A Supervision Information Platform for Intelligent Identification and Process-Oriented Management of Major and Critical Engineering Projects

Abstract

This invention relates to the field of supervision technology, specifically to a supervision information platform for intelligent identification and process-oriented management of critical engineering projects. It includes a construction scenario simulation unit, a critical node marking unit, a multi-scenario fusion and comparison unit, and an early warning data display unit. This invention facilitates standardized modeling and process demonstration, supporting construction planning, resource allocation, and progress management. It combines real-time data with AI models to dynamically identify risk points during construction, provide early warnings, reduce safety accidents, and improve identification accuracy and automation. Simultaneously, multi-source monitoring data fusion and image comparison enable automatic deviation identification. Based on historical stage models and early warning data, it can be used for project review, problem tracing, and construction optimization. Furthermore, intelligent push notifications and dynamic displays improve supervision response efficiency and transparency, while graphical displays facilitate communication and collaboration among personnel from different professions and positions.

Description

Technical Field

[0001] This invention relates to the field of supervision technology, and more specifically, to a supervision information platform for intelligent identification and process-oriented management of critical engineering projects. Background Technology

[0002] High-risk and hazardous engineering projects (full name: sub-projects with high risk) refer to sub-projects during construction that may lead to mass casualties or significant economic losses, such as foundation pit engineering, formwork engineering, hoisting and installation / dismantling engineering, scaffolding engineering, and demolition and blasting engineering. During the construction of high-risk and hazardous engineering projects, in conjunction with digital management trends, the process includes four stages: identification, planning, implementation, and acceptance.

[0003] Identification Phase: Before construction, the construction unit should mark critical and major engineering nodes based on the design documents, geological survey reports, and other analytical drawings; Planning Phase: The construction unit prepares a special construction plan, which is reviewed by the construction unit's technical head and signed by the chief supervising engineer; Implementation Phase: The project manager must provide instructions to all workers, conduct daily inspections, and the supervision unit will supervise on-site; Acceptance Phase: Acceptance is carried out in stages, with joint signatures from the construction, supervision, and monitoring units, and the special plan, monitoring report, acceptance records, etc. are compiled.

[0004] Currently, most supervisory information platforms for intelligent identification and process-oriented management of critical and major engineering projects rely on the experience of operators and inspections. However, subtle deviations during the construction of critical and major engineering projects can accumulate and lead to uncontrollable situations. Human supervision is ineffective and has low accuracy. Deploying a large number of sensors for monitoring is inconvenient for real-time viewing of construction progress and discrepancies, resulting in time-consuming and labor-intensive follow-up investigations. In view of this, we propose a supervisory information platform for intelligent identification and process-oriented management of critical and major engineering projects. This platform uses simulated construction scenarios and monitoring data feature comparison to output construction progress and feature differences, and intelligently displays the marked critical and major nodes. Summary of the Invention

[0005] The purpose of this invention is to provide a supervision information platform for intelligent identification and process-oriented management of critical engineering projects, in order to solve the following problems mentioned in the background art:

[0006] Insufficient awareness of construction status: In traditional construction management, managers lack an intuitive understanding of the dynamic changes in risks of critical and major projects, and find it difficult to accurately predict risks through drawings or schedules;

[0007] The identification of critical nodes is lagging: risk points rely on manual experience to mark them, which is easy to miss or misjudge, and the priority cannot be quantified, resulting in unreasonable resource allocation;

[0008] Difficulty in integrating multi-source data: Monitoring videos, BIM models and schedule data are isolated, making it impossible to compare the deviations between actual construction and simulated scenarios in real time;

[0009] Low early warning response efficiency: Traditional early warning relies on manual inspections or single sensors, resulting in delayed information transmission and a lack of visual display, which affects decision-making efficiency;

[0010] Uncontrollable personnel behavior: The inability to detect unauthorized personnel entering or workers violating safety procedures in real time leads to safety hazards;

[0011] To achieve the above objectives, the present invention provides a supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects, including a construction scenario simulation unit, a critical and major node marking unit, a multi-scenario fusion comparison unit, and an early warning data display unit;

[0012] The construction scenario simulation unit establishes a construction simulation scenario based on the real-world scenario of a critical and major project. The construction simulation scenario includes a standardized 3D model and a phased visualization demonstration diagram. The critical and major node marking unit marks critical and major nodes on the 3D model according to the construction stage and displays the critical and major nodes in order of hazard weight.

