Street-along shop treatment task distribution method and system based on multi-modal scheduling

Abstract

The invention relates to the technical field of digital data processing, in particular to a street-along shop management task distribution method and system based on multi-modal scheduling, and the method comprises the steps: recognizing the abnormal behaviors of street-along shops, and calculating a processing score based on the importance parameters of the abnormal behaviors of the street-along shops; setting a threshold value based on the processing score, and adopting an online warning mode if the processing score is lower than the threshold value; if yes, acquiring position information and identity information of processing personnel in a preset range near the street shop position; matching the identity information with the abnormal behavior types of the shops along the street in the target area, generating to-be-processed tasks of the processing personnel, and generating a to-be-processed task queue by integrating a plurality of to-be-processed tasks obtained by matching the same identity information; and the to-be-processed task queues are sorted based on the processing scores and the moving distances to obtain the task processing queues, so that tasks can be arranged more efficiently, and the processing efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital data processing, and in particular to a method and system for dispatching tasks for managing street shops based on multi-modal scheduling. Background Art

[0002] The management of street shops refers to the standardization and optimization of the business operations of shops on both sides of urban streets through a series of management measures and technical means, so as to improve the quality of urban environment and the quality of life of citizens. This includes the supervision and management of shops occupying public space, illegal operations, environmental sanitation, safety and order, etc.

[0003] With the rapid development of cities and the booming business activities, the number of shops that need to be supervised has exploded. These shops cover a wide range of formats, from traditional retail stores and restaurants to emerging new retail models that integrate online and offline, each with its own unique operating characteristics and regulatory needs. However, in sharp contrast, the number of personnel responsible for supervising and managing these shops has not increased accordingly, resulting in a serious shortage of staff, making it easy for pending tasks to pile up and not be effectively handled. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for distributing tasks for street shop management based on multimodal scheduling, aiming to arrange tasks more efficiently and improve processing efficiency.

[0005] To achieve the above-mentioned purpose, in a first aspect, the present invention provides a method for dispatching tasks for managing shops along the street based on multimodal scheduling, comprising acquiring street data in a target area and identifying abnormal behaviors of shops along the street, and calculating a processing score based on importance parameters of abnormal behaviors of shops along the street, wherein the importance parameters include occupancy duration, flow of people and occupied area; A threshold is set based on the processing score. If the score is below the threshold, an online warning is used. If the score exceeds the threshold, the location and identity information of the processing personnel within a preset range near the street shops are obtained. Match the identity information with the abnormal behavior types of shops along the street in the target area, generate pending tasks for the processing personnel, and generate a pending task queue by combining multiple pending tasks obtained by matching the same identity information; Sort the task queue to be processed based on the processing score and the moving distance to obtain a task processing queue; The information in the task processing queue is desensitized and encrypted and then sent to the designated personnel.

[0006] The specific steps of acquiring street data in the target area and identifying abnormal behaviors of shops along the street, and calculating processing scores based on importance parameters of abnormal behaviors of shops along the street include: Collect street map data in the target area based on GIS technology, wherein the street map data includes street layout and store locations; Obtain surveillance video data of the street; Identify whether a store has abnormal behavior based on computer recognition technology; When there is abnormal behavior, obtain the specific start and end time of each abnormal behavior to calculate the duration; Calculate the flow of people during abnormal behavior based on the area of ​​crowd images in surveillance video data; Calculate the actual area of ​​the road-occupied area based on surveillance video data; The handling score of each violation is calculated based on duration, traffic volume, actual area and corresponding weights.

[0007] The specific steps of identifying whether a store has abnormal behavior based on computer recognition technology include: Create a labeled dataset that contains normal street views in different scenarios and samples of street-side business operations; Use the labeled data set to train the random forest model to obtain an abnormal behavior recognition model; Based on the abnormal behavior recognition model, target detection is performed on each frame of the image to identify abnormal behavior.

[0008] The specific steps of calculating the flow of people during abnormal behavior based on the area of ​​the crowd image in the surveillance video data include: Use background subtraction method to build background model to distinguish fixed background and moving objects; Apply the YOLO recognition algorithm to identify moving objects and obtain face information; Defining the outline of each group of people using image segmentation technology based on facial information; The total number of pixels occupied by the crowd outlines is calculated and converted to obtain the crowd flow.

[0009] The specific steps of setting a threshold based on the processing score, using an online warning method for scores below the threshold, and obtaining the location information and identity information of processing personnel within a preset range near the location of shops along the street when the threshold is exceeded include: Set score thresholds for online warnings and offline executions; When the processing score is lower than the score threshold, an electronic warning will be sent to the relevant shop owner, explaining the violation and the correction requirements; when the processing score is higher than the score threshold, the street shop will be marked; According to the input street shop location, search for nearby processing personnel within the preset range; Obtain the identity information corresponding to the processing personnel in the identity database.

