School facility supervision method, device and system

Abstract

The invention discloses a school facility supervision method, device and system, and the method comprises the steps: enabling a sensor group to be assembled in equipment in a school, carrying out the real-time collection of the state data of the equipment, collecting the state data of the facilities in real time through the sensor group disposed on the facilities, guaranteeing the real-time performance and comprehensiveness of the data, and achieving the real-time monitoring of the facilities. And performing deep analysis on the preprocessed data by using a deep learning model, evaluating the health state of the facility, generating maintenance decision suggestions, and triggering corresponding response measures according to different abnormal levels until an administrator intervenes, thereby ensuring timely processing of abnormal conditions. According to the school facility supervision method, device and system, the maintenance process is optimized, the long-term maintenance cost is reduced, the operation efficiency of the facility is improved, the service life of the facility is prolonged, and the crack sensor and the temperature and humidity sensor group can be switched to different states in different use environments, so that the situation that the crack sensor and the temperature and humidity sensor group are easily damaged by environmental factors during use is avoided; the overall service life of the device is prolonged.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of monitoring devices, in particular to a school facility supervision method, device and system. BACKGROUND

[0002] In the facility management of educational institutions, the traditional supervision method relies on manual inspection. Although this method can achieve a certain degree of facility state inspection, it has obvious shortcomings in efficiency and real-time performance. In particular, when facing a large number of equipment and complex environment, manual inspection is difficult to ensure the comprehensiveness and timeliness of the inspection. Currently, most educational institutions use periodic manual inspection as the main facility supervision method. By developing an inspection plan, staff are arranged to conduct regular inspections of facilities to discover potential problems or faults. However, this method has obvious defects, including but not limited to: low efficiency of manual inspection, prone to missed inspection and delayed response; unable to monitor facility state changes in real time, resulting in delayed problem discovery; independent operation of different facility systems causes data silo phenomenon, lack of unified management platform; due to delayed fault discovery, maintenance costs increase, maintenance costs are high. SUMMARY

[0003] In view of the deficiencies of the prior art, the present application provides a school facility supervision method, device and system to solve the above technical problems.

[0004] To achieve the above purpose, the present application provides the following technical scheme: a school facility supervision method, comprising: S1, multi-source data acquisition: Assemble a sensor group in the equipment inside the school, and collect state data of the equipment in real time. Through the deployment of a sensor group on the facility, the state data of the facility is collected in real time to ensure the real-time and comprehensiveness of the data. S2, edge computing preprocessing: Preliminary cleaning, feature extraction and anomaly judgment are performed on the collected raw data at the local gateway. S3, cloud intelligent analysis: The preprocessed data is analyzed in depth using a deep learning model to evaluate the health status of the facility, generate maintenance decision suggestions, and realize intelligent analysis and decision-making. S4, multi-level early warning mechanism: According to the different levels of anomalies, trigger the corresponding response measures until the administrator intervenes to ensure timely handling of abnormal situations. S5, maintenance closed-loop management: Record the maintenance process and update the maintenance information to the equipment life cycle database to form a closed-loop management, optimize the maintenance process and reduce the long-term maintenance cost.

[0005] Preferably, in the S1, multi-source data acquisition step, the sensor group includes a vibration sensor, a temperature and humidity sensor, a pressure sensor and an image sensor.

[0006] Preferably, in the S3, cloud intelligent analysis step, the Horovod framework is used for multi-group GPU parallelism and mixed precision training.

[0007] Preferably, in the S5, maintenance closed-loop management step, a three-dimensional model of the facility is constructed by Unity3D, and real-time data driving is performed, and a fault mode is stored in cooperation with a Neo4j graph database.

[0008] A school facility supervision device adopts the above-mentioned school facility supervision method, comprising: a mobile vehicle body, mobile wheels and a controller, the controller is assembled on the top of the mobile vehicle body, the mobile wheels are uniformly assembled on the outer positive and negative sides of the mobile vehicle body, the top left and right sides of the mobile vehicle body are respectively assembled with a group of image sensors and a protective shell, the inner cavity of the protective shell is assembled with a linear track module, the output end of the linear track module is connected with a receiver, the output end of the receiver is connected with a group of temperature and humidity sensors, and the inner cavity top of the protective shell is connected with a closed door through a hinge.

