Safety patrol method, patrol device and patrol equipment

Abstract

The invention provides a safety patrol method, a patrol device and patrol equipment, and relates to the technical field of building safety, and the safety patrol method comprises the steps: responding to an operation instruction of an editable interface, configuring a multi-stage process template of building safety patrol, and generating a corresponding safety patrol task; the shooting device collects real-time patrol data of a construction site; identifying violation operation, equipment abnormity and environment risk in the real-time patrol data, and generating a corresponding risk level result; generating a visual alarm signal according to the risk level result, issuing the visual alarm signal to a mobile terminal, and sending the visual alarm signal to a monitoring center when the visual alarm signal is an emergency signal; monitoring response information of the mobile terminal, updating the state of the safety patrol task and recording an operation log of processing a risk level result by patrol personnel; if the mobile terminal does not respond within the preset time, the alarm information is sent to the mobile terminal again, the alarm information is sent to the monitoring center, and the patrol efficiency and the risk handling capacity can be improved.

Description

Technical Field

[0001] The present application relates to the field of building safety technology, and in particular to a safety inspection method, inspection device and inspection equipment. Background Art

[0002] In the construction industry, traditional safety inspection methods rely primarily on manual on-site inspections, paper records, and fragmented electronic management systems. For example, inspectors must manually complete inspection forms, record hazards by taking photos or videos, and then report this data to management level by level. Risk prediction and real-time response still rely on manual analysis, resulting in limited efficiency gains. Summary of the Invention

[0003] The main purpose of the present invention is to provide a safety inspection method, aiming to improve the real-time and accuracy of construction site safety supervision.

[0004] To achieve the above objectives, the present invention provides a safety inspection method, which includes: In response to an operation instruction on the editable interface, a multi-level process template for building safety inspections is configured, and template parameters of the multi-level process template are adjusted according to the operation instruction to generate corresponding safety inspection tasks, wherein the multi-level process template includes customer management, risk assessment, approval input, and supply chain management modules; According to the safety inspection task, real-time inspection data of the construction site is collected by a camera device of a mobile terminal of an inspector, wherein the real-time inspection data includes at least one of images, videos or sensor data; Identify illegal operations, equipment anomalies, and environmental risks in the real-time inspection data and generate corresponding risk level results; Generate a corresponding visual alarm signal based on the risk level result, and send the visual alarm signal to the mobile terminal that collects the real-time inspection data based on LoRa spread spectrum technology. When the visual alarm signal is an emergency signal, send the visual alarm signal to the monitoring center via the 5G / 4G network; Monitor the response information of the corresponding mobile terminal, update the status of the security inspection task and record the operation log of the inspector processing the risk level result; If the mobile terminal does not respond to the visual alarm signal within a preset time, the alarm information is sent to the mobile terminal again through the backup communication method, and the alarm information is sent to the monitoring center.

[0005] Optionally, in response to the operation instruction of the editable interface, a multi-level process template for building safety inspection is configured, specifically: Obtain a standardized process template library corresponding to the building type or project stage, wherein the standardized process template library includes risk management processes for residential buildings, commercial buildings, industrial plants, and different construction stages; In response to an operation instruction on the editable interface, adding or deleting a process node to the standardized process template library, and configuring a corresponding trigger condition for each modified process node, the trigger condition including an environmental threshold, an equipment status threshold, and a manual judgment rule; Bind the execution personnel role to each process node, and determine the operation permissions, task execution deadlines, and timeout alarm rules corresponding to the personnel role; The trigger conditions, operation permissions, task execution deadlines and timeout alarm rules are imported into the standardized process template library to generate the multi-level process template adapted to the current project.

[0006] Optionally, the triggering conditions corresponding to each modified process node are configured as follows: Obtain historical inspection data of the current project and a risk case database of similar projects; The risk type and trigger probability are calculated based on the historical inspection data and the risk case library of similar projects through a time series prediction model to generate corresponding prediction results; Associating the prediction result with the trigger condition; The step of importing the trigger conditions, operation permissions, task execution deadlines, and timeout warning rules into the standardized process template library to generate the multi-level process template adapted to the current project includes: Mapping the trigger condition to a corresponding process node to automatically trigger the process node; When the trigger conditions are mapped to corresponding process nodes, task execution permissions are assigned to different roles according to the operation permissions to generate and determine the task execution deadlines and timeout alarm rules corresponding to the task execution permissions to generate the multi-level process template.

[0007] Optionally, adjusting the template parameters of the multi-level process template according to the operation instruction to generate a corresponding safety inspection task is specifically: Obtaining the engineering progress data, resource allocation table, historical risk distribution map, and material scheduling plan of the current project, and extracting the first parameter in the multi-level process template, the first parameter including the contract performance period, the hidden danger weight coefficient, and the budget allocation ratio; Determine the deviation between the engineering progress data of the current project and the contract performance period, the matching degree between the resource allocation table and the budget allocation ratio, and the correlation between the historical risk distribution map and the hidden danger weight coefficient; Determining a deviation threshold range of the first parameter according to the deviation value, the matching degree, and the correlation; Pushing the deviation threshold range of the first parameter to the editable interface, monitoring the parameter adjustment operation on the editable interface, updating the template parameters of the multi-level process template, and associating and binding the updated template parameters with the material scheduling plan; The template parameters bound to the material scheduling plan are logically checked, and after successful verification, a safety inspection task including task priority, inspection route planning and resource allocation instructions is generated.

[0008] Optionally, determining a deviation threshold range of the first parameter according to the deviation value, the matching degree, and the correlation includes: generating an initial deviation threshold corresponding to the deviation value; Calculating the priority weight of resource allocation corresponding to the matching degree by a weighted calculation method, and mapping it to the initial deviation threshold to adjust and generate a second deviation threshold; The hidden danger probability distribution corresponding to the correlation is predicted by a Bayesian probability model and integrated into a second deviation threshold to determine the deviation threshold range of the first parameter.

[0009] Optionally, the updated template parameters are associated and bound with the material scheduling plan, specifically: Through graph modeling technology, a topological relationship is established between the updated template parameters and the key nodes in the material scheduling plan; Embed a real-time monitoring trigger in the topological relationship, and automatically trigger a dynamic check of the template parameters when the status of a key node in the material scheduling plan changes; And adjust the template parameters in real time according to the results of dynamic verification.

