Method and system for monitoring environmental sanitation events in operation of outdoor unmanned sweeper

Abstract

The invention belongs to the technical field of intelligent environmental sanitation and unmanned sweeper operation quality monitoring, and discloses an environmental sanitation event monitoring method and system in outdoor unmanned sweeper operation, and the system comprises an unmanned sweeper camera terminal, a communication module, an image storage module, a garbage recognition module, an environmental sanitation event management module and a scheduling and supplementary sweeping task module. During cleaning operation, acquiring operation result images of a cleaned road section and an obstacle avoiding road section, synchronously recording longitude and latitude, time and a line number, performing garbage detection and classification after uploading, and generating an environmental sanitation event record carrying position and operation information; and the event priority is calculated according to the road section event statistics and the road section weight, a supplementary scanning task work order is generated and issued to the unmanned sweeper terminal, the event processing state is updated in combination with the supplementary scanning result image, and automatic discovery and closed-loop management of the outdoor unmanned sanitation operation result are realized.

Description

Technical Field

[0001] This invention belongs to the field of intelligent sanitation and unmanned sweeper operation quality monitoring technology, specifically a method and system for monitoring sanitation events during outdoor unmanned sweeper operations. Background Technology

[0002] With the advancement of smart city and smart sanitation construction, outdoor unmanned sweeping vehicles are gradually being used to replace or assist manual road sweeping in scenarios such as municipal roads, industrial parks, scenic spots and campuses. Existing unmanned sweeping vehicles mostly rely on autonomous driving and multi-sensor fusion to achieve route planning, automatic sweeping and obstacle avoidance. The platform can also receive structured data such as operating trajectory, operation time and equipment status. However, for whether the operation results themselves meet the standards and whether there are still garbage residues on the road after sweeping, the existing systems mostly rely on manual on-site spot checks or a small amount of fixed video monitoring, which makes it difficult to reflect the operation quality in a timely and comprehensive manner.

[0003] In existing technologies, even if some unmanned sweeping vehicles are equipped with forward or surround-view cameras, the image data is mainly used for obstacle avoidance or remote video viewing. They usually lack specific photo-taking strategies for swept sections and obstacle-avoidance sections, and they also lack the ability to automatically identify and count various types of small-target garbage such as leaves, tissues, cigarette butts, branches, cigarette boxes, cardboard boxes, aluminum cans, milk tea cups, and bottles from the images of the operation results. The platform generally cannot manage the identification results by matching them with information such as latitude and longitude, time, and route number.

[0004] Meanwhile, existing sanitation management platforms rely mainly on manual reporting and manual dispatching for incident handling and re-sweeping scheduling. They cannot automatically generate standardized sanitation incident records containing elements such as latitude and longitude, route, and project number based on the operation results of unmanned sweepers. They also cannot automatically generate re-sweeping tasks based on these records and coordinate with subsequent vehicles or manual teams. This can easily lead to delayed problem detection and re-sweeping arrangements relying on experience-based judgment, which is not conducive to improving the quality visualization and management refinement of outdoor unmanned sanitation operations. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for monitoring sanitation incidents during outdoor unmanned sweeping vehicle operations, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a sanitation event monitoring system for outdoor unmanned sweeping vehicle operations, used to collect operation result images and process them in combination with location information and operation information during the unmanned sweeping vehicle's road sweeping operation, converting the garbage identification results into structured sanitation event data, and on this basis, performing road segment analysis and task scheduling, thereby forming a closed-loop monitoring process of sanitation event discovery, analysis, re-sweeping task generation and execution feedback; The system uses an unmanned sweeper terminal as the front-end data acquisition unit. The unmanned sweeper terminal is installed on an outdoor road sweeping vehicle and is equipped with a front-facing camera and at least one set of side or rear cameras. During the sweeping operation, the unmanned sweeper terminal acquires images of the road surface based on the current operation status. At the same time, it acquires the latitude and longitude of the current location, the acquisition time, the task number, and the route number. The system associates the above location information and operation information with the corresponding operation result images to provide basic data for subsequent analysis of garbage residue by road segment and time dimension. The communication module serves as the communication channel between the vehicle and the platform. It establishes long-term connections with multiple unmanned sweeper terminals, receives image upload requests for operation results from the unmanned sweeper terminals, and includes information such as latitude and longitude, collection time, operation task number, and route number carried with the request. At the same time, it sends supplementary sweeping task instructions to the unmanned sweeper terminals. By maintaining long-term connections and a heartbeat mechanism, the communication module stably transmits the data collected by the unmanned sweeper during operation to the platform and reliably sends the control instructions generated by the platform back to the corresponding vehicles, providing a communication foundation for subsequent event analysis and task scheduling. The image storage module is connected to the communication module. After receiving the request to upload the operation result image, it generates the corresponding image upload address, receives the operation result image uploaded by the unmanned sweeper terminal, and stores the operation result image in the object storage. At the same time, the image storage module associates the image storage path with the corresponding device identifier, latitude and longitude, collection time, operation task number and route number in the data record, so that each operation result image can be associated with a specific vehicle, specific time and specific road segment, providing a traceable data source for garbage identification results and subsequent sanitation event records. The waste identification module is connected to the image storage module. It reads the operation result image from the object storage, performs waste target detection and classification on the ground area in the image, and outputs identification result data containing waste type, waste quantity and identification confidence. By centrally executing waste identification processing on the platform, images collected by different unmanned sweeper terminals can be converted into structured waste information in a unified format, so that the various types of waste residue can participate in subsequent event judgment and statistical analysis in the form of data. The sanitation event management module is connected to the waste identification module and the communication module. Based on the identification results data output by the waste identification module, as well as the latitude and longitude, collection time, route number, project number, and vehicle number of the corresponding operation result image, it generates sanitation event records according to the pre-set waste residue judgment conditions. When the number of waste categories in the identification results is greater than zero and the identification confidence level is not lower than the set threshold, the sanitation event management module judges the operation result as having waste residue and generates a sanitation event record. The sanitation event record saves at least the event number, image identifier or image storage path, waste category, waste quantity, latitude and longitude, shooting time, route number, project number, vehicle number, and event processing status, ensuring that each waste residue situation can be accurately associated with a specific image and specific operation condition, providing a basis for subsequent tracking and processing according to event status. Based on this, the sanitation event data analysis and decision-making module is connected to the sanitation event management module and the scheduling and re-sweeping task module. It reads road segment information, time information, and garbage identification information from the sanitation event records. Within a preset time range, it performs statistics and calculations on the occurrence of sanitation events on each road segment to obtain the event frequency of each road segment. Combined with the weight value configured for each road segment, it calculates factors such as garbage category, garbage quantity, identification confidence, and event frequency of each road segment to generate event priority information corresponding to each sanitation event. Through this module, it is possible to identify road segments with more concentrated garbage residue problems and road segments with higher importance in the road system from multiple road segments within the same project scope, providing a quantitative basis for the sorting of re-sweeping tasks and resource allocation. The scheduling and re-sweeping task module is connected to the sanitation event management module, the sanitation event data analysis and decision-making module, and the communication module. It generates re-sweeping task work orders based on sanitation event records and corresponding event priority information. The re-sweeping task work order records the task number, target road section information, and operation parameters. It sends the re-sweeping task instruction to the target unmanned sweeper terminal or manual operation terminal through the communication module, and at the same time receives re-sweeping task execution feedback information. By referencing event priority information in the re-sweeping task work order, the scheduling and re-sweeping task module can sort multiple pending events within the same time period, so that the order of re-sweeping task issuance is consistent with the degree of garbage residue on the road section and the importance of the road section. The data storage and caching module is connected to the communication module, image storage module, garbage recognition module, sanitation event management module, sanitation event data analysis and decision-making module, and scheduling and re-cleaning task module. It is used to centrally store basic equipment information, sanitation event records, re-cleaning task work orders, work task trajectories, and recognition result data, and maintain the online status of equipment and summary information of sanitation events within a preset time range. By managing the above data uniformly on the platform, the system can directly read the required data from the data storage and caching module during monitoring and display, and can also perform statistical analysis of historical data by road segment and by time. The monitoring and maintenance interface module is connected to the data storage and caching module. It reads the location, operation status, sanitation event records and re-sweeping task work orders of the unmanned sweeper from the data storage and caching module. The interface displays the vehicle location and operation route in map form, and displays the sanitation events and re-sweeping task execution status of different road sections in list form. It also provides the ability to view the details of individual sanitation events and configure re-sweeping tasks. Through this interface, managers can view the operation trajectory of the unmanned sweeper, the amount of garbage residue on the road, and the re-sweeping execution process in the same system. The event priority information provided by the sanitation event data analysis and decision-making module is used for manual review and scheduling decisions. Through the cooperation of the above modules, the system of the present invention establishes a closed-loop process within the same platform, from image acquisition of unmanned sweeper operation results, garbage identification and processing, sanitation event generation, event analysis based on road segment information and event frequency, to the generation, issuance and feedback of re-sweeping tasks, and re-sweeping results. This transforms the monitoring of the operation quality of outdoor unmanned sweepers from manual spot checks to continuous monitoring based on image data and event data, which is beneficial for allocating sweeping and re-sweeping operations according to road segment needs in complex road networks and multi-vehicle scenarios.

