Intelligent well lid monitoring management system with state monitoring

Abstract

This invention discloses an intelligent manhole cover monitoring and management system with status monitoring capabilities. The system involves intelligent monitoring of manhole covers, including: establishing a monitoring and control relationship for the area where the manhole cover is located; identifying abnormal trends based on real-time status information; locating candidate manhole covers and confirming local structures in abnormal road areas; and extracting the structural changes of the manhole cover relative to the road surface. The real-time status information, structural confirmation results, and structural changes are then used for disturbance decoupling to filter out road disturbances and identify real risk events. These real risk events are then introduced into the control link to generate corresponding control states, further executing local control, hazard mitigation linkage, regional manhole group collaborative defense, and automatic reset control. This not only improves the accuracy of identifying and handling abnormal manhole cover structures, water accumulation, and hazardous environments, but also enhances the risk diffusion suppression capability in complex scenarios and the overall safety of urban underground facility management.

Description

Technical Field

[0001] This invention relates to the field of intelligent manhole cover monitoring technology, and in particular to an intelligent manhole cover monitoring and management system with status monitoring capabilities. Background Technology

[0002] With the increasing development and utilization of urban underground space, underground drainage, gas, electricity, and communication pipelines are becoming increasingly dense. As a crucial component connecting surface roads and underground manholes, the operational status of manhole covers directly affects road traffic safety, the safety of underground facilities, and the stability of the city's lifeline system. Existing intelligent manhole cover monitoring and management technologies typically utilize sensors for vibration, tilt angle, displacement, water level, gas, temperature, and humidity detection deployed inside the manhole cover or manhole. These sensors, combined with wireless communication and control terminals, enable manhole cover status data collection, anomaly alarms, and remote monitoring. Simultaneously, focusing on manhole cover target identification and structural assessment, a technical approach based on road point cloud candidate localization, local structured light point cloud acquisition, road surface plane extraction, and manhole cover edge geometric recognition has emerged. This approach identifies manhole cover locations and extracts information on structural changes relative to the road surface.

[0003] However, existing technologies still have the following two shortcomings: First, existing manhole cover status monitoring systems typically use single-item anomalies such as vibration, tilt angle, displacement, water level, or gas as the basis for alarms or control triggers. They lack a technical mechanism that incorporates real-time status information, local structural confirmation results, and structural changes of the manhole cover relative to the road surface into the pre-control judgment stage. Therefore, in scenarios involving vehicle crushing, construction impacts, or short-term road disturbances, road disturbances are easily misidentified as real risks to the manhole cover, leading to false triggering of locking, warnings, or linkage controls. Second, most existing technologies remain at the level of single manhole cover status monitoring, single-point alarms, or single-time structural detection. They have not yet further integrated the manhole cover structural confirmation results, the underground environment status, and the correlation between the manhole group and the surrounding area to form an integrated closed-loop control mechanism that combines local control, hazard mitigation linkage, and regional collaborative defense. Therefore, in scenarios involving water accumulation, hazardous gas propagation, and risk spillover from key traffic areas, it is difficult to achieve hierarchical collaborative handling from the source manhole cover to associated manhole covers and linkage equipment. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent manhole cover monitoring and management system with status monitoring to solve the problems of high false triggering rate of road disturbances, insufficient ability to screen real risks, and incomplete regional collaborative prevention and control in existing methods.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides an intelligent manhole cover monitoring and management system with status monitoring capabilities, comprising:

[0008] The relationship establishment module is used to establish monitoring and control relationships for the area where the manhole cover is located;

[0009] The trend recognition and positioning module is used to identify abnormal trends based on the real-time status information of manhole covers, and to locate candidate manhole covers in the road area where the manhole covers are located after identifying abnormal trends.

[0010] The structure confirmation and extraction module is used to obtain local structural information around the candidate manhole cover location, confirm the structure of the candidate manhole cover and extract the structural change of the manhole cover relative to the road surface;

[0011] The disturbance decoupling module is used to input real-time status information, structural confirmation results, and structural changes into the disturbance decoupling process to distinguish between road disturbance events and real risk events.

[0012] The status generation module is used to filter real risk events into the control link, and to perform risk fusion on the events entering the control link to generate corresponding control statuses.

[0013] The linkage control module is used to execute local control of the manhole cover based on the control status; execute hazard mitigation linkage control when a hazardous environment is determined to exist; and execute collaborative joint defense control of associated manhole covers when the control status reaches the joint defense trigger condition.

[0014] The recovery module is used to release control and restore normal monitoring status after the manhole cover status, environmental status, and joint defense status have returned to normal.

[0015] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps for establishing a monitoring and control relationship for the area where the manhole cover is located are as follows:

[0016] Collect basic attribute information of manhole covers and generate basic files for a single manhole cover. Connect the monitoring unit and the execution unit to the controller to form a corresponding integrated manhole cover monitoring and control object.

[0017] A regional well group association table is established based on the underground connectivity and road adjacency relationships. Control parameters are initialized according to the regional attributes of the well cover. After the parameters are written and the equipment is online for verification, the corresponding well cover object is set to a controllable and active state.

[0018] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps for identifying abnormal trends based on real-time manhole cover status information are as follows:

[0019] Collect real-time status information corresponding to activated manhole cover objects, synchronously acquire and benchmark correct vibration, tilt angle, displacement, water level, gas, temperature and humidity and electronic lock status, and update the current abnormal duration by combining the abnormal duration register value cached by the edge controller, and generate the original state vector at the current moment.

[0020] Based on the joint change relationship of vibration, tilt angle, displacement and abnormal duration in the original state vector, the real-time abnormal trend of the manhole cover is identified, and the corresponding manhole cover is judged to be in a suspected abnormality triggered state when the structural abnormality trend condition is met.

[0021] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the step of locating candidate manhole covers in the road area where the manhole cover is located after identifying an abnormal trend is as follows:

[0022] Obtain the original road point cloud of the road area where the manhole cover is located, corresponding to the suspected abnormal trigger state of the structure, and sequentially perform statistical outlier filtering, connected component filtering, ground segmentation and common domain filtering to obtain candidate search point cloud that retains only the common domain road ground.

[0023] The candidate search point cloud is rasterized according to the preset slice size and sampling interval, and a single-channel intensity ground map is constructed based on the intensity information of neighboring points;

[0024] Perform fully convolutional spatial classification inference, candidate window clustering, and activation center localization on the single-channel intensity ground map to generate a set of candidate manhole cover centers.

[0025] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps of acquiring local structural information around the candidate manhole cover location, confirming the structure of the candidate manhole cover, and extracting the structural change of the manhole cover relative to the road surface are as follows:

[0026] A local structure sampling window is established around the center of the candidate manhole cover to obtain the local structure light point cloud of the corresponding candidate manhole cover location. Voxel downsampling, median filtering, statistical filtering and radius filtering are performed in sequence to obtain the purified local structure point cloud.

[0027] Based on the purified local structural point cloud, the road reference plane is extracted and the out-of-surface point set is separated. The out-of-surface point set is then clustered to obtain candidate structure clusters.

[0028] Perform manhole cover edge feature discrimination on candidate structure clusters, generate manhole cover structure confirmation marks, and extract the structural change of the confirmed manhole cover structure clusters relative to the road reference surface.

[0029] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps for distinguishing between road disturbance events and actual risk events are as follows:

[0030] The original state vector, manhole cover structure confirmation flag, and structural change quantity are merged into a disturbance decoupling discrimination input group, and the event type discrimination is performed in the order of first judging manhole cover structure abnormal events, then judging underground environmental risk events, and finally judging road disturbance events.