[0010] Optionally, identifying illegal operations, equipment anomalies, and environmental risks in the real-time inspection data and generating corresponding risk level results includes: Invoke a pre-trained violation identification model to perform frame-level analysis on the photos or videos in the real-time patrol data to generate corresponding analysis results; When the analysis result indicates a violation, the facial features and violation timestamp of the violator are extracted to generate structured data including the employee ID and violation type. Obtain equipment operation data and environmental monitoring data at the construction site; The structured data, the equipment operation data and the environmental monitoring data are input into a risk level assessment model to generate a corresponding risk level result.

[0011] Optionally, generating a corresponding visual warning signal from the risk level result is specifically: Matching a preset alarm template library based on the risk level results, the alarm template library includes equipment failure alarm templates, environmental exceeding standard alarm templates and personnel violation alarm templates; Perform spatial enhancement processing on the alarm signal, superimpose the coordinates of the hidden danger location on the BIM model of the construction site, and mark the relationship with the affected equipment in the editable interface; If it is an environmental risk warning, the link to the historical similar accident handling case and the emergency material inventory status are displayed in the corresponding display position of the editable interface; And according to the risk level results, warning icons with different colors and different flashing frequencies are generated on the editable interface.

[0012] In addition, to achieve the above-mentioned purpose, the present invention also provides a patrol device, which includes: a memory, a processor, and a safety patrol program stored in the memory and executable on the processor, wherein the safety patrol program is configured to implement the safety patrol method described above.

[0013] In addition, to achieve the above-mentioned purpose, the present invention also provides an inspection equipment, including the inspection device as described above.

[0014] The embodiment of the present invention configures a multi-level process template for building safety inspections in response to operating instructions on an editable interface, and adjusts template parameters of the multi-level process template according to the operating instructions to generate corresponding safety inspection tasks, wherein the multi-level process template includes customer management, risk assessment, approval investment and supply chain management modules, and then collects real-time inspection data of the construction site through the shooting device of the inspector's mobile terminal according to the safety inspection task, wherein the real-time inspection data includes at least one of images, videos or sensor data, and then identifies illegal operations, equipment anomalies and environmental risks in the real-time inspection data, generates corresponding risk level results, and generates corresponding visual alarm signals based on the risk level results, and sends the visual alarm signals to The mobile terminal corresponding to the real-time inspection data is collected, and when the visual alarm signal is an emergency signal, the visual alarm signal is sent to the monitoring center through the 5G / 4G network, and the response information of the corresponding mobile terminal is monitored, the status of the safety inspection task is updated, and the operation log of the inspector's handling of the risk level results is recorded. Finally, if the mobile terminal does not respond to the visual alarm signal within the preset time, the alarm information is sent to the mobile terminal again through the backup communication method, and the alarm information is sent to the monitoring center. By configuring multi-level process templates, real-time data collection, intelligent risk analysis and closed-loop alarm management, dynamic safety supervision of the entire life cycle of the building is realized, which significantly improves the inspection efficiency and risk handling capabilities, ensures the accuracy and timeliness of building safety management, and reduces the accident rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0017] Figure 1 A schematic diagram of a safety inspection method flow chart according to an embodiment of the present invention; Figure 2 A schematic flow chart of a safety inspection method according to another embodiment of the present invention; Figure 3 A schematic flow chart of a safety inspection method according to another embodiment of the present invention; Figure 4 A schematic flow chart of a safety inspection method according to another embodiment of the present invention; Figure 5A schematic flow chart of a safety inspection method according to another embodiment of the present invention; Figure 6 A schematic flow chart of a safety inspection method according to another embodiment of the present invention; Figure 7 A schematic flow chart of a safety inspection method according to another embodiment of the present invention; Figure 8 This is a flow chart of a safety inspection method according to another embodiment of the present invention.

[0018] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments, and the well-known modules, units and their connections, links, communications or operations are not shown or described in detail. In addition, the described features, architectures or functions can be combined in any way in one or more embodiments. It should be understood by those skilled in the art that the various embodiments described below are only for illustration and are not intended to limit the scope of protection of the present invention. It can also be easily understood that the modules or units or processing methods in the various embodiments described herein and shown in the drawings can be combined and designed according to various different configurations. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0020] The definitions of various nouns or methods in the following embodiments, except for those that are logically untenable, are generally based on the broad concepts that can be implemented under the premise of the disclosure in the embodiments. Under such an understanding, the various specific subordinate specific definitions of the nouns or methods should be regarded as the inventive content of the present invention, and should not be narrowly understood or interpreted in a biased manner on the grounds that the specification does not disclose such specific definitions. Similarly, under the premise that it can be logically implemented, the order of the steps in the method is flexible and changeable, and the specific subordinate specific definitions of the broad concepts of various nouns or methods are all within the scope of protection of the present invention.

[0021] In the area of ​​safety supervision in the construction industry, traditional safety inspection methods rely primarily on manual on-site inspections, paper records, and decentralized electronic management systems. For example, inspectors must manually fill out inspection forms, record hidden dangers by taking photos or videos, and then report the data to management level by level. This method suffers from cumbersome processes and delayed responses. In particular, risk prediction and real-time response still rely on manual analysis, resulting in low efficiency and the easy omission of key risk points. In addition, traditional methods lack a unified multi-level process management mechanism, making it impossible to dynamically adjust inspection strategies based on building type or project stage, and it is difficult to achieve automated identification of risk data and graded alerts. In terms of communication methods, traditional systems typically use a single network to transmit alarm signals. When mobile terminals fail to respond in a timely manner, there is a lack of effective backup communication mechanisms, which may delay the handling of emergency situations.

[0022] The main solution of the embodiment of the present application is: by responding to the operation instructions of the editable interface, configuring a multi-level process template for building safety inspection, and adjusting the template parameters of the multi-level process template according to the operation instructions to generate corresponding safety inspection tasks, wherein the multi-level process template includes customer management, risk assessment, approval investment and supply chain management modules, and then according to the safety inspection tasks, collecting real-time inspection data of the construction site through the shooting device of the mobile terminal of the inspector, wherein the real-time inspection data includes at least one of images, videos or sensor data, and then identifying illegal operations, equipment anomalies and environmental risks in the real-time inspection data, generating corresponding risk level results, and generating risk level results. The corresponding visual alarm signal is sent to the mobile terminal corresponding to the real-time patrol data based on LoRa spread spectrum technology. When the visual alarm signal is an emergency signal, the visual alarm signal is sent to the monitoring center through the 5G / 4G network, and the response information of the corresponding mobile terminal is monitored, the status of the safety patrol task is updated, and the operation log of the patrol personnel handling the risk level results is recorded. Finally, if the mobile terminal does not respond to the visual alarm signal within the preset time, the alarm information is sent to the mobile terminal again through the backup communication method, and the alarm information is sent to the monitoring center, by configuring multi-level process templates, real-time data collection, intelligent risk analysis and closed-loop alarm management.

