Anomaly detection method and system for express sorting based on DWS

Abstract

The present application provides a method and system for detecting anomalies in express sorting based on DWS, which relates to the field of anomaly detection technology. The method includes: obtaining basic configuration information, setting sorting index configuration index feature constraints, supervising and training an adaptive supervision model, establishing a communication connection with DWS, extracting and cross-channel interactive proofreading, identifying abnormal indicator data, combining image sensing devices for continuous response detection of equipment collaboration, and executing adaptive early warning. This application can solve the technical problems that the prior art lacks effective identification of abnormal data, lacks equipment connectivity and continuous response monitoring, resulting in low efficiency in abnormal positioning, and further affects the results of abnormal detection. It uses an adaptive supervision model for real-time monitoring and anomaly detection, and timely discovers and solves equipment failures and docking problems through equipment collaboration and continuous response monitoring, thereby improving sorting accuracy and customer satisfaction.

Description

Technical Field

[0001] The present application relates to the field of anomaly detection technology, and in particular to a DWS-based express sorting anomaly detection method and system. Background Art

[0002] With the booming development of e-commerce and the trend of globalization, the logistics industry has gained unprecedented development opportunities. The volume of express delivery business has shown explosive growth. As a key link in the logistics system, the efficiency and accuracy of the sorting system directly affect the efficiency and quality of the entire logistics process. Therefore, improving the automation and intelligence level of express sorting and reducing errors and anomalies in the sorting process have become urgent issues to be solved in the logistics industry.

[0003] At present, there are many types of anomalies that may occur during express sorting, such as damaged labels, damaged packages, sorting errors, etc. These anomalies not only affect logistics efficiency, but may also cause economic losses and reputation risks to express companies.

[0004] In summary, the existing technology lacks effective identification of abnormal data, lacks equipment connectivity and continuous response monitoring, resulting in low efficiency in anomaly positioning, which further affects the technical problems of anomaly detection results. Summary of the invention

[0005] The purpose of this application is to provide a DWS-based express sorting anomaly detection method and system to solve the technical problems that the existing technology lacks effective identification of abnormal data, lacks equipment connectivity and continuous response monitoring, resulting in low efficiency in anomaly positioning, and further affects the results of anomaly detection. An adaptive supervision model is used for real-time monitoring and anomaly detection, and through equipment collaboration and continuous response monitoring, equipment failures and docking problems are discovered and resolved in a timely manner.

[0006] In view of the above problems, the present application provides a method and system for detecting anomalies in express sorting based on DWS.

[0007] In the first aspect, the present application provides a DWS-based express sorting anomaly detection method, which is implemented by a DWS-based express sorting anomaly detection system, wherein the method includes: interacting with the basic configuration information of the target sorting center, wherein the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information; setting sorting indicators based on sorting scenarios, and configuring indicator feature constraints, wherein the indicator feature constraints have a tolerance interval based on the feature critical expansion, including a single-piece separation dimension and a sorting dimension; based on the basic configuration information, combined with the indicator feature constraints, supervised training of an adaptive supervision model, wherein the adaptive supervision model includes a parallel scanning reference channel The adaptive supervision model establishes a communication connection with the DWS system; uploads express scanning information to the DWS system, and receives sensor collection data based on detection equipment, extracts effective indicator features and cross-channel interactive proofreading based on the adaptive supervision model, and identifies abnormal indicator data, and the abnormal indicator data is identified with a unique label; based on the equipment docking information, combined with the image sensor device, performs continuous response monitoring of equipment coordination, and identifies abnormal docking information; based on the abnormal alarm system, executes adaptive early warning based on the abnormal indicator data and the abnormal docking information, wherein the abnormal alarm system includes a misclassification alarm mode and an operation early warning mode.