[0007] In a preferred embodiment, in order to take into account the recording of operation results under both normal cleaning conditions and obstacle avoidance conditions, the unmanned sweeper terminal is equipped with a targeted operation result image acquisition strategy. After the unmanned sweeper completes the cleaning of a section of the road along the preset cleaning route, the rear camera acquires the operation result image of the cleaned section of the road to record the road surface condition after regular cleaning. When obstacle avoidance occurs during the cleaning process, the front camera and the side or rear camera acquire the road surface images before and after the obstacle avoidance, respectively, to cover the road surface area before and after the obstacle avoidance trajectory. When the unmanned sweeper terminal acquires the above operation result images, it stores the latitude and longitude, acquisition time, operation task number and the corresponding route number one by one with the corresponding image, and reports them together when it sends an image upload request to the platform. This enables the platform to distinguish between normal cleaning areas and obstacle avoidance areas in subsequent analysis, and to determine the garbage residue and missed cleaning situations in combination with specific road segment and time information. In another preferred embodiment, to ensure the continuity of communication between the vehicle and the platform and the validity of the collected data, the communication module is equipped with connection management and heartbeat management functions. The connection management function is used to read the basic information of the device from the data storage and cache module when it receives a connection request from the unmanned sweeper terminal, and to verify the device's unique identifier, project number, and route number. A long connection is established for the device only when the device is in the registered state and its project and route match. The heartbeat management function is used to receive heartbeat messages periodically sent by the unmanned sweeper terminal after the long connection is established. When no heartbeat message is received from a certain device within a preset number of consecutive heartbeat cycles, the online status of the device is updated to offline and the corresponding long connection is closed. In this way, the platform only maintains the communication channel for sweeping vehicles in normal state, avoiding interference from abnormal devices to the analysis of sanitation events and the scheduling of supplementary sweeping tasks. When centrally managing system data, the data storage and caching module adopts a hierarchical structure. The relational database stores basic equipment information, sanitation event records, re-sweeping task orders, road segment weight configurations, and work task trajectories. It records equipment attributes, event fields, task fields, and road segment configurations in structured tables, facilitating queries and associations using event numbers, task numbers, or road segment identifiers as keys. The non-relational database stores the recognition results data corresponding to the work result images, as well as equipment status logs and operation logs. Recognition results, status changes, and operation records are saved in document form for tracing the recognition and control processes. Object storage is used for... The system organizes and stores image files of operation results and rescanning results according to project number, route number, date and equipment number, and provides image storage paths to relational databases to establish a stable reference relationship between structured records and actual image files. The cache is used to save equipment online status, summary information of sanitation events within a preset time range, road segment weight configuration and summary information of identification results. Data is obtained based on the cache during monitoring display and task calculation, reducing the access pressure on the underlying database. Through the above layered design, the system selects the appropriate storage medium according to different data types and access characteristics to achieve collaborative management of image files, event records and log information. Regarding the generation and display of sanitation incidents, the sanitation incident management module has refined the incident judgment and recording process. When the quantity of a certain type of waste in the identification result data is greater than zero, and the corresponding identification confidence level is not lower than the preset confidence threshold, the sanitation incident management module generates a sanitation incident record according to the latitude and longitude, shooting time, and route information of the operation result image, and writes the record into the sanitation incident record table in the relational database. At the same time, the sanitation incident management module writes the incident number, waste type, waste quantity, and location summary information into the cache for the monitoring and operation and maintenance interface module to read when displaying the sanitation incident list. When the monitoring and operation and maintenance interface module displays the details of a single sanitation incident, it retrieves the corresponding operation result image from the object storage according to the image storage path in the sanitation incident record, and displays the image and incident information together in the interface, so that managers can verify and handle the incident by combining the image information when viewing the incident. When conducting comprehensive analysis and ranking of sanitation incidents, the sanitation incident data analysis and decision-making module includes a road segment incident statistics unit, a road segment weight configuration unit, and a priority calculation unit. The road segment incident statistics unit counts the number of sanitation incidents occurring on the same road segment or under the same route number within a preset time range, and classifies the road segment into low-frequency, medium-frequency, and high-frequency segments based on the statistical results, reflecting the frequency of garbage residue problems on each road segment. The road segment weight configuration unit sets an integer road segment weight value for each cleaning route or road segment, reflecting the importance of that road segment in the road network. The priority calculation unit calculates the priority based on... The basic event level is determined by the garbage category, garbage quantity, and identification confidence level in the sanitation event record. Based on the low-frequency, medium-frequency, and high-frequency road segments determined by the road segment event statistics unit, coefficients corresponding to the frequency of occurrence are set. The basic event level, road segment weight value, and the coefficients are multiplied to obtain the priority value of the sanitation event, and this priority value is written into the corresponding sanitation event record. Through the above calculation method, when determining the processing order of multiple sanitation events in the same time period, the severity of the garbage itself, the importance of the road segment, and the frequency of event occurrence can be considered simultaneously, providing a unified quantitative basis for the scheduling and re-sweeping task module. Regarding the generation and execution of supplementary cleaning tasks, the scheduling and supplementary cleaning task module works in conjunction with the communication module. When generating a supplementary cleaning task work order, the scheduling and supplementary cleaning task module reads the latitude and longitude range of the target road segment from the corresponding sanitation event record and, in conjunction with the operation mode configuration, generates a supplementary cleaning task instruction including the task number, the latitude and longitude range of the target road segment, and operation power parameters. This instruction is then sent to the corresponding unmanned sweeper terminal via a long connection of the communication module. If no confirmation information containing the task number is received from the unmanned sweeper terminal within a preset time, the scheduling and supplementary cleaning task module cancels the supplementary cleaning task. The command is resent. After the number of resentments reaches the preset limit, the corresponding cleaning task work order is marked as failed to be issued, and the issuance result is written to the data storage and caching module so that it can be prompted on the monitoring interface and allowed for manual intervention. After receiving the cleaning task execution result information from the unmanned cleaning vehicle terminal, the scheduling and cleaning task module writes the cleaning task execution status and execution time into the cleaning task work order, so that the process of each cleaning task from generation and issuance to completion is fully recorded in the system and corresponds to the previous sanitation event records and subsequent cleaning result images.

[0008] In one embodiment of the present invention, the sanitation event monitoring system in outdoor unmanned sweeping vehicle operation forms a sanitation event monitoring method for the unmanned sweeping vehicle operation process through the coordinated operation of various functional modules. The method uses the unmanned sweeping vehicle terminal as the data acquisition end, and the communication module, image storage module, garbage recognition module, sanitation event management module, sanitation event data analysis and decision-making module, and scheduling and re-sweeping task module on the platform side as the processing end, and is executed according to the following process; During the connection establishment phase, after the unmanned sweeper terminal is powered on, it sends a connection request to the communication module. The request carries the device's unique identifier, project number, and line number. The communication module reads the corresponding device basic information from the data storage and caching module and verifies the device's unique identifier, project number, and line number item by item. Only when the device is in a registered state, the project number matches the project number recorded in the device's basic information, and the device status is available, a long connection is established with the unmanned sweeper terminal. After the long connection is established, the unmanned sweeper terminal sends heartbeat messages to the communication module at preset time intervals. If the communication module does not receive a heartbeat message within a preset number of consecutive heartbeat cycles, it updates the device's online status to offline and disconnects the corresponding long connection. Through the above connection management and heartbeat management mechanism, the vehicles participating in the monitoring method are limited to authorized devices in normal status, providing a stable communication foundation for subsequent data transmission and command issuance. During the data collection phase, the unmanned sweeper terminal collects images of the cleaning results based on the cleaning route and obstacle avoidance during the sweeping operation. After completing a preset cleaning section, the terminal controls the rear camera to capture images of the road surface after the section has been cleaned. When obstacle avoidance occurs, the terminal controls the front camera and the side or rear camera to capture images of the road surface before and after the obstacle avoidance. Each time an image of the cleaning results is collected, the terminal records the latitude and longitude, collection time, task number, and route number, and associates the image with the recorded information. Through this collection strategy, subsequent analysis can accurately map each image to a specific road segment and specific working condition, providing a locational and task basis for determining sanitation incidents and defining re-sweeping areas. During the image upload and storage phase, the unmanned sweeper terminal sends an image upload request to the communication module through an established long connection. The request includes the associated latitude and longitude, acquisition time, job task number, and route number. The communication module forwards the image upload request to the image storage module, which generates the corresponding image upload address and returns it to the unmanned sweeper terminal. The unmanned sweeper terminal uploads the job result image to the object storage based on the upload address. After receiving and storing the job result image, the image storage module associates the image storage path with the corresponding device identifier, latitude and longitude, acquisition time, job task number, and route number in the data record and writes it to the data storage and caching module. Through this process, the platform establishes a correspondence between image files and structured metadata, facilitating subsequent queries and image backtracking based on device, time, and road conditions. During the waste identification stage, the waste identification module reads the operation result image from the object storage, performs waste target detection and classification on the ground area in the image, obtains the waste category, waste quantity and identification confidence of each type of waste, and combines the above information with the image label to form the identification result data. The identification result data is uniformly entered into the subsequent event management stage, so that the operation result images uploaded by different vehicles are transformed into structured waste data that can participate in statistical analysis after being processed by consistent identification rules. During the sanitation event generation phase, the sanitation event management module generates sanitation event records based on the latitude and longitude, shooting time, route number, project number, and vehicle number corresponding to the identification result data and the operation result image, according to preset garbage residue judgment conditions. When the quantity of a certain type of garbage in the identification result data is greater than zero and the corresponding identification confidence is not lower than the preset confidence threshold, the sanitation event management module judges the operation result as having garbage residue, generates a sanitation event record, and writes the sanitation event record into the data storage and caching module. By centrally storing image identifiers, garbage information, and road segment information in the sanitation event record, the system manages garbage residue at the event level, rather than just recording it piecemeal for a single identification result, which facilitates subsequent processing and statistics according to the event status. During the event analysis and priority calculation phase, the sanitation event data analysis and decision-making module reads sanitation event records within a preset time range from the data storage and caching module, and performs statistics and calculations on the sanitation event situation of each road segment. Specifically, the sanitation event data analysis and decision-making module counts the number of sanitation events occurring on the same road segment or under the same route number within a preset time range, divides the road segment into low-frequency, medium-frequency, and high-frequency road segments based on the statistical results, and sets an integer road segment weight value for each cleaning route or road segment to reflect the importance of the road segment in the road network. Based on this, the sanitation event data analysis and decision-making module determines the basic event level according to the garbage category, garbage quantity, and identification confidence level in the sanitation event records. It determines the coefficient corresponding to the frequency of occurrence based on whether the road segment is divided into low-frequency, medium-frequency, or high-frequency road segments. The basic event level, road segment weight value, and the coefficient are multiplied to obtain the priority value of the sanitation event, and this priority value is written into the corresponding sanitation event record. Through the above calculation method, when sorting multiple sanitation events to be processed, the severity of garbage, the importance of the road segment, and the frequency of event occurrence in that road segment are considered simultaneously, providing a unified quantitative indicator for the generation and scheduling of subsequent cleaning tasks.