[0009] The bottom of the mobile vehicle body is assembled with a protective bottom shell, the inner cavity top of the protective bottom shell is connected with a first air cylinder, the output end of the first air cylinder is connected with a crack sensor, the inner cavity of the protective bottom shell is assembled with a partition plate on both sides, and the inner cavity of the first air cylinder is connected with an L-shaped closed bottom plate on both sides through a hinge.

[0010] The guiding mechanism comprises a hollow column connected with the protective bottom shell, a piston rod slidably connected in the hollow column and connected with the crack sensor, a bend pipe in communication with the outside of the hollow column, and a stretchable air bag in communication with the other end of the bend pipe and sleeved on the outside of the crack sensor.

[0011] Preferably, the top of the L-shaped sealing bottom plate is connected with a second spring, the other end of the second spring is connected with the inner cavity of the protective bottom shell through a bolt, the second spring can pull the L-shaped sealing bottom plate to return after being driven to rotate by the piston rod, and the outer part of the piston rod is provided with a cross plate for driving the L-shaped sealing bottom plate to move.

[0012] Preferably, the inner cavity of the piston rod is embedded with a limiting plate, the outer part of the limiting plate is in sliding connection with the inner cavity of the hollow column, and the top of the hollow column is provided with an exhaust hole, the limiting plate is used for improving the smoothness of the movement of the piston rod, so that the piston rod moves linearly in the inner cavity of the hollow column.

[0013] Preferably, the bottom of the image sensor group is provided with an electronic turntable, the bottom of the electronic turntable is provided with a pneumatic cylinder, the bottom of the pneumatic cylinder is connected with the moving vehicle body through a bolt, and the outer parts of the electronic turntable and the pneumatic cylinder are connected with the controller through lines, the electronic turntable can drive the image sensor group to rotate, the pneumatic cylinder can drive the image sensor group to stretch and contract, and the pneumatic cylinder and the electronic turntable are controlled by the controller.

[0014] Preferably, the bottom of the moving vehicle body is provided with a vehicle lamp, the inner cavity of the protective shell is provided with a first spring at the top, the other end of the first spring is connected with the sealing door, the first spring can pull the sealing door to return after movement, and the outer part of the receiver is provided with a cross plate for driving the sealing door to move.

[0015] A school facility supervision system adopts the school facility supervision method, and comprises a multi-source perception layer, an edge computing unit, a cloud analysis platform and a decision response module, the multi-source perception layer comprises an integrated vibration sensor array, a temperature and humidity sensor array and an image sensor array, the edge computing unit is configured with a local gateway of an FPGA acceleration chip, the cloud analysis platform is loaded with a server cluster of an LSTM neural network, and the decision response module realizes a hierarchical early warning and maintenance work order generation system.

[0016] In summary, the school facility supervision method, device and system provided by the application have the following beneficial effects: The school facility supervision method, device and system, through the sensor group deployed on the facility, real-time collection of the state data of the facility, realizes real-time monitoring and comprehensive coverage of the facility state data, improves the supervision efficiency and timeliness, through the local gateway, the collected raw data is preliminarily cleaned, feature extraction and abnormality judgment are carried out, the data processing speed is improved, the network transmission pressure is reduced, the real-time of data analysis is ensured, through the deep analysis of the preprocessed data by using the deep learning model, the health state of the facility can be accurately evaluated, scientific maintenance decision basis is provided, the intelligent level of supervision is enhanced, through the triggering of corresponding response measures according to the different abnormal levels, the facility abnormal situation is effectively responded, the occurrence of accidents is avoided, the reliability of safety management is improved, through recording the maintenance process and updating the maintenance information to the equipment life cycle database, a closed-loop management is formed, the maintenance process is optimized, the long-term maintenance cost is reduced, and the operation efficiency and service life of the facility are improved.

[0017] The school facility supervision method, device and system, through the mobile wheels, the sensor distance can be approached, thereby improving the data receiving speed, and the air detection of the area without the installed sensor is carried out, the data coverage area is improved, and the protection bottom shell and the protection shell are additionally arranged to protect the crack sensor and the temperature and humidity sensor group, respectively, the crack sensor and the temperature and humidity sensor group can be switched to different states in different use environments, thereby avoiding damage caused by environmental factors during use, and improving the overall service life of the device.