[0008] In the second aspect, the present application also provides a DWS-based express sorting anomaly detection system for executing the DWS-based express sorting anomaly detection method as described in the first aspect, wherein the system includes: a basic configuration information acquisition module, the basic configuration information acquisition module is used to exchange the basic configuration information of the target sorting center, the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information; a feature constraint configuration module, the feature constraint configuration module is used to set sorting indicators based on sorting scenarios, and configure indicator feature constraints, wherein the indicator feature constraints have a tolerance range based on the feature critical expansion, including a single-piece separation dimension and a sorting dimension; a supervision model training module, the supervision model training module is used to supervise the training of an adaptive supervision model based on the basic configuration information and in combination with the indicator feature constraints, and the adaptive supervision model includes parallel scanning parameters. The adaptive supervision model establishes a communication connection with the DWS system; an abnormal data identification module, which is used to upload the express scanning information to the DWS system, and receive the sensor collection data based on the detection equipment, extract the effective indicator features and cross-channel interactive proofreading based on the adaptive supervision model, and identify the abnormal indicator data, and the abnormal indicator data is identified with a unique label; an abnormal docking information identification module, which is used to perform continuous response monitoring of equipment coordination based on the equipment docking information and combined with the image sensor device, and identify the abnormal docking information; an adaptive early warning module, which is used to execute an adaptive early warning based on the abnormal indicator data and the abnormal docking information based on the abnormal alarm system, wherein the abnormal alarm system includes a misclassified alarm mode and an operation early warning mode.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] By interacting with the basic configuration information of the target sorting center, the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information; setting sorting indicators based on sorting scenarios, configuring indicator feature constraints, wherein the indicator feature constraints have a tolerance interval based on the critical expansion of the feature, including the single-piece separation dimension and the sorting dimension; based on the basic configuration information, combined with the indicator feature constraints, supervise the training of the adaptive supervision model, the adaptive supervision model includes a parallel scanning reference channel and multiple actual inspection control channels, and the adaptive supervision model establishes a communication connection with the DWS system; uploading the express scanning information to the DWS system, and receiving the sensor collection data based on the detection equipment, based on the adaptive supervision model, extracting effective indicator features and cross-channel interactive proofreading, marking The system can identify abnormal indicator data, wherein the abnormal indicator data is identified with a unique number; based on the device docking information, combined with the image sensor device, continuous response monitoring of device coordination is performed to identify abnormal docking information; based on the abnormal alarm system, an adaptive early warning based on the abnormal indicator data and the abnormal docking information is executed, wherein the abnormal alarm system includes a misclassification alarm mode and an operation early warning mode, which effectively solves the technical problems that the prior art lacks effective identification of abnormal data, lacks device connectivity and continuous response monitoring, resulting in low abnormal positioning efficiency, and further affects the abnormal detection results. The system uses an adaptive supervision model to perform real-time monitoring and abnormal detection, and through device coordination and continuous response monitoring, timely discovers and solves equipment failures and docking problems, thereby improving sorting accuracy and customer satisfaction.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0013] Figure 1 This is a flowchart of the express sorting anomaly detection method based on DWS in this application;

[0014] Figure 2This is a structural diagram of the express sorting anomaly detection system based on DWS in this application.

[0015] Description of reference numerals:

[0016] Basic configuration information acquisition module 11, feature constraint configuration module 12, supervision model training module 13, abnormal data identification module 14, abnormal docking information identification module 15, and adaptive early warning module 16. DETAILED DESCRIPTION

[0017] This application provides a DWS-based express sorting anomaly detection method and system to solve the technical problems that the existing technology lacks effective identification of abnormal data, lacks equipment connectivity and continuous response monitoring, resulting in low efficiency in anomaly positioning, which further affects the results of anomaly detection. It uses an adaptive supervision model for real-time monitoring and anomaly detection, and through equipment collaboration and continuous response monitoring, it timely discovers and solves equipment failures and docking problems, thereby improving sorting accuracy and enhancing customer satisfaction.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] Embodiment 1

[0020] Please refer to the attached Figure 1 The present application provides a method for detecting anomalies in express sorting based on DWS, wherein the method is applied to an express sorting anomaly detection system based on DWS, and the method specifically comprises the following steps:

[0021] Step 1: basic configuration information of the target sorting center is exchanged, and the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information;

[0022] Specifically, equipment assembly information includes, but is not limited to, automated sorting equipment: such as conveyor belts, scanners, robotic arms and other equipment models, quantities, installation locations and operating status. Storage equipment: such as shelves, stackers, shuttles, etc. Auxiliary equipment: such as surveillance cameras, security alarm devices, etc. Equipment docking information, data interface: data transmission interface between different devices, such as network interface, serial communication, etc. Control protocol: communication protocol for collaborative work between devices, such as communication protocol between PLC programmable logic controllers, to ensure that devices can be linked according to preset rules. Regional distribution information mainly describes the spatial layout and regional division within the sorting center. This includes, cargo storage area: storage areas divided according to the type, size, weight and other characteristics of the cargo, as well as the capacity and utilization rate of each area. Sorting operation area: including manual sorting area and automated sorting area, the operation process, equipment configuration and staffing of each area. Auxiliary areas: such as office areas, rest areas, maintenance areas, etc. The setting of these areas helps to improve employee work efficiency and job satisfaction.

[0023] Step 2: Set sorting indicators based on sorting scenarios and configure indicator feature constraints, wherein the indicator feature constraints have a tolerance range based on feature critical expansion, including single-piece separation dimension and sorting dimension;

[0024] Specifically, the sorting dimension indicators include sorting accuracy, which is expressed as the proportion of packages correctly assigned to the corresponding destination. Sorting speed, the number of packages processed by the sorting system per unit time. Determine its critical value. For example, the critical value of the separation success rate may be 98%. These critical values ​​are set according to the actual situation and business needs of the sorting center. Based on the characteristic critical value, a tolerance interval is set to deal with uncertainties and errors in the sorting process. This interval can be adjusted according to historical data, equipment performance and management requirements. Separation success rate tolerance interval, for example, based on the critical value of 98%, a tolerance interval of ±1% is set, that is, the actual separation success rate between 97% and 99% is acceptable. Sorting accuracy tolerance interval: Similarly, based on the critical value of 99%, a tolerance interval of ±0.5% is set to ensure the accuracy of sorting. In different scenarios, the corresponding expansion steps are different. For example, for peak periods, the expansion is larger and the tolerance interval is larger to improve efficiency. The relative accuracy is weakened to a certain extent. For example, anomalies that may have a greater impact on the results are detected;