[0009] During the generation and issuance phase of the re-sweeping task, the scheduling and re-sweeping task module generates a re-sweeping task work order based on the garbage information, road segment information, and the aforementioned event priority values ​​in the sanitation event record. When generating the re-sweeping task work order, the scheduling and re-sweeping task module reads the latitude and longitude range of the target road segment from the corresponding sanitation event record, determines the operation power parameters in combination with the operation mode configuration, writes the task number, the latitude and longitude range of the target road segment, and the operation power parameters into the re-sweeping task work order, constructs the re-sweeping task instruction, and sends it to the corresponding unmanned sweeper terminal or manual operation terminal through the long connection of the communication module. At the same time, the scheduling and re-sweeping task module receives the re-sweeping task execution result information returned by the unmanned sweeper terminal or manual operation terminal, writes the re-sweeping task execution status and execution time into the re-sweeping task work order, so that the process of each re-sweeping task from generation and issuance to execution completion forms a complete record on the platform and corresponds to the corresponding sanitation event record.

[0010] In the closed-loop stage of the re-sweeping results, after the re-sweeping task is completed, the unmanned sweeping vehicle terminal collects and uploads the re-sweeping result image again. The garbage identification module detects and classifies the re-sweeping result image to obtain the garbage category, garbage quantity, and identification confidence level after re-sweeping. The sanitation event management module updates the corresponding sanitation event record based on the identification result of the re-sweeping result image and the information of the target road section to be swept, and updates the event processing status from pending status to processed status or other preset status. By associating the re-sweeping result with the original sanitation event record, this invention forms a closed-loop process within the system from operation result image acquisition, garbage identification, sanitation event generation, event analysis and priority calculation, re-sweeping task generation and distribution, to re-sweeping result confirmation and event status update. This enables the garbage residue situation during the outdoor unmanned sweeping vehicle operation to be discovered, scheduled, and verified on a sanitation event basis.

[0011] The beneficial effects of this invention are as follows: 1. This invention combines the cleaning route and obstacle avoidance actions at the terminal side of the unmanned sweeper to perform image acquisition of the operation results for the cleaned road sections and the road sections before and after obstacle avoidance. During acquisition, latitude and longitude, acquisition time, operation task number and the corresponding route number are recorded simultaneously, and uploaded to the object storage through the communication module and image storage module. Based on this, the garbage recognition module detects and classifies various small target garbage such as leaves, tissues, cigarette butts, branches, cigarette boxes, cardboard boxes, aluminum cans, milk tea cups, and bottles, generating recognition result data that corresponds one-to-one with spatial location, time and operation task. This allows the platform to directly evaluate the operation quality based on whether there is still garbage residue on the road surface, making up for the shortcomings of relying solely on structured data such as trajectory, duration and equipment status, which cannot reflect the cleaning effect.

[0012] 2. This invention establishes a sanitation event management module and a sanitation event data analysis and decision-making module on the platform side. It uniformly and solidifies the identified garbage category, quantity, corresponding latitude and longitude, shooting time, route number, project number, and vehicle number into sanitation event records. The number of events occurring within a certain time range is statistically analyzed by road segment or route, and event priority values ​​are calculated based on road segment weight configuration. The scheduling and re-sweeping task module generates re-sweeping task work orders based on these priorities, which are then sent to unmanned sweeper terminals or manual operation terminals via the communication module. The re-sweeping result images are re-identified, and the processing status is written back to the event record and work order record. This unified data link connects the closed loop of problem discovery, event formation, work order generation, and result verification, reducing reliance on manual reporting and manual dispatching.

[0013] 3. This invention adopts a microservice architecture and a long-connection communication mechanism in its system architecture, breaking down the capabilities of unmanned sweeper terminal access, operation result image uploading, garbage identification service, sanitation event management, and re-sweeping task scheduling into multiple business services. A data storage and caching module is composed of relational databases, non-relational databases, caches, and object storage, respectively carrying different types of data such as basic equipment information, sanitation event records, re-sweeping task work orders, identification result data, and equipment online status. Combined with a monitoring and maintenance interface module for unified display of map and list views, the system can maintain the stability and scalability of operation result reporting, event querying, and task scheduling even when multiple projects, multiple routes, and multiple unmanned sweepers are operating simultaneously, facilitating integration and deployment within existing sanitation information platforms. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the overall workflow of the system of the present invention; Figure 2 This is a flowchart of the sanitation event generation process of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] like Figures 1 to 2As shown, this embodiment of the invention provides a sanitation event monitoring system for outdoor unmanned sweeping vehicle operations. The system uses the unmanned sweeping vehicle terminal as the front-end acquisition unit and multiple functional modules on the platform side as processing and display units. It forms a continuous processing flow around the stages of sweeping operation result image acquisition, garbage identification, sanitation event generation and analysis, and re-sweeping task issuance.

[0017] In this embodiment, the unmanned sweeper terminal is installed on an outdoor road sweeper vehicle. The terminal is equipped with a front-facing camera and at least one set of side or rear cameras. When performing sweeping operations, the unmanned sweeper terminal collects images of the road surface operation results based on the vehicle's current location and operation status. At the same time, it acquires the latitude and longitude information of the current location, the collection time, the operation task number, and the route number. The above location information and operation information are associated and stored with the corresponding operation result images to provide basic data for subsequent analysis of garbage residue by road segment and time dimension. The communication module is located on the platform side and establishes a long connection with multiple unmanned sweeper terminals. The communication module receives image upload requests for the operation results sent by each unmanned sweeper terminal. The upload request carries information such as the latitude and longitude, collection time, operation task number and the route number. At the same time, the communication module also serves as a channel for issuing supplementary sweeping tasks. It receives supplementary sweeping task instructions generated by the platform and forwards them to the corresponding unmanned sweeper terminal or manual operation terminal, realizing two-way communication between the vehicle and the platform. The image storage module is connected to the communication module. After receiving a request to upload an image of the operation result, it generates the corresponding image upload address and returns the upload address to the unmanned sweeper terminal. The unmanned sweeper terminal sends the operation result image to the platform through the upload address. After receiving the operation result image, the image storage module stores the image file in the object storage and associates the image storage path with the corresponding device identifier, latitude and longitude, acquisition time, operation task number and the route number in the data record and writes it into the data storage and cache module, so that each image can be associated with a specific device and a specific road segment. The waste identification module is connected to the image storage module and is used to read the operation result image from the object storage, perform waste target detection and classification on the ground area in the image, identify different types of waste, and count the quantity of each type of waste and the corresponding identification confidence. The waste identification module combines the waste type, waste quantity and identification confidence with the image label to form structured identification result data, which is used by the subsequent sanitation event management and sanitation event data analysis and decision-making module. The sanitation event management module is connected to the waste identification module and the communication module respectively. It is used to convert the identification result data into manageable sanitation event records. Specifically, the sanitation event management module determines whether there is waste residue in the operation based on the identification result data output by the waste identification module, as well as the latitude and longitude, collection time, route number, project number and vehicle number of the corresponding operation result image, according to the pre-set waste residue judgment conditions. If waste residue is determined to exist, a sanitation event record is generated. The sanitation event record shall at least save the event number, image identifier or image storage path, waste type, waste quantity, latitude and longitude, shooting time, route number, project number, vehicle number and event processing status, so that the waste residue situation can be recorded and tracked at the event level. The sanitation event data analysis and decision-making module is connected to the sanitation event management module and the scheduling and re-cleaning task module. It is used to analyze the garbage situation of road sections based on sanitation event records and generate event priority information. The sanitation event data analysis and decision-making module reads the sanitation event records of each road section from the data storage and caching module, obtains the time information and garbage identification information related to the road section, calculates the sanitation event situation of each road section within a preset time range, and calculates the corresponding event priority information for each sanitation event by combining the weight value configured for the road section and factors such as garbage type, garbage quantity and identification confidence in the sanitation event. The event priority information is then written into the corresponding sanitation event record to provide a basis for the sorting of subsequent re-cleaning tasks. The scheduling and re-sweeping task module is connected to the sanitation event management module, the sanitation event data analysis and decision-making module, and the communication module. It is used to generate re-sweeping task work orders based on sanitation event data and complete task issuance and feedback recording. The scheduling and re-sweeping task module reads target road segment information, garbage information, and event priority information from the sanitation event records. It sorts the sanitation events to be processed according to the event priority information, generates a re-sweeping task work order containing task number, target road segment information, and operation parameters, and sends the re-sweeping task instruction to the corresponding unmanned sweeper terminal or manual operation terminal through the communication module. After the re-sweeping task is completed, the scheduling and re-sweeping task module receives the re-sweeping task execution feedback information from the terminal, writes the execution status and execution time in the feedback into the re-sweeping task work order, so that each re-sweeping task is associated with the previous sanitation event records and the subsequent re-sweeping results in the system. The data storage and caching module is connected to the communication module, image storage module, garbage recognition module, sanitation event management module, sanitation event data analysis and decision-making module, and scheduling and re-cleaning task module, respectively. It is used to uniformly manage various types of data generated during system operation. In the data storage and caching module, the relational database is used to store structured data such as basic equipment information, sanitation event records, re-cleaning task work orders, and operation task trajectories. The non-relational database is used to store recognition result data, equipment status logs, and operation logs. The object storage is used to store operation result images and re-cleaning result images, and establishes reference relationships with records in the relational database through image storage paths. The cache is used to save the online status of the equipment and the summary information of sanitation events and road segment weight configuration information within a preset time range, providing fast access capabilities for monitoring display and calculation processing. The monitoring and maintenance interface module is connected to the data storage and caching module to present the aforementioned data to management personnel in a visual manner. The monitoring and maintenance interface module reads the location, operation status, sanitation event records, and re-sweeping task work orders of the unmanned sweeper from the data storage and caching module. The interface displays the vehicle location and operation route in map form, and displays sanitation events and re-sweeping task information for each road segment in list form. It also provides the ability to view the details of individual sanitation events and configure re-sweeping tasks. Management personnel can centrally monitor the operation quality of the unmanned sweeper based on the sanitation event information and event priority information displayed on the interface, and adjust the re-sweeping task arrangement as needed.