[0023] In this embodiment, for ease of description, the following description is made with the patrol device as the execution entity.

[0024] This application provides a solution that addresses the cumbersome processes, delayed responses, and limited communication capabilities of traditional methods by configuring multi-level process templates to generate inspection tasks, collecting and analyzing construction site data in real time, issuing alarm signals based on a multi-network communication mechanism, and ensuring reliable response through backup communications. This solution improves inspection process efficiency, enables real-time risk identification and graded alarms, and strengthens communication redundancy. This approach enables dynamic safety supervision throughout the building lifecycle, significantly improving inspection efficiency and risk management capabilities, ensuring the accuracy and timeliness of building safety management, and reducing the incidence of accidents.

[0025] To this end, the present invention proposes a security patrol method; it can be understood that a patrol device for storing and executing the following method is provided in the patrol equipment, and the patrol device can be implemented by a main controller, such as an MCU (Microcontroller Unit), a DSP (Digital Signal Process), an FPGA (Field Programmable Gate Array), a SOC (System On Chip), etc.

[0026] In existing technologies, safety oversight in the construction industry primarily relies on manual on-site inspections and paper records. Inspectors are required to manually fill out forms and take photos to record potential hazards, which are then reported to management level by level. Traditional electronic management systems suffer from fragmented processes, and risk prediction and real-time response still rely on manual analysis, resulting in low efficiency. For example, at large construction sites, the complex working environment makes paper records easily lost, and decentralized systems hinder multi-departmental collaboration, making it difficult to promptly address sudden risks.

[0027] To solve the above problems, traditional methods have three core defects: lack of unified standards for process management, insufficient efficiency in data collection and analysis, and unreliable transmission of alarm information. To address these problems, we first build a configurable multi-level process template to integrate modules such as customer management and risk assessment to achieve dynamic adaptation of standard processes and project characteristics; secondly, we use mobile terminals to collect multimodal data in real time, and combine it with intelligent recognition models to improve risk discovery efficiency; finally, we establish a dual-channel communication mechanism to ensure basic alarm transmission through LoRa, and give priority to superimposing emergency signals on 5G / 4G, while setting backup communication trigger conditions to ensure alarm accessibility. This series of designs breaks through traditional single-point solutions and forms a complete closed-loop management system. Based on the above, refer to Figure 1 In one embodiment of the present invention, the safety inspection method includes steps S100-S600, wherein: S100, in response to an operation instruction of the editable interface, configuring a multi-level process template for building safety inspection, and adjusting template parameters of the multi-level process template according to the operation instruction to generate a corresponding safety inspection task; S200: According to the safety inspection task, real-time inspection data of the construction site is collected by a camera device of an inspector's mobile terminal; S300, identifying illegal operations, equipment anomalies, and environmental risks in the real-time inspection data, and generating corresponding risk level results; S400, generating a corresponding visual alarm signal based on the risk level result, and sending the visual alarm signal to the mobile terminal corresponding to the real-time inspection data based on LoRa spread spectrum technology, and when the visual alarm signal is an emergency signal, sending the visual alarm signal to the monitoring center via the 5G / 4G network; S500: Monitor the response information of the corresponding mobile terminal, update the status of the security inspection task and record the operation log of the inspection personnel processing the risk level result; S600: If the mobile terminal does not respond to the visual alarm signal within a preset time, an alarm message is sent to the mobile terminal again through a backup communication method, and an alarm message is sent to the monitoring center.

[0028] The multi-level process template includes modules for customer management, risk assessment, approval and investment, and supply chain management. It represents a dynamically configurable building safety management framework that uses a visual programming interface to add, delete, and modify process nodes, supporting the integration of customer needs, risk assessment standards, and other factors into a unified management platform. Real-time inspection data includes at least one of images, videos, or sensor data. Real-time inspection data is collected using mobile terminals with integrated cameras, environmental sensors, and other devices, such as industrial-grade tablets with waterproof and dustproof features for on-site recording. Risk level identification utilizes a pre-trained image recognition model and a modified violation detection algorithm based on YOLOv5 to automatically identify missing safety equipment and abnormal equipment status. Visual alarm signals can be generated through BIM model coordinate mapping, for example, highlighting the location of hidden dangers within a 3D building model. Backup communication methods can utilize satellite communication modules or mesh network devices, automatically switching transmission paths when the primary communication channel fails.

[0029] Among them, when project managers configure the process template through the editable interface, the inspection device automatically links the contract terms of the customer management module with the inspection standards of the risk assessment module. After the inspection task is generated, the mobile terminal automatically pushes the inspection route based on GPS positioning, and the video data collected on-site is uploaded to the cloud after preliminary analysis by the edge computing device. When the scaffolding fixing bolts are identified as missing, the inspection device automatically marks the hidden danger location and assesses it as a secondary risk. It broadcasts an alarm to all terminals within a range of 200 meters through the LoRa base station, and transmits the alarm information to the headquarters monitoring screen via the 5G network. After receiving an unconfirmed alarm, the monitoring center automatically activates the backup communication channel to send a text message reminder to the project leader's mobile phone to ensure a closed-loop risk management.

[0030] Compared with existing technologies, traditional methods use independent risk assessment systems and manual inspection processes, resulting in data silos and response delays. This embodiment achieves cross-departmental business integration through multi-level process templates, eliminating the time lost in the circulation of paper work orders. A dual-channel communication mechanism improves the arrival rate of alarm information in complex building environments compared to a single wireless transmission method. An intelligent recognition model replaces manual screening of image data, reducing the time it takes to discover hidden dangers from hours to minutes. The provision of backup communication trigger conditions effectively solves the problem of alarm loss caused by signal obstruction in existing technologies.

[0031] This embodiment implements standardized management of the building safety inspection process, improving the efficiency of cross-departmental collaboration; ensures the real-time and accuracy of risk identification at the construction site; establishes a hierarchical alarm transmission mechanism to ensure the accessibility of key information in scenarios with limited communication conditions; forms a complete risk management supervision chain, and achieves accountability traceability through operation logs.

