and a judgment threshold. The detection target could be, for example, the kitchen floor or tables and chairs in the dining area, and the detection attribute could be, for example, cleanliness or tidiness. The detection model can detect each input video frame and output the location and attribute values ​​of the detected target in that frame, such as a floor dirt score or table / chair offset. The attribute value is compared with a pre-configured threshold. If the value exceeds the threshold, the frame is determined to be a violation event frame.

[0030] The detection model can be a deep learning-based computer vision model, such as a convolutional neural network (CNN), or a multimodal language model with cross-modal understanding capabilities. In some embodiments, the detection model may include multiple sub-models, each capable of detecting a non-compliance condition. For example, multiple sub-models include, but are not limited to, sub-models for detecting the service quality of store employees, sub-models for detecting store hygiene issues, sub-models for detecting queue conditions, and sub-models for detecting the cleanliness of store shelves. Furthermore, each sub-model can be associated with one or more cameras within the store to acquire video from the associated cameras and perform corresponding video detection based on the acquired video. That is, the video captured by each camera is sent to a specific sub-model; for example, video from a camera covering the kitchen area is sent to a sub-model for detecting store hygiene issues, video from a camera covering the dining area is sent to a sub-model for detecting the service quality of store employees, and video from a camera covering the checkout area is sent to a sub-model for detecting queue conditions.

[0031] In one implementation, the detection model can be a locally deployed model at the store (hereinafter referred to as the local model) or a cloud-deployed model (hereinafter referred to as the cloud model). A video frame is only uploaded to the management platform for inspection personnel to view if it is detected as a violation event frame by either the local model or the cloud model. This effectively reduces the workload of inspection personnel. In another implementation, the detection model can also include both a local model and a cloud model. The local model can first detect violation event frames from the inspection video, and then transmit the violation event frames to the management platform. The management platform can input the violation event frames into the cloud model for secondary detection. If the secondary detection result still determines that the video frame is a violation event frame, then the video frame is sent to the inspection personnel. The detection accuracy of the cloud model can be higher than that of the local model. For example, the cloud model can be a model with a larger scale, richer training data, and / or more advanced architecture than the local model. This embodiment uses a cloud-edge collaborative approach to filter violation event frames, which can effectively improve the detection accuracy of violation event frames.

[0032] In step S12, the inspection result data can be used to indicate target sub-areas with rectification items. Rectification items indicate violations. For example, a rectification item could be "tables and chairs not arranged neatly" or "the floor not clean." Furthermore, rectification items can include type information, location information, severity, and time of occurrence, so that store staff can understand the details of the rectification task when carrying out rectification. The type information indicates the type of rectification item; for example, it can be divided into a first type related to safety and / or a second type related to service standards. The location information can be the target sub-area where the rectification item was detected, or it can be further refined to the location coordinates of the rectification item within the target sub-area. The severity is a quantitative rating of the scope of impact and urgency of the rectification item based on preset rules, used to determine the priority of the rectification task. The time of occurrence is the precise timestamp of the first identification or recording of the rectification item.

[0033] After identifying the target sub-area with rectification items, rectification tasks can be generated for these items. These tasks guide the correction of violations. For example, if the violation is "tables and chairs are not neatly arranged," the rectification task can directly include the information about the violation, i.e., "tables and chairs are not neatly arranged," or it can include information about the corresponding handling method, such as "please arrange the tables and chairs neatly." Furthermore, the rectification task can also include images of the violation, such as the violation event frames in the aforementioned embodiments. These images can be automatically identified by the detection model or obtained by inspectors taking screenshots of the inspection images while reviewing them. After generating the rectification task, it can be distributed to stores. For example, it can be distributed to fixed terminal devices deployed within the store, such as industrial control tablets, industrial control computers, smart POS terminals, self-service kiosks, etc., or to the mobile terminals of store employees, such as mobile phones, tablets, and laptops.

[0034] In step S13, the store's cameras can be automatically invoked to capture images, hereinafter referred to as rectification images. These rectification images can be used to determine the effectiveness of the rectification task. To achieve the above automatic invocation, the shooting area corresponding to the rectification task must first be determined. Although the inspection result data indicates the target sub-area where rectification items exist, a rectification item may not only appear in a single sub-area. For example, assuming the rectification item is the service quality of store staff, and the target sub-area where this rectification item is detected is the checkout area, this does not mean that the service quality problem only exists in the checkout area, but may exist in the entire customer area. Therefore, in order to comprehensively detect the effectiveness of the rectification task and thus improve the comprehensiveness and accuracy of subsequent rectification effect tracking, the shooting area is inherited from the target store area associated with the target sub-area configured in the inspection task. Here, inheritance means that the target store area associated with the target sub-area pre-configured in the specified inspection task is directly used as the shooting area corresponding to the rectification task, without resetting or additionally entering new area information. For example, if the checkout area is a target sub-area with rectification items, and the target store area associated with the checkout area includes the customer area, then the customer area can be determined as the shooting area corresponding to the rectification task. Since the aforementioned store areas are already configured in the designated inspection tasks, the store areas configured in the designated inspection tasks can be directly inherited into the rectification tasks. This method not only automates the determination of the shooting area, but also automatically identifies the updated store areas as the shooting areas for the rectification tasks when the store areas configured in the inspection tasks are updated, eliminating the need for users to manually change the shooting areas and reducing the complexity of manual operations.