[0018] In a preferred embodiment, the unmanned sweeper terminal 10 is installed on the outdoor unmanned sweeper. The terminal 10 is equipped with a front-facing camera and at least one set of side or rear cameras. Before operation, the unmanned sweeper loads a preset sweeping route according to the scheduling task. The preset sweeping route is divided into several sweeping sections during task configuration. As the vehicle travels along the preset sweeping route, the unmanned sweeper terminal 10 determines the current sweeping section based on the vehicle's position and the range of the sweeping section. When the sweeping operation of the current sweeping section is detected to be completed, the terminal 10 controls the rear-facing camera to take pictures of the road surface of the sweeping section to obtain the operation result image of the corresponding sweeping section. During the cleaning operation, when the vehicle control system detects obstacles such as fixed roadblocks or temporarily parked vehicles in front of the vehicle and issues an obstacle avoidance control signal, the unmanned sweeper enters the obstacle avoidance operation state. When the obstacle avoidance starts, the unmanned sweeper terminal 10 controls the forward camera to collect the road surface image before the obstacle avoidance, and controls the side or rear camera to collect the road surface image around the vehicle or the cleaned area before the obstacle avoidance. After the obstacle avoidance operation is completed and the vehicle returns to the preset cleaning route, the forward camera and the corresponding side or rear camera are controlled again to collect the road surface image after the obstacle avoidance, thereby forming the operation result image corresponding to the obstacle avoidance operation. During the collection of normal cleaning operation results and obstacle avoidance operation results, the unmanned sweeper terminal 10 obtains the latitude and longitude information of the current location through the vehicle positioning module each time it collects an operation result image. It records the image collection time through the internal clock of the terminal and reads the operation task number and the route number from the current operation task information. The terminal 10 associates the latitude and longitude, collection time, operation task number and the route number with the corresponding operation result image. When constructing an operation result image upload request, the terminal 10 sends the operation result image upload request to the communication module through a long connection. The upload request carries the latitude and longitude, collection time, operation task number and the route number corresponding to the operation result image. This allows the platform to associate the image with the specific cleaning section and the corresponding operation task according to the above one-to-one correspondence when receiving the operation result image.

[0019] In a preferred embodiment, the communication module 20 is located on the platform server side and is used to maintain continuous communication with multiple unmanned sweeper terminals 10. The communication module 20 includes a connection management unit and a heartbeat management unit. The connection management unit is communicatively connected to the data storage and caching module. The heartbeat management unit works in cooperation with the connection management unit to maintain the long connection status and online status of each unmanned sweeper terminal 10. Specifically, after the unmanned sweeper terminal 10 is powered on and completes local initialization, the unmanned sweeper terminal 10 sends a connection request to the communication module 20 through a preset network access method. The connection request carries the terminal's unique device identifier, project number, and line number. After receiving the connection request, the communication module 20 calls the connection management unit to read the device basic information corresponding to the device unique identifier from the data storage and cache module. The device basic information includes at least the device unique identifier, project number, line number, device registration status, and device status. The connection management unit compares the device unique identifier, project number, and line number carried in the connection request with the corresponding fields in the device basic information item by item. Only when the device is in the registered state, the project number in the connection request matches the project number recorded in the device basic information, the line number in the connection request matches the line number recorded in the device basic information, and the device status is available, is the verification considered successful. If the verification passes, the connection management unit allocates connection session resources to the unmanned sweeper terminal 10 and establishes a long connection with the terminal, binding the long connection to the device's unique identifier. At the same time, the online status of the corresponding device is updated to online in the data storage and caching module. If the verification fails, the connection management unit does not establish a long connection for the terminal, and can return an error code indicating the reason for rejection to the unmanned sweeper terminal 10, while keeping the device's online status as offline, thereby preventing unauthorized devices or devices with mismatched project numbers and line numbers from accessing the system. After a long connection is established, the heartbeat management unit begins to monitor the unmanned sweeper terminal 10 corresponding to the long connection online. The unmanned sweeper terminal 10 sends heartbeat messages to the communication module 20 at preset time intervals. The heartbeat message carries at least a unique device identifier to identify the source of the message. In one specific embodiment, the time interval can be set to thirty seconds. When the heartbeat management unit receives a heartbeat message from a certain unmanned sweeper terminal 10, it records the latest heartbeat time of the terminal and keeps the online status of the terminal online. When the heartbeat management unit does not receive a heartbeat message from a certain unmanned sweeper terminal 10 within a preset number of consecutive heartbeat cycles, it determines that the long connection of the terminal is abnormal or the terminal is offline. It updates the online status of the corresponding device in the data storage and cache module to the offline status and calls the connection management unit to close the long connection corresponding to the unmanned sweeper terminal 10, releasing the session resources occupied by the long connection. In one specific embodiment, the preset number of consecutive heartbeat cycles can be set to three heartbeat cycles. When no heartbeat message is received for three consecutive heartbeat cycles, the device is updated to the offline status and an offline alarm notification is triggered. Through the cooperation of the connection management unit and the heartbeat management unit, the communication module 20 performs joint verification of the device's unique identifier, project number, and line number before establishing a long connection, ensuring that the unmanned sweeper terminal 10 connected to the system is consistent with the project configuration and line configuration. After the long connection is established, abnormal disconnection is detected through periodic heartbeat messages, and if no heartbeat message is received within a preset number of consecutive heartbeat cycles, the terminal's online status is updated to offline status and the long connection is closed, so that subsequent work result image uploading, sanitation event recognition, and re-sweeping task distribution are processed based on the device in the online state.

[0020] In a preferred embodiment, the data storage and caching module is set as a data management unit on the platform side, which is used to perform hierarchical storage and fast access to the equipment information, operation result images, garbage identification results, sanitation event records and re-sweeping task work orders reported by the unmanned sweeper terminal. The data storage and caching module includes four parts: relational database, non-relational database, object storage and cache. Different types of storage media carry data with different characteristics and establish association relationships between the parts. The relational database is used to store highly structured business data with clear relationships. In one specific embodiment, the relational database includes a basic equipment information table, a sanitation event record table, a sweeping task work order table, a road segment weight configuration table, and a work task trajectory table. The basic equipment information table stores basic equipment information such as the unique identifier of the unmanned sweeping vehicle terminal, project number, route number, vehicle model, registration time, and project to which it belongs. The sanitation event record table stores the event number, corresponding image storage path, garbage type and quantity, collection point latitude and longitude, shooting time, route number, project number, vehicle number, event level, and event priority. The system includes information such as level values ​​and event processing status; the supplementary sweeping task work order table stores information such as work order number, associated sanitation event number, assigned vehicle or team number, task status, creation time, and completion time; the road segment weight configuration table stores information such as route number, road segment identifier, road segment weight value, and configuration effective time; and the operation task trajectory table stores structured data related to the operation trajectory, such as vehicle number, task number, trajectory point latitude and longitude, timestamp, vehicle speed, and operation status. Through foreign keys or field relationships between the above tables, the relational database can support joint queries on equipment, events, work orders, and trajectory information by project, route, vehicle, and task dimensions. Non-relational databases are used to store data with low structure or variable number of fields. In one specific embodiment, the non-relational database includes a set for storing the recognition results corresponding to the operation result images, a set for storing equipment status logs, and a set for storing operation logs. The recognition result set stores a complete recognition result document for each operation result image, which records information such as the category, location coordinates, recognition confidence, image resolution, and recognition model version number of all detected garbage targets. The equipment status log set stores real-time status data reported by the unmanned sweeper terminal through the communication module, including status records formed in time sequence such as battery level, sweeper speed, water tank level, sensor status, positioning information, and communication quality. The operation log set records operations performed by maintenance personnel in the monitoring and maintenance interface module, such as event confirmation, task scheduling, and parameter configuration, including information such as operator account, operation time, operation type, and operation object identifier. The recognition result set and the sanitation event record table are associated through image identifiers or event numbers, so that when it is necessary to view the detailed recognition results of a certain event, the complete recognition result document in the non-relational database can be located based on the event record. Object storage is used to save large, unstructured files such as operation result image files and re-sweeping result image files. In one specific embodiment, object storage organizes storage paths according to project number, route number, date, and equipment number. Operation result image files collected by the unmanned sweeper terminal and re-sweeping result image files uploaded after performing re-sweeping tasks are stored as image files in the corresponding paths. After receiving the image files uploaded by the terminal, the image upload and storage module writes the file storage path returned by object storage into the sanitation event record table in the relational database. The storage path is used as the image path field and associated with the event record. When displaying the event list, the monitoring and maintenance interface module can read the image path from the sanitation event record table and read the corresponding image from object storage for display, thereby realizing the association access between structured event data and files in object storage. The cache is used for temporary storage of data with high access frequency and relatively short update cycles to reduce the pressure of direct access to relational and non-relational databases. In one specific embodiment, the cache is used to save at least the online status of the equipment, summary information of sanitation events within a preset time range, road segment weight configuration, and summary information of identification results. The online status of the equipment can be written to the cache when the communication module completes the connection establishment or detects offline, so that the monitoring and maintenance interface module can display the online status of the equipment in real time. The summary information of sanitation events can include event number, location summary, event level, and current processing status, so as to quickly display the event list in the monitoring interface. The road segment weight configuration can be written to the cache when the configuration changes, so that the sanitation event data analysis and decision-making module can directly obtain the currently valid road segment weight value from the cache when calculating the event priority. The summary information of identification results can include the number of events and the distribution of garbage types for each project or route within a preset time range, so as to support the statistical view in the front-end interface. Through the division of labor and cooperation among relational databases, non-relational databases, object storage, and caching, the data storage and caching module carries structured data such as basic equipment information, sanitation event records, re-sweeping task work orders, road segment weight configurations, and work task trajectories within the same system; it also carries complete identification results and log-type semi-structured data; and file-type data such as work result images and re-sweeping result images. Furthermore, it stores online status and event summary information in the cache, supporting the access needs of the sanitation event management module, sanitation event data analysis and decision-making module, scheduling and re-sweeping task module, and monitoring and maintenance interface module for different types of data.