[0010] The specific steps of matching the identity information with the abnormal behavior types of shops along the street in the target area, generating pending tasks for the processing personnel, and generating a pending task queue by combining multiple pending tasks obtained by matching the same identity information include: Match the corresponding processing function information in the database based on the abnormal behavior of shops along the street; Match the identity information with the processing function information within a preset range to obtain multiple tasks to be processed; A queue of pending tasks is generated by combining multiple pending tasks matched with the same identity information.

[0011] The specific steps of reordering the task queue to be processed based on the priority and the shortest moving distance to obtain the task processing queue include: Calculate the actual driving distance based on the coordinates of the processing personnel's current location and the location where the pending task queue occurs; Calculate a composite score for each pending task based on the actual distance traveled and the processing score; Arrange all pending tasks in descending order according to the calculated comprehensive scores and divide them into high priority sequence and low priority sequence; The priorities of the tasks in the low-priority sequence are increased based on the delay time of the tasks in the low-priority sequence and then reordered.

[0012] The specific steps of increasing the priority of the tasks in the low priority sequence based on the delay time of the tasks in the low priority sequence and then reordering the tasks include: Setting a hysteresis coefficient for each low priority task, wherein the hysteresis coefficient increases as the task hysteresis time increases; The composite score is corrected based on the hysteresis factor; Based on the revised priority scores, the entire queue of pending tasks is periodically reordered.

[0013] The specific steps of desensitizing and encrypting the information in the task processing queue and sending it to the designated person include: Replace and desensitize the sensitive information of the shops along the street in the task processing queue, including names and contact information; Use the AES algorithm to encrypt the desensitized data; Use HTTPS protocol to transmit encrypted data to designated personnel through electronic channels.

[0014] In a second aspect, the present invention also provides a task dispatching system for managing shops along the street based on multimodal scheduling, including a shop abnormal behavior recognition module, a personnel matching module, a task queue production module, a queue priority processing module and a task dispatching module; The shop abnormal behavior identification module is used to obtain street data in the target area and identify abnormal behaviors of shops along the street, and calculate processing scores based on importance parameters of abnormal behaviors of shops along the street, wherein the importance parameters include occupancy duration, traffic volume and occupied area; The personnel matching module is used to set a threshold based on the processing score, and an online warning is used if the score is below the threshold; when the score exceeds the threshold, the location information and identity information of the processing personnel within a preset range near the location of the street shops are obtained; The task queue production module is used to match the identity information with the abnormal behavior types of shops along the street in the target area to obtain tasks to be processed, and to generate a queue of tasks to be processed by combining multiple tasks to be processed obtained by matching the same identity information; The queue priority processing module is used to sort the task queue to be processed based on the processing score and the moving distance to obtain a task processing queue; The task dispatching module is used to desensitize and encrypt the information in the task processing queue and send it to the designated personnel.

[0015] The method and system for distributing tasks for street shops management based on multi-modal scheduling of the present invention, First, the street data in the target area is obtained, and the abnormal behavior of shops along the street is accurately identified using advanced recognition technology. During this process, the system analyzes multiple factors to calculate the importance parameters of each abnormal behavior, including the duration of occupancy, the flow of people affected, and the occupied area, and calculates a comprehensive processing score based on this.

[0016] Set the corresponding threshold based on the calculated processing score. For those anomalies with lower importance (i.e., the processing score is lower than the preset threshold), the system will automatically take the form of online warnings for preliminary processing. This method not only saves human resources, but also quickly conveys rectification requirements to the relevant shops. For more serious anomalies that exceed the threshold, more direct intervention measures are required. At this time, the system will automatically search and obtain the location information and identity information of law enforcement officers or managers available within a preset range near the target shop for rapid response and processing.

[0017] Then, intelligent matching is performed based on the acquired identity information and the specific abnormal behavior types of shops along the street, and a list of pending tasks for each handler is generated. If the same handler is assigned multiple tasks, these tasks will be integrated into a queue of pending tasks to achieve more efficient resource management and task scheduling. In order to further improve the efficiency of task execution, the system will also sort the queue of pending tasks based on the processing score and the moving distance required for the handler to arrive at the site to form the final task processing order. Doing so will not only ensure that the most urgent problems are solved first, but also minimize the movement costs of the handlers and improve overall work efficiency. In order to ensure information security and personal privacy, the system will desensitize and encrypt the relevant information in the task processing queue before sending it to the designated handler. This step is not only a respect for personal privacy, but also an important guarantee for the security of information transmission, ensuring that sensitive information will not be leaked during the transmission process. This has achieved effective supervision and governance of urban street shops and improved the level of intelligence and efficiency of urban management. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a flow chart of the method for dispatching tasks for street shops management based on multimodal scheduling of the present invention.

