IPC equipment remote supervision platform based on Internet of Things card

Abstract

The invention discloses an IPC device remote supervision platform based on an Internet of Things card, and relates to the technical field of IPC supervision, and the platform comprises an IPC remote cooperation unit construction module which is interconnected with a plurality of IPC devices; the IPC node data set reading module is used for reading a plurality of IPC node data sets; the abnormal risk detection module is used for determining abnormal risk nodes; the monitoring abnormity linkage tracking strategy determining module is used for determining a monitoring abnormity linkage tracking strategy; an equipment exception collaborative operation and maintenance strategy acquisition module for acquiring an equipment exception collaborative operation and maintenance strategy; and the optimization management module performs optimization management. According to the invention, the technical problem that the monitoring stability and reliability of the IPC equipment are insufficient due to the fact that effective intelligent cooperative management and control and accurate operation and maintenance cannot be carried out on the plurality of IPC equipment in the prior art is solved, and the technical effects of realizing intelligent cooperative management and control and accurate operation and maintenance on the plurality of IPC equipment and improving the monitoring stability and reliability of the IPC equipment are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of IPC supervision, and particularly to a remote supervision platform for IPC devices based on Internet of Things cards. Background Art

[0002] In the current security and industrial monitoring fields, IPC (Internet Protocol Camera) devices are widely used in various scenarios, such as urban streets, factory workshops, intelligent buildings, etc., and shoulder the heavy responsibility of real-time image acquisition and key area monitoring. Traditional IPC device monitoring technologies mostly adopt an isolated management mode, where each IPC device operates independently and lacks an efficient coordination mechanism among them. This leads to frequent problems when facing complex and changeable monitoring requirements. On the one hand, when a device malfunctions, due to the lack of linkage and cooperation, the efficiency of fault diagnosis and repair is low, and it is also easy to cause monitoring blind spots, greatly affecting the reliability of monitoring. On the other hand, during the process of tracking abnormal events, each device cannot intelligently cooperate to adjust the monitoring strategy, cannot flexibly allocate resources according to the real-time situation, and it is difficult to achieve precise monitoring, thus greatly reducing the monitoring stability.

[0003] There is a technical problem in the prior art that multiple IPC devices cannot be effectively and intelligently coordinated and controlled and precisely maintained, resulting in insufficient monitoring stability and reliability of IPC devices. Summary of the Invention

[0004] The present application provides a remote supervision platform for IPC devices based on Internet of Things cards, which is used to solve the technical problem in the prior art that multiple IPC devices cannot be effectively and intelligently coordinated and controlled and precisely maintained, resulting in insufficient monitoring stability and reliability of IPC devices.

[0005] In view of the above problems, the present application provides a remote supervision platform for IPC devices based on Internet of Things cards.

[0006] The present application provides a remote supervision platform for IPC devices based on Internet of Things cards, and the platform includes: IPC Remote Collaboration Unit Construction Module, which is used to interconnect multiple IPC devices to construct an IPC remote collaboration unit. Among them, each IPC device is built-in with an Internet of Things card; IPC Node Dataset Reading Module, which is used to read multiple IPC node datasets according to the IPC remote collaboration unit. Among them, each IPC node dataset includes the monitoring video data and device status data corresponding to each IPC device; Abnormal Risk Inspection Module, which is used to perform abnormal risk inspection on the multiple IPC node datasets according to the three channels of IPC abnormal risk inspection to determine the monitoring abnormal risk nodes and device abnormal risk nodes; Monitoring Abnormal Linkage Tracking Strategy Determination Module, which is used to make abnormal linkage tracking decisions on the IPC remote collaboration unit according to the monitoring abnormal risk nodes to determine the monitoring abnormal linkage tracking strategy; Device Abnormal Collaborative Operation and Maintenance Strategy Acquisition Module, which is used to make task collaborative operation and maintenance decisions on the IPC remote collaboration unit according to the device abnormal risk nodes to obtain the device abnormal collaborative operation and maintenance strategy; Optimization Management Module, which is used to optimize and manage the IPC remote collaboration unit according to the monitoring abnormal linkage operation and maintenance strategy and the device abnormal collaborative operation and maintenance strategy.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: IPC Remote Collaboration Unit Construction Module, which is used to interconnect multiple IPC devices to construct an IPC remote collaboration unit; IPC Node Dataset Reading Module, which is used to read multiple IPC node datasets according to the IPC remote collaboration unit; Abnormal Risk Inspection Module, which performs abnormal risk inspection on the multiple IPC node datasets to determine the monitoring abnormal risk nodes and device abnormal risk nodes; Monitoring Abnormal Linkage Tracking Strategy Determination Module, which is used to make abnormal linkage tracking decisions on the IPC remote collaboration unit according to the monitoring abnormal risk nodes to determine the monitoring abnormal linkage tracking strategy; Device Abnormal Collaborative Operation and Maintenance Strategy Acquisition Module, which makes task collaborative operation and maintenance decisions on the IPC remote collaboration unit to obtain the device abnormal collaborative operation and maintenance strategy; Optimization Management Module, which is used to optimize and manage the IPC remote collaboration unit according to the monitoring abnormal linkage operation and maintenance strategy and the device abnormal collaborative operation and maintenance strategy. It achieves the technical effect of realizing intelligent collaborative control and precise operation and maintenance of multiple IPC devices, and improving the monitoring stability and reliability of IPC devices. Description of the Drawings

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0009] Figure 1 The structural schematic diagram of the IPC device remote supervision platform based on the Internet of Things card provided by the embodiment of the present application.

[0010] Figure 2 The process schematic diagram of the abnormal risk inspection module in the IPC device remote supervision platform based on the Internet of Things card provided by the embodiment of the present application.

[0011] Explanation of reference numerals: IPC remote collaboration unit construction module 10, IPC node data set reading module 20, abnormal risk inspection module 30, monitoring abnormal linkage tracking strategy determination module 40, device abnormal collaboration operation and maintenance strategy acquisition module 50, optimization management module 60. Specific implementation manners

[0012] The present application provides an IPC device remote supervision platform based on the Internet of Things card, which is used to solve the technical problem that in the prior art, multiple IPC devices cannot be effectively and intelligently collaboratively controlled and precisely operated and maintained, resulting in insufficient monitoring stability and reliability of the IPC devices.

[0013] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0014] Embodiment, as Figure 1 shown, the present application provides an IPC device remote supervision platform based on the Internet of Things card, and the platform includes: IPC remote collaboration unit construction module 10, which is used to interconnect multiple IPC devices to construct an IPC remote collaboration unit, wherein each IPC device is built with an Internet of Things card.

[0015] Specifically, the core function of the IPC remote collaboration unit construction module 10 is to use the Internet of Things card to realize the interconnection and intercommunication of multiple IPC devices, and construct an IPC remote collaboration unit. The Internet of Things card provides a stable network connection for each IPC device, ensuring that the device can transmit data in real time and receive remote instructions. During the construction process, this module will identify the identities of each IPC device and configure the network, enabling them to work collaboratively under a unified architecture, forming an organic whole, and laying a foundation for subsequent centralized management and data interaction.

[0016] The IPC node dataset reading module 20 is used to read multiple IPC node datasets according to the IPC remote collaboration unit, where each IPC node dataset includes the monitoring video data and device status data corresponding to each IPC device.

[0017] Specifically, after the IPC remote collaboration unit is started, the IPC node dataset reading module 20 first activates the Internet of Things cards built into each IPC device, triggers the device authentication mechanism, and ensures that only legal and properly operating IPC devices can access the platform. This authentication process involves a series of complex encryption algorithms and device unique identifier comparisons, effectively preventing illegal device intrusion. Once the verification is passed, the module starts to collect the video and status data of the IPC devices. For the monitoring video data, a real-time compression algorithm is used to significantly reduce the data volume without significantly degrading the video quality, improving the transmission efficiency; at the same time, high-strength encryption technology is used to encrypt the video data to ensure the security and privacy of the data. The device status data is also strictly encrypted and then transmitted to the IPC remote collaboration unit through the Internet of Things card stably and efficiently together with the compressed video data. In the IPC remote collaboration unit, these processed data are further integrated to form an IPC node dataset corresponding to each IPC device, including monitoring video data and device status data, providing a comprehensive, accurate and secure data basis for subsequent abnormal risk inspection, policy decision-making and other operations.

[0018] The abnormal risk inspection module 30 is used to perform abnormal risk inspection on the multiple IPC node datasets according to the three channels of IPC abnormal risk inspection, and determine the monitoring abnormal risk nodes and device abnormal risk nodes.