[0032] This embodiment configures a multi-level process template for building safety inspections in response to operating instructions on an editable interface, and adjusts template parameters of the multi-level process template according to the operating instructions to generate corresponding safety inspection tasks, wherein the multi-level process template includes customer management, risk assessment, approval investment and supply chain management modules, and then collects real-time inspection data of the construction site through the shooting device of the inspector's mobile terminal according to the safety inspection task, wherein the real-time inspection data includes at least one of images, videos or sensor data, and then identifies illegal operations, equipment anomalies and environmental risks in the real-time inspection data to generate corresponding risk level results, so as to generate corresponding visual alarm signals for the risk level results, and sends the visual alarm signals to the mobile terminal corresponding to the real-time inspection data based on LoRa spread spectrum technology, and when the visual alarm signal is an emergency signal, the The 5G / 4G network transmits visual alarm signals to the monitoring center, monitors the responses of corresponding mobile terminals, updates the status of safety inspection tasks, and records the operation log of inspectors' risk level processing results. Finally, if the mobile terminal does not respond to the visual alarm signal within a preset time, the alarm information is re-sent to the mobile terminal via a backup communication method, and an alarm information is sent to the monitoring center. By configuring multi-level process templates, real-time data collection, intelligent risk analysis, and closed-loop alarm management, the system generates inspection tasks, collects and analyzes construction site data in real time, issues alarm signals based on multi-network communication mechanisms, and combines backup communications to ensure response reliability. This solves the problems of traditional methods such as cumbersome processes, delayed responses, and single communication. It improves inspection process efficiency, realizes real-time risk identification and graded alarms, and strengthens communication redundancy. It realizes dynamic safety supervision throughout the building life cycle, significantly improves inspection efficiency and risk management capabilities, ensures the accuracy and timeliness of building safety management, and reduces the incidence of accidents.

[0033] Optionally, refer to Figure 2 Another embodiment of the present invention provides a safety inspection method based on the above Figure 1 In the illustrated embodiment, in response to the operation instructions of the editable interface, a multi-level process template for building safety inspection is configured, specifically steps S110-S140, wherein: S110, obtaining a standardized process template library corresponding to the building type or project stage; S120, in response to an operation instruction on the editable interface, adding or deleting a process node to the standardized process template library, and configuring a corresponding trigger condition for each modified process node; S130: Bind an executor role to each process node, and determine the operation authority, task execution deadline, and timeout alarm rules corresponding to the executor role; S140: Import the trigger conditions, operation permissions, task execution deadlines, and timeout alarm rules into the standardized process template library to generate the multi-level process template adapted to the current project.

[0034] Among them, the standardized process template library includes risk management processes for residential, commercial buildings, industrial plants and different construction stages. The standardized process template library refers to a pre-established database containing risk management processes for various building types and construction stages. It can be implemented by combining a relational database with classification tags to provide a basic framework for different projects; the trigger conditions include environmental thresholds, equipment status thresholds and manual judgment rules. The trigger conditions refer to the environment, equipment or manual judgment standards that need to be met when the process node is started. It can be implemented by combining a threshold setting module with a rule engine to convert manual experience into automatically executable conditional judgments; the executor role refers to the job responsibilities associated with the process node, which can be implemented using a role authority matrix and an RBAC model to clarify the operating scope and timeliness requirements of different positions.

[0035] First, a basic template matching the current building type is retrieved from the standardized process template library, such as a template for the construction phase of an industrial plant. Process nodes are then adjusted using an editable interface, such as adding a height work inspection node to the risk assessment module. Trigger conditions are configured for this node, such as automatically triggering an inspection task when wind speed exceeds a set threshold. The node is then associated with a safety officer role, with the task required to be completed within 24 hours and an alarm triggered if it exceeds the set threshold. Finally, the configured node parameters are imported into the template library to create a customized template with a multi-level approval process.

[0036] This embodiment implements the dynamic addition and deletion of process nodes through an editable interface, and combines the trigger condition configuration with the role permission binding, so that the same template library can quickly generate different processes adapted to the main construction phase of commercial buildings or the equipment installation phase of industrial plants, while maintaining the consistency of the core risk management logic. This embodiment solves the problem that fixed process templates cannot adapt to the needs of diverse projects, and realizes rapid customized configuration based on standardized templates, so that process nodes can be dynamically adjusted according to specific project characteristics, and at the same time ensures the standardization and timeliness of task execution by binding role permissions and timeliness rules.

[0037] Optionally, refer to Figure 3 Another embodiment of the present invention provides a safety inspection method based on the above Figure 2 In the embodiment shown, corresponding trigger conditions are configured for each modified process node, specifically steps S121-S123, wherein: S121. Obtain historical inspection data of the current project and a risk case database of similar projects; S122. Calculate the risk type and trigger probability based on the historical inspection data and the risk case library of similar projects using a time series prediction model to generate corresponding prediction results; S123, associating the prediction result with the trigger condition; The step of importing the trigger conditions, operation permissions, task execution deadlines, and timeout warning rules into the standardized process template library to generate the multi-level process template adapted to the current project includes steps S141-S142, wherein: S141. Mapping the trigger condition to a corresponding process node to automatically trigger the process node; S142. When mapping the trigger conditions to corresponding process nodes, assign task execution permissions to different roles according to the operation permissions to generate and determine the task execution deadlines and timeout alarm rules corresponding to the task execution permissions to generate the multi-level process template.

[0038] Among them, the time series prediction model refers to a prediction model built based on time series analysis methods. It can be implemented using the ARIMA model or LSTM neural network and is used to capture the temporal correlation and periodicity of risk events. Trigger condition mapping refers to establishing a relationship between risk prediction results and process node execution logic. It can be implemented using a rule engine or event-driven architecture to automatically link risk events with process actions. The risk case library of similar projects is a structured database that stores historical risk events from different engineering projects. It can be built using relational databases or knowledge graph technology to provide cross-project risk model references.