[0020] Figure 2 It is a flowchart of the present invention for acquiring street data in a target area, identifying abnormal behaviors of shops along the street, and calculating processing scores based on importance parameters of abnormal behaviors of shops along the street.

[0021] Figure 3 It is a flow chart of the present invention for identifying whether a shop has abnormal behavior based on computer recognition technology.

[0022] Figure 4 The present invention is a flow chart for calculating the flow of people when abnormal behavior occurs based on the area of ​​a crowd image in monitoring video data.

[0023] Figure 5 This is a flowchart of the present invention that sets a threshold based on the processing score, uses an online warning method if the score is below the threshold, and obtains the location information and identity information of the processing personnel within a preset range near the location of the street shops when the score exceeds the threshold.

[0024] Figure 6It is a flowchart of the present invention that matches identity information with abnormal behavior types of shops along the street in a target area, generates pending tasks for processing personnel, and generates a pending task queue by combining multiple pending tasks obtained by matching the same identity information.

[0025] Figure 7 The present invention reorders the queue of pending tasks based on priority and the shortest moving distance to obtain a flow chart of the task processing queue.

[0026] Figure 8 It is a flowchart of the present invention for improving the priority of low-priority sequence tasks based on the delay time of the low-priority sequence tasks and then re-ordering them.

[0027] Fig. 9 This is a flow chart of the present invention that desensitizes and encrypts the information in the task processing queue and then sends it to a designated person. DETAILED DESCRIPTION

[0028] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0029] First embodiment:

[0030] See also Figure 1 to Figure 9 The present invention provides a method for distributing tasks for street shops management based on multi-modal scheduling, comprising: S101 acquires street data in a target area and identifies abnormal behaviors of shops along the street, and calculates a processing score based on importance parameters of the abnormal behaviors of shops along the street, wherein the importance parameters include occupancy duration, flow of people, and occupied area; The specific steps include: S201 collects street map data in the target area based on GIS technology, wherein the street map data includes street layout and store locations; Use geographic information system technology to obtain street layout and store location information in the target area. This includes but is not limited to street names, widths, lengths, and specific addresses and business scopes of stores. GIS technology can create high-precision electronic maps, laying the foundation for subsequent data analysis.

[0031] S202 obtains surveillance video data of the street; Obtain real-time or historical video data of the street from surveillance cameras installed at key locations. These video materials are the key basis for identifying whether there are abnormal behaviors in shops along the street. Ensuring the quality of video data is crucial to improving recognition accuracy.

[0032] S203 identifies whether the store has abnormal behavior based on computer recognition technology; The specific steps include: S301 creates a labeled data set, which includes normal street views in different scenarios and samples of street-occupying business behaviors; Collect and organize a sample data set containing normal street views in different scenarios and samples of street-side business operations. This data set needs to be carefully labeled to clearly mark which situations are normal and which are abnormal behaviors.

[0033] S302 uses the labeled data set to train the random forest model to obtain an abnormal behavior recognition model; Random forest is a powerful machine learning algorithm, especially suitable for classification tasks. By training it with the above labeled dataset, an effective abnormal behavior recognition model can be built. The model is able to learn the rules for distinguishing normal and abnormal behaviors from a large number of features.

[0034] S303 performs target detection on each frame of image based on the abnormal behavior recognition model to identify abnormal behavior.

[0035] The trained model is applied to the actual surveillance video stream, and each frame is analyzed to identify any possible abnormal behavior. This process involves complex object detection technology, which aims to accurately capture every subtle violation.

[0036] S204 When there is abnormal behavior, obtain the specific start time and end time of each abnormal behavior to calculate the duration; If abnormal behavior is confirmed in step S203, the system needs to accurately record the start and end time of the behavior. Specifically, when the detection algorithm first identifies abnormal behavior such as occupying the road for business, the system will automatically mark the current time as the start time of the abnormal behavior (t start As the surveillance video continues to be analyzed, until the abnormal behavior completely disappears or is corrected, the system will record the timestamp at this time again as the end time of the abnormal behavior (t end ). The difference between these two time points is the duration of the abnormal behavior. (T), the calculation formula is as follows:

[0037] This step is critical to assessing the severity of the violation, as longer duration generally means greater impact on public space.

[0038] S205 calculates the flow of people during abnormal behavior based on the area of ​​the crowd image in the surveillance video data; In order to more comprehensively assess the impact of abnormal behavior, in addition to considering the duration, it is also necessary to estimate the density of people in the affected area. The specific steps include: S401 uses a background difference method to establish a background model to distinguish between a fixed background and a moving object; Background subtraction technology is used to process surveillance videos to distinguish fixed background from moving objects. Background subtraction is an effective computer vision method that extracts foreground targets, i.e. moving objects, by comparing the differences between consecutive frames. This method can effectively remove static background interference and focus on the dynamically changing parts.