[0019] Specifically, the abnormal risk inspection module 30 deeply analyzes multiple IPC node datasets using the three channels of IPC abnormal risk inspection. This module first accurately extracts the monitoring video data and device status data of specific IPC devices from the dataset through the device status data extraction unit, and then inputs them into the corresponding detection and prediction channels respectively. In the monitoring abnormal risk coefficient acquisition unit, according to the working mechanism of the IPC monitoring abnormal risk detection channel, the monitoring abnormal risk detection branch of specific IPC devices is activated. This branch includes an advanced first monitoring abnormal recognition model and a first monitoring abnormal risk prediction model. Inputting the monitoring video data into the recognition model can obtain accurate monitoring abnormal recognition results, and then inputting this result into the prediction model can generate the first monitoring abnormal risk coefficient. The acquisition of the device abnormal risk coefficient involves the collaborative work of multiple subunits. The device expected deviation coefficient acquisition unit first performs expected deviation detection on the device status data according to the IPC device status recording unit. By collecting the IPC device scenario information (including control parameters and environmental parameters) and IPC device attributes (specification model parameters), retrieving the normal state sample set from the device status recording unit, and then analyzing to obtain the device expected state, and then performing deviation detection on the device status data accordingly, generating a device expected deviation matrix, and performing deviation depth evaluation to obtain the device expected deviation coefficient. The abnormal risk prediction gating coefficient acquisition unit inputs the device expected deviation coefficient into the abnormal risk prediction gating unit to obtain the first abnormal risk prediction gating coefficient. Finally, the device abnormal risk coefficient output unit, based on the above gating coefficient and expected deviation matrix, uses multiple device abnormal risk prediction models in the device abnormal risk prediction unit, and through a series of operations such as activating the model, obtaining the prediction coefficient, calculating the risk prediction incentive coefficient, and weighted calculation, outputs the first device abnormal risk coefficient. The IPC risk abnormal verification unit verifies the obtained risk coefficients according to the set monitoring abnormal risk threshold and device abnormal risk threshold, and finally determines the monitoring abnormal risk nodes and device abnormal risk nodes, so as to accurately locate the IPC devices that may have abnormalities.

[0020] The monitoring abnormal linkage tracking strategy determination module 40 is used to make an abnormal linkage tracking decision on the IPC remote collaboration unit according to the monitoring abnormal risk nodes, and determine the monitoring abnormal linkage tracking strategy.

[0021] Specifically, the monitoring anomaly linkage tracking strategy determination module 40 conducts a series of intelligent decisions on the determined monitoring anomaly risk nodes. Based on these risk nodes, the monitoring anomaly warning signal generation unit timely generates monitoring anomaly warning signals, providing trigger conditions for subsequent operations. The monitoring anomaly trend prediction result determination unit deeply analyzes the monitoring anomaly recognition results corresponding to the monitoring anomaly risk nodes according to the warning signals, and uses data analysis algorithms to conduct trend prediction to determine the monitoring anomaly trend prediction results, such as the development direction of abnormal situations, the possible affected range, etc. The abnormal monitoring range fitting result acquisition unit determines the abnormal monitoring range fitting result through precise calculations and model fitting according to the prediction results. Based on this result, the associated IPC device set acquisition unit conducts a comprehensive IPC device association identification on the IPC remote collaboration unit, screens out the device set associated with the abnormal situation, and quickly loads its corresponding IPC monitoring configuration set. Finally, according to the monitoring anomaly trend prediction results, the monitoring configuration optimization unit conducts targeted optimization on the IPC monitoring configuration set, such as adjusting monitoring parameters, switching monitoring modes, etc., so as to generate a practical monitoring anomaly linkage tracking strategy, realizing efficient response and collaborative processing of monitoring anomaly situations.

[0022] The device anomaly collaborative operation and maintenance strategy acquisition module 50 is used to make task collaborative operation and maintenance decisions on the IPC remote collaboration unit according to the device anomaly risk nodes, and obtain the device anomaly collaborative operation and maintenance strategy.

[0023] Specifically, the device anomaly collaborative operation and maintenance strategy acquisition module 50 plays a role when facing device anomaly risk nodes. The device anomaly warning signal generation unit generates device anomaly warning signals based on the risk nodes, triggering subsequent operation and maintenance processes. The collaborative monitoring task acquisition unit collects the detailed monitoring task parameters of the device anomaly risk nodes according to the warning signals, and then determines the collaborative monitoring tasks. The monitoring task scheduling space establishment unit conducts dynamic task adjustment on the IPC remote collaboration unit according to the collaborative monitoring tasks, reasonably allocates resources, and establishes a scientific and reasonable monitoring task scheduling space. The monitoring quality optimization unit uses optimization algorithms to optimize the monitoring quality within this scheduling space. By adjusting various parameters and task allocation schemes, it ensures that the best monitoring effect can still be maintained in case of device anomalies, and finally generates the device anomaly collaborative operation and maintenance strategy to ensure the stable operation of IPC devices and the continuous effectiveness of monitoring.

[0024] The optimization management module 60 is used to optimize the management of the IPC remote collaboration unit according to the monitoring anomaly linkage operation and maintenance strategy and the device anomaly collaborative operation and maintenance strategy.

[0025] Specifically, the optimization management module 60 is the control center of the entire platform. It comprehensively optimizes and manages the IPC remote collaboration unit according to the monitoring anomaly linkage operation and maintenance strategy and the device anomaly collaborative operation and maintenance strategy. This module comprehensively considers various measures in the two strategies, such as adjusting device parameters, optimizing network configuration, reallocating tasks, etc., to ensure that the IPC remote collaboration unit can maintain an efficient and stable operating state when dealing with various abnormal situations. In addition, the platform is also equipped with an abnormal operation and maintenance integration module. When there is a complex situation where the same IPC device is involved in both the monitoring anomaly risk node and the device anomaly risk node, this module can skillfully integrate the two strategies to generate an abnormal operation and maintenance integration strategy, thereby performing more accurate and effective optimization management on the IPC remote collaboration unit, further improving the platform's ability to handle complex abnormal situations, and ensuring the reliability and security of the entire IPC device monitoring.

[0026] In a possible implementation manner, as Figure 2 shown, the anomaly risk inspection module 30 includes: A device status data extraction unit, configured to extract the first monitoring video data and the first device status data corresponding to the first IPC device according to the multiple IPC node data sets.

[0027] A monitoring anomaly risk coefficient acquisition unit, configured to input the first monitoring video data into the IPC monitoring anomaly risk detection channel to obtain a first monitoring anomaly risk coefficient.

[0028] A device anomaly risk coefficient acquisition unit, configured to input the first device status data into the IPC device anomaly risk prediction channel to obtain a first device anomaly risk coefficient.

[0029] An IPC risk anomaly verification unit, configured to input the first monitoring anomaly risk coefficient and the first device anomaly risk coefficient into the IPC risk anomaly verification channel, where the IPC risk anomaly verification channel includes a monitoring anomaly risk threshold and a device anomaly risk threshold.

[0030] A device anomaly risk node addition unit, configured to add the first IPC device to the monitoring anomaly risk node if the first monitoring anomaly risk coefficient is greater than or equal to the monitoring anomaly risk threshold, and add the first IPC device to the device anomaly risk node if the first device anomaly risk coefficient is greater than or equal to the device anomaly risk threshold.

[0031] Specifically, the three channels of IPC anomaly risk inspection cover the IPC monitoring anomaly risk detection channel, the IPC device anomaly risk prediction channel, and the IPC risk anomaly verification channel. The three work together to deeply analyze the operating conditions of the IPC device from different dimensions.

[0032] The device status data extraction unit undertakes the key data extraction task in the IPC device remote supervision platform based on the Internet of Things card. When multiple IPC node data sets converge, it quickly locates the data related to the first IPC device, that is, any one of the multiple IPC devices. By using data retrieval and recognition algorithms, it accurately filters out the first monitoring video data corresponding to the first IPC device from the complex data structure. These video data contain the real-time picture information of the area monitored by the device, such as rich content like the activities of people in the scene and the changes in the state of objects. At the same time, it also accurately extracts the first device status data, which cover the key parameters of the operation of the IPC device itself, including aspects such as the temperature, power, network connection status, lens focal length, and working status of the image sensor of the device. Through this meticulous data extraction process, it provides indispensable basic data for the subsequent comprehensive abnormal risk assessment of the first IPC device. Whether in the monitoring abnormal risk detection or the device abnormal risk prediction link, these accurately extracted data will become the key analysis basis, ensuring that the entire supervision platform can timely and accurately grasp the operation status of the IPC device and the monitoring scene information, thus effectively guaranteeing the stable operation of the monitoring and the smooth completion of the monitoring task.