[0035] In some embodiments, the designated inspection task and rectification task are stored and managed using preset task data objects. The task data object is a data structure in a computer system used to store and manage task-related information. It can be implemented as a collection of key-value pairs (such as a JSON object), records in a relational database table, or a class instance in an object-oriented programming language. The data object for the designated inspection task includes a task identifier, a sub-region identifier, and an in-store area identifier associated with the sub-region identifier. Based on this, the above inheritance process can be implemented as follows: When generating a rectification task, a rectification task data object is created. The rectification task data object has an inheritance field, which stores the source task identifier, pointing to the task identifier of the designated inspection task. When determining the shooting area corresponding to the rectification task, the task identifier of the designated inspection task pointed to by the source task identifier in the inheritance field is determined. A data object containing this task identifier is determined. The sub-region identifier is obtained from the determined data object, and the in-store area identifier associated with the determined sub-region identifier is obtained. The in-store area corresponding to the obtained in-store area identifier is determined as the shooting area.

[0036] For example, the task data object for a designated inspection task records a sub-area identifier called "Cashier," whose associated in-store area identifier is "Customer Area." When the inspection results determine that there are rectification items at the cashier and a rectification task is generated, the newly created rectification task data object stores the source task identifier through inherited fields. When subsequently determining the shooting area, the task data object for the designated inspection task is located based on this source task identifier. The sub-area identifier of the cashier is read from it, and then the in-store area identifier of the customer area associated with that sub-area identifier is obtained, thus determining the customer area as the shooting area for the rectification task.

[0037] Furthermore, after determining the shooting area, adjustments can be made to the shooting area based on the instructions from inspection personnel. These adjustments can include adding or deleting sub-areas within the shooting area. This method ensures that the determined shooting area better meets the inspection requirements.

[0038] In addition, the shooting parameters corresponding to the rectification task can be determined. Shooting parameters may include, but are not limited to, at least some of the following: shooting cycle, shooting duration, shooting frequency, first shooting time, shooting start time, shooting angle, and focal length. The shooting cycle refers to the total time span of image capture required to track the long-term effects of a specific rectification item. For example, a shooting cycle of 7 days means that images will be continuously captured from the area to be rectified over the next 7 days. Shooting duration refers to the duration of a single continuous image capture each time the camera starts shooting. For example, it can be set to 30 minutes per shooting session to ensure that enough images are captured for detection. Shooting frequency refers to the time interval between two consecutive camera shooting sessions, such as capturing an image every 15 minutes. The first shooting time refers to the time when the first image used to determine the rectification effect is captured; for example, it can be set to 3 days after the rectification task is generated. The shooting start time refers to the time when shooting begins within each operating cycle. For example, assuming a one-day operating cycle, 12:00 noon each day can be set as the shooting start time, thus starting image capture at 12:00 noon each day. Shooting parameters can be manually set by management personnel or automatically determined based on specific information. It is understood that the numerical values ​​in the embodiments of this application are illustrative and not intended to limit this application. The following provides examples of several implementation methods for automatically determining shooting parameters.

[0039] In some embodiments, the type of rectification item can be determined, and the shooting parameters corresponding to the rectification task can be determined based on the type of rectification item. As mentioned above, the type of rectification item may include a first type related to safety and / or a second type related to service specifications. For example, a first type of rectification item may be an employee not wearing a hat while working, or the temperature of a refrigerator exceeding the standard. A second type of rectification item may be, for example, an untidy table in the dining area or a dirty kitchen floor. Since the first type of rectification item carries a higher risk and is often more likely to recur than the second type, the shooting cycle for the first type of rectification item can be set longer than that for the second type, and / or the shooting frequency for the first type of rectification item can be set higher than that for the second type.

[0040] In some embodiments, the shooting parameters include the shooting period. The rectification item may include the object to be rectified. If the object to be rectified includes store employees, the shooting parameters corresponding to the rectification task are determined. Specifically, this may include determining the employee's working hours based on the store's employee schedule and determining the shooting period to include the employee's working hours. In this way, the rectification effect can be tracked specifically for particular employees. For example, assuming a service quality problem is detected with employee A, then employee A's working hours can be determined based on the employee schedule. Assuming employee A's working hours include Monday to Friday, then the shooting period including Monday to Friday can be determined.