[0039] S402 applies the YOLO recognition algorithm to recognize the moving object to obtain face information; For the identified moving objects, the YOLO object detection algorithm is further used to identify the face information in them. YOLO is particularly outstanding in real-time scenarios due to its fast and accurate characteristics. Through this process, it is possible to determine which moving objects are pedestrians, providing a basis for further analysis.

[0040] S403 defines the outline of each group of people using image segmentation technology based on face information; Next, image segmentation technology is used to accurately delineate the contour boundaries of each individual based on the previously acquired facial information. This step helps to accurately calculate the size of the space occupied by each person in the crowd, thereby improving the accuracy of subsequent crowd flow estimation.

[0041] S404 calculates the total number of pixels occupied by the crowd outline and converts it into the crowd flow.

[0042] Count the total number of pixels covered by the crowd outline and convert them into actual traffic flow values ​​according to the preset proportional relationship. For example, if you know the actual area corresponding to each unit pixel, you can use simple mathematical operations to get a rough estimate of the number of people, which reflects the density of people during abnormal behavior. First, determine the number of all pixels covered by the crowd outline in the image, denoted as P. You need to know the actual area represented by each unit pixel, which depends on factors such as image resolution, shooting distance, and camera angle. Let the actual area corresponding to each unit pixel be A pixel (e.g. square meters per pixel).

[0043] Based on actual or statistical data, estimate the average area A occupied by a person when standing. person(Also in square meters.) This value can be adjusted according to the specific application scenario. For example, the per capita floor space is smaller in crowded occasions and larger in loose occasions.

[0044] Using the above information, we can roughly estimate the number of people N in the crowd using the following formula:

[0045] Assume that in a specific scenario, after measurement, it is known that: The total number of pixels covered by the crowd outline is P = 5000 pixels; The actual area A corresponding to each unit pixel pixel =0.04m 2 / pixel (this means that each pixel represents 0.04 square meters of actual ground area); Average area per person A person =0.5m 2 ; Then, the flow of people in this scenario NN can be calculated as follows: N = 5000 × 0.04 / 0.5 = 400 This means that there are approximately 400 people in this area.

[0046] S206 calculates the actual area of ​​the road-occupied area based on the surveillance video data; In order to quantify the specific impact of abnormal behavior on public space, it is also necessary to measure the actual area occupied by the road-occupied area. With the help of computer vision technology, the size information of the road-occupied part can be extracted from the surveillance video and its area can be calculated accordingly.

[0047] First, perform necessary preprocessing on the frames of the surveillance video, such as removing noise and adjusting contrast, to improve the accuracy of subsequent processing. Use computer vision algorithms (such as edge detection, contour detection, or more advanced deep learning models) to identify the boundaries of objects or areas occupying the road. The goal of this step is to separate the occupied area of ​​interest from the background. In order to convert the distance in the image from pixels to the physical distance in the real world, you need to know the pixel representation of a known length in the image. For example, you can place a reference object of standard size within the monitoring range and then measure its pixel length in the image to establish a proportional relationship between pixels and actual length. Let this proportional factor be K, which is the actual length represented by each pixel (meters / pixel).

[0048] For the occupied area, the area can be estimated by contour approximation or polygon approximation. First, find all the contour points of the occupied area, and then use the Shoelace formula (also known as Gauss's area formula) to calculate the polygon area:

[0049] Among them, (x n ,y n ) are the vertex coordinates of the polygon.

[0050] Then convert it to actual area through the scale factor K:

[0051] S207 calculates the processing score of each violation based on the duration, the flow of people, the actual area and the corresponding weights.

[0052] Combining the above indicators, including the duration of abnormal behavior (T), the flow of people in the affected area (P), and the actual area of ​​the occupied area (A), combined with their respective importance weights (W T , W P , W A ), an overall processing score (Score) can be calculated. The formula is as follows:

[0053] The sum of all weights should be equal to 1 to ensure the standardization of the scoring system. This score can help management departments more objectively assess the severity of different violations and formulate corresponding corrective measures or punishment measures.

[0054] S102 sets a threshold based on the processing score, and uses an online warning method if the score is below the threshold; if the score exceeds the threshold, obtains the location information and identity information of the processing personnel within a preset range near the location of the street shops; The specific steps include: S501 sets the score thresholds for online warning and offline execution; A reasonable score threshold is set based on historical data, the severity of violations, and the actual needs of urban management. This threshold is used to distinguish which violations can be resolved through electronic warnings and which require on-site intervention by field personnel. This threshold is not fixed, but can be flexibly adjusted according to actual conditions (such as seasonal changes, holidays, etc.).