[0025] Step 3: Based on the basic configuration information and in combination with the indicator feature constraints, supervise and train an adaptive supervision model, wherein the adaptive supervision model includes a parallel scanning reference channel and multiple actual inspection control channels, and the adaptive supervision model establishes a communication connection with the DWS system;

[0026] Specifically, the adaptive supervision model adopts a parallel structure, including a scanning reference channel and multiple actual inspection control channels. The scanning reference channel is responsible for generating standard sorting processes and data templates based on basic configuration information and indicator feature constraints as a reference benchmark; while the actual inspection control channel is responsible for collecting the actual operation data of the sorting center in real time. The model obtains basic configuration data such as equipment assembly information, equipment docking information and regional distribution information through the scanning reference channel, and extracts key features in combination with indicator feature constraints. These features include equipment operating status, sorting efficiency, abnormal event incidence, etc., which are used to build a standard sorting process template. The adaptive supervision model is supervised and trained using historical data and actual operation data collected in real time. By continuously adjusting model parameters and optimizing algorithms, the model can accurately identify abnormal situations in the sorting process and give corresponding warnings or processing suggestions. A stable communication connection is established between the adaptive supervision model and the DWS system through API interfaces, database connections, etc., so as to obtain data from the sorting center in real time and realize real-time data transmission and sharing.

[0027] Step 4: Upload the express scanning information to the DWS system, receive the sensor data collected by the detection equipment, extract effective indicator features and perform cross-channel interactive proofreading based on the adaptive supervision model, and identify abnormal indicator data, which is identified with a unique label;

[0028] Specifically, during the sorting process, a scanning device is used to scan each express delivery to obtain its unique identification code, destination, weight, size and other related information. The express delivery information collected by the scanning device is transmitted to the DWS system in real time through a secure network channel. After receiving the express delivery scan information, the DWS system stores it in the database and organizes and manages it according to the preset data structure. Sensors and detection equipment, such as weight sensors, dimension measuring instruments, video monitoring, etc., are installed at key locations in the sorting center to collect various data in real time during the sorting process. These detection equipment will continuously generate data, including equipment operating status, sorting efficiency, package status, etc. The DWS system cleans, filters and formats the received sensor data to ensure data quality and consistency. The adaptive supervision model extracts key indicator features related to the sorting process from the DWS system according to preset algorithms and rules, such as sorting speed, accuracy, equipment failure rate, etc. The scanning reference channel in the model interacts with the actual inspection control channel for proofreading. The scanning reference channel provides a standard data template, and the actual inspection control channel provides real-time collected data. The model verifies whether the actual sorting process meets expectations by comparing the data of these two channels. By comparing and analyzing the extracted indicator features, the adaptive supervision model can identify abnormal data that does not conform to the normal pattern. These abnormal data may include sorting errors, equipment failures, package damage, etc. For the detected abnormal indicator data, the system generates a unique label for it. This label can be based on a combination of factors such as timestamp, abnormality type, and related device ID to ensure that each abnormal data has a unique identifier. In the DWS system, the abnormal indicator data with a unique label is marked for subsequent query, analysis, and processing.

[0029] Step 5: Based on the device docking information, combined with the image sensor device, continuous response monitoring of device coordination is performed to identify abnormal docking information;

[0030] Specifically, first, collect and integrate the docking information of all relevant equipment, which includes but is not limited to the communication protocols, data transmission interfaces, control logic, etc. between the equipment. Deploy image sensing devices, such as high-definition cameras, infrared sensors, etc., near the key equipment in the sorting center to capture the operating status and docking status of the equipment in real time. These devices need to be configured according to the actual situation to ensure that clear images and accurate data can be captured. Ensure that data can be transmitted in real time and accurately by monitoring the data transmission interfaces between the equipment. If data transmission interruptions or delays are found, it is marked as abnormal docking information. Monitor the collaborative work between the equipment according to the preset control logic. If the equipment fails to respond or operate according to the preset logic, it is considered an abnormal docking. For example, slow response and docking delays result in failure to dock the front and back, docking deviation, etc.

[0031] Step six: Based on the abnormal alarm system, execute adaptive warning based on the abnormal indicator data and the abnormal connection information, wherein the abnormal alarm system includes a misclassification alarm mode and an operation warning mode.

[0032] Specifically, the misclassification alarm mode focuses on the problem of misclassification in the sorting process. When the adaptive supervision model identifies abnormal indicator data, especially indicators related to sorting accuracy, the misclassification alarm mode will be triggered. The adaptive supervision model processes and analyzes the sorting data collected in real time. Once abnormal indicator data is detected, such as the sorting error rate exceeds the preset threshold, the system will automatically enter the misclassification alarm mode. Once the misclassification alarm mode is entered, the alarm information will be sent to relevant managers and operators through sound and light alarms, SMS notifications, email reminders, etc. The alarm information includes the specific content, time and location of the abnormal indicator data, and related equipment information. After receiving the alarm information, managers and operators should respond quickly to verify and handle the abnormal indicator data. Possible measures include adjusting sorting equipment parameters, retraining operators, optimizing sorting processes, etc. The operation warning mode focuses on monitoring and warning the overall operation status of the sorting center. When the adaptive supervision model detects abnormal docking information or abnormal equipment operation status, the operation warning mode will be activated. The docking status between devices is monitored in real time through device docking information and image sensing devices. Once abnormal docking information is detected, such as device communication interruption, docking position deviation, etc., the system will automatically enter the operation warning mode.