[0041] In some embodiments, the rectification task also includes a deadline. Setting a deadline establishes a clear time window for the rectification task, improving its execution efficiency. The deadline can be determined based on the type of rectification item corresponding to the task. For example, the deadline for a rectification task related to a first type of safety-related rectification item can be earlier than the deadline for a rectification task related to a second type of service specification-related rectification item. This differentiated deadline setting allows store staff to prioritize specific types of rectification tasks.

[0042] When a rectification task includes a deadline, the initial shooting time of the cameras covering the coverage area can be determined based on the deadline, ensuring that the cameras capture images upon reaching that deadline. The initial shooting time can be between, after, or the same as the deadline. For example, if the deadline is 48 hours after the rectification task is created, the initial shooting time can be set to one hour before the deadline, i.e., 47 hours after the task was created. This allows for early tracking of rectification progress and assessment of potential delays. Alternatively, the initial shooting time can be set to one hour after the deadline, i.e., 49 hours after the task was created. This allows for checking for overdue rectifications. Or, the initial shooting time can be set to the same as the deadline, i.e., 48 hours after the task was created.

[0043] In some embodiments, the shooting parameters include the shooting start time. It is also possible to obtain the store's closing time within at least one business cycle after the initial shooting time, and determine the shooting start time for the corresponding business cycle based on the closing time within at least one business cycle. The shooting start time within any business cycle is earlier than the closing time within that business cycle. Optionally, the shooting start time within a business cycle is within a preset time range before the closing time of that business cycle. For example, assuming a day is considered a business cycle and the store's closing time is 17:00, the shooting start time can be determined to be 30 minutes before the closing time, i.e., 16:30. Some rectification tasks are often related to the store's business hours; for example, stores typically dispose of waste materials before closing. In this way, the shooting start time can be flexibly set according to the store's business hours.

[0044] In some embodiments, the shooting parameters include the shooting angle and / or focal length. The location of the rectification item within the target sub-region can also be determined, and the shooting angle and / or focal length corresponding to the rectification task can be determined based on this location. Specifically, a pre-trained object detection model can be used to detect the image within the area where the rectification item is located to obtain the position coordinates of the rectification item within the target sub-region. For example, when there is a dirty area on the ground, the detection model can output the coordinates of the center point of the bounding box of the dirty area; when there are misplaced items, the detection model can output the key point coordinates of the misplaced items. After obtaining the location of the rectification item, the location can be converted into the horizontal and vertical angles that the camera needs to rotate, based on the camera's calibration parameters and installation position, so that the camera can be aligned with the location of the rectification item. Simultaneously, a suitable focal length value can be calculated based on the distance of the object in the rectification item from the camera and the actual size of the object in the rectification item. If the object in the rectification item is far away or small in size, the focal length is increased to bring it closer for shooting, ensuring clear details; if the object in the rectification item is large in size or close in distance, the focal length is appropriately reduced to obtain a wider field of view, facilitating observation of the overall appearance of the rectification item and its surrounding environment.

[0045] In some embodiments, historical shooting records of the target store area can also be obtained. These historical shooting records include the effective shooting parameters used when performing historical rectification tasks related to the target store area, and the quality scores of the images captured using those effective shooting parameters. The rectification items in the historical rectification tasks are of the same type as those in the current rectification task, for example, both being service standard-related or floor cleaning-related. Shooting parameters corresponding to images whose image quality scores meet preset scoring conditions can be filtered from the historical shooting records and used as the initial shooting parameters for the current rectification task. The preset scoring conditions could be, for example, a score higher than a preset value. For example, if a historical rectification task performed on a customer area used a parameter combination of a 12mm focal length and a shooting angle of 0 degrees horizontally and 30 degrees vertically, and the captured image quality score was 9 points, then that parameter combination would be used as the initial shooting parameters for the current rectification task.

[0046] Then, images captured under the initial shooting parameters can be obtained. Based on the acquired images, the execution effect of the rectification task and the confidence level of the execution effect can be determined. The execution effect and confidence level of the rectification task can be obtained through a pre-trained rectification effect evaluation model. If the confidence level is less than a preset confidence threshold, such as 0.7, the initial shooting parameters can be adjusted. Adjustable parameters include, but are not limited to, the camera's focal length (e.g., increasing the focal length to obtain clearer close-ups) and the shooting angle (e.g., adjusting the horizontal or vertical angle to cover the location of the rectification item). The adjusted shooting parameters can be determined as the shooting parameters corresponding to the rectification task and updated in the historical shooting record for reference by subsequent similar rectification tasks. For time-related parameters such as shooting cycle, shooting duration, shooting frequency, first shooting time, and shooting start time, the methods described in the aforementioned embodiments can be used, such as dynamic calculation based on the rectification item type, employee shift schedule, and task deadline.