[0039] Among them, by extracting the risk occurrence time, type and processing records from historical inspection data, combined with the risk statistical characteristics of similar projects in the case library of similar projects, the time series prediction model is used to calculate the risk occurrence probability and potential impact range in different construction stages. The prediction results are converted into conditional judgment rules. For example, when the probability of high-altitude work violations in a certain area exceeds the threshold, the inspection process node of the area is automatically triggered. The triggering conditions are embedded in the execution logic of the process node. When the sensor monitoring data or manual inspection results meet the preset conditions, the inspection device automatically activates the subsequent approval or rectification tasks. At the same time, task processing personnel are dynamically assigned according to role permissions. For example, high-risk alarms are automatically distributed to the safety supervisor account, and a deadline rule of responding within two hours is set. After the timeout, an upgraded alarm is sent to the next level of management.

[0040] This embodiment establishes a data-driven trigger condition generation mechanism by integrating historical data with an external case library, enabling process nodes to respond to changes in risk probabilities in real time and eliminating the lag caused by manual records. This embodiment achieves a dynamic match between task assignment and personnel capabilities by combining permissions with dynamically generated task execution deadlines. This embodiment can dynamically optimize process trigger rules based on actual engineering data, reduce the process configuration error rate caused by manual experience deviations, and at the same time shorten risk response time through an automated trigger mechanism to avoid the expansion of safety hazards caused by manual operation delays. And through the dynamic binding of permissions and task deadlines, it ensures that high-risk tasks are assigned first to personnel with processing authority, thereby improving task execution efficiency and responsibility traceability.

[0041] Optionally, refer to Figure 4 Another embodiment of the present invention provides a safety inspection method based on the above Figure 1 In the embodiment shown, the template parameters of the multi-level process template are adjusted according to the operation instructions to generate corresponding safety inspection tasks, specifically steps S150-S190, wherein: S150, obtaining the engineering progress data, resource allocation table, historical risk distribution map, and material scheduling plan of the current project, and extracting the first parameter in the multi-level process template; S160: Determine the deviation between the engineering progress data of the current project and the contract performance period, the matching degree between the resource allocation table and the budget allocation ratio, and the correlation between the historical risk distribution map and the hidden danger weight coefficient; S170: Determine a deviation threshold range of the first parameter according to the deviation value, the matching degree, and the correlation; S180: Pushing the deviation threshold range of the first parameter to the editable interface, monitoring parameter adjustment operations on the editable interface, updating template parameters of the multi-level process template, and associating and binding the updated template parameters with the material scheduling plan; S190, and logically verify the template parameters bound to the material scheduling plan, and after successful verification, generate a safety inspection task including task priority, inspection route planning and resource allocation instructions.

[0042] The first parameters include the contract fulfillment cycle, the hidden danger weight coefficient, and the budget allocation ratio. Project progress data refers to the time nodes and engineering quantity data for actual project completion. It can be collected using progress management software and used to detect deviations between actual progress and the contract fulfillment cycle. The resource allocation table records the allocation plan for manpower, equipment, and materials. It can be exported from the ERP system and used to assess the alignment between budget allocation ratios and actual resource utilization. The first parameters refer to the standardized parameters preset in the process template. These parameters are extracted from the process template using a parameter extraction algorithm and serve as a baseline reference for dynamic adjustment. The deviation threshold range refers to the reasonable range within which parameters are allowed to be dynamically adjusted. This is dynamically generated using a weighted calculation method and a Bayesian probability model to define the permissible range of parameter fluctuations. The material scheduling plan refers to the time nodes involved in material transportation and equipment deployment. It utilizes data from logistics management inspection devices and is linked to template parameters to ensure synchronization of resource supply with task requirements. Logical verification involves using a rule engine to verify the logical consistency of parameters with the material plan to prevent task conflicts or resource waste caused by parameter errors.

[0043] By collecting project progress data and comparing it with the contract fulfillment cycle, the system automatically detects whether the project is behind schedule or ahead of schedule, generating deviation values ​​as a basis for adjusting the contract cycle. Comparing the resource allocation table with the budget allocation ratio identifies deviations from the original plan and allows budget parameters to be adjusted accordingly. Correlation analysis between historical risk distribution maps and hidden danger weight coefficients dynamically adjusts hidden danger weights for different areas, ensuring that template parameters better align with the actual risk distribution. Deviation thresholds are pushed to an editable interface, allowing operators to adjust parameters based on quantitative indicators. For example, when the project progress deviation exceeds a threshold, the contract fulfillment cycle parameters are automatically adjusted. The updated template parameters are topologically linked to key nodes in the material scheduling plan using graph modeling technology. When the material supply status changes, triggers automatically verify the template parameters for compatibility, re-adjusting the parameters if verification fails. Finally, based on the logically verified template parameters, tasks are generated that include priority sorting, inspection route optimization, and resource instructions, ensuring that inspection tasks remain synchronized with project dynamics.

[0044] This embodiment achieves intelligent calibration of template parameters through automated data association and dynamic calculation of deviation thresholds. By topologically binding this to the material scheduling plan, it forms a closed-loop control system for parameter adjustment and resource supply, addressing the rigidity of plans caused by fixed parameters in traditional methods. This embodiment dynamically matches template parameters to project realities, ensuring that inspection task priorities, route planning, and resource allocation instructions are automatically adjusted based on project progress, resource changes, and risk distribution. This effectively avoids resource waste and task execution deviations caused by parameter lags or errors.

[0045] Optionally, refer to Figure 5 Another embodiment of the present invention provides a safety inspection method based on the above Figure 4 In the illustrated embodiment, determining the deviation threshold range of the first parameter based on the deviation value, the matching degree, and the correlation includes steps S171-S173, wherein: S171. Generate an initial deviation threshold corresponding to the deviation value; S172. Calculate the priority weight of resource allocation corresponding to the matching degree by a weighted calculation method, and map it to the initial deviation threshold to adjust and generate a second deviation threshold; S173. Predict the hidden danger probability distribution corresponding to the correlation through a Bayesian probability model, and integrate it into a second deviation threshold to determine the deviation threshold range of the first parameter.

[0046] The initial deviation threshold refers to the benchmark parameter range generated based on the deviation between the project progress and the contract performance cycle. It can be implemented using the linear interpolation method or the sliding window statistical method to quantify the contract performance constraints. The resource allocation priority weight refers to the dynamic coefficient calculated based on the matching degree between the resource allocation table and the budget allocation ratio. It can be implemented using the hierarchical analysis method or the entropy weight method to reflect the impact of resource scarcity on the threshold. The hidden danger probability distribution refers to the risk occurrence probability model calculated through historical risk data and similar cases. It can be implemented using Monte Carlo simulation or Markov chain to predict the statistical laws of hidden danger occurrence.