[0033] Furthermore, the present application also includes:

[0034] Determine a stable business scenario based on the basic configuration information, and determine a standard indicator characteristic value based on the stable business scenario;

[0035] Setting a tolerance interval, and performing bidirectional expansion processing on the characteristic value of the standard indicator as the characteristic constraint of the indicator, wherein the bidirectional expansion processing is an adjustment for the sorting accuracy and sorting tolerance;

[0036] The actual sorting scenario is read, the stable business scenario is checked to determine the scenario difference, and the tolerance interval is adjusted.

[0037] Specifically, collect the basic configuration information of the sorting center, including equipment performance, personnel configuration, site layout, etc. Based on the basic configuration information, analyze and determine stable business scenarios, which are usually situations where the sorting volume is stable, the operation is standardized, and the equipment is in good condition. For stable business scenarios, determine a series of standard indicator characteristic values, which should reflect key performance such as sorting efficiency, accuracy, and fault tolerance. According to the actual needs and operating experience of the sorting center, set the initial tolerance interval, which should be able to accommodate a certain degree of indicator fluctuations. Perform bidirectional external expansion on the standard indicator characteristic values, that is, consider the adjustment of sorting accuracy and sorting fault tolerance at the same time. If higher sorting accuracy is required, the tolerance interval is appropriately narrowed; if increasing fault tolerance is considered, the tolerance interval is appropriately expanded. For example, forward expansion, that is, taking more precision as the control standard, so that even if there is a certain deviation in the actual control, it is still in a normal state. Backward expansion, that is, improving fault tolerance, if there is a certain deviation, as long as it does not reach the sorting impact, it is considered normal. In different scenarios, the corresponding tolerance intervals are different. If the business volume is small, the accuracy can be appropriately increased and the speed can be slowed down. After two-way expansion processing, the final indicator characteristic constraints are obtained, which will be used as performance standards in actual sorting operations. The system reads the actual operation data of the sorting center in real time, including sorting volume, error rate, equipment status, etc. The actual sorting scenario is compared with the stable business scenario to determine the scenario difference, that is, the difference between the actual operation data and the standard indicator characteristic value under the stable business scenario. According to the size and nature of the scenario difference, the tolerance interval is dynamically adjusted. If the actual operation data is better than the standard indicator characteristic value, you can consider narrowing the tolerance interval to improve the requirements; if the actual operation data is poor, you need to expand the tolerance interval to accommodate more fluctuations. Based on the analysis of the scenario difference and the adjustment of the tolerance interval, optimize the sorting strategy, including adjusting equipment parameters, optimizing personnel configuration, improving operating procedures, etc., to improve the overall performance of the sorting center.

[0038] Furthermore, step three of this application also includes:

[0039] Establishing a communication connection between the DWS system and the scanning reference channel;

[0040] Training a first real inspection control channel based on a first detection type, wherein the first detection type is laser point cloud data;

[0041] Training a second real inspection control channel based on a second detection type, the second detection type being weight data;

[0042] The scanning reference channel is operated in parallel with the first real inspection control channel and the second real inspection control channel to generate the adaptive supervision model, wherein a lateral connection is established between the scanning reference channel and the first real inspection control channel, and between the scanning reference channel and the second real inspection control channel.

[0043] Specifically, first, it is necessary to determine the protocol used for communication between the DWS system and the scanning reference channel, such as TCP / IP, HTTP, etc. Develop the corresponding data receiving and sending interface in the DWS system for data transmission with the scanning reference channel. At the same time, configure the scanning reference channel so that it can send data to the DWS system according to the agreed protocol and format. After the interface development and configuration are completed, a communication test is performed to ensure that the DWS system can correctly receive and parse the data sent by the scanning reference channel. Collect the point cloud data generated by the laser scanning device and perform preprocessing, such as filtering and noise reduction, to improve the data quality. Extract key features, such as object shape, size, position, etc., from the processed point cloud data. Use the extracted features to train the model of the first real inspection control channel so that it can accurately identify and analyze anomalies in the laser point cloud data. Collect the weight data of each package during the sorting process to ensure the accuracy and completeness of the data. Clean and format the collected weight data to remove outliers and erroneous data. Train the model of the second real inspection control channel based on the preprocessed weight data so that it can monitor and identify weight anomalies in real time. The scanning reference channel, the first inspection control channel, and the second inspection control channel are parallelized so that they can run and process data simultaneously. An adaptive supervision model is generated based on the standard data of the scanning reference channel and the real-time data of each inspection control channel. The model can automatically identify and warn of abnormal situations, and continuously optimize the supervision strategy based on historical data and real-time feedback. Lateral connections are established between the scanning reference channel and the first inspection control channel, and between the scanning reference channel and the second inspection control channel, so that data interaction and collaborative processing can be carried out when necessary. These lateral connections can be implemented based on event triggering or periodic synchronization.