[0021] In a preferred embodiment, the sanitation event management module 50 is located on the platform server side and is connected to the waste identification module 40, the relational database, and the cache and object storage. It is used to convert the identification results into manageable sanitation event records and provide event information with associated images for front-end display. Specifically, after completing target detection and classification of the operation result image, the waste identification module 40 generates identification result data and sends it to the sanitation event management module 50. The identification result data includes at least the following fields: image identifier, corresponding equipment unique identifier, waste category list, quantity of each waste category, and identification confidence level of each category. After receiving the identification result data, the sanitation event management module 50 judges the identification result data according to the preset waste residue judgment conditions: when the quantity of any waste category in the identification result data is greater than zero and the corresponding identification confidence level is not lower than the preset confidence threshold, it is determined that there is waste residue in the operation result image. The preset confidence threshold can be set in the system configuration and can be adjusted according to different projects or different road environments. If it is determined that there is residual garbage, the sanitation event management module 50 combines the identification result data and the operation metadata associated with the operation result image to generate a sanitation event record. The operation metadata includes the latitude and longitude of the operation result image when it was collected, the shooting time, the route number, the project number, and the vehicle number. The operation metadata can be provided by the image upload and storage module 30 or the data storage and caching module 70. The sanitation event management module 50 generates a unique event number for the sanitation event, and assembles the event number, image identifier or image storage path, garbage type and quantity, latitude and longitude of the collection point, shooting time, route number, project number, vehicle number, event level, event processing status, and other fields into a structured event record, and writes the event record into a pre-established sanitation event record table in a relational database. For example, in a specific embodiment, the relational database can be implemented using a MySQL database. To improve the response speed of front-end display and statistical queries, the sanitation event management module 50, while writing complete event records to the relational database, extracts event number, garbage category, garbage quantity, and location summary information such as latitude and longitude and route number from the event records to generate event summary data, which is then written to a cache. The event summary information stored in the cache can be maintained according to a preset time range (e.g., the most recent few hours or the most recent day) to support high-frequency queries by the monitoring and maintenance interface module 90. When displaying sanitation events, the monitoring and maintenance interface module 90 can first retrieve event summary data from the cache to draw event locations on the map view and display information such as event number, waste type, quantity, and occurrence time in the list view. When maintenance personnel select to view detailed information of a specific sanitation event, the monitoring and maintenance interface module 90 reads the complete event record corresponding to that event from the relational database based on the event number or image storage path in the event summary, and initiates an access request to the object storage based on the image storage path in the event record to retrieve the corresponding job result image file from the object storage. In the front-end interface, the job result image is displayed in association with the event's waste type, quantity, location, and time information, allowing maintenance personnel to intuitively view the actual cleaning effect of the corresponding road section for that sanitation event.

[0022] In a preferred embodiment, the sanitation event data analysis and decision-making module 100 is deployed on the platform server side and is communicatively connected to the sanitation event management module 50, the scheduling and re-sweeping task module 60, and the caching and database module 70. It is used to perform statistical analysis and priority calculation on the generated sanitation event records, and to use the calculated event priorities for the generation and adjustment of re-sweeping task work orders. The sanitation event data analysis and decision-making module 100 includes a road segment event statistics unit, a road segment weight configuration unit, and a priority calculation unit. The road segment event statistics unit is used to count the number of sanitation events occurring on the same road segment or under the same route number within a preset time range, and to classify the road segments according to the statistical results. In a specific embodiment, the road segment event statistics unit retrieves event data from the sanitation event record table within a specified time range (e.g., the last seven days) from the cache and database module 70, and counts the total number N of sanitation events for each route or each road segment using the route number and road segment identifier as grouping keys. Based on the statistically obtained N, the road segment event statistics unit divides the road segments into low-frequency road segments, medium-frequency road segments, and high-frequency road segments. For example, the following classification rules can be adopted: when N is less than 3 times / week, the corresponding road segment is marked as a low-frequency road segment; when N is greater than or equal to 3 times / week and less than 7 times / week, the corresponding road segment is marked as a medium-frequency road segment; when N is greater than or equal to 7 times / week, the corresponding road segment is marked as a high-frequency road segment. The road segment event statistics unit saves the number of events and the classification results corresponding to each route or each road segment in the cache and database module 70 for subsequent reading by the priority calculation unit. The segment weight configuration unit is used to set an integer segment weight value for each cleaning route or segment. In a specific embodiment, the system configures an integer weight value W in the range of 1 to 100 for each route or segment during the deployment or operation phase. The larger the weight value, the higher the importance of the route or segment. The segment weight configuration unit writes the configured segment weight value into the segment weight configuration table in the relational database and saves the currently effective segment weight configuration in the cache. Operation and maintenance personnel can adjust the weight value of a certain route or segment through the monitoring and operation and maintenance interface module 90. After receiving the adjustment request, the segment weight configuration unit updates the corresponding record in the relational database and the cache so that the new segment weight configuration takes effect in the subsequent event priority calculation. The priority calculation unit calculates the priority value for each sanitation event record and writes the calculation result into the corresponding event record. In one specific embodiment, the priority calculation unit first reads fields such as the garbage category, garbage quantity, and identification confidence level corresponding to the event from the sanitation event record. Based on the impact of garbage type on cleaning, the size of garbage quantity, and the level of identification confidence, it determines the basic event level E corresponding to the event. The basic event level can be set to an integer in the range of 1 to 10. Subsequently, the priority calculation unit reads the segment weight value W of the corresponding route or segment from the cache and database module 70, as well as the segment classification result given by the segment event statistics unit. Based on whether the segment is divided into a low-frequency segment, a medium-frequency segment, or a high-frequency segment, it determines the frequency coefficient F. For example, the frequency coefficient F of a low-frequency segment is 1.0, the frequency coefficient F of a medium-frequency segment is 1.2, and the frequency coefficient F of a high-frequency segment is 1.5. The priority calculation unit calculates the priority value of the sanitation event according to the following formula: P=E×W×F Where P is the priority value of sanitation incidents, E is the basic incident level, W is the road segment weight value, and F is the frequency coefficient; After calculating the priority value P of a sanitation event, the priority calculation unit writes the value into the priority field of the corresponding record in the sanitation event record table of the relational database. It can also write some priority information into the cache so that the scheduling and re-sweeping task module 60 can read it quickly. During system operation, the priority calculation unit can perform the above priority calculation process on newly added or unprocessed sanitation event records according to a preset period or when a new event occurs, so as to ensure that the priority value of the event to be processed is updated as the number of events and the road conditions change. After receiving priority data from the sanitation event data analysis and decision-making module 100, the scheduling and cleaning task module 60 reads the sanitation event records to be processed from the relational database or cache. It sorts the sanitation events to be processed using the priority value P as one of the sorting criteria, and puts the sanitation events with higher priority values ​​in the order of generating and executing cleaning tasks. When generating cleaning task work orders, the scheduling and cleaning task module 60 generates or adjusts the cleaning task work orders in sequence according to the sorting results, so that the cleaning tasks corresponding to sanitation events located on high-frequency road sections or road sections with large weight values ​​and large amounts of garbage or high basic event levels are given priority for execution, so that cleaning resources are more concentrated on areas with more prominent garbage problems or more important road sections. Through the above implementation methods, the sanitation event data analysis and decision-making module 100, based on the event records provided by the sanitation event management module 50, introduces a road segment event statistics unit that counts the frequency of road segment events according to time windows, a road segment weight configuration unit that configures integer weights according to the importance of road segments, and a priority calculation unit that calculates the event priority value P based on the basic event level E, the road segment weight W, and the frequency coefficient F. This enables the system to quantitatively sort different road segments and different types of sanitation events. The scheduling and re-sweeping task module 60 uses this priority value to sort and generate re-sweeping task work orders, realizing priority re-sweeping processing for high-frequency problem road segments and high-weight road segments, which is beneficial for targeted arrangement of re-sweeping resources in unmanned sweeper operation scenarios.