[0047] In step S14, after determining the shooting parameters, these parameters can be sent to the cameras covering the shooting area to update their shooting parameters. For example, assuming the shooting area is a customer area, which includes a cashier area and a dining area, and a camera covering the cashier area and a camera covering the dining area are deployed in the store, the shooting parameters can be sent to the camera covering the cashier area and the camera covering the dining area respectively to update their shooting parameters. The shooting parameters of cameras covering different sub-areas can be configured with the same parameter value or different parameter values.

[0048] In step S15, the camera covering the shooting area can capture rectification images under the updated shooting parameters. The management platform can obtain these rectification images and distribute both the images and rectification tasks to the inspection personnel. The inspection personnel can determine the rectification items based on the rectification tasks and assess the effectiveness of the rectification tasks by viewing the rectification images. Furthermore, the rectification images can be input into a detection model, which will determine the execution result of the rectification task based on the images. Further, if the detection model fails to determine the execution result of the rectification task, both the rectification images and the rectification tasks will be distributed to the inspection personnel for manual determination. This method, combining automatic model review with manual review, can accurately, comprehensively, and efficiently determine the execution result of the rectification tasks. The detection model can be the cloud-based detection model described in the previous embodiments, and it can be a multimodal language model or other types of models.

[0049] In some embodiments, after the inspectors determine the effectiveness of the rectification task based on the rectification images, they can also receive rectification result data characterizing the effectiveness of the rectification task. This rectification result data indicates whether the rectification is successful or unsuccessful. If the rectification result data indicates that the rectification is successful, the rectification task can be closed directly. If the rectification result data indicates that the rectification is unsuccessful, the rectification task can be reissued to the store, and the step of determining the store area configured in the specified inspection task as the shooting area corresponding to the rectification task can be returned, thereby starting a new round of rectification process.

[0050] In some embodiments, rectification result data is also used to indicate areas where rectification has failed. If the rectification result data indicates that rectification has failed, the shooting area can be narrowed down to match the areas where rectification has failed as indicated by the rectification result data. Specifically, the initial shooting area typically covers a large area, such as including both the dining area and the cashier. After receiving the rectification result data, it can be parsed to determine which specific sub-areas still failed rectification. For example, if the dining area rectification failed while the cashier rectification passed, the rectification result data would indicate the dining area as the area that failed rectification. In this case, the shooting area can be narrowed down from the original combined dining area and cashier to only cover the dining area. Through this narrowing method, subsequent images no longer include images of the cashier, but only images of the dining area, thereby improving the targeting of rectification result tracking and the efficiency of rectification effect judgment.

[0051] In some embodiments, rectification result data representing the execution effect of rectification tasks can also be obtained based on images. The images and rectification result data are then correlated, and a rectification history record is generated based on the correlated images and rectification result data. A visualized inspection report is then generated based on the rectification history record. The rectification report can record at least some of the following information: the time the inspection image was taken, the type of violation in the inspection image, the inspection result data, the rectification task, the completion time of the rectification task, the execution result of the rectification task, the rectification image, the time the rectification image was taken, the number of rectifications, and the type of rectification item. Reviewers who verify the execution results of the rectification tasks can view the inspection report to confirm the store's rectification status.

[0052] In actual operation, the store's regional layout may change due to renovations, functional adjustments, or other reasons. If the rectification task directly uses the regional information recorded in the inspection task, subsequent rectification verification may fail or be misidentified once the region is merged or split. To address this, this embodiment introduces a regional configuration version number, which is an incrementing value bound to the store's regional layout. The version number is automatically adjusted each time a store region is added, deleted, or modified.

[0053] Specifically, both inspection and rectification tasks are configured with regional configuration version numbers. The regional configuration version number in the rectification task is inherited from that in the inspection task, and it corresponds to the regional layout in the store. The regional layout can be obtained in several ways. For example, it can receive regional layout modification commands submitted by administrators through the configuration interface; or it can interface with the store design system to automatically synchronize the regional layout; or it can perform scene recognition based on images captured by cameras to automatically detect changes in the regional layout; or it can synchronize the regional layout from the store asset management system.

[0054] Before determining the effectiveness of the rectification task, a consistency comparison can be made between the regional configuration version number in the rectification task and the regional configuration version number in the inspection task. If they do not match, the regional layout corresponding to the regional configuration version number in the inspection task can be analyzed to determine whether the target store area exists. If it does, the shooting area of ​​the rectification task can be updated to the corresponding area in that regional layout; otherwise, the alternative area with the highest similarity to the target store area in that regional layout can be found, and the target shooting area can be updated to that alternative area.