[0044] Furthermore, the present application also includes:

[0045] Conduct big data retrieval to determine sorting record data;

[0046] Divide the sorting record data based on the detection type, and perform single-channel independent training of the scanning reference channel and the first actual inspection control channel and the second actual inspection control channel;

[0047] Based on the sorting record data, channel linkage training is performed, and verification is performed based on detection accuracy and channel consistency to obtain the adaptive supervision model that meets the convergence conditions.

[0048] Specifically, retrieve historical sorting record data from the DWS system or related database, including scanning data, equipment docking information, weight data, laser point cloud data, etc. Clean the collected data to remove duplicate, erroneous or invalid data to ensure the accuracy and completeness of the data. Perform preprocessing operations such as formatting, standardization or normalization on the data as needed to facilitate subsequent analysis and training. Classify the sorting record data according to the detection type, such as laser point cloud data, weight data, etc. For different types of detection data, extract corresponding features, such as shape features of point cloud data, distribution features of weight data, etc. For the scanning reference channel, the laser point cloud data of the first actual inspection control channel, and the weight data of the second actual inspection control channel, perform independent training of single-channel models respectively. Use appropriate optimization algorithms such as gradient descent, stochastic gradient descent, etc. to optimize the parameters of the model to improve the performance of the model. Fusion the data of different channels to form a multi-channel joint data set. Based on the multi-channel joint data set, the linkage model training of the scanning reference channel and the actual inspection control channel is carried out. During the training process, the mutual influence and synergy between different channels are considered. Use the test data set to evaluate the trained adaptive supervision model and calculate the model's detection accuracy, recall rate, F1 value and other indicators. Check whether the output results between the scan reference channel and the actual inspection control channel are consistent to ensure the accuracy and reliability of the model. According to the preset convergence conditions such as the number of training rounds, loss function value, etc., determine whether the model meets the convergence requirements. If it does, obtain the final adaptive supervision model; if not, continue to adjust the model parameters or optimize the algorithm until the convergence conditions are met.

[0049] Furthermore, step five of this application also includes:

[0050] Setting up a diversion mechanism and a docking management mechanism, wherein the diversion mechanism includes a primary diversion based on address, a secondary diversion based on courier type, and a temporary storage diversion based on address time limit, and the docking management mechanism is used for docking management between the upper computer and the lower computer;

[0051] Based on the image sensing device, continuous response monitoring of equipment coordination is performed, and monitoring data is transmitted back to a transfer file;

[0052] Based on the diversion mechanism and the docking management mechanism, the monitoring data is verified and the abnormal docking information is screened.

[0053] Specifically, use a scanning device to read the address information on the package. Match the read address information with the preset address database. According to the matching results, assign the package to the corresponding sorting area or conveyor line. Perform preliminary classification according to the destination address of the package, and classify packages with the same or similar addresses into one category for subsequent processing. After the package is diverted once, identify its express type label or information. According to the express type, assign the package to different processing areas or priority queues again. Further classify according to the type of express, such as ordinary items, expedited items, special items, etc., to meet the needs of different customers. Analyze the time limit requirements in the address database. For addresses with specific time limit requirements, set up corresponding temporary storage areas. According to the destination address and current time of the package, determine whether temporary storage is required and perform corresponding operations. For some addresses with special time limit requirements, temporarily store the package in a specific area to ensure on-time delivery. Formulate a unified data transmission protocol and interface standard. Establish a stable communication connection to ensure real-time data transmission and rapid response to instructions. Design a verification mechanism to verify the transmitted data to ensure the accuracy and integrity of the data. Ensure that information transmission and instruction execution between upper computers such as control centers and lower computers such as sorting equipment and conveyor lines are accurate. Use image sensors deployed in the sorting center to monitor the operating status and docking status of the equipment in real time. Extract information such as the movement trajectory and docking position of the equipment through image processing and analysis technology. Send the monitoring data back to the transfer file or database in real time for subsequent analysis and processing. Verify the monitoring data in combination with the requirements of the diversion mechanism and docking management mechanism. Check whether the operating status of the equipment is consistent with the preset diversion rules and docking requirements. Identify any abnormal status or non-compliant behavior, such as equipment failure, docking error, etc. Filter out abnormal docking information based on the verification results. Record the abnormal information in detail, including the time, location, and equipment involved. Classify and process the abnormal information according to the preset processing strategy, such as sending alarm notifications and automatically adjusting equipment parameters.

[0054] Furthermore, the present application also includes:

[0055] Collecting separation data based on a single-piece separation device, wherein the single-piece separation device is arranged at the front end of the sorting device;

[0056] Identify the separation data, perform automatic separation detection on the express, and determine the separation status of the individual items, wherein the automatic separation standard at least includes overlap determination, adjacent item spacing determination, and automatic sorting determination;

[0057] Based on the single-piece separation state, an alarm is given based on the abnormal alarm system.