[0023] In a preferred embodiment, the scheduling and cleaning task module 60 and the communication module 20 work together to convert the sanitation event priority results output by the sanitation event data analysis and decision-making module 100 into executable cleaning task work orders, and realize the issuance, confirmation and result feedback of cleaning task instructions through a long connection with the unmanned sweeper terminal 10. Specifically, when the sanitation event data analysis and decision-making module 100 determines that a certain sanitation event requires supplementary cleaning, the scheduling and supplementary cleaning task module 60 generates a supplementary cleaning task work order record in the supplementary cleaning task work order table of the relational database, assigns a unique task number to the supplementary cleaning task, and when generating the supplementary cleaning task work order, the scheduling and supplementary cleaning task module 60 determines the latitude and longitude range of the target area for supplementary cleaning based on the latitude and longitude of the collection point, the route number, and the road segment range in the sanitation event record, and selects matching operation power parameters from the preset operation parameter configuration according to the garbage type, garbage quantity, and road segment type corresponding to the sanitation event. The operation power parameters may include the sweeping brush speed parameters, the vacuum fan power parameters, and the water spray flow rate. The scheduling and supplementary cleaning task module 60 encapsulates the task number, the latitude and longitude range of the target road segment, and the operation power parameters into a supplementary cleaning task instruction, and writes the corresponding task number, target road segment information, operation parameters, and initial task status into the supplementary cleaning task work order record; After a cleaning task order is generated, the scheduling and cleaning task module 60 sends the cleaning task instruction to the communication module 20 through an interface call. The communication module 20 forwards the cleaning task instruction to the corresponding unmanned cleaning vehicle terminal 10 based on the established long connection. After receiving the cleaning task instruction, the unmanned cleaning vehicle terminal 10 parses the instruction data and returns confirmation information containing the task number to the communication module 20 to indicate that the cleaning task instruction has been successfully received. When sending the cleaning task instruction to the communication module 20, the scheduling and cleaning task module 60 starts a timeout countdown and listens for confirmation information from the corresponding unmanned cleaning vehicle terminal 10 within a preset time. When the confirmation information containing the instruction task number is received within the preset time, the scheduling and cleaning task module 60 updates the distribution status in the cleaning task order to "distribution successful". In a specific embodiment, the preset time can be set to the order of several seconds, such as ten seconds. If no confirmation information containing the task number is received from the corresponding unmanned sweeper terminal 10 within a preset time, the scheduling and resweeping task module 60 considers that the instruction may have failed and resends the resweeping task instruction. During resending, the same resweeping task instruction with the same task number is sent through the communication module 20 using the long connection between the unmanned sweeper terminal 10, and the timeout count is restarted. In a specific embodiment, the scheduling and resweeping task module 60 configures a maximum number of resends for each resweeping task instruction, such as three times. When the number of resends reaches the preset maximum and no confirmation information containing the task number is received, the issuance status of the corresponding resweeping task work order is updated to issuance failure, and the reason for issuance failure is written into the resweeping task work order record in the relational database. At the same time, the issuance result of the work order is updated in the data storage and caching module 70 so that the monitoring and maintenance interface module 90 can display it on the interface. During the process of successfully issuing the re-sweeping task instruction and executing the re-sweeping operation by the unmanned sweeping vehicle terminal 10, after completing the re-sweeping task, the terminal 10 sends the re-sweeping task execution result information to the platform through the communication module 20. The execution result information includes at least fields such as task number, execution status, and re-sweeping completion time. After receiving the re-sweeping task execution result information, the scheduling and re-sweeping task module 60 locates the corresponding work order record in the re-sweeping task work order table according to the task number, updates the re-sweeping task execution status field to execution success or execution failure, and writes the execution completion time into the work order record. In the case of execution failure, the scheduling and re-sweeping task module 60 can also write the updated re-sweeping task work order status into the data storage and cache module 70 to support the real-time display of the re-sweeping task execution status by the monitoring and maintenance interface module 90.

[0024] In a preferred embodiment, the sanitation event monitoring method for outdoor unmanned sweeping vehicle operations of the present invention is applied to the sanitation event monitoring system for outdoor unmanned sweeping vehicle operations in any of the foregoing embodiments. It is completed collaboratively by modules such as the unmanned sweeping vehicle terminal 10, communication module 20, image storage module 30, waste recognition module 40, sanitation event management module 50, sanitation event data analysis and decision-making module 100, scheduling and re-sweeping task module 60, and data storage and caching module 70. Specifically, it includes the following steps: First, during the unmanned sweeper terminal deployment phase, after the unmanned sweeper terminal 10 powers on and completes local initialization, it reads the device unique identifier, project number, and line number from its local configuration. It then sends a connection request to the communication module 20 via a pre-configured network address, carrying the device unique identifier, project number, and line number. Upon receiving the connection request, the communication module 20 reads the device basic information corresponding to the device unique identifier from the data storage and caching module 70. This basic information includes at least the device unique identifier, project number, line number, and device status. The communication module 20 verifies the device unique identifier, project number, and line number carried in the connection request against the corresponding fields in the device basic information. Only if the device recorded in the device basic information is registered... Only when the available status, project number, and line number are consistent will a long connection be established with the unmanned sweeper terminal 10. The long connection can be implemented using a bidirectional communication channel such as WebSocket. At the same time as establishing the long connection, the online status of the corresponding device is written to the data storage and cache module 70. After the long connection is established, the unmanned sweeper terminal 10 sends a heartbeat message to the communication module 20 at a preset time interval. When the communication module 20 receives the heartbeat message, it updates the latest heartbeat time of the terminal and maintains its online status. When no heartbeat message is received from a terminal within a preset number of consecutive heartbeat cycles, the communication module 20 determines that the terminal is offline, updates the status of the device in the data storage and cache module 70 to offline, and disconnects the corresponding long connection, thereby realizing connection establishment and online status maintenance. Secondly, during the cleaning operation, the unmanned sweeper terminal 10 runs on the designated route according to the cleaning task issued by the platform. It executes the operation result image acquisition strategy according to the current operation status. When the vehicle completes the cleaning of a section of the road along the preset cleaning route, the unmanned sweeper terminal 10 controls the rear-facing camera to acquire the operation result image of the cleaned section of the road. When the vehicle performs obstacle avoidance operations during the operation, such as bypassing fixed obstacles or temporarily parked vehicles on the road, the unmanned sweeper terminal 10 acquires the road image before bypassing the obstacle using the forward-facing camera, and acquires the road image after bypassing the obstacle using the side or rear-facing camera after bypassing the obstacle, so as to cover the bypassed area. For each acquired operation result image, the unmanned sweeper terminal 10 obtains the current location latitude and longitude through the vehicle positioning device, records the acquisition time through the terminal's internal clock, and reads the operation task number and the route number from the current operation task information. It then associates the operation result image with the latitude and longitude, acquisition time, operation task number, and route number locally to form an image record for uploading. Then, during the image upload phase, the unmanned sweeper terminal 10 sends an image upload request to the communication module 20 via the aforementioned long connection. The image upload request carries an image identifier and the unique device identifier, latitude and longitude, acquisition time, job task number, and route number associated with the image. The communication module 20 forwards the image upload request to the image storage module 30. The image storage module 30 generates an image upload address for object storage services based on the image upload request, such as generating an upload path or pre-signed address for MinIO object storage, and returns the upload address to the unmanned sweeper terminal 10. The unmanned sweeper terminal 10 uploads the job result image file to the object storage according to the image upload address. After confirming successful upload, the image storage module 30 writes the image storage path returned by the object storage along with the corresponding device identifier, latitude and longitude, acquisition time, job task number, and route number into the record of the data storage and cache module 70, such as writing it into the image path field of the intermediate record table or sanitation event record table in a relational database, thereby establishing a correspondence between the image file and the job metadata at the data level. Next, in the waste identification stage, the waste identification module 40 reads the corresponding operation result image from the object storage according to the image storage path provided by the image storage module 30 or the message mechanism, performs waste detection and classification on the ground area in the image, and uses a pre-trained target detection model to identify waste targets in the image, count the quantity of each waste category, and calculate the corresponding identification confidence, thereby obtaining identification result data including waste category, waste quantity, and identification confidence. The identification result data, along with image labels and equipment labels, is provided to the sanitation event management module 50, and the complete identification result document can be saved in a non-relational database; Subsequently, during the sanitation event generation stage, the sanitation event management module 50 obtains the recognition result data output by the waste recognition module 40, and reads information such as latitude and longitude, shooting time, route number, project number, and vehicle number corresponding to the operation result image from the data storage and caching module 70. It then generates a sanitation event record according to preset waste residue judgment conditions. In one specific embodiment, the preset waste residue judgment conditions can be set as follows: when the quantity of any waste category in the recognition result data is greater than zero and the recognition confidence is not lower than a preset confidence threshold, it is determined that waste residue exists, and a sanitation event needs to be generated. When the judgment conditions are met, the sanitation event management module 50 assigns an event number to the event and writes the event number, image storage path, waste category, waste quantity, recognition confidence, latitude and longitude, shooting time, route number, project number, vehicle number, and event processing status into the sanitation event record table in the data storage and caching module 70, realizing structured storage of sanitation events. If the judgment conditions are not met, the recognition result can be saved as an operation record without generating a sanitation event. Based on this, the sanitation event data analysis and decision-making module 100 periodically or when a new event occurs reads sanitation event records within a preset time range from the data storage and caching module 70, performs statistical analysis and calculations on the sanitation events on each route or section, and generates an event priority value for each sanitation event by combining the results of the aforementioned section event statistics unit and section weight configuration unit. During priority calculation, the sanitation event data analysis and decision-making module 100 determines the basic event level based on the garbage category, garbage quantity, and identification confidence level in the sanitation event records, and calculates the event priority value based on the weight configuration of the corresponding section or route and the frequency of event occurrence within the preset time range. The priority value is then written into the corresponding sanitation event record, enabling the subsequent scheduling module to sort events according to a unified priority index. During the task generation and distribution phase, the scheduling and re-sweeping task module 60 reads the sanitation event records in the pending state, comprehensively evaluates them according to the aforementioned event priority values, garbage type, road segment information, etc., and generates a re-sweeping task work order. The scheduling and re-sweeping task module 60 assigns a unique task number to each re-sweeping task work order, determines the latitude and longitude range of the target road segment based on the latitude and longitude and route information in the sanitation event records, selects the operation power parameters based on the garbage quantity and garbage type, and combines the task number, the latitude and longitude range of the target road segment, and the operation power parameters to form a re-sweeping task instruction. This instruction is sent to the corresponding unmanned sweeper terminal 10 or manual operation terminal through the long connection established by the communication module 20, and the distribution status and distribution time of the work order are recorded in the data storage and caching module 70. The scheduling and re-sweeping task module 60 simultaneously listens for the re-sweeping task execution result information returned by the unmanned sweeper terminal. Upon receiving the execution result information containing the task number, the execution status and execution time in the re-sweeping task work order are updated to the latest status. Finally, in the stage of checking the re-sweeping results and updating the closed-loop event, after completing the re-sweeping task, the unmanned sweeping vehicle terminal 10 collects the re-sweeping result image in the same way as in step two, and uploads the re-sweeping result image to the object storage through the communication module 20 and the image storage module 30 in the same way as in step three. The data storage and caching module 70 establishes the association between the re-sweeping result image and the corresponding task number and event number. The garbage identification module 40 detects and classifies the re-sweeping result image to obtain the garbage identification result data after re-sweeping. The sanitation event management module 50 judges whether the garbage at the location corresponding to the original sanitation event has been cleared based on the identification result of the re-sweeping result. When the quantity of each garbage category in the identification result is zero or lower than the preset threshold, the processing status in the corresponding sanitation event record is updated to processed, and the re-sweeping completion time and the executing vehicle information are recorded. At the same time, the re-sweeping task work order status is updated to executed and completed, thus forming a closed-loop processing flow from the collection of operation results, garbage identification, sanitation event generation, re-sweeping task issuance to the checking of re-sweeping results.