[0013] The multi-scene fusion comparison unit is used to fuse multi-scene monitoring data of critical and major projects in real time to form VR images. The VR images are divided into multiple feature images by the area where the critical and major nodes are located, and the phased visualization demonstration map of the feature images is output in sequence according to the order. The similarity between the feature images and the phased visualization demonstration map is calculated.

[0014] The early warning data display unit is used to preset a difference threshold, receive feature images with similarity lower than the difference threshold, issue early warning signals, and dynamically display the reasons for the early warning of critical nodes.

[0015] As a further improvement to this technical solution, the construction scenario simulation unit includes a stage division module, a model building module, and a dynamic demonstration module;

[0016] The phase division module is used to divide the entire project into multiple construction phases using project management data and construction process knowledge. Each phase corresponds to a construction status, including structural status, equipment layout, personnel distribution, and risk points.

[0017] The model building module is used to construct a standardized 3D model based on BIM technology, and generates a standard model for each stage output by the stage division module.

[0018] The dynamic demonstration module is used to arrange standardized 3D models of each stage according to the timeline to form a dynamic process demonstration diagram, which supports timeline playback, stage jump, multi-view switching, and animation simulation.

[0019] As a further improvement to this technical solution, the critical node marking unit includes a node marking module and a hazard quantification and sorting module;

[0020] The node marking module is used to establish a risk factor database, extract the construction content of the 3D model of the current construction stage from the model building module, input the construction content into the risk factor database, and output the risk factors of all critical nodes in the 3D model of the current stage.

[0021] The risk quantification and ranking module is used to preset the quantification dimension and scoring rules of each risk factor to establish a risk factor scoring rule table. It sequentially receives multiple risk factors output by the node marking module into the risk factor scoring rule table, outputs weighted scoring values ​​to define risk values, and uses bubble sort to sort multiple risk factors.

[0022] As a further improvement to this technical solution, the critical node marking unit also includes a model highlighting module. The model highlighting module is used to receive the risk factors of all critical nodes in the current stage of the 3D model from the node marking module, and to receive the sorting of multiple risk factors output by the hazard quantification and sorting module. It sets a color mapping relationship where the higher the hazard value, the darker the color, and highlights multiple risk factors in the 3D model according to the color mapping relationship.

[0023] As a further improvement to this technical solution, the multi-scene fusion comparison unit includes a multi-source data acquisition module, an actual highlight display module, and a feature matching module;

[0024] The multi-source data acquisition module is used to divide the 3D model of the final stage into multiple monitoring areas, and allocate several cameras to each area so that the monitoring range of multiple cameras covers the entire monitoring area. The module also matches the camera installation positions in the 3D model with the actual locations of the critical engineering projects to obtain actual VR images.

[0025] The actual highlighting display module is used to calculate the actual risk factors of the 3D model risk factors in the VR image through the node marking module, and to highlight the actual risk factors using the model highlighting display module.

[0026] The feature matching module is used to sort the actual risk factors based on the hazard quantification and sorting module, and then use an image similarity algorithm to compare the features of the actual risk factors with the features of the risk factors in the 3D model, and output a similarity value. If the similarity value is greater than the matching threshold, a normal signal is output; if the similarity value is not greater than the matching threshold, a construction warning signal is output.

[0027] As a further improvement to this technical solution, the multi-scene fusion comparison unit also includes a personnel movement matching module. This module is used to pre-collect standard personnel features of staff, extract current personnel features from the multi-source data acquisition module, and compare the matching degree between the current personnel features and the standard personnel features, including the following postures:

[0028] If the output does not match, the warning data display unit will send a warning signal to irrelevant personnel.