[0018] The school facility supervision method, device and system, through the hollow column, piston rod, elbow and telescopic air bag, when the crack sensor is retracted to the inner cavity of the protection bottom shell, the crack sensor is automatically guided, the stability of the crack sensor during movement is improved, the gas in the inner cavity of the hollow column is pressed into the inside of the telescopic air bag through the elbow by the piston rod, the telescopic air bag is expanded, the crack sensor is protected, the service life of the crack sensor is improved, the crack sensor is prevented from being damaged by impact, and the sensor is arranged in the inner cavity of the hollow column, when the piston rod is reset, the signal is transmitted to the receiver, and then the temperature and humidity sensor group is driven to reset by the linear track module, so that the temperature and humidity sensor group is protected. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 It is the front view of the present application.

[0020] Figure 2 It is the plane view of the present application.

[0021] Figure 3 It is the front view of the protection shell of the present application.

[0022] Figure 4 is the front view of the protective bottom shell according to the present application.

[0023] Figure 5 is the front view of the guiding mechanism according to the present application.

[0024] Figure 6 is the schematic diagram of the work flow according to the present application.

[0025] Explanation of reference signs: 1, mobile vehicle body; 11, vehicle light; 2, mobile wheel; 3, controller; 4, image sensor group; 5, protective shell; 51, straight rail module; 52, receiver; 53, temperature and humidity sensor group; 54, first spring; 55, closing door; 6, protective bottom shell; 61, first air cylinder; 62, crack sensor; 63, partition; 64, L-shaped closing bottom plate; 65, second spring; 7, guiding mechanism; 71, hollow column; 72, piston rod; 73, elbow pipe; 74, telescopic air bag. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0027] The present application provides a technical solution, please refer to Figure 6 A school facility monitoring method, comprising: S1, multi-source data acquisition: Assemble the sensor group in the equipment inside the school, collect the state data of the equipment in real time, collect the state data of the facility in real time through the sensor group deployed on the facility, and ensure the real-time and comprehensiveness of the data; S2, edge computing preprocessing: Preliminary cleaning, feature extraction and abnormality judgment are performed on the collected raw data at the local gateway; S3, cloud intelligent analysis: The preprocessed data are deeply analyzed by using a deep learning model, the health status of the facility is evaluated, maintenance decision suggestions are generated, and intelligent analysis and decision are realized; S4, multi-level early warning mechanism: According to the different abnormality levels, the corresponding response measures are triggered until the administrator intervenes, so as to ensure the timely treatment of abnormal conditions; S5, maintenance closed-loop management: Record the maintenance process and update the maintenance information to the equipment life cycle database to form a closed-loop management, optimize the maintenance process and reduce the long-term maintenance cost.

[0028] In the multi-source data collection step S1, the sensor group includes a vibration sensor, a temperature and humidity sensor, a pressure sensor, and an image sensor. In the cloud intelligent analysis step S3, the Horovod framework is used for multi-GPU parallelism, and mixed precision training is performed. In the maintenance closed-loop management step S5, a three-dimensional model of the facility is constructed by Unity3D, and real-time data driving is performed. The Neo4j graph database is used to store fault patterns.

[0029] The edge gateway architecture is NVIDIA Jetson Xavier NX with a TensorRT acceleration engine, supporting multi-sensor data concurrent processing. The vibration sensor is a three-axis acceleration sensor ADXL357 that captures device vibration signals in real time. The temperature sensor is a temperature and humidity sensor and a gas sensor that form a micro-environment monitoring network to detect PM2.5 concentration changes and H2S leaks. The image sensor is an industrial camera that collects device operation video streams at a frame rate of 30fps. The optical flow algorithm is used to detect abnormal jitter of the robot arm. The sensor group can be deployed on classroom equipment such as projector motor shafts and air conditioner compressor bases, laboratory instruments such as pasted on centrifuge bearings and reaction kettle support frames, and sports equipment such as embedded in treadmill drive shafts and single bar connectors. In the edge computing preprocessing step S2, Kalman filtering is used to eliminate noise, and feature extraction such as mean, variance, and kurtosis in time domain features, FFT transform in frequency domain features, and YOLOv5 target detection in image features are performed. The abnormality judgment is based on the isolation forest algorithm, and the data set size is 100,000 multi-modal samples.