[0055] For example, suppose a store initially configured a kitchen area in its inspection tasks, and the area configuration version number recorded when creating the inspection task was v1. This version number is inherited by the rectification task. Subsequently, the store underwent renovations, splitting the kitchen area into a hot kitchen area and a cold kitchen area. The area configuration version number in the inspection task was updated to v2. When performing rectification verification based on the v1 version, a version number inconsistency was detected. The area layout corresponding to v1 was analyzed, and it was found that the kitchen area no longer existed. Therefore, the replacement area with the most similar function to the kitchen area was searched in the area layout with the v2 version number. For example, the hot kitchen area and the cold kitchen area could be used together as the new shooting area. By comparing version numbers and automatically adapting, the failure of rectification tasks or incorrect area positioning caused by store layout changes is reduced. This enables rectification tasks to adapt to dynamic environments and ensures that rectification verification is always based on a valid store area configuration.

[0056] Figure 2A and Figure 2B Taking the example that both the inspection images and rectification images are video, this diagram illustrates the process of setting the shooting area and shooting parameters, as well as viewing the inspection report. Figure 2AAs shown, the management platform interface displays a rectification item: "Employees have conflicts or arguments with customers." Rectification tasks can be automatically generated, assigned to reviewers (e.g., Zhang San), and a deadline (e.g., 5 days) can be set. It can also acquire shooting parameters, including shooting area, shooting cycle, shooting time, and shooting frequency. These shooting parameters can be determined based on the method described in any of the preceding embodiments. After determination, they can be displayed on the user interface and manually adjusted. Figure 2B As shown, a task list can be displayed on the workbench. For example, the task list in the image includes two tasks: one is a rectification task for Restaurant A, which has not yet been processed by the store staff and is marked "Pending Improvement" in the image; the other is a rectification task for Restaurant B, which has been completed by the store staff or has reached its deadline and needs to be reviewed by the reviewer, marked "Pending Review" in the image. For the latter task, the reviewer can click on the task to view its details, such as the improvement history, i.e., the rectification report, which includes video frames taken by the store's cameras at different times. These video frames can serve as evidence to prove that there are rectification items in the store, or to prove that the rectification items have been eliminated after rectification. In addition, the rectification report can also include the execution result of the rectification task, such as "Review Passed" indicating that the rectification was successful, and "Review Failed" indicating that the rectification was unsuccessful. Furthermore, the execution result of the rectification task can also include the reviewer's comments, such as "Significant progress has been made."

[0057] This application embodiment automatically captures and analyzes video frames through the system, reducing the frequency and time of inspections during the inspection process; in the rectification stage, it can reduce the time for manual review and the problem of inconsistent execution standards caused by different people's different execution efforts, thereby improving the authenticity and fairness of the improvement content.

[0058] The following is an example illustrating an overall process of an embodiment of this application. For example... Figure 3As shown, store inspections can be conducted first, either manually or automatically. During manual inspections, inspectors can physically visit the store to observe its condition. If rectification items are found, images can be captured and uploaded to the cloud management platform. During automatic inspections, images of the store can be captured by cameras, rectification items identified based on these images, and the captured images uploaded to the management platform. The images captured during these store inspections can be single photos or continuously recorded videos; this embodiment uses video recording as an example. The management platform can then generate rectification tasks and distribute them to the stores. Furthermore, evidence can be generated simultaneously with the rectification tasks, namely the images uploaded during automatic or manual inspections. Rectification tasks can be distributed to stores, for example, to in-store terminal devices or to the terminal devices of designated employees, who can be employees with attendance determined according to their work schedules. The management platform can determine the completion rate of store rectifications, such as by estimating whether the rectification task is completed based on whether the current time has reached the preset deadline. If so, the shooting area and shooting parameters corresponding to the rectification task can be set, and images collected by cameras covering the shooting area based on the above shooting parameters can be obtained, and the inspection personnel can be notified. The inspection personnel can automatically analyze the images based on artificial intelligence, or obtain the execution results of the rectification task manually. The following uses a restaurant as an example to illustrate the shooting parameters in some embodiments.

[0059] In some embodiments, multiple candidate parameter values ​​can be set for each shooting parameter from smallest to largest. For example, two candidate parameter values ​​can be set for each shooting parameter.

[0060] Some rectification items are characterized by high risk and recurrence. These items are often closely related to food safety and compliance, requiring continuous verification of improvement effects to prevent recurrence. Therefore, the rectification tasks for these items necessitate a high shooting frequency and long shooting duration. Thus, for such rectification tasks, shooting parameters should at least include shooting frequency and shooting duration, with both values ​​set to the larger of several candidate values. Scenarios requiring a high shooting frequency and long shooting duration include, but are not limited to, the following: Scene 1: Item requiring rectification: Employees were not wearing hats while operating the equipment. Shooting frequency: High shooting frequency, such as capturing a snapshot once every 10 minutes.

[0061] Shooting duration: Long shooting duration, covering daily peak passenger flow and key meal preparation periods, such as 11:00-13:30 noon and 17:30-20:30 in the afternoon.