[0047] Among them, when there is a deviation between the contract performance cycle and the project progress, the sliding window statistical method is first used to calculate the mean progress deviation of the past four weeks to generate the initial deviation threshold range. Then, based on the relationship between the equipment scheduling efficiency and the budget allocation ratio in the resource allocation table, the entropy weight method is used to calculate the priority weight of each resource item, and the weight value is multiplied by the upper limit of the initial deviation threshold to generate the second deviation threshold. Finally, the collapse accident probability data of similar projects are extracted from the historical database, and the posterior probability of the current hidden danger is calculated through the Bayesian model. The probability value is mapped to the second deviation threshold range for dynamic expansion to form the final deviation threshold range. This process realizes intelligent adjustment of the parameter range through the triple data fusion of rigid constraints of contract performance, dynamic weights of resource scheduling, and statistical laws of hidden danger occurrence.

[0048] This embodiment integrates resource priority weights and hidden danger probability distributions, allowing the threshold range to automatically shrink according to the degree of resource scarcity, while dynamically expanding in combination with hidden danger statistical laws, thereby avoiding misjudgments and improving risk capture sensitivity. This embodiment can generate a deviation threshold range that dynamically matches the actual state of the project, making the triggering conditions for contract performance monitoring, resource allocation adjustments, and hidden danger warnings in safety inspection tasks more in line with the actual project situation. For example, during construction during the rainy season, this method automatically relaxes the environmental monitoring threshold range by integrating historical collapse probability data to avoid frequent false alarms due to weather factors. At the same time, it dynamically adjusts the resource allocation deviation threshold according to the scheduling priority of the pumping equipment to ensure that key equipment is monitored in a timely manner.

[0049] Optionally, refer to Figure 6 Another embodiment of the present invention provides a safety inspection method based on the above Figure 4 In the embodiment shown, the updated template parameters are associated and bound with the material scheduling plan, including steps S181-S183, wherein: S181. Using graph modeling technology, establish a topological relationship between the updated template parameters and the key nodes in the material scheduling plan; S182. Embed a real-time monitoring trigger in the topological relationship, and automatically trigger a dynamic check of the template parameters when the status of a key node in the material scheduling plan changes; S183. Adjust template parameters in real time according to the results of dynamic verification.

[0050] Graph modeling technology involves expressing the relationship between parameters and material scheduling through a network structure composed of nodes and edges. This can be implemented using knowledge graphs or graph databases, such as using Neo4j to build the attributes and association rules between parameters and material nodes. Real-time monitoring triggers are state-aware modules embedded in topological relationships. They can be implemented using an event-driven architecture combined with a rule engine, such as configuring Apache Kafka to listen for state change events on material nodes. Dynamic verification involves logically verifying parameters based on changes in material status. This can be achieved by combining verification algorithms with real-time data stream processing technology, such as executing parameter verification logic through the Flink stream processing engine.

[0051] In the material scheduling plan for the project, there is a correlation between the concrete supply node and the construction progress threshold in the template parameters. When a concrete transport vehicle is delayed due to a malfunction, the real-time monitoring trigger captures the node status change event and immediately triggers the parameter verification process. The inspection device calls the preset verification rule library, for example, comparing the delay time with the construction progress threshold, and finds that the concrete pouring window period set by the original parameters cannot meet the adjusted arrival time. The new progress threshold is calculated through a dynamic verification algorithm, and the construction process schedule in the template parameters is automatically updated. The adjusted parameters are synchronized to the associated inspection task generation module to re-plan the safety inspection route.

[0052] Compared to existing technologies, traditional parameter adjustment relies on manual identification of material changes and parameter associations, for example, managers need to manually check shipping delay records against construction schedules. Existing technologies use linear association methods, such as setting simple time association formulas in spreadsheets, which cannot handle complex relationships with cross-node influences. This embodiment uses graph modeling to establish a multidimensional association network, for example, simultaneously associating multiple nodes such as concrete supply, scaffolding erection progress, and worker schedules, enabling dynamic adjustment of parameters across dimensions.

[0053] Through the aforementioned technical means, this embodiment addresses the issue of adjustment lag caused by insufficient correlation between template parameters and material scheduling. For example, it automatically adjusts safety inspection frequency and key areas when material supply anomalies occur. A real-time monitoring trigger mechanism avoids response delays associated with manual monitoring, for example, completing parameter verification within 5 minutes of an equipment failure. Dynamic parameter adjustment ensures real-time coordination between safety inspection tasks and material scheduling plans. For example, it automatically updates high-altitude work inspection routes based on changes in steel arrival times, optimizing resource allocation efficiency.

[0054] Optionally, refer to Figure 7 Another embodiment of the present invention provides a safety inspection method based on the above Figure 1 The embodiment shown, identifying illegal operations, equipment anomalies, and environmental risks in the real-time inspection data and generating corresponding risk level results, includes steps S310-S340, wherein: S310: Calling a pre-trained violation identification model to perform frame-level analysis on the photos or videos in the real-time patrol data to generate corresponding analysis results; S320: When the analysis result indicates a violation, extract the violator's facial features and violation timestamp to generate structured data including the employee ID and violation type. S330, obtaining equipment operation data and environmental monitoring data at the construction site; S340: Input the structured data, the equipment operation data, and the environmental monitoring data into a risk level assessment model to generate a corresponding risk level result.

[0055] Among them, the pre-trained violation identification model refers to a deep learning model constructed through multi-dimensional training data. It can be implemented by a composite training method that combines convolutional neural networks with helmet wearing detection, high-altitude work safety belt binding status recognition, and hot work permit verification rules. It is used to extract violation characteristics from images or videos. Among them, structured data refers to violation records integrated through standardized fields. It can be implemented by fuzzy processing of facial features and associating work numbers and timestamps. It is used to achieve traceability of violations while protecting privacy. Among them, the risk level assessment model refers to a comprehensive analysis algorithm that integrates multi-source data. It can be implemented by a dynamic weight allocation method of equipment operating parameters and environmental monitoring indicators. It is used to quantitatively assess the impact of risk events.