[0031] After completing the event type identification, a disturbance decoupling result object is generated, which includes the event type, the original state vector, the manhole cover structure confirmation flag, and the structural change amount.

[0032] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps for generating the corresponding control status are as follows:

[0033] Read the disturbance decoupling result object, first block the road disturbance event outside the control closed loop, and write the manhole cover structure abnormal event or the underground environment risk event into the effective control input object;

[0034] The system performs rule-based fusion on the structural risk information, water accumulation risk information, and hazardous environment risk information in the effective control input object to generate a unique control state among the structural instability state, water accumulation risk state, hazardous environment state, or composite risk state.

[0035] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps of performing local control of the manhole cover based on the control status are as follows:

[0036] Based on the unique control status, local control of this manhole cover is executed, and corresponding control is performed on the electronic lock, manhole warning, sampling frequency and manhole cover opening permission, generating a local control result object;

[0037] When the local control result indicates that the manhole cover is in a hazardous environment or a complex risk state, the ventilation device is invoked to perform the risk mitigation linkage control, and the changes in hazardous gas concentration, temperature and humidity are continuously collected during the risk mitigation process to form a risk mitigation monitoring sequence.

[0038] The step of executing hazard mitigation linkage control when a hazardous environment is determined to exist refers to determining the completion of hazard mitigation based on the continuous downward trend of the relevant state quantities of the hazardous environment according to the hazard mitigation monitoring sequence, and generating a hazard mitigation linkage result object containing a local control result object and a hazard mitigation completion flag.

[0039] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the step of executing collaborative joint defense control of the associated manhole covers when the control status reaches the joint defense trigger condition includes the following steps:

[0040] Read the local control result object and the risk mitigation linkage result object, determine the joint defense trigger type based on the unique control status and risk mitigation completion flag, and extract the corresponding associated well cover and associated equipment from the regional well group association table;

[0041] Based on the joint defense trigger type, collaborative joint defense control instructions are issued to associated manhole covers and associated equipment, generating regional collaborative joint defense result objects.

[0042] As a preferred embodiment of the intelligent manhole cover monitoring and management system with status monitoring described in this invention, the steps of executing the release of control and restoring the normal monitoring state are as follows:

[0043] Risk clearance is determined based on the recovery status of manhole cover condition, environmental condition, and joint prevention status. Control is lifted and routine monitoring is restored when normal conditions are restored.

[0044] The beneficial effects of this invention are as follows: By coordinating the real-time status information of manhole covers, the location results of candidate manhole covers, the confirmation results of local structures, and the structural changes of manhole covers relative to the road surface, this invention first decouples road disturbance events from actual risk events, and then only filters actual risk events into the control link, thereby effectively reducing false triggering problems caused by factors such as vehicle rolling and construction impact. At the same time, by combining this manhole cover control, hazardous environment risk mitigation linkage, and regional manhole group collaborative prevention and control, it realizes the improvement from single-point monitoring and alarm to hierarchical closed-loop management and control. This not only improves the accuracy of identifying and handling manhole cover structural anomalies, water accumulation, and hazardous environments, but also enhances the risk diffusion suppression capability in complex scenarios and the overall safety of urban underground facility management. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of an intelligent manhole cover monitoring and management system with status monitoring capabilities.

[0047] Figure 2 A flowchart for the process of establishing relationships and identifying and locating trends.

[0048] Figure 3 This is a flowchart illustrating the regional joint defense relationship.

[0049] Figure 4 A flowchart for establishing a monitoring and control relationship. Detailed Implementation

[0050] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0051] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0053] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides an intelligent manhole cover monitoring and management system with status monitoring, including the following steps:

[0054] S1, Relationship Establishment Module, is used to establish monitoring and control relationships for the area where the manhole cover is located;

[0055] S1.1: Collect basic attribute information of manhole covers and generate basic files for a single manhole cover. Connect the monitoring unit and the execution unit to the controller to form a corresponding integrated manhole cover monitoring and control object.

[0056] Specifically, taking a single manhole cover as the basic management unit, the basic attribute information corresponding to the manhole cover is collected and written into the basic file corresponding to the manhole cover to form a basic file for a single manhole cover. Following the principle that one manhole cover corresponds to one set of fixed monitoring devices and one set of fixed execution devices, vibration sensors, tilt sensors, displacement sensors, water level sensors, hazardous gas sensors, temperature and humidity sensors, electronic locks, audible and visual warning devices, ventilation execution devices, structured light camera acquisition units, and inspection point cloud acquisition units are registered for the manhole cover. The device address, communication port, installation location, and controller address of each device are recorded to generate a device mapping table. Each sensor and execution device in the device mapping table is connected to the corresponding acquisition port and execution port of the edge controller, and a fixed communication link is established between the sensor acquisition port, status register area, and execution output port. Then, a master-slave communication relationship is established between the edge controller and the central controller, enabling the central controller to uniformly read status data and issue control commands. This generates an integrated manhole cover monitoring and control object.

[0057] It should be noted that the basic attribute information includes at least: manhole cover number, latitude and longitude coordinates, road name, road grade, type of underground pipeline, manhole depth, manhole diameter, manhole cover material, manhole cover load-bearing capacity, management responsibility unit, and on-site power supply and communication conditions; the manhole cover number is generated using a fixed coding rule of "regional code + road code + sequential number".

[0058] S1.2: Establish a regional well group association table based on the underground connectivity and road adjacency, and initialize the control parameters according to the regional attributes of the well cover. After completing the parameter writing and equipment online verification, set the corresponding well cover object to a controllable and active state.

[0059] Specifically, taking the road area where the manhole cover is located as a unit, the relationships of all manhole cover objects within the same area are registered. First, based on the existing underground pipeline network map or on-site manhole chamber upstream and downstream connectivity information, the upstream and downstream associated manhole covers of each manhole cover are registered. Then, based on the road plane adjacency relationship, adjacent intersection manhole covers are registered, ultimately generating a regional manhole group association table. Parameters for each manhole cover object are initialized according to fixed rules based on road level and underground pipeline type. After completing object layering, vibration trigger thresholds, tilt trigger thresholds, displacement trigger thresholds, and structural change thresholds are written, and a parameter table is generated. The parameter table is written to the parameter register area of ​​the corresponding edge controller, and a parameter image is synchronously saved in the central controller. Subsequently, online detection commands are sent sequentially to each sensor and execution device in the device mapping table, the number of successfully responding devices is counted, and the online effectiveness rate is calculated. When When it is determined that all monitoring and execution devices for the manhole cover are connected, the manhole cover is marked as activated; when At that time, the device mapping table is rechecked and the device is reconnected; after completion, an activated integrated manhole cover monitoring and control object is obtained.

[0060] Online computing efficiency The expression is:

[0061]

[0062] In the formula, Indicates the number of devices that successfully responded. , Indicates the total number of devices that should be connected;

[0063] It should be noted that the relationship registration rules include: First, if manhole cover j and manhole cover i are located upstream of each other in the same pipe segment, then manhole cover j is recorded as the upstream associated manhole cover of manhole cover i; Second, if manhole cover j and manhole cover i are located downstream of each other in the same pipe segment, then manhole cover j is recorded as the downstream associated manhole cover of manhole cover i; Third, if manhole cover j and manhole cover i are located in adjacent manhole locations in the same road segment or within the warning range of the same intersection, then manhole cover j is recorded as the road adjacent manhole cover of manhole cover i.

[0064] The specific rules for parameter initialization are as follows: when the manhole cover is located on a main road and the underground pipeline type is a gas well, power well, or integrated utility tunnel access well, it is initialized as a high-alert object; when the manhole cover is located on a secondary road or the underground pipeline type is a drainage well or sewage well, it is initialized as a medium-alert object; when the manhole cover is located on a branch road and the underground pipeline type is a communication well or general maintenance well, it is initialized as a low-alert object.