[0030] Please refer to the drawings and Figure 2 A school facility monitoring device using the above-mentioned school facility monitoring method, comprising: a mobile vehicle body 1, mobile wheels 2, and a controller 3, the controller 3 is assembled on the top of the mobile vehicle body 1, the mobile wheels 2 are uniformly assembled on the outer positive and negative sides of the mobile vehicle body 1, the top left and right sides of the mobile vehicle body 1 are respectively equipped with an image sensor group 4 and a protective shell 5, the inner cavity of the protective shell 5 is equipped with a linear track module 51, the output end of the linear track module 51 is connected with a receiver 52, the output end of the receiver 52 is connected with a temperature and humidity sensor group 53, and the inner cavity of the protective shell 5 is equipped with a linear track module 51.

[0031] Please refer to Figure 3The mobile vehicle body 1 is a mobile vehicle body commonly known in the prior art, which includes collision sensors, infrared sensors, ultrasonic sensors and cliff sensors inside, the movement of the mobile vehicle body 1 is controlled through navigation positioning technology, and the inside of the mobile vehicle body 1 is equipped with a mobile execution mechanism, the mobile wheel 2 can support the mobile vehicle body 1 to move, the mobile wheel 2 is controlled to rotate through the internal motor, and the mobile wheel 2 is controlled through the mobile vehicle body 1, the controller 3 can receive and process signals of multiple groups of sensors, the controller 3 is composed of a high-performance processor, a memory module, a communication module and a storage device, and the inside of the controller 3 is equipped with a model for analyzing and judging data, the temperature and humidity sensor group 53 is an environmental sensor commonly known in the prior art, which is used for detecting the environment, the protective shell 5 is used for protecting the temperature and humidity sensor group 53, and the linear track module 51 is an electric guide rail commonly known in the prior art, which is used to drive the temperature and humidity sensor group 53 to move linearly. The closed door 55 is used to close the protective shell 5.

[0032] Please refer to Figure 4 The end of the closed bottom plate 64 opposite to the closed door 55 is designed as a round corner, so as to avoid interference between the closed bottom plate 64 and the closed door 55 during mutual rotation, and the end of the closed bottom plate 64 and the closed door 55 away from each other is L-shaped, so that the closed bottom plate 64 and the closed door 55 can only rotate towards one end.

[0033] The bottom of the mobile vehicle body 1 is equipped with a protective bottom shell 6, the top of the inner cavity of the protective bottom shell 6 is connected with a first air cylinder 61, the output end of the first air cylinder 61 is connected with a crack sensor 62, the inner cavity of the protective bottom shell 6 is equipped with a partition plate 63 on both sides, the inner cavity of the first air cylinder 61 is connected with an L-shaped closed bottom plate 64 on both sides through a hinge, and the inner cavity of the protective bottom shell 6 is connected with a guide mechanism 7 on both sides.

[0034] The protective bottom shell 6 is used for protecting the first air cylinder 61, the crack sensor 62 is used for driving the crack sensor 62 to move, and the crack sensor 62 is used for detecting the playground cracks.

[0035] Please refer to Figure 5 The guide mechanism 7 includes a hollow column 71 connected with the protective bottom shell 6, a piston rod 72 slidably connected in the inside of the hollow column 71, the piston rod 72 connected with the crack sensor 62, a bend pipe 73 communicated on the outside of the hollow column 71, a stretchable air bag 74 communicated at the other end of the bend pipe 73, and the stretchable air bag 74 sleeved on the outside of the crack sensor 62.

[0036] The hollow column 71 is used for guiding the piston rod 72, the movement of the piston rod 72 in the hollow column 71 can generate air flow, the air flow enters the inside of the telescopic air bag 74 through the elbow pipe 73, so that the crack sensor 62 is protected by the telescopic air bag 74 during use, prolonging the service life of the crack sensor 62, and the elbow pipe 73 is elastic and can be telescopic.