[0033] The monitoring anomaly risk coefficient acquisition unit plays a crucial role in the IPC device remote supervision platform based on IoT cards. Its core function is to accurately obtain the first monitoring anomaly risk coefficient by inputting the first monitoring video data into the IPC monitoring anomaly risk detection channel, thereby providing a key basis for evaluating the degree of anomaly risk in the monitoring footage. When the first monitoring video data is received, this unit first guides the data to the IPC monitoring anomaly risk detection channel. Inside this channel, there are multiple monitoring anomaly risk detection branches for different IPC devices and monitoring scenarios, and each branch is equipped with highly specialized first monitoring anomaly recognition models and first monitoring anomaly risk prediction models. After the data is input into the branch, the first monitoring anomaly recognition model immediately activates its powerful image analysis and behavior recognition functions. Using advanced deep learning algorithms, it carefully analyzes each frame of the video data, and can accurately identify various objects, people, and their behavior actions in the footage, such as quickly detecting the appearance of abnormal objects, abnormal gatherings of people, or abnormal behavior patterns. Based on these recognition results, the first monitoring anomaly recognition model generates detailed first monitoring anomaly recognition results, which not only include basic information such as the type, occurrence time, and location of the abnormal event, but may also include a preliminary severity assessment of the abnormal event. Subsequently, the first monitoring anomaly recognition results are input into the first monitoring anomaly risk prediction model. Based on a large amount of historical monitoring data and preset risk assessment algorithms, this model comprehensively considers various factors, such as the duration, frequency, scope of influence of the abnormal event, and its relationship with the surrounding environment, etc., to conduct a deep quantitative assessment of the anomaly risk. For example, for an abnormal object that appears briefly and has a small scope of influence, the model may give a relatively low risk score; while for an abnormal behavior that lasts for a long time, involves a large number of people, and may pose a threat to public safety, the model will calculate a higher risk score. Through this series of complex calculation and analysis processes, the first monitoring anomaly risk coefficient is finally generated. This coefficient intuitively reflects the degree of anomaly risk existing in the current monitoring footage in numerical form, provides an important quantitative basis for subsequent risk judgment and decision-making, helps the supervision platform timely and accurately discover potential safety hazards, and take corresponding measures for handling.

[0034] After receiving the first device status data, the device abnormal risk coefficient acquisition unit transmits it to the IPC device abnormal risk prediction channel. This channel consists of several important components, such as the IPC device status recording unit, the abnormal risk prediction gating unit, and the device abnormal risk prediction unit. They work together to complete the calculation of the device abnormal risk coefficient. First, the device expected deviation coefficient acquisition unit relies on the IPC device status recording unit to collect various information related to the first IPC device, including control parameters (such as the focal length of the camera, the aperture size, the pan-tilt rotation angle, etc.) and environmental parameters (such as the temperature, humidity, light intensity of the environment where the device is located, etc.), so as to obtain the first IPC device scenario information. At the same time, the specification model parameters of the first IPC device are also collected to comprehensively understand the inherent attributes of the device. With this information, the first device normal state sample set that matches the current device scenario and attributes is accurately retrieved from the IPC device status recording unit. Then, a central tendency analysis is performed on this sample set, such as calculating statistical indicators such as the mean and median, so as to determine the first device expected state, which represents the ideal value range of each parameter of the device under normal operating conditions. Subsequently, the deviation detection unit conducts a detailed deviation detection on the first device status data based on the first device expected state, comparing each parameter in the device status data with the expected state one by one, such as checking whether the device temperature exceeds the normal range and whether the power supply voltage is stable within a reasonable range, and quantifying these differences in matrix form to generate the first device expected deviation matrix. This matrix details the deviation degree and direction of the actual device state from the expected state in each parameter dimension, providing an important data basis for subsequent risk assessment. The deviation depth evaluation unit further conducts a deep analysis on the first device expected deviation matrix. It uses a preset evaluation algorithm, comprehensively considering the weights of the deviations of each parameter (determining the impact degree of different parameter deviations on the overall risk of the device according to the device operation principle and experience) and the degree of deviation, and calculates the first device expected deviation coefficient. This coefficient intuitively reflects the overall degree of the device state deviating from the normal range and is an important intermediate indicator for evaluating the device abnormal risk. The abnormal risk prediction gating coefficient acquisition unit inputs the first device expected deviation coefficient into the abnormal risk prediction gating unit, which, based on complex models and algorithms, calculates the first abnormal risk prediction gating coefficient according to the input expected deviation coefficient, combined with the device's historical operation data and risk statistical rules. This gating coefficient plays a role in dynamically adjusting the weights and sensitivities of the subsequent risk prediction model, ensuring that the risk prediction result can more accurately reflect the current actual risk status of the device. Finally, the device abnormal risk coefficient output unit works in the device abnormal risk prediction unit based on the first abnormal risk prediction gating coefficient and the first device expected deviation matrix.The device anomaly risk prediction unit includes multiple device anomaly risk prediction models for different types of device anomalies. The activated anomaly risk prediction model acquisition unit activates the features of these models according to the first anomaly risk prediction gating coefficient, and screens out multiple activated anomaly risk prediction models that are most relevant to the current device state and risk characteristics. The anomaly risk prediction coefficient acquisition unit inputs the first device expected deviation matrix into these activated models, and each model calculates an anomaly risk prediction coefficient according to its own algorithm and training data. The risk prediction incentive coefficient acquisition unit calculates the proportion based on multiple anomaly risk prediction accuracy parameters corresponding to multiple activated anomaly risk prediction models (these accuracy parameters are obtained through the testing and verification of the models on historical data and reflect the accuracy of the model prediction results), determines the weight of each model in the final result, that is, multiple risk prediction incentive coefficients. The weighted calculation unit performs weighted calculation on multiple anomaly risk prediction coefficients according to these incentive coefficients, comprehensively considers the prediction results and their credibility of each model, and finally obtains a comprehensive and accurate first device anomaly risk coefficient. This coefficient comprehensively reflects the degree of anomaly risk in the internal operation state of the IPC device, provides a key quantitative basis for subsequent judgment on whether the device needs maintenance, adjustment or warning, ensures that the supervision platform can timely detect potential problems of the device, and guarantees the stable and reliable operation of the IPC device.

[0035] The main function of the IPC risk anomaly verification unit is to input the obtained first monitoring anomaly risk coefficient and the first device anomaly risk coefficient into the IPC risk anomaly verification channel, and accurately judge whether there is an anomaly risk in the IPC device by comparing them with the preset monitoring anomaly risk threshold and device anomaly risk threshold. After receiving the first monitoring anomaly risk coefficient and the first device anomaly risk coefficient, the IPC risk anomaly verification unit immediately transmits these two coefficients to the IPC risk anomaly verification channel. The channel has preset scientific and reasonable monitoring anomaly risk thresholds and device anomaly risk thresholds, which are determined based on a large amount of historical monitoring data, device operation data, and industry experience, and are used to clearly distinguish the boundaries between normal and abnormal risk states. When performing verification, first compare the first monitoring anomaly risk coefficient with the monitoring anomaly risk threshold. If the first monitoring anomaly risk coefficient is greater than or equal to the monitoring anomaly risk threshold, it indicates that there is a relatively high anomaly risk in the monitoring screen, involving situations such as security threats, abnormal behaviors, or emergencies, and timely attention and handling are required. For example, when there are scenes such as a large number of people gathering abnormally, suspicious objects staying for a long time, or violent behaviors in the monitoring area, the monitoring anomaly risk coefficient will increase, and once it exceeds the threshold, the corresponding warning mechanism will be triggered. At the same time, compare the first device anomaly risk coefficient with the device anomaly risk threshold. If the first device anomaly risk coefficient is greater than or equal to the device anomaly risk threshold, it means that the operation state of the IPC device itself is abnormal, and there are problems such as hardware failures, software errors, unstable network connections, or abnormal device parameters, which will directly affect the normal monitoring function of the device and may even cause the device to stop working or data transmission to be interrupted. For example, situations such as too high device temperature, unstable power supply, or abnormal lens focusing will cause the device anomaly risk coefficient to rise, and when it reaches or exceeds the threshold, it indicates that the device needs maintenance, repair, or adjustment. Through this rigorous verification method, the IPC risk anomaly verification unit can accurately identify the anomaly risks of the IPC device in the monitoring screen and its own operation, providing a basis for subsequent targeted measures. Whether it is to start the monitoring anomaly linkage tracking strategy or implement the device anomaly collaborative operation and maintenance strategy, it depends on the accurate judgment of risks by this unit, so as to ensure that the entire IPC device monitoring always maintains an efficient and stable operation state, effectively ensuring the safety and normal order of the monitoring area.