[0065] S2, Trend Recognition and Positioning Module, is used to identify abnormal trends based on real-time status information of manhole covers, and to locate candidate manhole covers in the road area where the manhole covers are located after identifying abnormal trends.

[0066] S2.1: Collect real-time status information corresponding to the activated manhole cover object, synchronously acquire and benchmark correct the vibration, tilt angle, displacement, water level, gas, temperature and humidity and electronic lock status, and update the current abnormal duration by combining the abnormal duration register value cached by the edge controller, and generate the original state vector at the current moment.

[0067] Specifically, the activated integrated manhole cover monitoring and control object and its parameter table are invoked to read the corresponding vibration trigger threshold. Tilt angle trigger threshold Displacement trigger threshold and fixed sampling period The edge controller uses a fixed sampling period as the current state update period for the manhole cover object. At the same sampling time, it synchronously acquires data from the vibration sensor, tilt sensor, displacement sensor, water level sensor, hazardous gas sensor, temperature and humidity sensor, and electronic lock status interface corresponding to the current manhole cover object, obtaining the original measurement set at time t. It reads the vibration static reference, tilt installation reference, displacement closure reference, static liquid level reference, background gas reference, ambient temperature reference, and ambient humidity reference of the current manhole cover object, and performs reference correction on the original measurement set to obtain the current state quantity. After completing the calculation of the current state quantity, the edge controller first reads the abnormal duration register value cached by the manhole cover object in the previous sampling period. The current abnormal duration is updated based on whether the current structure-related state variables are in the abnormal range; the current state variables and the abnormal duration are written together into the original state vector at the current moment.

[0068] The benchmark calibration adopts a unified processing method of "subtracting the corresponding benchmark value from the original measurement value", and its representative expression is as follows:

[0069]

[0070] In the formula, This represents the original measurement value of a certain state variable at time t. This represents the reference value corresponding to the state variable. This represents the state quantity after reference correction.

[0071] Specifically, vibration intensity From the original vibration measurement value Compared with the static reference of vibration Correction yielded tilt offset From the original measurement value of the tilt angle With installation reference The corrected displacement offset From the original displacement measurement value With closed reference The corrected water level change Gas concentration change Temperature change and humidity change All values ​​are obtained by subtracting from their respective reference values; the vibration intensity is expressed in absolute value form, i.e.:

[0072]

[0073] The rules for updating the duration of an anomaly are as follows:

[0074]

[0075]

[0076] In the formula, This indicates the structural anomaly indicator at the current moment. This indicates the cumulative duration of the anomaly since the previous sampling time. This indicates the fixed sampling period for the current manhole cover object.

[0077] It should be noted that an abnormal duration register is set up in the edge controller to store the abnormal duration register value of the previous sampling period. The initial value is set to 0;

[0078] The value range is 1.5 to 2.5 times the average normal vibration of the manhole cover, preferably 2.0 times. The setting method is as follows: based on vibration amplitude data from normal operation samples, vehicle rolling disturbance samples, and actual structural anomaly samples of the manhole cover, offline statistical analysis combined with cross-validation is used to calculate the anomaly detection rate and false trigger rate under different vibration thresholds. The target is determined by maximizing the actual structural anomaly detection rate and minimizing the vehicle ballast false alarm rate. The range of values ​​and preferred values;

[0079] The value range is from 0.5 degrees to 2.0 degrees, preferably 1.0 degrees. The setting method is as follows: using the initial posture of the manhole cover after installation as a baseline, inclination offset data are collected for normal manhole covers, slightly shaking manhole covers, and genuinely tilted manhole covers. A stratified sample comparison and threshold scanning method is used to calculate the stability and sensitivity under different inclination thresholds, thereby determining the value that can identify continuous posture changes without excessively responding to instantaneous shaking. Range of values;

[0080] The value range is 2 mm to 10 mm, preferably 5 mm. The setting method is as follows: using the fully closed position of the manhole cover as the zero displacement reference, displacement offset data are collected under normal closed, loosely connected, rebounding, and open states of the manhole cover. A combination of historical sample playback and graded loading tests is used to calculate the opening recognition rate and false alarm rate under different displacement thresholds, thereby determining... The range of values ​​and preferred values.

[0081] S2.2: Based on the joint change relationship of vibration, tilt angle, displacement and abnormal duration in the original state vector, the real-time abnormal trend of the manhole cover is identified, and the corresponding manhole cover is judged as a suspected structural abnormality trigger state when the structural abnormality trend condition is met.

[0082] Specifically, based on the vibration intensity, tilt offset, displacement offset, and anomaly duration in the original state vector, a joint discrimination is performed to calculate the structural anomaly trend discrimination value at the current moment. ;

[0083] When satisfied

[0084]

[0085] And simultaneously satisfy

[0086]

[0087] When the current manhole cover is in a state of suspected structural abnormality, the manhole cover number and its surrounding road area are output as the input area for the next inspection point cloud data collection. When the above conditions are not met simultaneously, the current manhole cover is in a state of normal monitoring and normal sampling continues.

[0088] Calculate the structural anomaly trend discriminant at the current moment. The expression is:

[0089]

[0090] In the formula, This indicates the preset reference duration for the current manhole cover object's anomaly. , , and These represent the discrimination weights for vibration, tilt angle, displacement, and duration, respectively.

[0091] It should be noted that in some embodiments, The value range is 0.25 to 0.35, with 0.30 being preferred. The setting method is as follows: based on the labeled normal manhole cover samples, vehicle rolling disturbance samples, and real abnormal manhole cover structural samples, the constraint grid search combined with cross-validation method is used to perform offline traversal calculation of the vibration term weight, and the value range and preferred value are determined with the maximum structural abnormality identification rate and the lowest false trigger rate as the optimization objective.

[0092] The value range is 0.20 to 0.30, preferably 0.25. The setting method is as follows: taking the offset of the tilt sensor relative to the installation reference surface as input, the layered sampling verification method is used to compare the discrimination results of manhole cover lifting, manhole cover loosening and short-term road impact scenarios under different values, and select the weight range that can enhance the recognition of continuous attitude deviation and suppress the misjudgment of instantaneous shaking.

[0093] The value range is 0.20 to 0.30, preferably 0.25. The setting method is as follows: taking the offset of the displacement sensor relative to the closed reference position as input, the historical sample playback calculation method is used to perform group calculations on the manhole cover opening, loose connection, rebound and instantaneous vehicle ballast scenarios under different values, and the value range and preferred value are determined by taking the highest detection rate of real opening or loosening events and the lowest false alarm rate of instantaneous ballast as constraints.

[0094] The value range is 0.15 to 0.25, preferably 0.20. The setting method is as follows: taking the duration of the anomaly as input, the time window cumulative simulation calculation method is used to compare the separation effect of short-term disturbance events and continuous anomaly events under different values, and select the weight range that can effectively strengthen the characterization of continuous anomaly trend without weakening the dominant role of vibration, tilt and displacement terms.

[0095] The value range is 3 ~10 Preferably 5 Its setting method is: based on a fixed sampling period Time window cumulative statistics and simulation calculations were performed on instantaneous vehicle impacts, short-term construction disturbances, and real continuous abnormal events. The separability of short-term disturbances and continuous anomalies under different duration thresholds was statistically analyzed, and a continuous reference duration that can effectively distinguish between instantaneous disturbances and continuous anomaly trends was selected as the basis for analysis. The range of values ​​and preferred values.