[0055] S502: When the processing score is lower than the score threshold, an electronic warning is sent to the relevant shop owner to explain the violation and the correction requirements; when the processing score is higher than the score threshold, the street shop is marked; If the score is below the threshold, the system will automatically generate an electronic warning containing violation details and correction suggestions, and send it to the relevant store manager via email, SMS or special application message. This warning not only describes the violation in detail, but also provides clear correction deadlines and requirements, encouraging merchants to take the initiative to correct the problem.

[0056] When the processing score is higher than the threshold, the street shops involved in the violation will be marked in the system. These marks will be clearly displayed on the map, making it easier for the processing personnel in the subsequent steps to quickly locate and go to the processing.

[0057] S503 searches for nearby processing personnel within a preset range according to the input street shop location; Once it is determined that on-site processing is required, the system will automatically search for available processing personnel within a preset distance range (e.g. 500 meters to 1 kilometer) based on the specific location of the illegal shop. This involves using GPS or other positioning technologies to accurately locate the current location of the processing personnel and assess their feasibility to reach the designated location.

[0058] S504: Obtain identity information corresponding to the processing personnel from the identity database.

[0059] For the selected handlers, the system will extract relevant information from the pre-established identity database, including but not limited to name, contact information, job responsibilities, etc. This information helps managers better coordinate task allocation and ensure that the dispatched handlers have the corresponding qualifications and capabilities to solve the problem. In addition, providing the handlers' identity information also helps to enhance the transparency and sense of responsibility of the work and promote efficient communication and cooperation.

[0060] S103 matches the identity information with the abnormal behavior types of shops along the street in the target area, generates pending tasks for the processing personnel, and generates a pending task queue by combining multiple pending tasks obtained by matching the same identity information; The specific steps include: S601 matches corresponding processing function information in a database based on abnormal behaviors of shops along the street; After identifying abnormal behaviors of shops along the street, the system will classify these abnormal behaviors (such as occupying the road for business, illegal construction, etc.), and then query the pre-established behavior-role correspondence database. This database contains information on various types of violations and what specific skills or permissions are required to handle these problems. In this way, it can ensure that each abnormal behavior is assigned to a person with corresponding capabilities to handle it.

[0061] S602 matches the identity information with the processing function information within a preset range to obtain a plurality of tasks to be processed; After determining the processing functions required for each abnormal behavior, the next step is to match these functional requirements with the identity information of nearby dispatchable processing personnel. The identity information here includes but is not limited to name, contact information, professional skills, current work status, etc. The system will select the most suitable person to perform the task based on factors such as the processing personnel's location, professional background, and current workload.

[0062] S603 generates a queue of tasks to be processed by combining multiple tasks to be processed obtained by matching the same identity information.

[0063] For each selected handler, the system will aggregate all pending tasks matched based on their identity information to form an orderly task queue. This queue takes into account the priority of the task (such as urgency), geographical location (minimizing the moving distance) and other relevant factors (such as the estimated processing time). This not only helps handlers arrange their work schedules more effectively, but also makes resource utilization more rational.

[0064] If an officer has three different tasks to handle in the same area - clearing roadblocks, investigating noise complaints and checking building permits, the system will generate an optimized task order list for them based on the urgency and geographical location of these tasks, such as dealing with roadblocks first, followed by noise complaints, and finally checking building permits.

[0065] S104 sorts the task queue to be processed based on the processing score and the moving distance to obtain a task processing queue; The specific steps include: S701 calculates the actual travel distance according to the current location of the processing personnel and the coordinates of the location where the task queue to be processed occurs; The system needs to obtain the exact current location of each handler and the specific geographic location of each task to be processed. Using GPS data or map service API, the actual driving distance from the handler's current location to each task location can be accurately calculated. Considering factors such as traffic conditions and road types, these calculations should be as accurate as possible to provide reliable data support for subsequent steps.

[0066] S702 calculates a comprehensive score for each task to be processed based on the actual driving distance and the processing score; After obtaining the actual driving distance of each task, the next step is to combine the processing score to evaluate the overall priority of the task. Here you can design a formula to calculate the comprehensive score, for example: Comprehensive score = (α×handling score) + (β×actual driving distance) Among them, α and β are the importance weights of the processing score and the actual driving distance, respectively, and α+β=1. This calculation method takes into account both the severity of the task itself and the time cost required to perform the task, thereby helping to determine the most effective task processing order. In order to combine the two parameters, the processing score and the actual driving distance are normalized for calculation.