[0036] After receiving the first monitoring abnormality risk coefficient and the first equipment abnormality risk coefficient, the equipment abnormality risk node adding unit will compare the first monitoring abnormality risk coefficient with the monitoring abnormality risk threshold. If the first monitoring abnormality risk coefficient is greater than or equal to the monitoring abnormality risk threshold, this means that a more serious abnormal situation has appeared in the monitoring screen, such as suspicious personnel activities, abnormal objects, or emergencies, etc. These situations may pose a threat to the safety and normal order of the monitoring area. At this time, the equipment abnormality risk node adding unit will decisively add the first IPC device to the monitoring abnormality risk node set. The purpose of this operation is to centrally manage all IPC devices that may have monitoring abnormality risks, so as to subsequently initiate monitoring abnormality linkage tracking strategies for these devices, such as adjusting monitoring parameters, switching monitoring modes, notifying relevant personnel to pay real-time attention, etc., to ensure that abnormal situations in the monitoring screen can be responded to in a timely and effective manner to prevent potential risks from further expanding. At the same time, the unit will also compare the first equipment abnormality risk coefficient and the equipment abnormality risk threshold. If the first device abnormality risk coefficient is greater than or equal to the device abnormality risk threshold, this indicates that the operating status of the first IPC device itself is abnormal, which may be caused by device hardware failure (such as camera damage, sensor failure, etc.), software problems (such as program crashes, system errors, etc.) or unstable network connections. In this case, the device abnormality risk node addition unit will add the first IPC device to the device abnormality risk node set. This is done to facilitate the implementation of device abnormality collaborative operation and maintenance strategies for these devices, including arranging technicians to perform equipment maintenance, remotely diagnose and repair equipment failures, optimize equipment parameter configuration, etc., to ensure that the IPC device can resume normal operation and ensure the stability and reliability of monitoring.

[0037] In a possible implementation, the monitoring abnormality risk coefficient acquisition unit includes: A monitoring abnormality risk detection branch activation unit is used to activate the first monitoring abnormality risk detection branch corresponding to the first IPC device according to the IPC monitoring abnormality risk detection channel, wherein the first monitoring abnormality risk detection branch includes a first monitoring abnormality identification model and a first monitoring abnormality risk prediction model.

[0038] The monitoring anomaly recognition result acquisition unit is used to input the first monitoring video data into the first monitoring anomaly recognition model to obtain a first monitoring anomaly recognition result.

[0039] The monitoring anomaly risk coefficient generating unit is used to input the first monitoring anomaly identification result into the first monitoring anomaly risk prediction model to generate the first monitoring anomaly risk coefficient.

[0040] Specifically, the IPC monitoring abnormal risk detection channel includes multiple monitoring abnormal risk detection branches corresponding one-to-one with IPC devices, and each branch is optimized and configured according to the characteristics of specific devices and monitoring scenarios.

[0041] The monitoring abnormal risk detection branch activation unit accurately activates the first monitoring abnormal risk detection branch corresponding to the first IPC device according to the overall architecture and control logic of the IPC monitoring abnormal risk detection channel. This activation process is based on various factors such as the identification information of the device, the monitoring scenario setting, and the current risk monitoring focus, ensuring that only the detection branch related to the first IPC device and most suitable for the current situation is enabled, thus improving the accuracy and efficiency of detection.

[0042] The first monitoring abnormal risk detection branch internally integrates the first monitoring abnormal recognition model and the first monitoring abnormal risk prediction model, and the two complement each other to jointly complete the abnormal risk detection task. The monitoring abnormal recognition result acquisition unit inputs the first monitoring video data into the first monitoring abnormal recognition model, which is trained by using deep learning algorithms and a large amount of image data and has strong image understanding and pattern recognition capabilities. It can deeply analyze various objects, human behaviors, scene changes, etc. in the video data, and quickly identify abnormal situations that do not conform to the normal monitoring mode, such as unauthorized personnel intrusion, the appearance of abnormal objects, abnormal behavior actions (such as running, fighting, wandering, etc.), and environmental changes (such as fire smoke, flood inundation, etc.). Through complex calculations and feature comparisons, the model outputs detailed and accurate first monitoring abnormal recognition results, which not only cover basic information such as the type, occurrence time, and location of abnormal events, but may also include a preliminary assessment of the degree of abnormality, providing key basic data for subsequent risk prediction.

[0043] The monitoring abnormal risk coefficient generation unit inputs the first monitoring abnormal recognition result into the first monitoring abnormal risk prediction model, which comprehensively considers various factors in the abnormal recognition result, such as the duration, frequency, influence range, development trend of the abnormal event, and the correlation with the surrounding environment and other events, based on big data analysis and risk assessment algorithms, and conducts a comprehensive quantitative assessment of the abnormal risk. For example, for a short and isolated abnormal event, the model may give a relatively low risk score; while for an abnormal situation with a long duration, frequent occurrence, and likely to trigger a chain reaction, the model will calculate a higher risk score. After a series of complex calculations and model reasoning processes, the first monitoring abnormal risk coefficient is finally generated, and this coefficient accurately reflects the degree of abnormal risk existing in the current monitoring screen in an intuitive numerical form, providing an important decision-making basis for the entire supervision platform to judge the safety status of the monitoring area and take corresponding early warning and response measures.

[0044] In a possible implementation manner, the device anomaly risk coefficient acquisition unit includes: A device expected deviation coefficient acquisition unit, configured to perform expected deviation detection on the first device status data according to the IPC device status recording unit, and obtain a first device expected deviation matrix and a first device expected deviation coefficient.

[0045] An anomaly risk prediction gating coefficient acquisition unit, configured to input the first device expected deviation coefficient into the anomaly risk prediction gating unit to obtain a first anomaly risk prediction gating coefficient.

[0046] A device anomaly risk coefficient output unit, configured to output the first device anomaly risk coefficient according to the device anomaly risk prediction unit based on the first anomaly risk prediction gating coefficient and the first device expected deviation matrix.

[0047] Specifically, in the IPC device remote supervision platform based on the Internet of Things card, the IPC device anomaly risk prediction channel is a key part to ensure the stable operation of the device and early warning of potential faults. Through the collaborative cooperation of multiple functional units, it realizes the accurate prediction of the device anomaly risk. The IPC device anomaly risk prediction channel mainly consists of an IPC device status recording unit, an anomaly risk prediction gating unit, and a device anomaly risk prediction unit.

[0048] The device expected deviation coefficient acquisition unit plays a role. It performs expected deviation detection on the first device status data based on the massive historical data in the IPC device status recording unit. The IPC device status recording unit stores in detail the various parameter change ranges and modes of each IPC device in the normal operation state. The device expected deviation coefficient acquisition unit accurately retrieves the corresponding historical normal state sample set by collecting the control parameters (such as camera focal length adjustment, pan-tilt rotation speed, etc.), environmental parameters (such as the temperature and humidity of the environment where the device is located), and the specification model parameters of the device itself. Then, through in-depth central tendency analysis of this sample set, statistical indicators such as the mean and median are calculated to determine the first device expected state. Then, the first device status data is carefully compared with the expected state, and the deviation degree of each parameter is calculated to generate a first device expected deviation matrix, which comprehensively shows the differences between the actual state and the expected state of the device in each parameter dimension. At the same time, based on a specific algorithm, these deviation degrees are comprehensively evaluated to obtain a first device expected deviation coefficient, which intuitively reflects the overall degree of the current state of the device deviating from the normal range.

[0049] The core task of the abnormal risk prediction gating coefficient acquisition unit is to input the first device expected deviation coefficient into the neural network-based abnormal risk prediction gating unit, and then obtain the first abnormal risk prediction gating coefficient, providing a key parameter basis for the subsequent accurate assessment of the device abnormal risk. After obtaining the first device expected deviation coefficient, the unit immediately inputs it into the abnormal risk prediction gating unit. This neural network model has been deeply trained with a large amount of historical device operation data, constructing a highly complex and accurate mapping relationship. Its neural network architecture includes multiple carefully designed hidden layers, and each layer is interconnected by dense neurons, forming a powerful computing network. When the data is input, the first device expected deviation coefficient first enters the input layer of the neural network. Subsequently, the data is sequentially transmitted between the neurons of each hidden layer. Each neuron performs a weighted summation operation on the input data according to its unique connection weights and bias parameters. This weighted summation process can integrally target different dimensions of information according to the importance of data features. Then, the neuron performs a non-linear transformation on the weighted summation result through a pre-set activation function (such as ReLU, Sigmoid, etc.), thereby introducing non-linear factors and enhancing the model's ability to express complex data relationships. Through the calculation and information transmission of layer upon layer of neurons, the neural network can automatically mine and capture the deep feature information contained in the first device expected deviation coefficient, as well as the subtle correlation patterns between these features and the potential risks of the device. After complex operations of multiple layers of neurons, the neural network finally outputs the first abnormal risk prediction gating coefficient. This coefficient is not a simple numerical value, but comprehensively considers the potential risk significance of the device's current expected deviation coefficient under different operation scenarios and historical data backgrounds. As a dynamically adjusted key factor, it will play a key role in weight distribution and sensitivity regulation for each model in the device abnormal risk prediction unit during the subsequent calculation of the device abnormal risk coefficient. In this way, it is ensured that the finally generated device abnormal risk coefficient can closely fit the actual risk situation of the device, providing a solid and reliable basis for timely and accurate implementation of device maintenance and issuance of warning signals, thereby ensuring the stable operation of the IPC device and improving the security and reliability of the entire monitoring.