[0096] S2.3: Obtain the original road point cloud of the road area where the manhole cover is located, corresponding to the suspected abnormal trigger state of the structure, and sequentially perform statistical outlier filtering, connected component filtering, ground segmentation and common domain filtering to obtain candidate search point cloud that retains only the common domain road ground;

[0097] Specifically, when the current manhole cover enters a state of suspected structural anomaly, the central controller reads the manhole cover number and its corresponding installation location, road area, and associated relationship with the manhole group in the basic file, and calls the inspection point cloud acquisition unit to obtain the original road point cloud of the road area; the central controller performs nearest neighbor distance statistics on each point in the point cloud, and calculates the average Euclidean distance from the i-th point to its k nearest neighbors. And based on the mean of the average neighborhood distance of all points and standard deviation Perform outlier detection;

[0098] If satisfied

[0099]

[0100] Then the i-th point is identified as an outlier and deleted; where Indicates the outlier factor;

[0101] After filtering out outliers, connected component filtering is performed on the remaining point cloud. Points with an Euclidean distance less than the separation distance d are grouped into the same cluster, and the number of points in each cluster is counted. If the number of points in a cluster is less than the minimum cluster threshold C, then... minIf the entire point cloud cluster is deleted, the point cloud after connected component filtering is obtained. Then, CSF cloth simulation ground segmentation is performed on it, that is, the point cloud is first reversed along the Z-axis, and then cloth falling simulation is performed on the reversed point cloud to form a road digital terrain surface. Points that are less than the ground segmentation threshold from the digital terrain surface are determined as ground points, and the rest are determined as non-ground points. After the ground segmentation is completed, RoI common domain filtering is performed on the ground point cloud to obtain candidate search point clouds.

[0102] It should be noted that the RoI public domain filtering is as follows: the ground point cloud is spatially overlaid with the vector boundary of the municipal road maintenance area, and only the ground point cloud of the public domain within the boundary is retained, while the point clouds of the corresponding private area ground, building steps and green edge outside the boundary are deleted; when there is no vector boundary of the municipal road maintenance area, the ground point cloud after ground segmentation is directly used as the candidate search point cloud.

[0103] Outlier coefficient The value range can be set to 1.5 to 2.5, with 2 being preferred. The basis for this is: first, the mean and dispersion of the neighborhood distance of each point in the road point cloud are statistically analyzed, and then the noise point removal effect and the road main point retention effect are compared under different multiple conditions to determine the value.

[0104] Minimum cluster threshold C min The value range can be set to 5000~15000, with 10000 being preferred. The basis for this is: performing connected component clustering on the denoised road point cloud, and comparing the ghost pseudo-cluster deletion effect and the road main point cloud retention effect under different cluster size thresholds to determine the value.

[0105] The ground segmentation threshold can be set to a range of 0.2 meters to 0.4 meters, with 0.3 meters being preferred. This is determined by comparing the stability of ground point recognition and the misclassification of non-ground points under different point-to-ground distance thresholds during the ground segmentation process in fabric simulation.

[0106] S2.4: Rasterize the candidate search point cloud according to the preset slice size and sampling interval, and construct a single-channel intensity ground map based on the intensity information of neighboring points;

[0107] Specifically, the candidate search point cloud is sliced ​​into 50m×50m slices, with a 5m overlap between adjacent slices. Then, a regular grid is established with a ground sampling interval of 2.5cm, and neighborhood points are searched within a radius of 2.5cm for each grid center point (u,v). The intensity value of the grid is calculated based on the original intensity of the neighborhood points and their distance from the grid center. The above intensity assignment is repeated for all grid centers, and the pixel intensity values ​​I(u,v) of all grids are arranged into a two-dimensional pixel matrix in row and column coordinate order, thereby generating a single-channel intensity ground map.

[0108] Calculate the intensity value of the grid. The expression is:

[0109]

[0110] In the formula, This represents the original intensity value of the j-th neighboring point. This represents the weight of the j-th neighboring point. This represents the number of neighboring points.

[0111] It should be noted that the regular grid is established with the road plane reference point corresponding to the upper left corner of the current slice as the origin, and the grid row direction is parallel to the horizontal coordinate axis of the road plane coordinate system, and the grid column direction is parallel to the vertical coordinate axis of the road plane coordinate system.

[0112] The value ranges from 0 to 1, and the sum of the weights of all neighboring points within the same grid is 1. The basis for this is that the weights are normalized based on the distance of each neighboring point to the grid center, so that the points closer to the grid center receive greater weights.

[0113] S2.5: Perform fully convolutional spatial classification inference, candidate window clustering, and activation center localization on the single-channel intensity ground map to generate a set of candidate manhole cover centers.

[0114] Specifically, the single-channel intensity ground map is input into the candidate manhole cover location model, and the spatial classification map is output. ;in This indicates the classification score of the window corresponding to the current location as belonging to the manhole cover category; and performs candidate window filtering on it, selecting those that meet the criteria.

[0115]

[0116] The location is used as a high-confidence candidate window, and spatially adjacent high-confidence candidate windows are grouped into the same candidate cluster. This represents the candidate classification threshold; subsequently, the number of windows for each candidate cluster is counted, and when the number of windows for a candidate cluster is less than the minimum cluster size threshold, the candidate cluster is deleted.

[0117] For each retained candidate cluster, the activation map of the corresponding candidate window in the activation center localization layer is extracted, and the activation map is subjected to depth-dimensional averaging and normalization to obtain the activation center image coordinates of each candidate window within the cluster. Then, a weighted average is performed on the activation center image coordinates according to the candidate window classification score to obtain the image coordinate center corresponding to the candidate cluster. Based on the road plane reference origin coordinates and ground sampling interval of the current slice, the image coordinate center is mapped to the road plane candidate manhole cover center. The road plane candidate manhole cover centers corresponding to all candidate clusters are summarized to obtain the candidate manhole cover center set.

[0118] The image coordinate center is expressed as:

[0119]

[0120] In the formula, Indicates the first The classification score of each candidate window. Indicates the first The activation center image coordinates corresponding to each candidate window Indicates the first Image coordinate centers of candidate clusters;

[0121] The expression for the candidate manhole cover center on the road plane is:

[0122]

[0123] In the formula, This represents the road plane coordinates of the current slice's reference origin. Indicates the ground sampling interval. This represents the road plane coordinates of the center of the k-th candidate manhole cover.

[0124] It should be noted that the candidate manhole cover localization model is a fully convolutional neural network trained offline, comprising a convolutional feature extraction layer, a spatial classification output layer, and an activation center localization layer connected in sequence. The convolutional feature extraction layer extracts the texture and edge features of the manhole cover from the single-channel intensity ground map; the spatial classification output layer outputs the candidate manhole cover scores for each window position; and the activation center localization layer determines the center position of the candidate manhole cover based on the activation response corresponding to the candidate window. The candidate manhole cover localization model is obtained through offline training using intensity ground map samples with manhole cover center annotations, enabling the model to output a spatial classification map and candidate center coordinates from an input single-channel intensity ground map.

[0125] Candidate classification threshold The value range can be set to 0.5 to 0.9, with 0.9 being preferred. The basis for this is that candidate windows are screened under different threshold conditions for the spatial classification map, and the corresponding recall, precision and overall detection effect are compared before the value is determined.

[0126] The minimum cluster size threshold can be set from 2 to 16, with 12 being preferred. The rationale is that, in the spatial classification map, the values ​​that satisfy the threshold are first... Adjacent high-confidence windows are clustered into candidate clusters. The small pseudo-cluster removal effect and the real manhole cover candidate cluster retention effect are statistically analyzed under different cluster size thresholds. The parameter group corresponding to the maximum F2 value is taken as the optimal value. Thus, it is determined that when the value is 12 in the road surface scenario, it can better suppress scattered pseudo-clusters and retain real manhole cover candidate clusters.