[0058] Specifically, a single-piece separation device is deployed at the front end of the sorting device, and the single-piece separation device is initialized and calibrated to ensure that the device is in good condition and the parameters are accurate. The single-piece separation device is started and separation data is collected. These data may include separation time, separation success rate, package size, weight and other information. The collected separation data is transmitted to the data processing center in real time and stored for subsequent analysis and use. The collected separation data is preprocessed by cleaning, denoising and formatting to improve data quality. Features related to automatic separation of express delivery are extracted from the preprocessed data, such as the spacing between packages, the degree of overlap, and the sorting situation. Based on the extracted features, the preset automatic separation standards including overlap judgment, adjacent piece spacing judgment and automatic sorting judgment are applied to perform automatic separation detection of express delivery. The results of the automatic separation detection are output, including which packages meet the separation standards, which do not meet the standards, and the reasons for non-compliance. According to the results of the automatic separation detection, it is determined whether the single-piece separation state is normal. If abnormal conditions such as overlap, insufficient spacing or sorting errors are detected, it is determined to be an abnormal state. When it is determined to be an abnormal state, the abnormal alarm system is triggered. This can be achieved by sending alarm signals, displaying alarm information, or triggering sound and light alarms. Detailed alarm information is displayed in the alarm system, including the type of abnormality, time of occurrence, location, etc., so that operators can quickly locate and handle the abnormality.

[0059] Furthermore, the present application also includes:

[0060] Mark the express numbers based on single piece separation anomalies and misclassification, and perform re-sorting operations based on the return line;

[0061] Sort the express labels to be re-picked and configure the insertion nodes;

[0062] Based on the insertion node, reflux re-sorting is performed, and courier reflux and insertion monitoring of re-sorting is performed simultaneously.

[0063] Specifically, based on the single-piece separation status detection and automatic separation results, the express with abnormal or missorted single-piece separation is identified. The identified abnormal or missorted express is marked with a number for subsequent processing and tracking. Ensure that the return line is unobstructed and prepare the equipment and personnel required for re-sorting. Import the express marked as abnormal or missorted into the return line for re-sorting. Re-sort the express on the return line and re-check its separation status and sorting accuracy. Sort according to the characteristics of the express label such as destination, priority, etc. to optimize the re-sorting process. Configure an insertion node at the appropriate position of the return line to insert the express to be re-picked into the main sorting line. Re-sort the express on the return line at the insertion node to ensure its sorting accuracy. Simultaneously monitor the process of the express after re-sorting returning to the main sorting line, and track its insertion at the insertion node in real time. According to the monitoring results, timely feedback on the status of re-picking and return insertion, and adjust and optimize the process in time if there are any problems.

[0064] In summary, the express sorting anomaly detection method based on DWS provided in this application has the following technical effects:

[0065] By interacting with the basic configuration information of the target sorting center, the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information; setting sorting indicators based on sorting scenarios, configuring indicator feature constraints, wherein the indicator feature constraints have a tolerance interval based on the critical expansion of the feature, including the single-piece separation dimension and the sorting dimension; based on the basic configuration information, combined with the indicator feature constraints, supervise the training of the adaptive supervision model, the adaptive supervision model includes a parallel scanning reference channel and multiple actual inspection control channels, and the adaptive supervision model establishes a communication connection with the DWS system; uploading the express scanning information to the DWS system, and receiving the sensor collection data based on the detection equipment, based on the adaptive supervision model, extracting effective indicator features and cross-channel interactive proofreading, marking The system can identify abnormal indicator data, wherein the abnormal indicator data is identified with a unique number; based on the device docking information, combined with the image sensor device, continuous response monitoring of device coordination is performed to identify abnormal docking information; based on the abnormal alarm system, an adaptive early warning based on the abnormal indicator data and the abnormal docking information is executed, wherein the abnormal alarm system includes a misclassification alarm mode and an operation early warning mode, which effectively solves the technical problems that the prior art lacks effective identification of abnormal data, lacks device connectivity and continuous response monitoring, resulting in low abnormal positioning efficiency, and further affects the abnormal detection results. The system uses an adaptive supervision model to perform real-time monitoring and abnormal detection, and through device coordination and continuous response monitoring, timely discovers and solves equipment failures and docking problems, thereby improving sorting accuracy and customer satisfaction.

[0066] Embodiment 2

[0067] Based on the same inventive concept as the express sorting anomaly detection method based on DWS in the aforementioned embodiment, the present application also provides an express sorting anomaly detection system based on DWS, please refer to the attached Figure 2 , the system comprising:

[0068] A basic configuration information acquisition module 11, which is used to exchange basic configuration information of a target sorting center, wherein the basic configuration information includes equipment assembly information, equipment docking information, and regional distribution information;

[0069] A feature constraint configuration module 12, the feature constraint configuration module 12 is used to set sorting indicators based on sorting scenarios and configure indicator feature constraints, wherein the indicator feature constraints have a tolerance interval based on feature critical expansion, including a single piece separation dimension and a sorting dimension;