[0062] Filming area: The kitchen area, including the kitchen preparation area, the front-of-house food delivery area, and the food preparation station. These areas are all high-frequency areas for employee operations.

[0063] Scene 2: Item requiring rectification: Refrigerated display case temperature exceeds standard. Shooting frequency: High shooting frequency, such as taking a snapshot once every 5 minutes.

[0064] Shooting duration: Long shooting duration, such as 24-hour all-weather coverage, especially during low-temperature periods at night.

[0065] Filming area: The kitchen area, including the refrigerator temperature control panel, the inside of the refrigerator, and the kitchen equipment room. The operating status of the refrigerator is observed by filming these areas.

[0066] Other rectification items are characterized by long-term compliance, low risk, and low likelihood of relapse. These rectification items are mostly related to service standards and environmental cleanliness, and their stability after improvement is strong, requiring no continuous high-frequency monitoring. Rectification tasks for these items can be set with lower shooting frequencies and shorter shooting durations. Therefore, for these rectification tasks, shooting parameters should at least include shooting frequency and shooting duration, and both the shooting frequency and shooting duration values ​​can be set to the smaller candidate value from multiple candidate parameter values. Scenarios requiring lower shooting frequencies and shorter shooting durations include, but are not limited to, the following: Scene 3: Item requiring rectification: Tableware on the dining tables in the front of house is not neatly arranged. Shooting frequency: Low shooting frequency, such as taking one shot each before meals and after lunch break per day.

[0067] Shooting duration: Short shooting duration, such as each shoot lasting only 10 minutes, covering key moments of the table setting.

[0068] Shooting area: Lobby Scene 4: Item requiring rectification: Dirty kitchen floor Shooting frequency: Low shooting frequency, such as taking a snapshot once a day during the cleaning period after closing.

[0069] Shooting duration: Short shooting duration, lasting 20 minutes, covering the entire ground cleaning process.

[0070] Filming areas: the back kitchen passage and the entrance to the dishwashing area, these are areas prone to water stains.

[0071] Some rectification items require flexible setting of shooting times based on the restaurant's business hours, and dynamic adjustment of shooting strategies at different times to avoid ineffective snapshots. Therefore, for such rectification tasks, shooting parameters must at least include the shooting time. Scenarios requiring the setting of shooting times include, but are not limited to, the following: Scene 5: Rectification item: Waste materials were not disposed of Shooting time: such as 30 minutes before the restaurant closes for business.

[0072] Filming area: kitchen cutting station, such as a centralized abandoned workstation.

[0073] If the rectification is confirmed to be complete, the inspection personnel can be notified to conduct a review to confirm the execution result of the rectification task. The review process can be carried out through artificial intelligence or a manual channel. If the improvement is approved, the task is terminated; otherwise, the rectification task is reissued to the store.

[0074] Figure 4 A schematic diagram of the system architecture according to an embodiment of this application is shown. The system includes a management platform 301, a camera 302, a local model 303, a cloud model 304, an inspection personnel terminal 305, a human resource management system 306, a scheduling system 307, a store manager terminal 308, and a store employee terminal 309. The management platform 301 can automatically create inspection tasks, assign inspection personnel to these tasks, and distribute the tasks to the inspection personnel terminals 305. After receiving the task, the inspection personnel terminal 305 can send an image acquisition request to the management platform 301. In response to this request, the management platform 301 can send the inspection images captured by the camera 302 to the inspection personnel terminal 305 for the inspection personnel to view. The inspection personnel can generate inspection result data through the inspection personnel terminal 305 and upload it to the management platform 301.

[0075] If the inspection results indicate a sub-area with rectification items, the management platform 301 can generate rectification tasks for those items, then retrieve the information of the on-duty store staff from the scheduling system 307, and distribute the rectification tasks to the staff terminals 309 based on that information. Furthermore, the management platform 301 can also retrieve the organizational structure of the store staff from the human resources management system 306, determine the store manager, and then distribute the rectification tasks to the manager terminal 308, enabling the store manager to supervise the store staff in executing the rectification tasks.

[0076] In addition, the management platform 301 can push parameter configuration tasks to the inspection personnel's terminal 305, enabling the inspection personnel to configure the shooting parameters for the shooting area corresponding to the rectification task. The rectification task may also include a deadline, which can be automatically generated by the management platform 301 or manually generated by the inspection personnel. When the deadline is reached, the rectification task can be presumed to be completed. At this time, the management platform 301 can send the shooting parameters to the camera 302 covering the shooting area, allowing the camera 302 to capture images. Alternatively, after the store clerk completes the rectification, they can manually confirm on their terminal 309 and return the confirmation information to the management platform 301 to notify the management platform 301 that the rectification is complete.