[0037] The top of the L-shaped sealing bottom plate 64 is connected with the second spring 65, the other end of the second spring 65 is connected with the inner cavity of the protection bottom shell 6 through the bolt, the second spring 65 can pull the L-shaped sealing bottom plate 64 to reset after being driven to rotate by the piston rod 72, and the outside of the piston rod 72 is provided with a horizontal plate for driving the L-shaped sealing bottom plate 64 to move.

[0038] The inner cavity of the piston rod 72 is embedded with a limiting plate, the outer portion of the limiting plate is in sliding connection with the inner cavity of the hollow column 71, and the top of the hollow column 71 is provided with an exhaust hole, the limiting plate is used to improve the smoothness of the movement of the piston rod 72, so that the piston rod 72 moves linearly in the inner cavity of the hollow column 71.

[0039] The bottom of the image sensor group 4 is provided with an electronic turntable, the bottom of the electronic turntable is provided with an air cylinder, the bottom of the air cylinder is connected with the mobile vehicle body 1 through the bolt, and the outer portions of the electronic turntable and the air cylinder are connected with the controller 3 through the lines, the electronic turntable can drive the image sensor group 4 to rotate, the air cylinder can drive the image sensor group 4 to stretch and retract, and the air cylinder and the electronic turntable are controlled by the controller 3.

[0040] The bottom of the mobile vehicle body 1 is provided with a vehicle lamp 11, the inner cavity of the protection shell 5 is provided with a first spring 54 at the top, the other end of the first spring 54 is connected with the sealing door 55, the first spring 54 can pull the sealing door 55 to reset after movement, and the outer portion of the receiver 52 is provided with a horizontal plate for driving the sealing door 55 to move.

[0041] When the crack sensor 62 performs the retraction action, the precision guide system composed of the hollow column 71 and the piston rod 72 is immediately started. The hollow column 71 forms a low-friction sliding fit with the piston rod 72 through the built-in linear bearing, and cooperates with the multi-stage positioning boss design to correct the sensor motion trajectory in real time, control the traditional free retraction radial offset within ±0.05mm, and the buffer rubber ring configured at the end of the piston rod 72 can effectively absorb the retraction impact force, so that the sensor can still maintain millimeter-level positioning accuracy when moving at high speed, completely eliminating the sensor offset problem caused by vibration in the traditional structure. The hollow column 71 integrates a miniature magnetostrictive displacement sensor in the inner cavity, which cooperates with the self-developed PID control algorithm to monitor the retraction position of the piston rod 72 in real time, with a positioning accuracy of ±0.01mm. When the retraction in-place signal is detected, the system sends a trigger instruction to the receiver 52 through the 433MHz wireless transmission module, and drives the linear rail module 51 to smoothly retreat the temperature and humidity sensor group 53 at an acceleration of 0.3m / s².

[0042] A school facility supervision system adopts the above-mentioned school facility supervision method, comprising: a multi-source perception layer, an edge computing unit, a cloud analysis platform and a decision response module, the multi-source perception layer comprises an integrated vibration sensor array, a temperature and humidity sensor array and an image sensor array, the edge computing unit is configured with a local gateway of an FPGA acceleration chip, the cloud analysis platform is equipped with a server cluster of an LSTM neural network, and the decision response module realizes a hierarchical early warning and maintenance work order generation system.

[0043] The scheme first assembles the sensor group respectively, collects the state data of the facility in real time, ensures the real-time and comprehensiveness of the data, preliminarily cleanses, extracts features and judges abnormalities of the collected original data at the local gateway, improves the data processing efficiency, reduces the network transmission burden, uses a deep learning model such as an LSTM network to deeply analyze the preprocessed data, evaluates the health state of the facility, generates maintenance decision suggestions, realizes intelligent analysis and decision, triggers corresponding response measures such as LED prompts, APP notifications and administrator intervention according to different abnormal levels, ensures timely processing of abnormal conditions, records the maintenance process, and updates the maintenance information to the equipment life cycle database, forms a closed-loop management, optimizes the maintenance process, and reduces the long-term maintenance cost.