[0070] A supervision model training module 13, wherein the supervision model training module 13 is used to supervise and train an adaptive supervision model based on the basic configuration information and the indicator feature constraints, wherein the adaptive supervision model includes a parallel scanning reference channel and a plurality of actual inspection control channels, and wherein a communication connection is established between the adaptive supervision model and the DWS system;

[0071] An abnormal data identification module 14, which is used to upload express scanning information to the DWS system, receive sensor-collected data based on detection equipment, extract effective indicator features and perform cross-channel interactive proofreading based on the adaptive supervision model, and identify abnormal indicator data, wherein the abnormal indicator data is identified with a unique label;

[0072] An abnormal docking information identification module 15, the abnormal docking information identification module 15 is used to perform continuous response monitoring of device coordination based on the device docking information in combination with an image sensor device, and identify abnormal docking information;

[0073] The adaptive warning module 16 is used to perform adaptive warning based on the abnormal indicator data and the abnormal connection information based on the abnormal alarm system, wherein the abnormal alarm system includes a misclassified alarm mode and an operation warning mode.

[0074] Furthermore, the system further includes a tolerance interval setting module, and the tolerance interval setting module is used to:

[0075] Determine a stable business scenario based on the basic configuration information, and determine a standard indicator characteristic value based on the stable business scenario;

[0076] Setting a tolerance interval, and performing bidirectional expansion processing on the characteristic value of the standard indicator as the characteristic constraint of the indicator, wherein the bidirectional expansion processing is an adjustment for the sorting accuracy and sorting tolerance;

[0077] The actual sorting scenario is read, the stable business scenario is checked to determine the scenario difference, and the tolerance interval is adjusted.

[0078] Furthermore, the supervision model training module 13 in the system is also used for:

[0079] Establishing a communication connection between the DWS system and the scanning reference channel;

[0080] Training a first real inspection control channel based on a first detection type, wherein the first detection type is laser point cloud data;

[0081] Training a second real inspection control channel based on a second detection type, the second detection type being weight data;

[0082] The scanning reference channel is operated in parallel with the first real inspection control channel and the second real inspection control channel to generate the adaptive supervision model, wherein a lateral connection is established between the scanning reference channel and the first real inspection control channel, and between the scanning reference channel and the second real inspection control channel.

[0083] Furthermore, the system further comprises a linkage training module, wherein the linkage training module is used for:

[0084] Conduct big data retrieval to determine sorting record data;

[0085] Divide the sorting record data based on the detection type, and perform single-channel independent training of the scanning reference channel and the first actual inspection control channel and the second actual inspection control channel;

[0086] Based on the sorting record data, channel linkage training is performed, and verification is performed based on detection accuracy and channel consistency to obtain the adaptive supervision model that meets the convergence conditions.

[0087] Furthermore, the abnormal connection information identification module 15 in the system is also used for:

[0088] Setting up a diversion mechanism and a docking management mechanism, wherein the diversion mechanism includes a primary diversion based on address, a secondary diversion based on courier type, and a temporary storage diversion based on address time limit, and the docking management mechanism is used for docking management between the upper computer and the lower computer;

[0089] Based on the image sensing device, continuous response monitoring of equipment coordination is performed, and monitoring data is transmitted back to a transfer file;

[0090] Based on the diversion mechanism and the docking management mechanism, the monitoring data is verified and the abnormal docking information is screened.

[0091] Furthermore, the system further comprises a separation data acquisition module, wherein the separation data acquisition module is used for:

[0092] Collecting separation data based on a single-piece separation device, wherein the single-piece separation device is arranged at the front end of the sorting device;

[0093] Identify the separation data, perform automatic separation detection on the express, and determine the separation status of the individual items, wherein the automatic separation standard at least includes overlap determination, adjacent item spacing determination, and automatic sorting determination;

[0094] Based on the single-piece separation state, an alarm is given based on the abnormal alarm system.

[0095] Furthermore, the system further comprises an insertion node configuration module, wherein the insertion node configuration module is used to:

[0096] Mark the express numbers based on single piece separation anomalies and misclassification, and perform re-sorting operations based on the return line;

[0097] Sort the express labels to be re-picked and configure the insertion nodes;

[0098] Based on the insertion node, reflux re-sorting is performed, and courier reflux and insertion monitoring of re-sorting is performed simultaneously.