[0050] The device abnormal risk coefficient output unit, based on the first abnormal risk prediction gating coefficient and the first device expected deviation matrix, performs a series of precise algorithm operations in the device abnormal risk prediction unit to output the first device abnormal risk coefficient. First, according to the first abnormal risk prediction gating coefficient, the matching algorithm is used to screen out the activated abnormal risk prediction models from multiple professional models in the device abnormal risk prediction unit. Then, the first device expected deviation matrix is input into these activated models, and each model performs feature extraction, pattern recognition, and risk quantification assessment on the matrix data according to its internal risk assessment algorithm, and calculates the abnormal risk prediction coefficient. Next, according to the abnormal risk prediction accuracy parameters of each activated model (obtained by validating historical data), the weighted average algorithm is used to calculate the risk prediction incentive coefficient. Finally, the sum of the products of the abnormal risk prediction coefficient and the corresponding risk prediction incentive coefficient is calculated through the weighted summation algorithm to obtain the first device abnormal risk coefficient, which provides a precise quantitative key basis for device operation and maintenance decisions.

[0051] In a possible implementation manner, the device abnormal risk coefficient acquisition unit includes: The IPC device scenario information acquisition unit is used to collect the control parameters and environmental parameters corresponding to the first IPC device to obtain the first IPC device scenario information.

[0052] The IPC device attribute acquisition unit is used to collect the specification and model parameters corresponding to the first IPC device to obtain the first IPC device attribute.

[0053] The device normal state sample set acquisition unit is used to perform normal state sample retrieval on the IPC device state record unit according to the first IPC device scenario information and the first IPC device attribute to obtain the first device normal state sample set.

[0054] The device expected state acquisition unit is used to perform central tendency analysis according to the first device normal state sample set to obtain the first device expected state.

[0055] The deviation detection unit is used to perform deviation detection on the first device state data according to the first device expected state to generate the first device expected deviation matrix.

[0056] The deviation depth evaluation unit is used to perform deviation depth evaluation according to the first device expected deviation matrix to generate the first device expected deviation coefficient.

[0057] Specifically, in the IPC device remote monitoring platform based on IoT cards, the following units work together to comprehensively analyze the abnormal risks of IPC devices. The IPC device scenario information acquisition unit is responsible for collecting the control parameters corresponding to the first IPC device (such as the focal length adjustment of the camera, exposure time setting, pan / tilt rotation angle and speed, etc.) and environmental parameters (such as the temperature, humidity, light intensity, electromagnetic interference intensity, etc. of the environment where the device is located), and obtaining the first IPC device scenario information after integrating these parameters. This step lays the foundation for accurately evaluating the device status subsequently, because the operating status of the device is closely related to the scenario it is in.

[0058] The IPC device attribute acquisition unit focuses on collecting the specification and model parameters of the first IPC device (such as the resolution of the camera, sensor type, communication protocol of the device, hardware version number, etc.), so as to accurately obtain the first IPC device attributes. These attribute information clarify the inherent characteristics of the device and are crucial for determining the normal operation standards of the device.

[0059] The device normal state sample set acquisition unit comprehensively retrieves the IPC device status record unit based on the obtained first IPC device scenario information and first IPC device attributes. The IPC device status record unit stores a large amount of historical operation data. By matching the scenario information and device attributes, this unit screens out the normal operation status data of the current first IPC device under similar scenarios and the same attributes, and then obtains the first device normal state sample set. This process ensures that the selected samples are highly relevant and representative.

[0060] The device expected state acquisition unit conducts a central tendency analysis based on the first device normal state sample set, and uses statistical methods to calculate the mean, median, mode, etc. to determine the first device expected state. For example, for the device temperature parameter, the mean value of the temperature values in the normal state sample set is calculated to obtain the expected temperature value; for the communication delay parameter, the median is taken as the expected communication delay. These expected state values represent the ideal value ranges of each parameter when the device is operating normally.

[0061] The deviation detection unit conducts deviation detection on the first device status data according to the determined first device expected state. It compares each parameter in the device status data with the expected state one by one, quantifies these differences and organizes them in matrix form to generate the first device expected deviation matrix. For example, if the actual temperature of the device is higher than the expected temperature by a certain threshold, a positive deviation value is recorded at the corresponding position in the matrix; if the power supply voltage is lower than the expected voltage, a negative deviation value is recorded.

[0062] Finally, the deviation depth evaluation unit performs deviation depth evaluation based on the first device expected deviation matrix, using weighted summation, comprehensively considering the deviation degrees of various parameters and their influence weights on the overall operation of the device, and calculates the first device expected deviation coefficient. This coefficient intuitively reflects the depth of the current state of the device deviating from the normal expected state, providing a key basis for subsequent judgment of the device abnormal risk. Through this series of rigorous processes, the operation status of the IPC device can be accurately grasped, potential abnormal risks can be discovered in a timely manner, and the stable operation of the device can be ensured.

[0063] In a possible implementation manner, the device abnormal risk coefficient output unit includes: An activation abnormal risk prediction model acquisition unit, configured to perform feature activation on the device abnormal risk prediction unit according to the first abnormal risk prediction gating coefficient, and obtain a plurality of activated abnormal risk prediction models, where the device abnormal risk prediction unit includes a plurality of device abnormal risk prediction models.

[0064] An abnormal risk prediction coefficient acquisition unit, configured to input the first device expected deviation matrix into the plurality of activated abnormal risk prediction models, and obtain a plurality of abnormal risk prediction coefficients.

[0065] A risk prediction incentive coefficient acquisition unit, configured to perform proportion calculation according to the plurality of abnormal risk prediction accuracy parameters corresponding to the plurality of activated abnormal risk prediction models, and obtain a plurality of risk prediction incentive coefficients.

[0066] A weighted calculation unit, configured to perform weighted calculation on the plurality of abnormal risk prediction coefficients according to the plurality of risk prediction incentive coefficients, and obtain the first device abnormal risk coefficient.

[0067] Specifically, in the IPC device remote supervision platform based on the Internet of Things card, the following several units cooperate closely to accurately calculate the first device abnormal risk coefficient. The activation abnormal risk prediction model acquisition unit performs feature activation operations according to the first abnormal risk prediction gating coefficient. Assuming that the first abnormal risk prediction gating coefficient is a positive integer (for example, 4), this unit will accurately activate 4 models from the multiple device abnormal risk prediction models included in the device abnormal risk prediction unit, thereby obtaining 4 activated abnormal risk prediction models. These models are carefully designed and trained to predict and evaluate different types and degrees of device abnormal risks.

[0068] The abnormal risk prediction coefficient acquisition unit then inputs the first device expectation deviation matrix into these multiple activated abnormal risk prediction models. Each activated model will conduct in-depth analysis on the input expectation deviation matrix according to its own unique algorithm and internal logic. For example, the model will focus on factors such as the degree to which the key parameters of the device in the matrix deviate from the normal range, the duration of the deviation, and the mutual relationship between the parameters. Through a complex calculation process, it will finally output an abnormal risk prediction coefficient. In this way, each activated model will generate a corresponding abnormal risk prediction coefficient, thus obtaining multiple abnormal risk prediction coefficients.

[0069] The risk prediction incentive coefficient acquisition unit plays a key role in weight allocation in the entire IPC device abnormal risk prediction system. Its core task is to perform precise proportion calculation based on the multiple abnormal risk prediction accuracy parameters corresponding to the multiple activated abnormal risk prediction models, so as to obtain multiple risk prediction incentive coefficients. When receiving the abnormal risk prediction accuracy parameters corresponding to the activated abnormal risk prediction models, this unit will first comprehensively sort out and analyze these accuracy parameters. These accuracy parameters are obtained through repeated testing and strict verification of each model on a large amount of historical device operation data, and they accurately reflect the accuracy and reliability performance of each model in past risk predictions in a quantitative form. For example, an accuracy parameter indicates that a certain model has an accuracy rate of 85% when predicting abnormalities of a specific type of device, and another model may have an accuracy rate of 90% when predicting other types of abnormalities, etc. Then, using a professional proportion calculation algorithm, the abnormal risk prediction accuracy parameter of each model is compared and calculated with the sum of the accuracy parameters of all activated models. Specifically, for each activated abnormal risk prediction model, the calculation method of its corresponding risk prediction incentive coefficient is: the abnormal risk prediction accuracy parameter of this model divided by the sum of the abnormal risk prediction accuracy parameters of all activated models. Through such a calculation process, the relative importance weight of each activated model in the overall risk prediction can be determined, that is, multiple risk prediction incentive coefficients are obtained.