[0127] S3, Structure Confirmation and Extraction Module, is used to obtain local structural information around the candidate manhole cover location, confirm the structure of the candidate manhole cover and extract the structural change of the manhole cover relative to the road surface;

[0128] S3.1: Establish a local structure sampling window around the center of the candidate manhole cover, obtain the local structure light point cloud of the corresponding candidate manhole cover location, and perform voxel downsampling, median filtering, statistical filtering and radius filtering in sequence to obtain the purified local structure point cloud;

[0129] Specifically, the set of candidate manhole cover centers is read and combined with the standard manhole cover diameter recorded in the manhole cover basic file. Center of each candidate manhole cover With the center, and A local structure sampling window is established for the sampling radius; the structured light camera acquisition unit is invoked to acquire the original local structure point cloud within the corresponding local structure sampling window. ; For the original local structure point cloud Voxel downsampling is performed, dividing the local structural point cloud into voxel units with a side length of 0.5 cm. The average coordinates of all points within each voxel are used as the representative point of that voxel, resulting in the downsampled point cloud. ; for downsampling point clouds Perform median filtering to obtain the median-filtered point cloud. Point cloud after median filtering Perform statistical filtering to obtain point clouds. ; Point cloud after statistical filtering Perform radius filtering; if the number of points in a neighborhood of radius r is less than a preset threshold for the number of neighborhood points. If the point is identified as edge noise, it will be deleted, resulting in a cleaned local structure point cloud. .

[0130] Downsampling point cloud The generation process specifically includes: assuming a voxel contains m points. Then the coordinates of the centroids of all points within the voxel are:

[0131]

[0132] in, Represents the physical fitness coordinates. Represents the number of points within the voxel;

[0133] Replace all original points within the corresponding voxel with the centroid point, and combine the centroid points corresponding to all voxels to form a downsampled point cloud. ;

[0134] Median filtering specifically includes: for Any sampling point In point clouds Search for all neighboring points whose distance to the sampling point is less than the neighborhood radius d, and construct a spherical neighborhood:

[0135]

[0136] In the formula, Indicates sampling points A spherical neighborhood set of points is constructed with d as the center and d as the radius. Represents the three-dimensional coordinates of any neighborhood point in the spherical neighborhood N;

[0137] The median values ​​of the x, y, and z coordinates of each point within the spherical neighborhood are then taken to obtain the replacement point. :

[0138]

[0139] In the formula, , and Let x, y, and z represent the median values ​​of the x, y, and z coordinates of all points in the neighborhood, respectively.

[0140] By replacing the corresponding sampling points with alternative points, the point cloud after median filtering is obtained. ;

[0141] Statistical filtering specifically includes: Let the i-th point be... The set of distances to its nearest neighbors is the mean value. If the following conditions are met:

[0142]

[0143] Then delete point The point cloud after statistical filtering is obtained. ;in, This represents the average distance from the i-th point to its nearest neighbors. This represents the overall mean of the neighborhood distances of all points. This represents the population standard deviation of the mean neighborhood distance of all points. Indicates the standard deviation multiple.

[0144] It should be noted that mul is used to control the intensity of outlier removal. Since the point cloud after statistical filtering will continue to be used for road surface benchmark extraction, manhole cover structure confirmation, and calculation of the structural change of the manhole cover relative to the road surface in this invention, the value of mul should not be too small to avoid accidentally deleting effective structural points at the edge of the manhole cover and the settlement boundary, nor should it be too large to avoid the retention of random drift points and interference with subsequent plane fitting and structural change extraction. Based on this, mul is set to a value close to 2 and slightly less than 2 to achieve a balance between preserving the geometric features of the manhole cover edge and removing random outliers. The preferred value of mul is 1.8 to 1.9.

[0145] The neighborhood radius r is preferably set to 0.15 cm, and the threshold for the number of neighborhood points is... The preferred value is 12. The neighborhood radius setting is based on the following: the radius filtering is performed after voxel downsampling, median filtering, and statistical filtering. Its processing object is the local structural point cloud that has undergone density compression and preliminary removal of random noise. Therefore, the neighborhood range should be smaller than the geometric scale of the manhole cover edge structure and local undulation changes, and larger than the local distance between adjacent effective points after downsampling, so as to ensure that the effective points of the manhole cover edge still have neighborhood support within the neighborhood, while isolated drifting points on the outer edge are identified as noise points due to the lack of neighborhood support. The neighborhood point number threshold setting is based on the following: the points on the real manhole cover structure surface can form continuous surface support within the neighborhood range, while random drifting points, local suspended points, and sparse pseudo points on the outer edge usually do not have sufficient neighborhood point support. Therefore, using 12 as the neighborhood point number threshold can remove sparse edge noise while preserving the geometric contour of the manhole cover edge, thereby providing a stable input for subsequent road surface reference extraction, manhole cover structure confirmation, and extraction of the structural change of the manhole cover relative to the road surface.

[0146] S3.2: Based on the purified local structural point cloud, perform road reference plane extraction and out-of-surface point set separation, and perform clustering processing on the out-of-surface point set to obtain candidate structure clusters;

[0147] Specifically, for the purified local structural point cloud Perform RANSAC plane fitting to extract the reference surface of the road surface around the manhole cover, and represent the fitted road surface reference surface as follows:

[0148]

[0149] Where a, b, c, and d are the parameters of the road surface plane equation; preferably, the maximum distance between points in the plane is 0.75 cm, the number of fitting sampling points is 5, the maximum number of iterations is 500k, and small clusters with fewer than 5 points are removed. Based on the density distribution of the local structural point cloud, the smoothness characteristics of the road surface, and the number of manhole cover edge points, the parameters are determined through preliminary experiments to reduce the interference of noise points and discrete small clusters on the plane extraction results while ensuring the accuracy of the road surface reference plane fitting.

[0150] Based on the road surface reference plane, extract the purified local structural point cloud. Points that deviate from the road surface reference plane by more than a threshold are considered to form a set of points that deviate from the road surface reference plane.

[0151]

[0152] In the formula, Point The actual vertical distance to the road surface reference plane This represents the set of points that deviate from the road surface reference plane by more than 0.75 cm. Represents the coordinates of the i-th point in the set of points on the surface;

[0153] For the set of points on the surface Perform DBSCAN clustering to group densely distributed out-of-surface points into several candidate structure clusters; where the neighborhood radius of DBSCAN is one-third of the diameter of the current standard manhole cover.

[0154] S3.3: Perform manhole cover edge feature discrimination on candidate structure clusters, generate manhole cover structure confirmation marks, and extract the structural change of the confirmed manhole cover structure clusters relative to the road reference surface.

[0155] Specifically, sector fitting discrimination is performed for each candidate structure cluster; the center coordinates of the corresponding fitting circle are obtained by using the edge point set of the candidate structure cluster as the fitting object and the least squares method. and fitted radius And calculate the fitted radius and the standard manhole cover radius. The difference:

[0156]

[0157] in, Indicates the deviation of the fitted radius. The fitted radius represents the candidate structure cluster. This indicates the standard diameter of the current manhole cover;

[0158] The deviation of the fitted radius Compare with the allowable radius range of the standard manhole cover corresponding to the current manhole cover basic file; when If a candidate structure cluster falls within the allowable range and its edge points are continuously distributed along the fitted circle, the candidate structure cluster is determined to have manhole cover edge features, and a manhole cover structure confirmation flag is output. Otherwise output ,in Indicates whether the current candidate manhole cover has passed local geometry verification;

[0159] when At that time, the highest and lowest points are extracted from the candidate structure clusters that have been confirmed as belonging to the manhole cover structure, and the maximum deviation distance of the candidate structure cluster relative to the road reference surface is calculated. and minimum deviation distance This allows us to obtain the structural change of the manhole cover relative to the road surface. :

[0160] .