[0067] S703: arranging all pending tasks in descending order according to the calculated comprehensive scores, and dividing them into a high priority sequence and a low priority sequence; Using the comprehensive score calculated in the previous step, sort all pending tasks by score. Generally, tasks with higher scores are classified as high priority sequences, which means they should be handled first, while tasks with lower scores form low priority sequences and can be gradually addressed as resources permit. This approach helps focus on the most important and urgent issues while not neglecting other tasks that are relatively minor but still need attention.

[0068] S704: The priorities of the tasks in the low priority sequence are increased based on the delay time of the tasks in the low priority sequence, and then the tasks are reordered.

[0069] The specific steps include: S801 sets a hysteresis coefficient for each low priority task, wherein the hysteresis coefficient increases as the task hysteresis time increases; Each task in the low priority queue is assigned a hysteresis factor. This factor is based on the length of time that the task has not been processed since it was added to the queue. The longer the task waits to be processed, the greater the hysteresis factor. For example, you can set a base hysteresis factor, and every time a certain amount of time passes (such as 24 hours), the hysteresis factor increases proportionally to reflect the increasing urgency of the task.

[0070] Example formula: Assume that the basic hysteresis coefficient is C0, and the hysteresis coefficient increases k times every T hours. Then the hysteresis coefficient C can be expressed as: C=C0+k×T. This approach ensures that problems that have not been solved for a long time can receive higher attention and be solved in a timely manner.

[0071] S802 modifies the comprehensive score based on the hysteresis coefficient; By adding the hysteresis factor to the comprehensive score, the priority of tasks that have become more urgent due to long-term non-processing can be effectively increased. The adjusted comprehensive score calculation formula is: Comprehensive score = ((α×handling score)+(β×actual driving distance)) *C.

[0072] S803 periodically reorders the entire queue of pending tasks based on the revised priority scores.

[0073] The entire queue of pending tasks will be reordered periodically (e.g., hourly or daily) based on the revised priority scores. This step ensures that even tasks that were initially considered low priority can be reordered to a higher position after a period of time due to their increased slack, thus receiving more attention and faster response.

[0074] The regular reordering mechanism helps to maintain the flexibility and adaptability of the task management system, ensuring that all tasks can be properly arranged and processed according to the latest situation and needs. In addition, this mechanism can also help identify possible system bottlenecks or unreasonable resource allocation so that corresponding adjustments can be made.

[0075] S105 desensitizes and encrypts the information in the task processing queue and sends it to the designated person.

[0076] The specific steps include: S901: grading the encryption level of each task based on the confidential information of each task in the task processing queue to obtain the encryption level of each task; Evaluate the confidential information contained in each task in the task processing queue. The goal of this stage is to determine the encryption strength required for each task, which depends on the sensitivity of the information. For example, highly sensitive information such as personal information and financial data will be assigned a higher encryption level; while public information or data that does not involve personal privacy may be assigned a lower encryption level. In this way, it can be ensured that all data has received the appropriate level of protection.

[0077] S902 encrypts and desensitizes the task using a corresponding encryption method based on the encryption level; Based on the encryption level set for each task in the previous step, the system will select the appropriate encryption algorithm and technology to encrypt the information. For example, different encryption combinations can be used: high task + high encryption, low task + high encryption, high task + low encryption, low task + low encryption.

[0078] By distinguishing the importance and encryption level of tasks, limited resources can be allocated to high-priority and high-security tasks to ensure efficient completion of key tasks; low-priority and low-encryption-level tasks can be assigned to more people or processed automatically to avoid waste of resources. In addition, high-priority tasks combined with high encryption levels can effectively prevent data leakage and ensure the security of key information; low-priority tasks will not have a serious impact on overall security even if the encryption level is low, while avoiding the waste of resources caused by excessive encryption. This classification method can be flexibly adjusted according to the nature and needs of the task and is suitable for a variety of business scenarios.

[0079] At the same time, data desensitization will be implemented in this process, which means that any information that can directly or indirectly identify an individual will be removed or blurred to further protect privacy. In this way, even if the data is intercepted during transmission, it will be difficult for unauthorized third parties to interpret the true content of this information.

[0080] S903: Send the encrypted task information to the corresponding processing personnel according to the corresponding priority sequence.

[0081] The last step is to distribute the encrypted and desensitized task information to the corresponding processing personnel according to the pre-defined priority sequence. Ensure that important tasks can get the fastest response. Each task will be accurately sent to the person with the corresponding authority. These people are usually determined in advance and have all the credentials required to decrypt and view this information. This not only improves work efficiency, but also maximizes information security.