[0070] Finally, the weighted calculation unit conducts weighted calculation on the multiple abnormal risk prediction coefficients using the multiple risk prediction incentive coefficients. Each abnormal risk prediction coefficient will be multiplied by the corresponding risk prediction incentive coefficient, and then all the products are accumulated and summed. Through this weighted summation method, the prediction accuracy of each activated model and its contribution degree in the overall risk assessment are fully considered, and finally a comprehensive value, that is, the first device abnormal risk coefficient, is obtained. This coefficient accurately quantifies the degree of abnormal risk existing in the internal operation state of the IPC device, providing a reliable basis for judgment for the platform to timely decide whether it is necessary to maintain, adjust or take other corresponding measures for the device to ensure the stable operation of the IPC device and the normal work of monitoring.

[0071] In a possible implementation manner, the monitoring exception linkage tracking policy determination module 40 includes: A monitoring exception warning signal generation unit, configured to generate a monitoring exception warning signal based on the monitoring exception risk node.

[0072] A monitoring exception trend prediction result determination unit, configured to perform trend prediction based on the monitoring exception warning signal according to the monitoring exception recognition result corresponding to the monitoring exception risk node, and determine the monitoring exception trend prediction result.

[0073] An abnormal monitoring range fitting result acquisition unit, configured to perform monitoring range fitting according to the monitoring exception trend prediction result to obtain an abnormal monitoring range fitting result.

[0074] An associated IPC device set acquisition unit, configured to perform IPC device association recognition on the IPC remote collaboration unit according to the abnormal monitoring range fitting result, obtain an associated IPC device set, and load the IPC monitoring configuration set corresponding to the associated IPC device set.

[0075] A monitoring configuration optimization unit, configured to optimize the IPC monitoring configuration set according to the monitoring exception trend prediction result to generate the monitoring exception linkage tracking policy.

[0076] Specifically, in the IPC device remote supervision system based on the Internet of Things card, once a device detects an abnormal situation, it will quickly respond and start a full-range linkage tracking process, aiming to enhance the monitoring range and improve the efficiency of abnormal event handling. First, the monitoring exception warning signal generation unit is based on the rich information contained in the monitoring exception risk node. These risk nodes detail key elements such as the device identifier, abnormal type, and occurrence time of the abnormality. The unit uses built-in intelligent algorithms to deeply analyze this data and quickly generate a monitoring exception warning signal. This signal is reported immediately to inform that an abnormality has occurred and resources need to be allocated immediately to deal with it.

[0077] After receiving the monitoring anomaly warning signal, the monitoring anomaly trend prediction result determination unit retrieves a large amount of historical data related to the current anomaly type from the big data repository. This data covers various details of past similar anomaly events, such as the movement trajectory of the abnormal object, the duration and variation law of the abnormal behavior, the impact of environmental factors on the development of the anomaly, etc. Using the clustering analysis algorithm, these historical data are classified according to different features, such as clustering according to the category of the abnormal object, the scene environment where it appears, the time period when it occurs, etc., so as to quickly locate the historical sample set most similar to the current anomaly. Then, the time series analysis algorithm is introduced. For the selected similar historical sample set, the focus is on the change trend of the anomaly event over time. For example, in the case of an unknown object breaking into the monitoring area, analyze the change in the moving speed of such objects, the frequency of direction change, and the probability distribution of possible staying areas in different time periods in the past similar scenarios. At the same time, combined with the recurrent neural network (RNN) in deep learning, the long-term dependence relationship in the time series data is processed. Taking the initial state information of the current anomaly, such as the initial position, appearance time, and initial moving direction of the unknown object, as the input, the model makes dynamic predictions based on the laws learned from the historical data. The neurons inside the model continuously iterate and calculate, considering the mutual influence of various complex factors, and output multi-dimensional prediction results such as the possible subsequent movement path of the abnormal object, the change trend of the speed, and the probability distribution of the staying area. Finally, through the Bayesian decision theory, comprehensively consider the results obtained by the above various algorithms, and weigh the credibility of different prediction results. For example, according to the historical data statistics, the accuracy rates of the prediction results of different algorithms in the current environment are obtained, corresponding weights are assigned, and multiple prediction results are fused and optimized, so as to accurately determine the monitoring anomaly trend prediction result, providing highly forward-looking guidance for subsequent monitoring range fitting, device association scheduling, and configuration optimization.

[0078] Once the abnormal monitoring range fitting result acquisition unit receives the monitoring abnormal trend prediction result, it uses the data parsing algorithm to deeply mine the key information therein, accurately extracts the moving speed, direction and possible stopping points of abnormal objects, or accurately judges the core dynamic elements such as the starting position and diffusion rate of abnormal behaviors such as smoke diffusion and crowd gathering. Based on this information, a dynamic model is constructed. For abnormal objects moving in a straight line, vector operations and straight line extension algorithms are used. Starting from the initial coordinates of the object, combined with the speed direction and the preset tracking duration, the position coordinates of multiple subsequent time nodes are gradually calculated at fixed time intervals, and then a complete movement trajectory is outlined to determine the possible spatial range it may pass through. In the face of abnormal situations with diffusion characteristics, a diffusion simulation algorithm is adopted. Considering the real-time collected environmental parameters such as wind direction, wind speed, air circulation, etc., as well as the laws of similar abnormal diffusion in historical data, a three-dimensional diffusion model is constructed. As time progresses, the equation is continuously solved to dynamically present the spreading trend of diffusion objects such as smoke and crowd in space, and accurately define its influence area. Finally, optimization and correction are carried out with the help of Geographic Information System (GIS) technology. Detailed digital maps are retrieved, occlusion elements such as buildings and obstacles in the monitoring area are identified through spatial analysis algorithms, the monitoring line-of-sight obstruction situation is judged using ray tracing algorithms, and then the blind areas caused by occlusion are accurately calculated in combination with the principle of triangulation, and these areas are excluded from the initially delimited range. Finally, the abnormal monitoring range fitting result is finely fitted, providing a solid foundation guarantee for the subsequent efficient allocation of associated IPC devices and achieving non-blind-spot monitoring.

[0079] After the abnormal monitoring range fitting result is generated and output, the associated IPC device set acquisition unit immediately starts an efficient processing process. This unit takes the high-precision abnormal monitoring range fitting result as the key input and establishes a close data interaction link with the IPC remote collaboration unit. Using the area matching algorithm based on spatial geometric calculation, the three-dimensional space range information precisely defined by the fitting result is compared in detail with the deployment position parameters such as the longitude and latitude coordinates and installation height of each IPC device stored in the IPC remote collaboration unit. Through strict coordinate operations, spatial relationship quantities such as the Euclidean distance and azimuth angle of each device's position relative to the abnormal range boundary are accurately calculated, and based on this, multiple IPC devices located inside the potential risk area or capable of effectively covering this area according to parameters such as their viewing range and optical focal length are precisely selected. These strictly selected devices together form the associated IPC device set. After successfully identifying the associated IPC device set, this unit quickly activates the data loading module, and according to the preset data retrieval path and index rules, quickly retrieves the IPC monitoring configuration set that precisely matches it from the core database. This configuration set is stored in the form of structured data and covers a series of key technical indicators such as the resolution setting value of the camera optical imaging module, the frame rate parameter, the shooting angle control amount of the pan-tilt mechanical structure, and the bandwidth allocation strategy of the data transmission link, constituting the core configuration plan to ensure the high-performance operation of the device in a specific monitoring scenario.