[0161] S4, Disturbance Decoupling Module, is used to input real-time status information, structural confirmation results and structural changes into the disturbance decoupling process to distinguish between road disturbance events and real risk events;

[0162] S4.1: Combine the original state vector, manhole cover structure confirmation flag, and structural change quantity into a disturbance decoupling discrimination input group, and perform event type discrimination in the order of first judging manhole cover structure abnormal events, then judging underground environmental risk events, and finally judging road disturbance events;

[0163] Specifically, the edge controller merges the original state vector, the manhole cover structure confirmation flag, and the structural change quantity to form the disturbance decoupling discrimination input group for the current moment; it determines whether the current event is a manhole cover structure abnormality event; when the displacement offset or tilt angle offset in the original state vector is the dominant state quantity of the current abnormality, and the manhole cover structure confirmation flag satisfies... At the same time, the structural change of the manhole cover relative to the road surface satisfies When an event is determined to be an abnormal event in the manhole cover structure, the edge controller writes the determination result into the decoupling result object at the current moment and stops judging subsequent downhole environmental risk events and road disturbance events. When the current event is not determined to be an abnormal event in the manhole cover structure, the edge controller continues to judge whether the current event is an downhole environmental risk event. When at least one of the changes in liquid level, hazardous gas concentration, temperature, or humidity in the original state vector changes continuously in the same direction relative to its baseline value and does not meet the determination conditions for an abnormal event in the manhole cover structure, the edge controller determines the current event to be an downhole environmental risk event. When the current event does not meet the determination conditions for an abnormal event in the manhole cover structure or the determination conditions for an downhole environmental risk event, the edge controller determines the current event to be a road disturbance event.

[0164] It should be noted that, The threshold for structural change is 2.5 cm to 3.5 cm, preferably 3.0 cm. This is based on the fact that when the height difference between the manhole cover and the road surface reaches about 3 cm in road engineering, it will have a significant impact on the comfort and safety of vehicle passage. Therefore, 3.0 cm is used as the trigger value for structural change of manhole cover.

[0165] Displacement offset or tilt offset is the dominant state quantity of the current anomaly, which means that the combined change of displacement offset and tilt offset is greater than the vibration intensity. The reason for adopting this judgment method is that road disturbance is usually dominated by a sudden increase in vibration, while the real structural anomaly is usually manifested as a continuous dominant change in displacement and tilt, accompanied by the establishment of local geometry and the fulfillment of the structural change conditions.

[0166] Maintaining a continuous change in the same direction relative to its baseline value means that the corresponding state quantity always maintains the same direction of change in the continuous sampling sequence corresponding to the current duration of the anomaly τ, without repeated switching between positive and negative directions; this judgment method is adopted because the focus of identifying downhole environmental risks is on whether the environmental state forms a continuous evolution, rather than whether a certain instantaneous value exceeds a single threshold.

[0167] The road disturbance event corresponds to: the current anomaly is mainly manifested as a short-term dominant change in vibration intensity, coupled with the result object. The output is the input object for subsequent control input filtering and control state generation. When dealing with road disturbance events, only the decoupling result object is considered. It is reserved for subsequent continuous sampling updates, and will not be included in subsequent control loops.

[0168] S4.2: After completing the event type identification, generate a disturbance decoupling result object containing the event type, the original state vector, the manhole cover structure confirmation flag, and the structural change amount.

[0169] S5, Status Generation Module, is used to filter real risk events into the control link, and to perform risk fusion on the events entering the control link to generate corresponding control statuses.

[0170] S5.1: Read the disturbance decoupling result object, first block the road disturbance event outside the control closed loop, and write the manhole cover structure abnormal event or the underground environment risk event into the effective control input object;

[0171] Specifically, the edge controller reads the disturbance decoupling result object and performs control link filtering. When the event type is a road disturbance event, the edge controller does not write the current event into the control link, but only performs high-sampling verification. That is, the sampling period of the current manhole cover object is adjusted to half of the original sampling period. Within the verification window, based on the shortened sampling period, the real-time status acquisition, abnormal trend identification, candidate manhole cover location, local structure verification, and event type re-judgment are repeatedly performed on the current manhole cover. If it is still not determined to be a manhole cover structural abnormality event or an underground environment risk event after the verification window ends, the original sampling period is restored, and the current control input filtering ends. When the event type is a manhole cover structural abnormality event or an underground environment risk event, the edge controller determines that the current event belongs to a real risk event and writes it into the valid control input object.

[0172] S5.2: Perform rule-based fusion on the structural risk information, water accumulation risk information and hazardous environment risk information in the effective control input object to generate a unique control state among the structural instability state, water accumulation risk state, hazardous environment state or composite risk state.

[0173] Specifically, after obtaining a valid control input object, the edge controller generates a unique control state in a fixed sequence. First, it determines whether it is a composite risk state. When the event type corresponding to the valid control input object is a manhole cover structure abnormality event, and at least one of the liquid level change, hazardous gas concentration change, temperature change, or humidity change in the original state vector changes continuously in the same direction, the edge controller outputs a composite risk state. Alternatively, when the event type corresponding to the valid control input object is a downhole environmental risk event, and the manhole cover structure confirmation flag is satisfied... At the same time, the structural change satisfies At the same time, the edge controller also outputs a composite risk state as the control state;

[0174] If the current event does not meet the criteria for determining a composite risk state, continue to determine whether it is a structural instability state; if the event type corresponding to the effective control input object is a manhole cover structural anomaly event, and the manhole cover structural confirmation flag is met... At the same time, the structural change satisfies At that time, the edge controller outputs a control state indicating structural instability.

[0175] When the current event does not meet the conditions for determining the composite risk state and the structural instability state, the determination of whether it is a water accumulation risk state continues; when the event type corresponding to the effective control input object is a downhole environment risk event, and the liquid level change remains in the same direction and continuously changes, while the change in hazardous gas concentration, temperature change and humidity change do not simultaneously form a continuous change in the same direction, the edge controller outputs the control state as a water accumulation risk state.

[0176] When the current event does not meet the criteria for determining the composite risk state, structural instability state, and water accumulation risk state, the edge controller outputs a control state of hazardous environment. The hazardous environment state corresponds to the current event being determined as a downhole environmental risk event, and is not a water accumulation risk state caused by continuous changes in liquid level. In this way, the downhole hazardous environment characterized by changes in hazardous gas concentration, temperature, and humidity is uniformly classified into the hazardous environment state.

[0177] After completing the above sequential determination, the edge controller writes the final control state into the control result object at the current moment; when the current event is filtered out as a road disturbance event, no control result object is generated, but only the verification sampling result is retained for subsequent continuous monitoring and updates.

[0178] S6, the linkage control module, is used to execute local control of the manhole cover based on the control status; execute hazard mitigation linkage control when a hazardous environment is determined to exist; and execute collaborative linkage control of associated manhole covers when the control status reaches the joint defense trigger condition.

[0179] S6.1: Execute local control of this manhole cover based on the unique control status, and perform corresponding control on electronic lock, manhole warning, sampling frequency and manhole cover opening permission, and generate local control result object;

[0180] Specifically, the unique control status is read, and the corresponding local control for this manhole cover is executed based on the unique control status, generating a local control result object, including the unique control status, electronic lock status, manhole warning status, sampling frequency status, and manhole cover opening permission status.