[0025] In a preferred embodiment, in step one of the aforementioned method for monitoring sanitation events during outdoor unmanned sweeper operations, the communication module is located on the platform server side and connected to the data storage and caching module. This module is used to verify basic device information and maintain online status when the unmanned sweeper terminal initiates a connection request. Specifically, after the unmanned sweeper terminal powers on, completes local initialization, and sends a connection request carrying the device's unique identifier, project number, and line number to the communication module, the communication module first queries the local cache for the device's basic information corresponding to the unique identifier. If the local cache misses the request, it queries the distributed cache for the device's basic information. If the distributed cache also misses the request, it reads the corresponding record from the device's basic information table in the relational database. The device's basic information includes at least the following fields: device unique identifier, project number, line number, device registration status, and device status. The communication module compares the device's unique identifier, project number, and line number carried in the connection request with the corresponding fields in the device's basic information. Only when the device recorded in the device's basic information table is registered, the project number in the connection request matches the project number in the device's basic information, and the device status is available, is the connection request deemed valid, and a long connection is established with the corresponding unmanned sweeper terminal. If the verification fails, the long connection is rejected, and the device's status in the data storage and caching module remains offline or inactive. After the long connection is established, the unmanned sweeper terminal sends heartbeat messages to the communication module according to a preset heartbeat cycle. The heartbeat message carries at least the device's unique identifier to identify the message source. When the module receives a heartbeat message, it updates the latest heartbeat time of the corresponding device and keeps the online status of the device online in the data storage and caching module. When no heartbeat message is received from a certain device within a preset number of consecutive heartbeat cycles, the communication module determines that the device is offline, updates the corresponding device status in the data storage and caching module to offline, disconnects the long connection corresponding to the device, and releases the occupied session resources. Through the above-mentioned verification method of local caching, distributed caching and relational database query, as well as the online status maintenance mechanism based on heartbeat messages, the configuration consistency of the unmanned sweeper terminals connected to the system in terms of project number and route number is ensured, while avoiding long-term occupation of long connection resources by terminals without heartbeat for a long time. In the same embodiment, for step six of the aforementioned sanitation event monitoring method, the sanitation event data analysis and decision-making module is used to perform road segment event statistics and event priority calculation based on sanitation event records, so as to provide a basis for subsequent supplementary cleaning task scheduling. Specifically, the sanitation event data analysis and decision-making module reads event data from the sanitation event record table from the data storage and cache module within a preset time range (e.g., the most recent week or the most recent month), groups and statistically analyzes the events according to the route number and road segment identifier, and obtains the number of times N of sanitation events occur under the same road segment or the same route number. Based on the statistically obtained N, the sanitation event data analysis and decision-making module divides each road segment into low-frequency road segments, medium-frequency road segments, and high-frequency road segments. For example, in a specific embodiment, when N is less than 3 times per week, the road segment is marked as a low-frequency road segment; when N is greater than or equal to 3 times per week and less than 7 times per week, the road segment is marked as a medium-frequency road segment; when N is greater than or equal to 7 times per week, the road segment is marked as a high-frequency road segment, and writes the above classification results into the data storage and cache module for use in priority calculation. Meanwhile, the sanitation event data analysis and decision-making module reads the corresponding road segment weight value for each cleaning route or road segment from the road segment weight configuration table or cache through the road segment weight configuration unit. The road segment weight value is in integer form and can be set to a value in the range of 1 to 100, used to represent the importance of each road segment in the overall cleaning task. For each sanitation event record in the pending state, the priority calculation unit reads fields such as garbage category, garbage quantity, and identification confidence from the sanitation event record, and calculates the priority based on the degree of impact of garbage type on the road surface environment. The quantity of garbage and the level of confidence in its identification determine the basic event level E. The basic event level can be set in the configuration as an integer range from 1 to 10. At the same time as determining the basic event level, the priority calculation unit divides the road segment to which the event belongs into a low-frequency road segment, a medium-frequency road segment, or a high-frequency road segment according to the road segment event statistics results, and selects the corresponding frequency coefficient F from the preset frequency coefficient table. For example, the frequency coefficient F of the low-frequency road segment is 1.0, the frequency coefficient F of the medium-frequency road segment is 1.2, and the frequency coefficient F of the high-frequency road segment is 1.5. After obtaining the basic event level E, road segment weight W, and frequency coefficient F, the priority calculation unit calculates the sanitation event priority value P according to the preset priority calculation formula, which can be: P = E × W × F Wherein, P is the priority value of the sanitation event, E is the basic event level, W is the road segment weight value, and F is the frequency coefficient. The priority calculation unit writes the calculated priority value P into the priority field of the corresponding sanitation event record, and can write some priority data into the cache for quick reading by the scheduling and re-sweeping task module. When generating a cleanup task order, the scheduling and cleanup task module reads the sanitation event records in the pending state from the data storage and caching module. It sorts all pending events using the priority value P in the priority field as one of the sorting criteria. During the sorting process, events belonging to high-frequency road sections or with corresponding road section weight values ​​usually have higher priority values ​​P, thus placing them at the front of the queue in the sorting results. The scheduling and cleanup task module generates cleanup task orders for events or adjusts the execution order of existing orders according to the sorting results, so that sanitation events corresponding to high-frequency problem road sections or key road sections with high weight are processed first.