[0080] Based on the prediction data obtained in real time from the monitoring anomaly trend prediction result determination unit, the monitoring configuration optimization unit performs in parallel deep optimization operations. If it is predicted based on the deep learning model and time series analysis that the abnormal object will show a high-speed movement trend, the optimization unit will send a frame rate increase instruction to the cameras in the associated IPC device set through the device control instruction set to ensure that the image acquisition frequency meets the requirements for capturing rapidly changing scenes and avoid the loss of key images. If it is determined based on the abnormal diffusion model that there is a trend for the abnormal behavior to spread to the surrounding space, the optimization unit will call the multi-camera collaborative scheduling algorithm, comprehensively consider the geometric relationships between the field of view angles, installation positions of each camera and the target area, and precisely adjust the shooting angles of the cameras to achieve seamless and omnidirectional optical coverage of the diffusion area. At the data transmission level, combining the quantization index of the emergency level of the abnormal event and the evaluation parameter of the scene complexity, the optimization unit uses the dynamic bandwidth allocation algorithm to re-plan and allocate the data transmission link bandwidth from the associated IPC device set to the monitoring center. For key devices in the abnormal core area or covering key monitoring nodes, higher bandwidth quotas are preferentially given, and through technical means such as traffic shaping and priority queues, it is ensured that the monitoring data collected in real time can be quickly transmitted to the monitoring center in a low-latency and highly reliable manner, and finally an adaptive, precise and efficient monitoring anomaly linkage tracking strategy is integrated and generated to drive the associated IPC devices to perform their respective duties and cooperate closely according to their optimized configuration parameters, building a solid and reliable technical defense line for dealing with various abnormal events.

[0081] In a possible implementation manner, the device abnormal collaborative operation and maintenance strategy acquisition module 50 includes: A device abnormal warning signal generation unit, configured to generate a device abnormal warning signal based on the device abnormal risk node.

[0082] A collaborative monitoring task acquisition unit, configured to collect the monitoring task parameters of the device abnormal risk node based on the device abnormal warning signal to obtain a collaborative monitoring task.

[0083] A monitoring task scheduling space establishment unit, configured to perform dynamic task adjustment on the IPC remote collaborative unit according to the collaborative monitoring task and establish a monitoring task scheduling space.

[0084] A monitoring quality optimization unit, configured to perform monitoring quality optimization according to the monitoring task scheduling space and generate the device abnormal collaborative operation and maintenance strategy.

[0085] Specifically, the device anomaly warning signal generation unit activates an efficient response mechanism based on the device anomaly risk nodes. This unit relies on the big data storage and retrieval system to quickly retrieve all-round data of the device anomaly risk nodes, covering the device physical address, fault diagnosis code, fault trigger time, and associated environmental sensor data (such as temperature, humidity, electromagnetic interference intensity) and real-time device operation parameters (such as CPU usage rate, memory occupancy, port traffic), etc. Using a deep neural network model, after training with a large number of fault samples, it accurately identifies the abnormal feature combinations and generates device anomaly warning signals in real time, which are broadcast to the bus in a specific data format (such as JSON format, containing key information such as warning level, unique identifier of the faulty device, etc.) to trigger subsequent emergency processes.

[0086] The collaborative monitoring task acquisition unit is instantly activated upon receiving the warning signal. It uses high-speed data acquisition and parsing technology to lock in the monitoring task parameters associated with the device anomaly risk nodes. Through a deep scan of the task queue, it extracts the geographical coordinate range of the monitoring area being executed by the faulty device (accurate to the longitude and latitude sub-intervals), the resolution required for the image acquisition task (adaptively set according to the complexity of the scene, such as high-definition 1080P or ultra-high-definition 4K), the preset frequency of the data transmission task (dynamically adjusted according to the network bandwidth, ranging from 1 frame per second to 30 frames per second), and the special monitoring requirements for the monitoring target (such as the tracking sensitivity for moving objects, visibility enhancement parameters under low light at night), etc. After data cleaning and integration, it generates collaborative monitoring tasks according to the standardized task template, providing precise guidance for subsequent scheduling.

[0087] The monitoring task scheduling space establishment unit takes the collaborative monitoring task as the input and establishes a strong interaction link with the IPC remote collaboration unit, starting a complex dynamic task adjustment program. Using the distributed resource scheduling algorithm, it real-time monitors the resource status of each neighboring device within the IPC remote collaboration unit, including the CPU idle cycle, remaining GPU computing power, available memory space, remaining storage device capacity, combined with factors such as the topological distance between devices (calculated through the network routing hop count and signal strength attenuation model) and the load balancing coefficient of the device (considering the historical task volume and the current task processing progress comprehensively). Using the integer programming model, it reasonably disassembles the original tasks of the faulty device and distributes them to the optimal neighboring device combination. At the same time, it reconstructs the data transmission path based on software-defined network (SDN) technology, optimizes the cooperation process between devices, and creates a monitoring task scheduling space that meets the requirements of real-time performance and stability, ensuring the continuity of monitoring.

[0088] Based on the monitoring task scheduling space, the monitoring quality optimization unit conducts fine optimization using multi-objective optimization algorithms. For the monitoring video quality, it adopts adaptive image processing algorithms, dynamically adjusts the camera exposure parameters (such as shutter speed and aperture size) according to the light intensity feedback from the light sensor, and switches the image resolution mode based on the scene analysis results (from the energy-saving mode with low resolution to the detail capture mode with high resolution). For the data transmission quality, combined with data such as bandwidth utilization rate and signal strength feedback by the real-time network monitoring module, it uses adaptive flow control and retransmission mechanisms. When the bandwidth is tight, it gives priority to ensuring the transmission of monitoring data in key areas, and adopts efficient HARQ (Hybrid Automatic Repeat reQuest) technology to reduce the packet loss rate and improve the transmission reliability. By iteratively optimizing the above key indicators, it generates a device anomaly collaborative operation and maintenance strategy highly adapted to the current fault scenario, and issues it to each device participating in operation and maintenance in the form of a structured instruction set to drive efficient operation. When the faulty device is repaired and resumes normal operation, with the help of the status synchronization protocol, it quickly establishes a connection with the faulty device, and uses the incremental data synchronization technology to accurately push the status update data missed during its offline period (such as network topology changes, new task allocation rules, device parameter optimization plans) to the faulty device, enabling it to seamlessly rejoin the collaborative network and quickly resume to the best collaborative operation and maintenance state, ensuring the robustness and reliability of the remote supervision of the entire IPC device.

[0089] In a possible implementation manner, the optimization management module 60 includes: An abnormal operation and maintenance integration module, configured to, if the monitoring abnormal risk node and the device abnormal risk node are the same IPC device, integrate the monitoring abnormal linkage operation and maintenance strategy and the device abnormal collaborative operation and maintenance strategy, generate an abnormal operation and maintenance integration strategy, and optimize and manage the IPC remote collaboration unit according to the abnormal operation and maintenance integration strategy.

[0090] Specifically, when the abnormal operation and maintenance integration module detects that the monitored abnormal risk node and the device abnormal risk node point to the same IPC device, the module immediately starts the in-depth integration process. First, it simultaneously retrieves the generated monitored abnormal linkage operation and maintenance strategy and the device abnormal collaborative operation and maintenance strategy. These two sets of strategies respectively contain detailed solutions for dealing with abnormalities at different levels of the device. For the monitored abnormal linkage operation and maintenance strategy, it covers detailed settings such as precisely scheduling associated IPC devices according to the abnormal trend prediction results, optimizing the monitoring configurations of each device to achieve efficient tracking, including multi-camera collaborative angle adjustment based on the prediction of the moving path of abnormal objects by a deep learning model, dynamically allocating transmission resources according to the real-time network bandwidth, and other key technical means; while the device abnormal collaborative operation and maintenance strategy focuses on how to quickly transfer tasks and ensure the continuity of monitoring when a device failure occurs, such as using a distributed resource scheduling algorithm to disassemble the tasks of the faulty device to adjacent devices and ensuring the picture quality of the device undertaking the task through adaptive image processing. The abnormal operation and maintenance integration module uses an intelligent integration algorithm to comprehensively analyze these two sets of strategies. On the one hand, it identifies the duplicate or conflicting instruction parts and makes choices according to the preset priority rules. For example, if the two have different requirements for the bandwidth allocation of the same device at a certain moment, the final allocation plan is determined according to the principle of ensuring the transmission of key monitored data; on the other hand, it explores the complementary and enhancing links, combining the precise abnormal prediction advantage in the monitored abnormal linkage operation and maintenance strategy with the efficient task transfer and resource reallocation expertise in the device abnormal collaborative operation and maintenance strategy. Through in-depth integration and optimization, an abnormal operation and maintenance integration strategy is generated. Subsequently, the module conducts a full-range of optimized management on the IPC remote collaborative unit according to this integration strategy. It sends precise instructions to each relevant device, adjusts the collaboration process between devices, and reconfigures the resource parameters to ensure that when multiple abnormalities overlap in the same IPC device, it can still maintain a stable and efficient operating state, maximizing the smooth execution of the monitoring task and the reliability of monitoring.