[0082] Second embodiment: The present invention provides a task dispatching system for managing shops along the street based on multimodal scheduling, comprising a shop abnormal behavior identification module, a personnel matching module, a task queue production module, a queue priority processing module and a task dispatching module; the shop abnormal behavior identification module is used to obtain street data in a target area and identify abnormal behaviors of shops along the street, and calculate a processing score based on importance parameters of abnormal behaviors of shops along the street, wherein the importance parameters include occupation duration, flow of people and occupied area; the personnel matching module is used to set a threshold based on the processing score, and an online warning is adopted if the threshold is lower than the threshold; when the threshold is exceeded, the location information and identity information of the processing personnel within a preset range near the location of the shop along the street are obtained; the task queue production module is used to match the identity information with the abnormal behavior types of shops along the street in the target area to obtain tasks to be processed, and generate a queue of tasks to be processed by combining multiple tasks to be processed obtained by matching the same identity information; the queue priority processing module is used to sort the queue of tasks to be processed based on the processing score and the moving distance to obtain a task processing queue; the task dispatching module is used to send the information in the task processing queue to a designated person after desensitizing and encrypting it.

[0083] In this embodiment, the shop abnormal behavior identification module collects and analyzes street data in the target area to identify abnormal behaviors that may occur in shops along the street. These abnormal behaviors may include illegal occupation of public space, failure to operate at specified hours, etc. This module evaluates the importance of each abnormal behavior by calculating a comprehensive processing score, which is calculated based on multiple parameters, such as occupation duration, affected traffic flow, and occupied area. This quantitative approach helps to more accurately determine the urgency and scope of the problem.

[0084] The personnel matching module sets a threshold based on the processing score calculated above. For those situations with lower importance (i.e., the processing score is lower than the preset threshold), the system will automatically take the form of online warnings for preliminary processing. For more serious abnormal situations that exceed the threshold, the module will search for the location and identity information of law enforcement officers or managers available within a preset range around the target store. This step ensures that the most suitable person to solve the problem can be found quickly.

[0085] The task queue generation module matches the identity information determined in the above steps with the specific abnormal behavior type to generate a series of tasks to be processed. If the same person is assigned multiple tasks, these tasks will be combined into a queue of pending tasks for unified scheduling and management.

[0086] The queue priority processing module will sort the generated task queue according to the processing score and the distance that the executor needs to travel to the site to form the final task processing order. This not only takes into account the urgency of the problem, but also takes into account the effective use of resources, striving to achieve the highest processing efficiency in the shortest time.

[0087] The task dispatching module desensitizes and encrypts the information in the ordered task processing queue and sends it to the designated executor. Such measures protect personal privacy and ensure the security of information transmission.

[0088] In summary, this system has greatly improved the response speed and handling effect of urban management departments on abnormal behaviors of street shops through multi-dimensional data analysis and intelligent task allocation mechanism, and promoted the refined management and sustainable development of the city.

[0089] What is disclosed above is only a preferred embodiment of the present invention, and it certainly cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for distributing tasks for street shops management based on multimodal scheduling, characterized in that: include: Obtain street data in the target area and identify abnormal behaviors of shops along the street, and calculate processing scores based on importance parameters of abnormal behaviors of shops along the street, wherein the importance parameters include occupancy duration, traffic volume, and occupied area; A threshold is set based on the processing score. If the score is below the threshold, an online warning is used. If the score exceeds the threshold, the location and identity information of the processing personnel within a preset range near the street shops are obtained. Match the identity information with the abnormal behavior types of shops along the street in the target area, generate pending tasks for the processing personnel, and generate a pending task queue by combining multiple pending tasks obtained by matching the same identity information; Sort the task queue to be processed based on the processing score and the moving distance to obtain a task processing queue; The information in the task processing queue is desensitized and encrypted and then sent to the designated personnel.

2. The method for distributing tasks for street shops management based on multimodal scheduling as claimed in claim 1, characterized in that: The specific steps of acquiring street data in the target area, identifying abnormal behaviors of shops along the street, and calculating processing scores based on importance parameters of abnormal behaviors of shops along the street include: Collect street map data in the target area based on GIS technology, wherein the street map data includes street layout and store locations; Obtain surveillance video data of the street; Identify whether a store has abnormal behavior based on computer recognition technology; When there is abnormal behavior, obtain the specific start and end time of each abnormal behavior to calculate the duration; Calculate the flow of people during abnormal behavior based on the area of ​​crowd images in surveillance video data; Calculate the actual area of ​​the road-occupied area based on surveillance video data; The handling score of each violation is calculated based on duration, traffic volume, actual area and corresponding weights.

3. The method for distributing tasks for street shops management based on multimodal scheduling as claimed in claim 2 is characterized in that: The specific steps of identifying whether a shop has abnormal behavior based on computer recognition technology include: Create a labeled dataset that contains normal street views in different scenarios and samples of street-side business operations; Use the labeled data set to train the random forest model to obtain an abnormal behavior recognition model; Based on the abnormal behavior recognition model, target detection is performed on each frame of the image to identify abnormal behavior.