[0077] After rectification is completed, the management platform 301 can send the shooting parameters to the camera 302, so that the camera 302 can capture images according to the shooting parameters. The captured images will be stored in the management platform 301, which can then send the images and rectification tasks to the inspection personnel terminal 305, so that the inspection personnel can view the images through the inspection personnel terminal 305 and determine the rectification effect based on the image content. The inspection personnel terminal 305 can also send feedback information on the rectification effect to the management platform 301, such as whether the rectification is successful or unsuccessful. If the rectification is successful, the management platform 301 can end the rectification task; otherwise, the management platform 301 can reissue the rectification task to the store clerk terminal 309 and the store manager terminal 308, update the task deadline, re-acquire the shooting parameters, and re-execute the aforementioned process.

[0078] This application also provides a computer device, which includes at least a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the methods described in any of the foregoing embodiments.

[0079] Figure 5 This illustration shows a more specific hardware structure diagram of a computer device provided in an embodiment of this application. The device may include: a processor 401, a memory 402, an input / output interface 403, a communication interface 404, and a bus 405. The processor 401, memory 402, input / output interface 403, and communication interface 404 are interconnected internally via the bus 405.

[0080] The processor 401 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The processor 401 may also include a graphics card, such as an Nvidia Titan X graphics card or a 1080Ti graphics card.

[0081] The memory 402 can be implemented in the form of read-only memory (ROM), random access memory (RAM), static storage device, dynamic storage device, etc. The memory 402 can store the operating system and other applications. When the technical solutions provided in the embodiments of this application are implemented by tools or firmware, the relevant program code is stored in the memory 402 and is called and executed by the processor 401.

[0082] Input / output interface 403 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0083] Communication interface 404 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0084] Bus 405 includes a pathway for transmitting information between various components of the device, such as processor 401, memory 402, input / output interface 403, and communication interface 404.

[0085] It should be noted that although the above-described device only shows the processor 401, memory 402, input / output interface 403, communication interface 404, and bus 405, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this application, and not necessarily all the components shown in the figures.

[0086] This application provides a computer program product, including a computer program that, when executed by a processor, implements the methods described in any embodiment of this application.

[0087] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the methods described in any of the foregoing embodiments.

[0088] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by computer devices. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0089] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. When implementing the embodiments of this application, the functions of each module can be implemented in one or more tools and / or hardware. Alternatively, some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0090] The above description is only a specific implementation of the embodiments of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the embodiments of this application, and these improvements and modifications should also be considered as the protection scope of the embodiments of this application.

Claims

1. A store inspection method characterized by comprising: The method includes: Receive inspection result data for a specified inspection task, wherein the specified inspection task is configured with at least one in-store area to be inspected in the store, a sub-area associated with the at least one in-store area, and inspection items for the sub-area. Based on the inspection results data, target sub-areas with rectification items are identified; rectification tasks are generated for the rectification items and distributed to the stores; Determine the shooting area and shooting parameters corresponding to the rectification task; wherein, the shooting area is inherited from the target store area associated with the target sub-area configured in the inspection task; The shooting parameters are sent to the cameras covering the shooting area to update the shooting parameters of the corresponding cameras; Acquire images captured by the camera covering the shooting area under updated shooting parameters; the images are used to determine the effectiveness of the rectification task.

2. The method of claim 1, wherein, The method further includes: Multiple candidate inspection tasks are obtained; the task information of each candidate inspection task is pre-configured with the triggering conditions for that candidate inspection task. If the store meets the triggering conditions of any candidate inspection task, the candidate inspection task will be determined as the designated inspection task for the store.

3. The method of claim 1, wherein, The receiving of inspection result data for a specified inspection task includes: The detection model obtains video frames containing violation events detected in the original video, which are then used as violation event frames; wherein, the original video is the video collected when the specified inspection task is performed; The violation event frame is sent to the inspection personnel, and the inspection result data generated and uploaded by the inspection personnel based on the violation event frame is obtained.

4. The method according to claim 1, characterized in that, Determine the shooting parameters corresponding to the rectification task, including: Determine the type of the rectification item; The shooting parameters corresponding to the rectification task are determined based on the type of the rectification item.

5. The method of claim 4, wherein, The types of rectification items include a first type related to safety and / or a second type related to service specifications; the shooting parameters include shooting cycle and / or shooting frequency. The shooting cycle for the first type of rectification items is longer than that for the second type of rectification items, and the shooting frequency for the first type of rectification items is higher than that for the second type of rectification items.

6. The method of claim 1, wherein, The shooting parameters include the shooting cycle; the rectification items include information about the rectification targets. When the rectification targets include the store's employees, the shooting parameters corresponding to the rectification task are determined, including: Obtain the employee shift schedule for the aforementioned store; Based on the employee work schedule, determine the working hours of the employees included in the rectification targets; The shooting cycle is defined as including the working period.