[0091] It should be noted that the above sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above specific embodiments of this specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0092] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0093] This specification and the drawings are merely illustrative of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. The IPC device remote supervision platform based on the Internet of Things card is characterized in that The platform includes: An IPC remote collaboration unit construction module for interconnecting multiple IPC devices to construct an IPC remote collaboration unit, where each IPC device is built-in with an IoT card; An IPC node dataset reading module for reading multiple IPC node datasets according to the IPC remote collaboration unit, where each IPC node dataset includes the monitoring video data and device status data corresponding to each IPC device; An abnormal risk inspection module for performing abnormal risk inspection on the multiple IPC node datasets according to the three channels of IPC abnormal risk inspection to determine monitoring abnormal risk nodes and device abnormal risk nodes; A monitoring abnormal linkage tracking strategy determination module for making an abnormal linkage tracking decision on the IPC remote collaboration unit according to the monitoring abnormal risk nodes to determine a monitoring abnormal linkage tracking strategy; A device abnormal collaborative operation and maintenance strategy acquisition module for making a task collaborative operation and maintenance decision on the IPC remote collaboration unit according to the device abnormal risk nodes to obtain a device abnormal collaborative operation and maintenance strategy; An optimization management module for optimizing and managing the IPC remote collaboration unit according to the monitoring abnormal linkage operation and maintenance strategy and the device abnormal collaborative operation and maintenance strategy.

2. The IPC device remote supervision platform based on the IoT card according to claim 1, characterized in that, The three channels of IPC abnormal risk inspection include an IPC monitoring abnormal risk detection channel, an IPC device abnormal risk prediction channel, and an IPC risk abnormal verification channel. The abnormal risk inspection module includes: A device status data extraction unit for extracting the first monitoring video data and the first device status data corresponding to the first IPC device according to the multiple IPC node datasets; A monitoring abnormal risk coefficient acquisition unit for inputting the first monitoring video data into the IPC monitoring abnormal risk detection channel to obtain a first monitoring abnormal risk coefficient; A device abnormal risk coefficient acquisition unit for inputting the first device status data into the IPC device abnormal risk prediction channel to obtain a first device abnormal risk coefficient; An IPC risk abnormal verification unit for inputting the first monitoring abnormal risk coefficient and the first device abnormal risk coefficient into the IPC risk abnormal verification channel, where the IPC risk abnormal verification channel includes a monitoring abnormal risk threshold and a device abnormal risk threshold; A device abnormal risk node addition unit for adding the first IPC device to the monitoring abnormal risk node if the first monitoring abnormal risk coefficient is greater than or equal to the monitoring abnormal risk threshold, and adding the first IPC device to the device abnormal risk node if the first device abnormal risk coefficient is greater than or equal to the device abnormal risk threshold.

3. The IPC device remote supervision platform based on the IoT card according to claim 2, wherein, The IPC monitoring abnormal risk detection channel includes multiple monitoring abnormal risk detection branches corresponding to the multiple IPC devices. The monitoring abnormal risk coefficient acquisition unit includes: The monitoring anomaly risk detection branch activation unit is used to activate the first monitoring anomaly risk detection branch corresponding to the first IPC device according to the IPC monitoring anomaly risk detection channel, where the first monitoring anomaly risk detection branch includes a first monitoring anomaly recognition model and a first monitoring anomaly risk prediction model; The monitoring anomaly recognition result acquisition unit is used to input the first monitoring video data into the first monitoring anomaly recognition model to obtain a first monitoring anomaly recognition result; The monitoring anomaly risk coefficient generation unit is used to input the first monitoring anomaly recognition result into the first monitoring anomaly risk prediction model to generate the first monitoring anomaly risk coefficient.

4. The IPC device remote supervision platform based on the Internet of Things card according to claim 2, characterized in that The IPC device anomaly risk prediction channel includes an IPC device status recording unit, an anomaly risk prediction gating unit, and a device anomaly risk prediction unit. The device anomaly risk coefficient acquisition unit includes: The device expected deviation coefficient acquisition unit is used to perform expected deviation detection on the first device status data according to the IPC device status recording unit to obtain a first device expected deviation matrix and a first device expected deviation coefficient; The anomaly risk prediction gating coefficient acquisition unit is used to input the first device expected deviation coefficient into the anomaly risk prediction gating unit to obtain a first anomaly risk prediction gating coefficient; The device anomaly risk coefficient output unit is used to output the first device anomaly risk coefficient based on the first anomaly risk prediction gating coefficient and the first device expected deviation matrix according to the device anomaly risk prediction unit.

5. The IPC device remote supervision platform based on the Internet of Things card according to claim 4, characterized in that, The device expected deviation coefficient acquisition unit includes: The IPC device scenario information acquisition unit is used to collect the control parameters and environmental parameters corresponding to the first IPC device to obtain the first IPC device scenario information; The IPC device attribute acquisition unit is used to collect the specification and model parameters corresponding to the first IPC device to obtain the first IPC device attribute; The device normal state sample set acquisition unit is used to perform normal state sample retrieval on the IPC device status recording unit according to the first IPC device scenario information and the first IPC device attribute to obtain a first device normal state sample set; The device expected state acquisition unit is used to perform central tendency analysis according to the first device normal state sample set to obtain a first device expected state; The deviation detection unit is used to perform deviation detection on the first device status data according to the first device expected state to generate the first device expected deviation matrix; The deviation depth evaluation unit is used to perform deviation depth evaluation according to the first device expected deviation matrix to generate the first device expected deviation coefficient.

6. The IPC device remote supervision platform based on the IoT card according to claim 4, wherein, The device anomaly risk coefficient output unit includes: The activation anomaly risk prediction model acquisition unit is used to perform feature activation on the device anomaly risk prediction unit according to the first anomaly risk prediction gating coefficient to obtain multiple activation anomaly risk prediction models, where the device anomaly risk prediction unit includes multiple device anomaly risk prediction models; Anomaly risk prediction coefficient acquisition unit, configured to input the first device expected deviation matrix into the multiple activated anomaly risk prediction models to obtain multiple anomaly risk prediction coefficients; Risk prediction incentive coefficient acquisition unit, configured to perform proportion calculation according to the multiple anomaly risk prediction accuracy parameters corresponding to the multiple activated anomaly risk prediction models to obtain multiple risk prediction incentive coefficients; Weighted calculation unit, configured to perform weighted calculation on the multiple anomaly risk prediction coefficients according to the multiple risk prediction incentive coefficients to obtain the first device anomaly risk coefficient.

7. The IPC device remote supervision platform based on the IoT card according to claim 1, characterized in that, The monitoring anomaly linkage tracking strategy determination module includes: Monitoring anomaly warning signal generation unit, configured to generate a monitoring anomaly warning signal based on the monitoring anomaly risk node; Monitoring anomaly trend prediction result determination unit, configured to perform trend prediction based on the monitoring anomaly warning signal according to the monitoring anomaly recognition result corresponding to the monitoring anomaly risk node to determine the monitoring anomaly trend prediction result; Anomaly monitoring range fitting result acquisition unit, configured to perform monitoring range fitting according to the monitoring anomaly trend prediction result to obtain an anomaly monitoring range fitting result; Associated IPC device set acquisition unit, configured to perform IPC device association recognition on the IPC remote collaboration unit according to the anomaly monitoring range fitting result to obtain an associated IPC device set, and load the IPC monitoring configuration set corresponding to the associated IPC device set; Monitoring configuration optimization unit, configured to optimize the IPC monitoring configuration set according to the monitoring anomaly trend prediction result to generate the monitoring anomaly linkage tracking strategy.

8. The IPC device remote supervision platform based on the IoT card according to claim 1, wherein The device anomaly collaborative operation and maintenance strategy acquisition module includes: Device anomaly warning signal generation unit, configured to generate a device anomaly warning signal based on the device anomaly risk node; Collaborative monitoring task acquisition unit, configured to collect the monitoring task parameters of the device anomaly risk node based on the device anomaly warning signal to obtain a collaborative monitoring task; Monitoring task scheduling space establishment unit, configured to perform dynamic task adjustment on the IPC remote collaboration unit according to the collaborative monitoring task to establish a monitoring task scheduling space; Monitoring quality optimization unit, configured to perform monitoring quality optimization according to the monitoring task scheduling space to generate the device anomaly collaborative operation and maintenance strategy.

9. The IPC device remote supervision platform based on the Internet of Things card according to claim 1, characterized in that, The platform further includes an anomaly operation and maintenance integration module, configured to, if the monitoring anomaly risk node and the device anomaly risk node are the same IPC device, integrate the monitoring anomaly linkage operation and maintenance strategy and the device anomaly collaborative operation and maintenance strategy to generate an anomaly operation and maintenance integration strategy, and perform optimization management on the IPC remote collaboration unit according to the anomaly operation and maintenance integration strategy.

Images