4. The method for distributing tasks for managing shops along the street based on multimodal scheduling as claimed in claim 3 is characterized in that: The specific steps of calculating the flow of people when abnormal behavior occurs based on the area of ​​the crowd image in the monitoring video data include: Use background subtraction method to build background model to distinguish fixed background and moving objects; Apply the YOLO recognition algorithm to identify moving objects and obtain face information; Based on facial information, image segmentation technology is used to define the outline of each group of people; The total number of pixels occupied by the crowd outlines is calculated and converted to obtain the crowd flow.

5. The method for distributing tasks for managing shops along the street based on multimodal scheduling as claimed in claim 4 is characterized in that: The specific steps of setting a threshold based on the processing score, using an online warning method for scores below the threshold, and obtaining the location information and identity information of the processing personnel within a preset range near the location of the street shops when the threshold is exceeded include: Set score thresholds for online warnings and offline executions; When the processing score is lower than the score threshold, an electronic warning will be sent to the relevant shop owner, explaining the violation and the correction requirements; when the processing score is higher than the score threshold, the street shop will be marked; According to the input street shop location, search for nearby processing personnel within the preset range; Obtain the identity information corresponding to the processing personnel in the identity database.

6. The method for distributing tasks for managing shops along the street based on multimodal scheduling as claimed in claim 5, characterized in that: The specific steps of matching the identity information with the abnormal behavior types of shops along the street in the target area, generating pending tasks for processing personnel, and generating a pending task queue by combining multiple pending tasks obtained by matching the same identity information include: Match the corresponding processing function information in the database based on the abnormal behavior of shops along the street; Match the identity information with the processing function information within a preset range to obtain multiple tasks to be processed; A queue of pending tasks is generated by combining multiple pending tasks matched with the same identity information.

7. The method for distributing tasks for managing shops along the street based on multimodal scheduling as claimed in claim 6, characterized in that: The specific steps of reordering the task queue to be processed based on the priority and the shortest moving distance to obtain the task processing queue include: Calculate the actual driving distance based on the coordinates of the processing personnel's current location and the location where the pending task queue occurs; Calculate a comprehensive score for each pending task based on the actual travel distance and processing score; Arrange all pending tasks in descending order according to the calculated comprehensive scores and divide them into high priority sequence and low priority sequence; The priorities of the tasks in the low-priority sequence are increased based on the delay time of the tasks in the low-priority sequence and then reordered.

8. The method for distributing tasks for managing shops along the street based on multi-modal scheduling as claimed in claim 7, characterized in that: The specific steps of increasing the priority of the tasks in the low-priority sequence based on the delay time of the tasks in the low-priority sequence and then reordering the tasks include: Setting a hysteresis coefficient for each low priority task, wherein the hysteresis coefficient increases as the task hysteresis time increases; The composite score is corrected based on the hysteresis factor; Based on the revised priority scores, the entire queue of pending tasks is periodically reordered.

9. The method for distributing tasks for managing shops along the street based on multi-modal scheduling as claimed in claim 8, characterized in that: The specific steps of desensitizing and encrypting the information in the task processing queue and sending it to the designated person include: Replace and desensitize the sensitive information of the shops along the street in the task processing queue, including names and contact information; Use the AES algorithm to encrypt the desensitized data; Use HTTPS protocol to transmit encrypted data to designated personnel through electronic channels.

10. A system for distributing tasks for managing shops along the street based on multimodal scheduling, adopting a method for distributing tasks for managing shops along the street based on multimodal scheduling as described in any one of claims 1 to 9, characterized in that: It includes the store abnormal behavior identification module, personnel matching module, task queue production module, queue priority processing module and task dispatching module; The shop abnormal behavior identification module is used to obtain street data in the target area and identify abnormal behaviors of shops along the street, and calculate processing scores based on importance parameters of abnormal behaviors of shops along the street, wherein the importance parameters include occupancy duration, traffic volume and occupied area; The personnel matching module is used to set a threshold based on the processing score, and an online warning is used if the score is below the threshold; when the score exceeds the threshold, the location information and identity information of the processing personnel within a preset range near the location of the street shops are obtained; The task queue production module is used to match the identity information with the abnormal behavior types of shops along the street in the target area to obtain tasks to be processed, and to generate a queue of tasks to be processed by combining multiple tasks to be processed obtained by matching the same identity information; The queue priority processing module is used to sort the task queue to be processed based on the processing score and the moving distance to obtain a task processing queue; The task dispatching module is used to desensitize and encrypt the information in the task processing queue and send it to the designated personnel.

Images