[0181] S6.2: When the local control result object indicates that the manhole cover is in a hazardous environment or a complex risk state, the ventilation device is invoked to perform the risk mitigation linkage control, and the changes in hazardous gas concentration, temperature and humidity are continuously collected during the risk mitigation process to form a risk mitigation monitoring sequence.

[0182] When the only control state is structural instability, the edge controller sends a locking command to the electronic lock, setting the electronic lock to the locked state; sends a high-level warning activation command to the wellhead audible and visual warning device, setting the wellhead warning state to the high-level warning state; and switches the vibration sampling, tilt sampling, and displacement sampling in the wellhead area to high-frequency sampling mode, setting the sampling frequency state to high-frequency sampling state.

[0183] When the only control state is the water accumulation risk state, the edge controller sends an alarm activation command to the wellhead alarm, setting the wellhead alarm state to the alarm state; switches the water level sampling cycle in the well to the high-frequency sampling mode, setting the sampling frequency state to the high-frequency sampling state; and sets the manhole cover opening permission state to the frozen state.

[0184] When the only control state is a hazardous environment state, the edge controller sends a locking command to the electronic lock, setting the electronic lock to the locked state; sends a high-level warning opening command to the wellhead audible and visual warning device, setting the wellhead warning state to the high-level warning state; and sets the well cover opening permission state to the locked state.

[0185] When the only control state is the composite risk state, the edge controller simultaneously executes the local control actions corresponding to the structural instability state and the hazardous environment state, setting the electronic lock state to the locked state, the wellhead warning state to the high-level warning state, the sampling frequency state to the high-frequency sampling state, and the well cover opening permission state to the locked state.

[0186] After completing local control, the edge controller outputs the local control result object;

[0187] The risk mitigation linkage control will continue only if the sole control state is a hazardous environment state or a complex risk state.

[0188] S6.3: Based on the continuous decline trend of relevant state quantities of hazardous environment according to the risk elimination monitoring sequence, determine the completion of risk elimination and generate a risk elimination linkage result object containing local control result objects and risk elimination completion indicators.

[0189] Specifically, the edge controller sends a start command to the ventilation device to put the well chamber into ventilation and hazard mitigation mode and records the hazard mitigation start time; subsequently, in subsequent sampling cycles, it continuously reads the changes in hazardous gas concentration, temperature, and humidity in the original state vector and forms a hazard mitigation monitoring sequence.

[0190] When the self-hazard removal process starts, the edge controller determines that the current well chamber has completed hazard removal if the following conditions are met for m consecutive sampling cycles:

[0191] First, the absolute value of the change in the concentration of hazardous gas at the current moment within m consecutive sampling periods. Not greater than the corresponding value of the previous sampling period ,Right now Secondly, the absolute value of the temperature change at the current moment within m consecutive sampling periods. and absolute value of humidity change Each is not greater than the corresponding value of the previous sampling period. and ,Right now and ;

[0192] When the above-mentioned conditions for risk mitigation are met, the edge controller stops the operation of the ventilation device and generates a risk mitigation completion flag; then, the local control result object and the risk mitigation completion flag are written into the risk mitigation linkage result object.

[0193] S6.4: Read the local control result object and the risk mitigation linkage result object, determine the joint defense trigger type based on the unique control status and risk mitigation completion flag, and extract the corresponding associated well covers and associated equipment from the regional well group association table;

[0194] Specifically, when the only control state is the water accumulation risk state, the central controller determines that the source manhole cover has reached the trigger condition for joint prevention of water accumulation, and reads the set of upstream associated manhole covers directly corresponding to the source manhole cover from the regional manhole group association table;

[0195] When the only control state is a hazardous environment state and the hazard elimination is not completed, the central controller determines that the source manhole cover has reached the hazardous environment joint defense trigger condition, and reads the downstream associated manhole cover set and the adjacent connected manhole cover set directly corresponding to the source manhole cover from the regional manhole group association table;

[0196] When the only control state is structural instability and the source manhole cover is registered in the basic file as a manhole cover in a key intersection, in front of a school, in front of a hospital, or in a main road intersection area, the central controller determines that the source manhole cover has reached the trigger condition for joint prevention of traffic risk, and reads the set of road adjacent manhole covers corresponding to the source manhole cover in the regional manhole group association table, as well as the set of roadside warning devices corresponding to the source manhole cover.

[0197] When the only control state is a composite risk state, the central controller will handle the joint prevention of structural risks and the joint prevention of environmental risks separately.

[0198] When the local control results corresponding to the composite risk state include structural instability control actions, read the set of adjacent manhole covers and the set of roadside warning devices.

[0199] When the risk mitigation linkage result object corresponding to the composite risk state exists and the risk mitigation is not completed, read the set of manhole covers associated with the joint prevention and control of dangerous environments.

[0200] When the local control results corresponding to the composite risk state include water accumulation risk control actions, read the upstream associated manhole cover set.

[0201] S6.5: Issue collaborative joint defense control commands to associated manhole covers and associated equipment based on the joint defense trigger type, and generate regional collaborative joint defense result objects.

[0202] Specifically, when there is a risk of water accumulation, the central controller sends a high-frequency water level monitoring command to the upstream associated manhole cover set, which switches the water level sampling cycle of the corresponding upstream associated manhole cover to high-frequency sampling mode and sets its monitoring status to water accumulation joint prevention and monitoring status.

[0203] When the environment is in a hazardous state, the central controller sends a high-frequency environmental monitoring command to the set of manhole covers associated with the hazardous environment joint prevention and control, which switches the sampling of hazardous gas concentration, temperature and humidity of the corresponding manhole covers to high-frequency sampling mode and sets their monitoring status to the hazardous environment joint prevention and control early warning state.

[0204] When the structure is in an unstable state, the central controller sends a joint prevention warning command to the set of manhole covers adjacent to the road and sends a roadside warning activation command to the set of roadside warning devices, so that the manhole warning devices of the manhole covers adjacent to the road enter the joint prevention warning state, and at the same time, the corresponding roadside warning devices enter the joint prevention warning state.

[0205] When the situation is a complex risk state, joint prevention and control measures shall be implemented according to the corresponding state: traffic risk type, hazardous environment type, and water accumulation type.

[0206] After issuing the joint prevention and control command, the central controller writes the source manhole cover number, joint prevention trigger type, associated manhole cover set, associated device set, and current joint prevention and control status into the regional collaborative joint prevention result object.

[0207] S7, the recovery module, is used to release control and restore the normal monitoring status after the manhole cover status, environmental status, and joint defense status have returned to normal.

[0208] S7.1: Execute risk removal judgment based on the recovery status of manhole cover status, environmental status, and joint prevention status, and release control and restore regular monitoring status when normal conditions are restored.

[0209] Specifically, when n consecutive sampling periods satisfy:

[0210]

[0211] At that time, it was determined that the structure of the manhole cover had returned to normal.

[0212] When the liquid level change no longer increases after n consecutive sampling periods, and

[0213]

[0214] Furthermore, when each manhole cover in the associated manhole cover set does not re-enter the suspected structural anomaly trigger state for n consecutive sampling periods, and is not again identified as a manhole cover structural anomaly event or a downhole environment risk event, and the joint prevention and control status no longer expands, the joint prevention and control status of the area is determined to have returned to normal. When the risk elimination linkage control has been executed, the risk elimination is completed, and the downhole environment status is determined to have returned to normal.