[0056] The violation identification model uses a convolutional neural network to perform frame-by-frame detection on video streams, identifying violations such as not wearing a hard hat or fastening a seatbelt, and then verifies operational compliance against the hot work permit database. When a violation is detected, the inspection device automatically captures the relevant footage and extracts the violator's facial area for blurring, retaining the work number and timestamp to generate a structured record. Equipment operation data is collected in real time via IoT sensors, and environmental monitoring data includes parameters such as temperature, humidity, and gas concentration. After standardization, this data is input into the risk level assessment model, which calculates a comprehensive risk value based on preset weight coefficients and outputs three risk levels: low, medium, and high.

[0057] Existing methods rely on manual visual inspections of violations, failing to correlate device and environmental data in real time. Furthermore, violation records are often unstructured text. This embodiment utilizes an automated model to implement multi-dimensional data fusion analysis, preserving behavioral traceability while eliminating subjective errors associated with manual recording. Existing technologies typically directly store unprocessed facial images, posing a risk of privacy leakage. This embodiment, however, employs fuzzy processing to ensure the security of personal information.

[0058] Through the aforementioned technical means, this embodiment can automatically identify complex risk factors at construction sites, perform real-time correlation analysis between personnel behavior, equipment status, and environmental parameters, and generate standardized risk level assessment results. By using structured data storage, it not only meets the traceability requirements of safety supervision but also prevents the leakage of sensitive information. The integrated processing of multi-source data effectively improves the accuracy of risk assessments and provides a reliable basis for subsequent emergency response.

[0059] Optionally, refer to Figure 8 Another embodiment of the present invention provides a safety inspection method based on the above Figure 7 In the embodiment shown, the risk level result is used to generate a corresponding visual warning signal, specifically steps S410-S440, wherein: S410, matching a preset alarm template library according to the risk level result; S420: Perform spatial enhancement processing on the alarm signal, superimpose the coordinates of the hidden danger location on the BIM model of the construction site, and mark the association relationship with the affected equipment on the editable interface; S430: If it is an environmental risk warning, display a link to a historical similar accident handling case and the emergency material inventory status at a corresponding display position on the editable interface; S440: Generate warning icons with different colors and different flashing frequencies on the editable interface according to the risk level results.

[0060] The alarm template library is a pre-established collection of standardized alarm formats covering three types: equipment failure, environmental violations, and personnel violations. It can be implemented using XML or JSON data structures, allowing for automatic adaptation of alarm content to different risk types. Spatial enhancement processing is a technology that integrates two-dimensional hazard coordinates with three-dimensional BIM models. This technology can be implemented using a GIS coordinate conversion interface to precisely locate risk points within a three-dimensional scene. Hazard location coordinates refer to the latitude and longitude data of the hazard location, obtained through mobile device GPS or indoor positioning inspection devices. This technology can be implemented using Bluetooth beacons or UWB positioning technology, and is used to associate alarm signals with physical locations. The BIM model is a three-dimensional building information model of the construction site, constructed using Revit or Tekla software, for visualizing the construction scene structure. Historical accident handling case links are hyperlinks to historical accident handling records similar to the current environmental risk. This technology can be implemented using database association query technology to quickly retrieve reference cases. Emergency supply inventory status refers to the real-time quantity of protective equipment, firefighting equipment, and other supplies currently in the project's inventory. This information can be obtained through IoT sensors or inventory management inspection device interfaces to assess the adequacy of emergency response resources. Alarm icons of different colors and flashing frequencies refer to visual identifiers that are dynamically adjusted according to the risk level. They can be implemented using RGB color value encoding and timer controls to distinguish alarm priorities.

[0061] Among them, when the risk level result is generated, the inspection device first matches the corresponding equipment failure, environmental violation or personnel violation template from the alarm template library to ensure the uniform format of the alarm content. After the coordinates of the hidden danger location are obtained through the positioning module, they are spatially matched with the component coordinates in the BIM model to form an alarm mark point in the three-dimensional scene, and the affected equipment or area is marked with a line in the editable interface. For environmental risk alarms, the inspection device automatically retrieves the disposal records of similar accidents in the database, generates clickable case links, and calls the real-time data interface of the material management inspection device to display the inventory of emergency materials. The color of the alarm icon is divided into red, orange and yellow according to the risk level, and the flashing frequency increases as the risk level increases, allowing operators to quickly identify the severity of the alarm through visual features.

[0062] Compared with existing technologies, traditional methods rely on manual marking of hidden danger locations and alarm information is only displayed in text or single-color icons, which cannot be associated with three-dimensional spatial coordinates and historical disposal cases. This embodiment achieves three-dimensional precise positioning of alarm signals by superimposing hidden danger coordinates on the BIM model, solving the problem that traditional two-dimensional floor plans are difficult to intuitively display the scope of risk impact. In addition, the historical case links and material inventory status embedded in the environmental risk alarm avoid response delays caused by manual queries and provide direct data support for emergency response. The dynamically adjusted alarm icon overcomes the technical defect that a single visual identifier is difficult to distinguish priorities through a combination of color and flashing frequency.

[0063] This embodiment achieves precise spatial positioning of alarm signals at the construction site, allowing the relationship between risk points and affected equipment to be intuitively presented; by associating historical similar accident cases with real-time material inventory data, it provides operators with a reference for rapid disposal; through the dynamic combination of color and flashing frequency, alarm information of different risk levels is clearly distinguishable in the visual interface, facilitating the priority handling of high-risk events.

[0064] The present invention also proposes a patrol device, which includes: the patrol device includes: a memory, a processor, and a safety patrol program stored in the memory and executable on the processor, wherein the safety patrol program is configured to implement the safety patrol method described above.

[0065] It is worth noting that since the inspection device of the present invention is based on the above-mentioned safety inspection method, the embodiments of the inspection device of the present invention include all technical solutions of all embodiments of the above-mentioned safety inspection method, and the technical effects achieved are also exactly the same, which will not be repeated here.

[0066] The present invention further provides an inspection device, which includes the inspection device as described in the above embodiment.

[0067] It is worth noting that since the inspection equipment of the present invention is based on the above-mentioned inspection device, the embodiments of the inspection equipment of the present invention include all technical solutions of all embodiments of the above-mentioned inspection device, and the technical effects achieved are also exactly the same, which will not be repeated here.