[0026] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0027] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations, characterized in that: include: An unmanned sweeper terminal is installed on an outdoor road sweeper vehicle. The unmanned sweeper terminal includes a front-facing camera and at least one set of side or rear cameras. During the sweeping operation, the unmanned sweeper terminal collects images of the road surface operation results and simultaneously acquires the current location latitude and longitude, collection time, operation task number and route number. The communication module communicates with multiple unmanned sweeper terminals, establishes a long connection, receives the operation result image upload request sent by the unmanned sweeper terminal, as well as the associated latitude and longitude, acquisition time, operation task number and the route number, and sends a supplementary sweeping task instruction to the unmanned sweeper terminal. The image storage module is connected to the communication module. It generates an image upload address according to the operation result image upload request, receives the operation result image uploaded by the unmanned sweeper terminal, stores the operation result image in the object storage, and associates the image storage path with the corresponding device identifier, latitude and longitude, acquisition time, operation task number and the route number in the data record. The waste identification module is connected to the image storage module. It reads the operation result image from the object storage, performs waste target detection and classification on the ground area in the image, and generates identification result data including waste type, waste quantity and identification confidence level. The sanitation event management module is connected to the waste identification module and the communication module. Based on the identification result data and the latitude and longitude, collection time, route number, project number and vehicle number of the corresponding operation result image, it generates sanitation event records according to preset waste residue judgment conditions. The sanitation event records shall at least record the event number, image identifier or image storage path, waste category, waste quantity, latitude and longitude, shooting time, route number, project number, vehicle number and event processing status. The sanitation event data analysis and decision-making module is connected to the sanitation event management module and the scheduling and re-sweeping task module. It reads road segment, time and garbage information from the sanitation event records, performs calculations on the sanitation event situation of each road segment, and generates event priority information corresponding to each sanitation event. The scheduling and re-sweeping task module is connected to the sanitation event management module, the sanitation event data analysis and decision-making module, and the communication module. It generates a re-sweeping task work order based on the sanitation event record and the corresponding event priority information. The re-sweeping task work order records the task number, target road section information, and operation parameters. It also sends the re-sweeping task instruction to the corresponding unmanned sweeper terminal or manual operation terminal through the communication module and receives the re-sweeping task execution feedback information. The data storage and caching module is connected to the communication module, the image storage module, the garbage identification module, the sanitation event management module, the sanitation event data analysis and decision-making module, and the scheduling and re-cleaning task module. It is used to store basic equipment information, sanitation event records, re-cleaning task work orders, operation task trajectories and identification result data, and to cache the online status of the equipment and the summary information of recent sanitation events. The monitoring and maintenance interface module is connected to the data storage and caching module. It reads the location, operation status, sanitation event records and re-cleaning task work orders of the unmanned sweeper from the data storage and caching module, and displays the corresponding information in the form of a map and list in the interface. It also provides detailed viewing of sanitation events and re-cleaning task configuration operations.

2. The sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations according to claim 1, characterized in that: The image acquisition strategy for the operation results executed by the unmanned sweeper terminal includes: When a section of road is cleaned along a preset cleaning route, the system controls the rear camera to capture images of the completed cleaning process. When obstacle avoidance operations occur, the forward-facing camera and the side or rear camera are controlled to collect road surface images before and after the obstacle avoidance. The images of the operation results are associated one-to-one with the latitude and longitude, collection time, operation task number and route number of the unmanned sweeper terminal, and are reported together with the operation result image upload request.

3. The sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations according to claim 2, characterized in that: The communication module has connection management and heartbeat management functions, wherein: The connection management function is used to read basic device information from the data storage and caching module when a connection request is received from the unmanned sweeper terminal, and to verify the device unique identifier, project number and line number. When the verification is successful, a long connection is established with the unmanned sweeper terminal. The heartbeat management function is used to receive heartbeat messages periodically sent by the unmanned sweeper terminal after the long connection is established, and when no heartbeat message is received within a preset number of consecutive heartbeat cycles, the online status of the corresponding unmanned sweeper terminal is updated to offline and the corresponding long connection is closed.

4. The sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations according to claim 3, characterized in that: The data storage and caching module includes: Relational databases are used to store basic equipment information, sanitation event records, cleaning task work orders, road segment weight configurations, and work task trajectories; A non-relational database is used to store the recognition result data corresponding to the operation result image, as well as the equipment status log and operation log; Object storage is used to organize and store job result image files and rescan result image files according to project number, line number, date and equipment number, and to provide image storage paths to the relational database; Caching is used to cache the online status of devices, summary information of sanitation events within a preset time range, road segment weight configuration, and summary information of identification results.

5. The sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations according to claim 4, characterized in that: The sanitation incident management module is specifically used for: When the number of any waste category in the identification result data is greater than zero and the identification confidence is not lower than the preset confidence threshold, a sanitation event record is generated according to the latitude and longitude, shooting time and route information of the corresponding operation result image; The generated sanitation event records are written to the sanitation event record table in the relational database, and the event number, garbage category, garbage quantity, and location summary are written to the cache. When displaying sanitation events, the monitoring and maintenance interface module retrieves the corresponding operation result image from the object storage according to the image storage path in the sanitation event record, and displays it in association with the event information on the interface.

6. The sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations according to claim 5, characterized in that: The sanitation incident data analysis and decision-making module includes: The road segment event statistics unit is used to count the number of sanitation events occurring in the same road segment or under the same route number within a preset time range, and to divide the road segment into low-frequency road segments, medium-frequency road segments and high-frequency road segments based on the number of sanitation events occurring. The segment weight configuration unit is used to set an integer segment weight value for each cleaning route or segment; The priority calculation unit is used to determine the basic event level based on the garbage category, garbage quantity and identification confidence in the sanitation event record, set coefficients corresponding to the frequency of occurrence based on the low-frequency, medium-frequency and high-frequency road segments determined by the road segment event statistics unit, and multiply the basic event level, road segment weight value and the coefficients to obtain the sanitation event priority value, and write the sanitation event priority value into the corresponding sanitation event record. The scheduling and re-cleaning task module sorts the sanitation events to be processed according to the priority value of the sanitation events, and generates or adjusts the re-cleaning task work order according to the sorting result.

7. The sanitation incident monitoring system for outdoor unmanned sweeping vehicle operations according to claim 6, characterized in that: The scheduling and re-sweeping task module works in conjunction with the communication module. Specifically, the scheduling and re-sweeping task module is used for: When generating a cleaning task work order, a cleaning task instruction containing a task number, the latitude and longitude range of the target road section, and the operating power parameters is generated for the cleaning task, and sent to the corresponding unmanned cleaning vehicle terminal through the long connection of the communication module. If no confirmation information containing the task number is received from the unmanned sweeper terminal within a preset time, the resweeping task instruction is resent. After the number of resentments reaches a preset limit, the corresponding resweeping task work order is marked as a failure to be sent, and the sending result is written to the data storage and cache module. After receiving the supplementary sweeping task execution result information from the unmanned sweeping vehicle terminal, the supplementary sweeping task execution status and execution time are written into the supplementary sweeping task work order.

8. A method for monitoring sanitation incidents during outdoor unmanned sweeping vehicle operations, characterized in that: The method includes a sanitation incident monitoring system applied to the outdoor unmanned sweeper operation as described in any one of claims 1 to 7, and the specific steps are as follows: Step 1: After the unmanned sweeper terminal is powered on, it sends a connection request carrying the device's unique identifier, project number, and line number to the communication module. The communication module reads the device's basic information from the data storage and caching module, verifies the device's unique identifier, project number, and line number, and establishes a long connection with the unmanned sweeper terminal when the verification is successful. The long connection is maintained by receiving heartbeat messages. Step 2: During the cleaning operation, the unmanned cleaning vehicle terminal collects images of the operation results based on the cleaning route and obstacle avoidance, and records the latitude and longitude, collection time, task number and route number during collection, and associates the operation result images with the recorded information; Step 3: The unmanned sweeper terminal sends an image upload request to the communication module through the long connection. The communication module forwards the image upload request to the image storage module. The image storage module generates an image upload address and returns it to the unmanned sweeper terminal. The module receives the operation result image uploaded by the unmanned sweeper terminal, stores the operation result image in the object storage, and associates the image storage path with the corresponding device identifier, latitude and longitude, acquisition time, operation task number and the route number in the data record. Step four: The waste identification module reads the operation result image from the object storage, detects and classifies the ground waste in the image, and obtains the waste category, waste quantity and identification confidence level to form identification result data; Step 5: The sanitation event management module generates sanitation event records based on the latitude and longitude, shooting time, route number, project number, and vehicle number corresponding to the identification result data and the operation result image, according to the preset garbage residue judgment conditions, and writes the sanitation event records into the data storage and cache module. Step 6: The sanitation event data analysis and decision-making module performs statistics and calculations on the sanitation event situation of each road section based on the sanitation event records, and generates an event priority value for each sanitation event. Step 7: The scheduling and re-sweeping task module generates a re-sweeping task work order based on the garbage information, road section information and event priority value in the sanitation event record. It sends the re-sweeping task instruction to the unmanned sweeper terminal or the manual operation terminal through the communication module and receives the re-sweeping task execution result information. Step 8: After the re-sweeping task is completed, the unmanned sweeper terminal collects and uploads the re-sweeping result image. The garbage recognition module detects and classifies the re-sweeping result image. The sanitation event management module updates the processing status in the corresponding sanitation event record based on the recognition result of the re-sweeping result image.

9. A method for monitoring sanitation incidents during outdoor unmanned sweeping vehicle operations according to claim 8, characterized in that: In step one, after receiving the connection request from the unmanned sweeper terminal, the communication module queries the device's basic information from the local cache, distributed cache, and relational database in sequence. A long connection is established only when the device is in a registered state, the project number matches the project number recorded in the device's basic information, and the device status is available. If no heartbeat message is received for a preset number of consecutive heartbeat cycles, the device status is updated to offline and the long connection is disconnected.

10. A method for monitoring sanitation incidents during outdoor unmanned sweeping vehicle operations according to claim 9, characterized in that: In step six, the sanitation incident data analysis and decision-making module performs the following operations: Within a preset time range, the number of sanitation incidents occurring on the same road segment or under the same route number is counted, and the road segment is divided into low-frequency, medium-frequency, and high-frequency road segments based on the statistical results. Set an integer segment weight value for each cleaning route or segment; The basic event level is determined based on the garbage category, garbage quantity, and identification confidence level in the sanitation event record. The coefficient corresponding to the frequency is determined based on whether the road segment is divided into low-frequency, medium-frequency, or high-frequency road segments. The basic event level, road segment weight value, and the coefficient are multiplied to obtain the sanitation event priority value, and the sanitation event priority value is written into the corresponding sanitation event record. When generating a cleaning task order, the scheduling and cleaning task module uses the priority value of the sanitation event as one of the sorting criteria, and arranges the cleaning task orders corresponding to sanitation events that belong to high-frequency road sections or have a large road section weight value to be executed earlier.

Images