[0099] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The DWS-based express sorting anomaly detection method and specific examples in Example 1 are also applicable to the DWS-based express sorting anomaly detection system in this embodiment. Through the above detailed description of the DWS-based express sorting anomaly detection method, those skilled in the art can clearly understand the DWS-based express sorting anomaly detection system in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0100] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0101] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. The express sorting anomaly detection method based on DWS is characterized by: The method further comprises: Basic configuration information of the interactive target sorting center, wherein the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information; Set sorting indicators based on sorting scenarios and configure indicator feature constraints, wherein the indicator feature constraints have a tolerance range based on feature critical expansion, including single-piece separation dimension and sorting dimension; Based on the basic configuration information and in combination with the indicator feature constraints, supervise and train an adaptive supervision model, wherein the adaptive supervision model includes a parallel scanning reference channel and a plurality of actual inspection control channels, and the adaptive supervision model establishes a communication connection with the DWS system; Upload the express scanning information to the DWS system, receive the sensor data collected by the detection equipment, extract the effective indicator features and perform cross-channel interactive proofreading based on the adaptive supervision model, and identify the abnormal indicator data, and the abnormal indicator data is identified with a unique label; Based on the device docking information, combined with the image sensing device, continuous response monitoring of device coordination is performed to identify abnormal docking information; Based on the abnormal alarm system, an adaptive early warning based on the abnormal indicator data and the abnormal docking information is executed, wherein the abnormal alarm system includes a misclassification alarm mode and an operation early warning mode; The indicator feature constraint includes a single-piece separation dimension, and the method further includes: Collecting separation data based on a single-piece separation device, wherein the single-piece separation device is arranged at the front end of the sorting device; Identify the separation data, perform automatic separation detection on the express, and determine the separation status of the individual items, wherein the automatic separation standard at least includes overlap determination, adjacent item spacing determination, and automatic sorting determination; Based on the single-piece separation state, an alarm is given based on the abnormal alarm system.

2. The method according to claim 1, characterized in that The method further comprises: Determine a stable business scenario based on the basic configuration information, and determine a standard indicator characteristic value based on the stable business scenario; Setting a tolerance interval, and performing bidirectional expansion processing on the characteristic value of the standard indicator as the characteristic constraint of the indicator, wherein the bidirectional expansion processing is an adjustment for the sorting accuracy and sorting tolerance; The actual sorting scenario is read, the stable business scenario is checked to determine the scenario difference, and the tolerance interval is adjusted.

3. The method according to claim 1, characterized in that Supervising training of an adaptive supervision model, the method further comprising: Establishing a communication connection between the DWS system and the scanning reference channel; Training a first real inspection control channel based on a first detection type, wherein the first detection type is laser point cloud data; Training a second real inspection control channel based on a second detection type, the second detection type being weight data; The scanning reference channel is operated in parallel with the first real inspection control channel and the second real inspection control channel to generate the adaptive supervision model, wherein a lateral connection is established between the scanning reference channel and the first real inspection control channel, and between the scanning reference channel and the second real inspection control channel.

4. The method according to claim 3, characterized in that The method further comprises: Conduct big data retrieval to determine sorting record data; Divide the sorting record data based on the detection type, and perform single-channel independent training of the scanning reference channel and the first actual inspection control channel and the second actual inspection control channel; Based on the sorting record data, channel linkage training is performed, and verification is performed based on detection accuracy and channel consistency to obtain the adaptive supervision model that meets the convergence conditions.

5. The method according to claim 1, characterized in that The method of performing continuous response monitoring of device coordination further comprises: Setting up a diversion mechanism and a docking management mechanism, wherein the diversion mechanism includes a primary diversion based on address, a secondary diversion based on courier type, and a temporary storage diversion based on address time limit, and the docking management mechanism is used for docking management between the upper computer and the lower computer; Based on the image sensing device, continuous response monitoring of equipment coordination is performed, and monitoring data is transmitted back to a transfer file; Based on the diversion mechanism and the docking management mechanism, the monitoring data is verified and the abnormal docking information is screened.

6. The method according to claim 1, characterized in that The method further comprises: Mark the express numbers based on single piece separation anomalies and misclassification, and perform re-sorting operations based on the return line; Sort the express labels to be re-picked and configure the insertion nodes; Based on the insertion node, reflux re-sorting is performed, and courier reflux and insertion monitoring of re-sorting is performed simultaneously.

7. The express sorting anomaly detection system based on DWS is characterized by: For implementing the steps of the method according to any one of claims 1 to 6, the system comprises: A basic configuration information acquisition module, which is used to exchange basic configuration information of the target sorting center, and the basic configuration information includes equipment assembly information, equipment docking information and regional distribution information; A feature constraint configuration module, the feature constraint configuration module is used to set sorting indicators based on sorting scenarios and configure indicator feature constraints, wherein the indicator feature constraints have a tolerance range based on feature critical expansion, including single-piece separation dimensions and sorting dimensions; A supervision model training module, wherein the supervision model training module is used to supervise and train an adaptive supervision model based on the basic configuration information and the indicator feature constraints, wherein the adaptive supervision model includes a parallel scanning reference channel and a plurality of actual inspection control channels, and wherein a communication connection is established between the adaptive supervision model and the DWS system; An abnormal data identification module, which is used to upload express scanning information to the DWS system, receive sensor-collected data based on detection equipment, extract effective indicator features and perform cross-channel interactive proofreading based on the adaptive supervision model, and identify abnormal indicator data, wherein the abnormal indicator data is identified with a unique label; An abnormal docking information identification module, the abnormal docking information identification module is used to perform continuous response monitoring of device coordination based on the device docking information and in combination with an image sensing device, and identify abnormal docking information; An adaptive early warning module is used to execute an adaptive early warning based on the abnormal indicator data and the abnormal docking information based on an abnormal alarm system, wherein the abnormal alarm system includes a misclassified alarm mode and an operation early warning mode.

Images