[0044] The mobile vehicle body 1 is started to work, the external sensor data is received through the controller 3, the received data is processed and analyzed through the controller 3, and the analysis result is sent through the controller 3 for the staff to check, and the mobile wheel 2 can be started to drive the mobile vehicle body 1 to move, the distance from the sensor is approached, so that the data receiving speed is improved, and the mobile vehicle body 1 can monitor the air in the area without installing the sensor through the protection shell 5 when moving, and the first cylinder 61 is started to drive the crack sensor 62 to descend until the crack sensor 62 drives the L-shaped closed bottom plate 64 to rotate to remove the closure of the protection bottom shell 6, so that the crack of the playground can be scanned and detected, and the crack sensor 62 is guided and processed through the piston rod 72 when moving, so as to move in a straight line, and the piston rod 72 moves in the inner cavity of the hollow column 71 to generate airflow, the airflow enters the inner cavity of the telescopic air bag 74 through the elbow pipe 73, so that the telescopic air bag 74 expands, thereby protecting the optical lens of the crack sensor 62, and when the piston rod 72 moves in the inner cavity of the hollow column 71, the sensor assembled in the inner cavity of the hollow column 71 transmits the signal to the receiver 52, and then the linear track module 51 drives the temperature and humidity sensor group 53 to return, so as to protect the temperature and humidity sensor group 53.

[0045] Through the sensor group deployed on the facility, the state data of the facility is collected in real time, the real-time monitoring and comprehensive coverage of the facility state data are realized, the supervision efficiency and timeliness are improved, the collected original data is preliminarily cleaned, feature extraction and abnormality judgment are performed through the local gateway, the data processing speed is improved, the network transmission pressure is reduced, the real-time of data analysis is ensured, the health state of the facility is accurately evaluated through deep analysis of the preprocessed data by using the deep learning model, scientific maintenance decision basis is provided, the intelligent level of supervision is enhanced, corresponding response measures are triggered according to different abnormal levels, the facility abnormality is effectively responded, the occurrence of accidents is avoided, the reliability of safety management is improved, the maintenance process is recorded, and the maintenance information is updated to the equipment life cycle database, forming a closed-loop management, optimizing the maintenance process, reducing the long-term maintenance cost, and improving the operation efficiency and service life of the facility.

[0046] The mobile wheel 2 can move the mobile vehicle body 1, and then the distance of the sensor is approached, so that the data receiving speed is improved, the air in the area without installing the sensor is monitored, the data coverage area is improved, the crack sensor 62 and the temperature and humidity sensor group 53 are protected through the protection bottom shell 6 and the protection shell 5 respectively, the crack sensor 62 and the temperature and humidity sensor group 53 can be switched to different states in different use environments, so as to avoid being damaged by environmental factors during use, and improve the overall service life of the device.

[0047] By the hollow column 71, the piston rod 72, the elbow pipe 73 and the telescopic air bag 74, when the crack sensor 62 is retracted to the inner cavity of the protection bottom shell 6, the crack sensor 62 is automatically guided, and the stability of the crack sensor 62 is improved.

[0048] By the piston rod 72, the gas in the inner cavity of the hollow column 71 is pressed into the inside of the telescopic air bag 74 through the elbow pipe 73, the telescopic air bag 74 is inflated, then the crack sensor 62 is protected, the service life of the crack sensor 62 is improved, the crack sensor 62 is prevented from being damaged by impact, and at the same time, the sensor is assembled in the inner cavity of the hollow column 71, when the piston rod 72 is retracted, the signal is transmitted to the receiver 52, then the linear rail module 51 drives the temperature and humidity sensor group 53 to retract, so that the temperature and humidity sensor group 53 is protected.

[0049] It should be noted that the relational terms herein such as first and second, and the like, are used solely to distinguish one from another entity or action, without necessarily requiring or implying any actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0050] Although the embodiments of the present application have been shown and described, it is understood that various changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A school facility supervision method, characterized in that: include: S1. Multi-source data collection: Install sensor groups in devices within the school to collect real-time status data of the devices; S2. Edge computing preprocessing: Perform preliminary cleaning, feature extraction, and anomaly detection on the collected raw data at the local gateway; S3, Cloud-based Intelligent Analysis: Use deep learning models to conduct in-depth analysis of pre-processed data, assess the health status of facilities, generate maintenance decision recommendations, and achieve intelligent analysis and decision-making; S4. Multi-level early warning mechanism: Depending on the level of anomaly, corresponding response measures are triggered until the administrator intervenes to ensure timely handling of anomalies; S5. Maintain closed-loop management: Record the maintenance process and update the maintenance information to the equipment lifecycle database to form a closed-loop management, optimize the maintenance process, and reduce long-term maintenance costs.