[0068] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0069] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0070] Through the above description of the embodiments, those skilled in the art will clearly understand that the methods of the above embodiments can be implemented using software plus the necessary general-purpose hardware platform. Of course, hardware can also be used, but in many cases the former is a more preferred implementation method. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, or optical disk) as described above and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0071] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A safety inspection method, characterized in that: The safety inspection method includes: In response to an operation instruction on the editable interface, a multi-level process template for building safety inspections is configured, and template parameters of the multi-level process template are adjusted according to the operation instruction to generate corresponding safety inspection tasks, wherein the multi-level process template includes customer management, risk assessment, approval input, and supply chain management modules; According to the safety inspection task, real-time inspection data of the construction site is collected by a camera device of a mobile terminal of an inspector, wherein the real-time inspection data includes at least one of images, videos or sensor data; Identify illegal operations, equipment anomalies, and environmental risks in the real-time inspection data and generate corresponding risk level results; Generate a corresponding visual alarm signal based on the risk level result, and send the visual alarm signal to the mobile terminal that collects the real-time inspection data based on LoRa spread spectrum technology. When the visual alarm signal is an emergency signal, send the visual alarm signal to the monitoring center via the 5G / 4G network; Monitor the response information of the corresponding mobile terminal, update the status of the security inspection task and record the operation log of the inspector processing the risk level result; If the mobile terminal does not respond to the visual alarm signal within a preset time, the alarm information is sent to the mobile terminal again through the backup communication method, and the alarm information is sent to the monitoring center.

2. The safety inspection method according to claim 1, characterized in that: The multi-level process template for building safety inspection is configured in response to the operation instruction of the editable interface, specifically: Obtain a standardized process template library corresponding to the building type or project stage, wherein the standardized process template library includes risk management processes for residential buildings, commercial buildings, industrial plants, and different construction stages; In response to an operation instruction on the editable interface, adding or deleting a process node to the standardized process template library, and configuring a corresponding trigger condition for each modified process node, the trigger condition including an environmental threshold, an equipment status threshold, and a manual judgment rule; Bind the execution personnel role to each process node, and determine the operation permissions, task execution deadlines, and timeout alarm rules corresponding to the personnel role; The trigger conditions, operation permissions, task execution deadlines and timeout alarm rules are imported into the standardized process template library to generate the multi-level process template adapted to the current project.

3. The safety inspection method according to claim 2, characterized in that: The trigger conditions corresponding to each modified process node are configured as follows: Obtain historical inspection data of the current project and a risk case database of similar projects; The risk type and trigger probability are calculated based on the historical inspection data and the risk case library of similar projects through a time series prediction model to generate corresponding prediction results; Associating the prediction result with the trigger condition; The step of importing the trigger conditions, operation permissions, task execution deadlines, and timeout warning rules into the standardized process template library to generate the multi-level process template adapted to the current project includes: Mapping the trigger condition to a corresponding process node to automatically trigger the process node; When the trigger conditions are mapped to corresponding process nodes, task execution permissions are assigned to different roles according to the operation permissions to generate and determine the task execution deadlines and timeout alarm rules corresponding to the task execution permissions to generate the multi-level process template.

4. The safety inspection method according to claim 1, characterized in that: The template parameters of the multi-level process template are adjusted according to the operation instruction to generate the corresponding safety inspection task, specifically: Obtaining the engineering progress data, resource allocation table, historical risk distribution map, and material scheduling plan of the current project, and extracting the first parameter in the multi-level process template, the first parameter including the contract performance period, the hidden danger weight coefficient, and the budget allocation ratio; Determine the deviation between the engineering progress data of the current project and the contract performance period, the matching degree between the resource allocation table and the budget allocation ratio, and the correlation between the historical risk distribution map and the hidden danger weight coefficient; Determining a deviation threshold range of the first parameter according to the deviation value, the matching degree, and the correlation; Pushing the deviation threshold range of the first parameter to the editable interface, monitoring the parameter adjustment operation on the editable interface, updating the template parameters of the multi-level process template, and associating and binding the updated template parameters with the material scheduling plan; The template parameters bound to the material scheduling plan are logically checked, and after successful verification, a safety inspection task including task priority, inspection route planning and resource allocation instructions is generated.

5. The safety inspection method according to claim 4, characterized in that: The determining, according to the deviation value, the matching degree, and the correlation, a deviation threshold range of the first parameter includes: generating an initial deviation threshold corresponding to the deviation value; Calculating the priority weight of resource allocation corresponding to the matching degree by a weighted calculation method, and mapping it to the initial deviation threshold to adjust and generate a second deviation threshold; The hidden danger probability distribution corresponding to the correlation is predicted by a Bayesian probability model and integrated into a second deviation threshold to determine the deviation threshold range of the first parameter.

6. The safety inspection method according to claim 4, characterized in that: The updated template parameters are associated and bound with the material scheduling plan, specifically: Through graph modeling technology, a topological relationship is established between the updated template parameters and the key nodes in the material scheduling plan; Embed a real-time monitoring trigger in the topological relationship, and automatically trigger a dynamic check of the template parameters when the status of a key node in the material scheduling plan changes; And adjust the template parameters in real time according to the results of dynamic verification.

7. The safety inspection method according to claim 1, characterized in that: The identification of illegal operations, equipment anomalies and environmental risks in the real-time inspection data and the generation of corresponding risk level results include: Invoke a pre-trained violation identification model to perform frame-level analysis on the photos or videos in the real-time patrol data to generate corresponding analysis results; When the analysis result indicates a violation, the facial features and violation timestamp of the violator are extracted to generate structured data including the employee ID and violation type. Obtain equipment operation data and environmental monitoring data at the construction site; The structured data, the equipment operation data and the environmental monitoring data are input into a risk level assessment model to generate a corresponding risk level result.

8. The safety inspection method according to claim 7, characterized in that: The risk level result is used to generate a corresponding visual warning signal, specifically: Matching a preset alarm template library based on the risk level results, the alarm template library includes equipment failure alarm templates, environmental exceeding standard alarm templates and personnel violation alarm templates; Perform spatial enhancement processing on the alarm signal, superimpose the coordinates of the hidden danger location on the BIM model of the construction site, and mark the relationship with the affected equipment in the editable interface; If it is an environmental risk warning, the link to the historical similar accident handling case and the emergency material inventory status are displayed in the corresponding display position of the editable interface; And according to the risk level results, warning icons with different colors and different flashing frequencies are generated on the editable interface.

9. A patrol device, characterized in that: The inspection device includes: a memory, a processor, and a safety inspection program stored in the memory and executable on the processor, wherein the safety inspection program is configured to implement the safety inspection method according to any one of claims 1 to 8.

10. A patrol device, characterized in that: Comprising the inspection device as claimed in claim 9.

Images