7. The method of claim 1, wherein, The rectification task also includes a deadline; the method further includes: The first shooting time of the camera covering the shooting area is determined based on the task deadline. The first shooting time is sent to the camera covering the shooting area, so that the camera covering the shooting area can capture images of the shooting area when the first shooting time is reached.

8. The method of claim 7, wherein, The shooting parameters include the shooting start time; determining the shooting parameters corresponding to the rectification task includes: Obtain the closing time of the store within at least one business cycle after the initial shooting time; The shooting start time within a corresponding business cycle is determined based on the business end time within the at least one business cycle, and the shooting start time within any business cycle is earlier than the business end time within the business cycle.

9. The method of claim 1, wherein, The shooting parameters include shooting angle and / or focal length; Determine the shooting parameters corresponding to the rectification task, including: Determine the location of the rectification item in the target sub-region; Based on the location, determine the shooting angle and / or focal length corresponding to the rectification task.

10. The method of claim 1, wherein, The method further includes: Receive rectification result data that characterizes the execution effect of the rectification task; If the rectification result data indicates that the rectification has been successful, the rectification task is closed. If the rectification result data indicates that the rectification is unsuccessful, the rectification task will be reissued to the store, and the step of generating a rectification task for the rectification item will be returned.

11. The method of claim 10, wherein, The rectification result data is also used to indicate areas where rectification fails; the method further includes: If the rectification result data indicates that the rectification has failed, the shooting area will be reduced to match the area where the rectification has failed as indicated by the rectification result data.

12. The method of claim 1, wherein, The method further includes: Based on the image, obtain rectification result data characterizing the execution effect of the rectification task; Associate the image with the rectification result data; A rectification history record is generated based on the associated image and the rectification result data; A visualized inspection report is generated based on the rectification history.

13. The method according to claim 1, characterized in that, The designated inspection task and the rectification task are stored and managed using preset task data objects; the data object of the designated inspection task includes a task identifier, a sub-area identifier, and an in-store area identifier associated with the sub-area identifier; the inheritance process of the shooting area is as follows: When generating a rectification task, a rectification task data object is created. The rectification task data object is set with an inheritance field, which is used to store the source task identifier. The source task identifier points to the task identifier of the specified inspection task. When determining the shooting area corresponding to the rectification task, determine the task identifier of the specified inspection task pointed to by the source task identifier in the inheritance field; Determine the data object containing the data object, obtain the sub-area identifier from the determined data object, and obtain the in-store area identifier associated with the determined sub-area identifier; The store area corresponding to the obtained store area identifier is determined as the shooting area.

14. The method of claim 1, wherein, The method further includes: Obtain historical shooting records of the target store area; the historical shooting records include the effective shooting parameters used when performing historical rectification tasks related to the target store area and the quality scores of the images taken using the effective shooting parameters; the rectification items in the historical rectification tasks are of the same type as the rectification items in the rectification tasks. The shooting parameters corresponding to images whose image quality scores meet the preset scoring conditions are selected from the historical shooting records and used as the initial shooting parameters for the rectification task. Images captured under the initial shooting parameters are obtained, and the execution effect of the rectification task and the confidence level of the execution effect are determined based on the obtained images; If the confidence level is less than a preset confidence threshold, the initial shooting parameters are adjusted, the adjusted shooting parameters are determined as the shooting parameters corresponding to the rectification task, and the adjusted shooting parameters are updated in the historical shooting record.

15. The method of claim 1, wherein, The inspection task and the rectification task are also configured with a regional configuration version number. The regional configuration version number in the rectification task is inherited from the regional configuration version number in the inspection task, and the regional configuration version number in the rectification task corresponds to the regional layout in the store. The method further includes: Before determining the effectiveness of the rectification task, a consistency comparison is made between the regional configuration version number in the rectification task and the regional configuration version number in the inspection task. If they are inconsistent, analyze the area layout corresponding to the area configuration version number in the inspection task to determine whether the target store area exists. If so, update the shooting area of ​​the rectification task to the corresponding area in the area layout; Otherwise, find the alternative area in the area layout that has the highest similarity to the target store area, and update the target shooting area to the alternative area.

16. A computer readable storage medium having stored thereon a computer program, characterized in that, When executed by a processor, the computer program implements the method described in any one of claims 1 to 15.

17. A computer program product comprising a computer program, characterized in that, When executed by a processor, the computer program implements the method described in any one of claims 1 to 15.

18. A store tour system characterized by comprising: The system includes: A server for performing the method according to any one of claims 1 to 15; A terminal device is used to receive rectification tasks issued by the server; A camera is deployed within the store and covers at least one sub-area within the store. It is used to acquire updated shooting parameters issued by the server, capture images of the sub-area covered by the camera based on the acquired updated shooting parameters, and send the captured images back to the server to determine the execution effect of the rectification task.

Images