2. A school facility supervision method according to claim 1, characterized in that: In the multi-source data acquisition step S1, the sensor group includes a vibration sensor, a temperature and humidity sensor, a pressure sensor, and an image sensor.

3. A school facility supervision method according to claim 1, characterized in that: In the S3 and cloud-based intelligent analysis steps, the Horovod framework is used to perform multi-GPU parallel and mixed precision training.

4. A school facility supervision method according to claim 1, characterized in that: In the step S5, maintenance closed-loop management, a three-dimensional model of the facility is constructed using Unity3D, and real-time data is driven, and the Neo4j graph database is used to store fault modes.

5. A school facility monitoring device, using the school facility monitoring method according to any one of claims 1 to 4, comprising: A mobile body (1), mobile wheels (2) and a controller (3), wherein the controller (3) is mounted on the top of the mobile body (1), and the mobile wheels (2) are evenly mounted on the front and back sides of the outside of the mobile body (1), characterized in that: the left and right sides of the top of the mobile body (1) are respectively equipped with an image sensor group (4) and a protective shell (5), the inner cavity of the protective shell (5) is equipped with a linear track module (51), the output end of the linear track module (51) is connected to a receiver (52), the output end of the receiver (52) is connected to a temperature and humidity sensor group (53), and the inner cavity top of the protective shell (5) is connected to a closed door (55) via a hinge; The bottom of the mobile vehicle body (1) is equipped with a protective bottom shell (6), the top of the inner cavity of the protective bottom shell (6) is connected to a first cylinder (61), the output end of the first cylinder (61) is connected to a crack sensor (62), partitions (63) are equipped on both sides of the inner cavity of the protective bottom shell (6), both sides of the inner cavity of the first cylinder (61) are connected to L-shaped closed bottom plates (64) through hinges, and both sides of the inner cavity of the protective bottom shell (6) are connected to guide mechanisms (7); The guide mechanism (7) comprises a hollow column (71), the hollow column (71) being connected to the protective bottom shell (6), the interior of the hollow column (71) being slidably connected to a piston rod (72), the piston rod (72) being connected to the crack sensor (62), the exterior of the hollow column (71) being connected to a curved tube (73), the other end of the curved tube (73) being connected to a telescopic airbag (74), the telescopic airbag (74) being sleeved on the exterior of the crack sensor (62).

6. The school facility monitoring device according to claim 5, characterized in that: A second spring (65) is connected to the top of the L-shaped closed bottom plate (64), and the other end of the second spring (65) is connected to the inner cavity of the protective bottom shell (6) via a bolt.

7. The school facility monitoring device according to claim 5, characterized in that: A limiting plate is embedded in the inner cavity of the piston rod (72), the outer portion of the limiting plate is slidably connected to the inner cavity of the hollow column (71), and an exhaust hole is opened at the top of the hollow column (71).

8. The school facility monitoring device according to claim 5, characterized in that: The bottom of the image sensor group (4) is equipped with an electronic turntable, the bottom of the electronic turntable is equipped with a cylinder, the bottom of the cylinder is connected to the mobile body (1) through bolts, and the outside of the electronic turntable and the cylinder are connected to the controller (3) through lines.

9. The school facility monitoring device according to claim 5, characterized in that: The bottom of the mobile vehicle body (1) is equipped with a headlight (11), the top of the inner cavity of the protective shell (5) is equipped with a first spring (54), and the other end of the first spring (54) is connected to the closed door (55).

10. A school facility monitoring system, using the school facility monitoring method according to any one of claims 1 to 4, comprising: The multi-source perception layer, edge computing unit, cloud analysis platform and decision response module are characterized in that: the multi-source perception layer includes an integrated vibration sensor array, a temperature and humidity sensor array and an image sensor array; the edge computing unit is configured with a local gateway of an FPGA acceleration chip; the cloud analysis platform is equipped with a server cluster of an LSTM neural network; and the decision response module implements a hierarchical early warning and maintenance work order generation system.

Images