[0215] When the manhole cover structure, environmental conditions, and joint defense status are all restored to normal, the central controller executes the following in sequence: close the manhole warning; restore high-frequency sampling to regular sampling; close the ventilation device; release the regional manhole group collaborative joint defense control; restore the electronic lock to the regular control status and restore the manhole cover opening permission status from the locked or frozen status to the regular permission status.

[0216] Finally, a reset completion result object is generated.

[0217] In summary, this invention establishes a monitoring and control relationship for the area where the manhole cover is located. By combining real-time manhole cover status information, candidate manhole cover positioning, local structure confirmation, and extraction of structural changes of the manhole cover relative to the road surface, it first decouples road disturbance events from actual risk events. Then, it filters the actual risk events into the control link and generates corresponding control states, thereby effectively reducing false triggering problems caused by factors such as vehicle rolling and construction impact. At the same time, by combining local control of the manhole cover, linkage for hazardous environment mitigation, collaborative prevention and control of regional manhole groups, and automatic reset control after risk removal, it achieves an improvement from single-point monitoring and alarm to hierarchical closed-loop management. This not only improves the accuracy of identifying and handling manhole cover structural anomalies, water accumulation, and hazardous environments, but also enhances the risk diffusion suppression capability, linkage response capability, and underground facility operation safety assurance capability in complex urban road scenarios.

[0218] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An intelligent manhole cover monitoring and management system with status monitoring capabilities, characterized in that, include: The relationship establishment module is used to establish monitoring and control relationships for the area where the manhole cover is located; The trend recognition and positioning module is used to identify abnormal trends based on the real-time status information of manhole covers, and to locate candidate manhole covers in the road area where the manhole covers are located after identifying abnormal trends. The structure confirmation and extraction module is used to obtain local structural information around the candidate manhole cover location, confirm the structure of the candidate manhole cover and extract the structural change of the manhole cover relative to the road surface; The disturbance decoupling module is used to input real-time status information, structural confirmation results, and structural changes into the disturbance decoupling process to distinguish between road disturbance events and real risk events. The status generation module is used to filter real risk events into the control link, and to perform risk fusion on the events entering the control link to generate corresponding control statuses. The linkage control module is used to execute local control of the manhole cover based on the control status; execute hazard mitigation linkage control when a hazardous environment is determined to exist; and execute collaborative joint defense control of associated manhole covers when the control status reaches the joint defense trigger condition. The recovery module is used to release control and restore normal monitoring status after the manhole cover status, environmental status, and joint defense status have returned to normal.

2. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 1, characterized in that: The steps for establishing a monitoring and control relationship for the area where the manhole cover is located are as follows: Collect basic attribute information of manhole covers and generate basic files for a single manhole cover. Connect the monitoring unit and the execution unit to the controller to form a corresponding integrated manhole cover monitoring and control object. A regional well group association table is established based on the underground connectivity and road adjacency relationships. Control parameters are initialized according to the regional attributes of the well cover. After the parameters are written and the equipment is online for verification, the corresponding well cover object is set to a controllable and active state.

3. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 2, characterized in that: The steps for identifying abnormal trends based on real-time manhole cover status information are as follows: Collect real-time status information corresponding to activated manhole cover objects, synchronously acquire and benchmark correct vibration, tilt angle, displacement, water level, gas, temperature and humidity and electronic lock status, and update the current abnormal duration by combining the abnormal duration register value cached by the edge controller, and generate the original state vector at the current moment. Based on the joint change relationship of vibration, tilt angle, displacement and abnormal duration in the original state vector, the real-time abnormal trend of the manhole cover is identified, and the corresponding manhole cover is judged to be in a suspected abnormality triggered state when the structural abnormality trend condition is met.

4. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 3, characterized in that: The steps for locating candidate manhole covers in the road area where the manhole cover is located after identifying abnormal trends are as follows: Obtain the original road point cloud of the road area where the manhole cover is located, corresponding to the suspected abnormal trigger state of the structure, and sequentially perform statistical outlier filtering, connected component filtering, ground segmentation and common domain filtering to obtain candidate search point cloud that retains only the common domain road ground. The candidate search point cloud is rasterized according to the preset slice size and sampling interval, and a single-channel intensity ground map is constructed based on the intensity information of neighboring points; Perform fully convolutional spatial classification inference, candidate window clustering, and activation center localization on the single-channel intensity ground map to generate a set of candidate manhole cover centers.

5. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 4, characterized in that: The steps for obtaining local structural information around the candidate manhole cover location, confirming the structure of the candidate manhole cover, and extracting the structural change of the manhole cover relative to the road surface are as follows: A local structure sampling window is established around the center of the candidate manhole cover to obtain the local structure light point cloud of the corresponding candidate manhole cover location. Voxel downsampling, median filtering, statistical filtering and radius filtering are performed in sequence to obtain the purified local structure point cloud. Based on the purified local structural point cloud, the road reference plane is extracted and the out-of-surface point set is separated. The out-of-surface point set is then clustered to obtain candidate structure clusters. Perform manhole cover edge feature discrimination on candidate structure clusters, generate manhole cover structure confirmation marks, and extract the structural change of the confirmed manhole cover structure clusters relative to the road reference surface.

6. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 5, characterized in that: The steps to distinguish between road disturbance events and actual risk events are as follows: The original state vector, manhole cover structure confirmation flag, and structural change quantity are merged into a disturbance decoupling discrimination input group, and the event type discrimination is performed in the order of first judging manhole cover structure abnormal events, then judging underground environmental risk events, and finally judging road disturbance events. After completing the event type identification, a disturbance decoupling result object is generated, which includes the event type, the original state vector, the manhole cover structure confirmation flag, and the structural change amount.

7. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 6, characterized in that: The steps for generating the corresponding control state are as follows: Read the disturbance decoupling result object, first block the road disturbance event outside the control closed loop, and write the manhole cover structure abnormal event or the underground environment risk event into the effective control input object; The system performs rule-based fusion on the structural risk information, water accumulation risk information, and hazardous environment risk information in the effective control input object to generate a unique control state among the structural instability state, water accumulation risk state, hazardous environment state, or composite risk state.

8. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 7, characterized in that: The steps for implementing local control of the manhole cover based on the control status are as follows: Based on the unique control status, local control of this manhole cover is executed, and corresponding control is performed on the electronic lock, manhole warning, sampling frequency and manhole cover opening permission, generating a local control result object; When the local control result indicates that the manhole cover is in a hazardous environment or a complex risk state, the ventilation device is invoked to perform hazard mitigation linkage control, and during the hazard mitigation process, the changes in hazardous gas concentration, temperature, and humidity are continuously collected to form a hazard mitigation monitoring sequence. The step of executing hazard mitigation linkage control when a hazardous environment is determined to exist refers to determining the completion of hazard mitigation based on the continuous downward trend of the relevant state quantities of the hazardous environment according to the hazard mitigation monitoring sequence, and generating a hazard mitigation linkage result object containing a local control result object and a hazard mitigation completion flag.

9. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 8, characterized in that: When the control state reaches the joint defense trigger condition, the coordinated joint defense control of the associated manhole cover is executed, and the steps are as follows: Read the local control result object and the risk elimination linkage result object, determine the joint defense trigger type based on the unique control status and risk elimination completion flag, and extract the corresponding associated well cover and associated equipment from the regional well group association table; Based on the joint defense trigger type, collaborative joint defense control instructions are issued to associated manhole covers and associated equipment, generating regional collaborative joint defense result objects.

10. The intelligent manhole cover monitoring and management system with status monitoring as described in claim 9, characterized in that: The steps for releasing control and restoring normal monitoring status are as follows: Risk clearance is determined based on the recovery status of manhole cover condition, environmental condition, and joint prevention status. Control is lifted and routine monitoring is restored when normal conditions are restored.

Images