[0029] Second, if the output matches, the current personnel features are extracted from the dynamic flow diagram on the timeline based on the dynamic demonstration module, and defined as standard personnel movement features. At the same time, the current personnel movement features output by the multi-source data acquisition module on the timeline are extracted. The personnel movement features include the path and work area features of each personnel feature. The matching degree between the standard personnel movement features and the current personnel movement features is compared. If the output matches, a normal signal is issued. If the output does not match, a work warning signal is issued to the warning data display unit.

[0030] As a further improvement to this technical solution, the early warning data display unit includes a phased early warning module and a demonstration image retrieval module;

[0031] The staged early warning module is used to receive the construction early warning signal from the feature matching module, identify the construction stage matched by the construction early warning signal, and perform early warning operations according to the construction stage.

[0032] The demonstration diagram display module is used to display two warning reason diagrams in a split screen. The two warning reason diagrams include the actual risk factor VR image highlighted by the actual highlight display module in the current stage, and the dynamic process demonstration diagram output by the dynamic demonstration module before and after the current stage time node.

[0033] As a further improvement to this technical solution, the phased early warning module includes a flashlight and a voice broadcaster;

[0034] The number of flashlights is consistent with the number of multiple construction stages divided by the stage division module, and each construction stage is matched with one flashlight.

[0035] The voice broadcaster is used to preset reminder voices for each stage, identify the reminder voices that match the stage that issues the construction warning signal, and emit sound through a speaker.

[0036] As a further improvement to this technical solution, the early warning data display unit also includes an early warning feedback module. The early warning feedback module is used to establish a mapping relationship between early warning signal types and feedback terminals. The early warning signal types include construction early warning signals, unrelated personnel early warning signals, and work early warning signals. The module identifies the early warning signal type and feeds back the cause of the early warning to the matching terminal device via wireless communication.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] This intelligent identification and process-oriented management platform for critical and major engineering projects uses construction scenario simulation units to visualize the entire construction process, enhancing managers' awareness of the construction status. It can be used for new employee training and construction plan briefings, improving training efficiency and comprehension. Furthermore, the platform uses critical and major node marking units to mark critical and major nodes on a 3D model according to construction stages, displaying these nodes in a hazard weight order. This facilitates standardized modeling and process demonstrations, supporting construction planning, resource allocation, and progress management. It combines real-time data with AI models to dynamically identify risk points during construction, providing early warnings, reducing safety accidents, and improving accuracy and automation. Simultaneously, multi-source monitoring data fusion and image comparison enable automatic deviation identification. Based on historical stage models and early warning data, it can be used for project review, problem tracing, and construction optimization.

[0039] Meanwhile, intelligent push notifications and dynamic displays improve the efficiency and transparency of supervision responses, while graphical displays facilitate communication and collaboration among personnel from different professions and positions. Attached Figure Description

[0040] Figure 1 This is a block diagram illustrating the overall structural principle of the present invention;

[0041] Figure 2 This is a schematic diagram of the personnel movement matching module of the present invention.

[0042] The meanings of the labels in the diagram are as follows:

[0043] 100. Construction scenario simulation unit; 200. Critical and major node marking unit; 300. Multi-scenario fusion and comparison unit; 400. Early warning data display unit. Detailed Implementation

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] Example 1

[0046] Please see Figures 1-2 As shown, this embodiment provides a supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects, including a construction scenario simulation unit 100, a critical and major node marking unit 200, a multi-scenario fusion comparison unit 300, and an early warning data display unit 400.

[0047] The construction scenario simulation unit 100 establishes a construction simulation scenario based on the real-world scenario of a critical and major engineering project. The construction simulation scenario includes a standardized 3D model and phased visualization demonstration diagrams, realizing a visual display of the entire construction process, improving the management personnel's perception of the construction status, and can be used for new employee training and construction plan briefing, improving training efficiency and comprehension. The critical and major node marking unit 200 marks critical and major nodes on the 3D model according to the construction stage and displays critical and major nodes in a sorted order based on hazard weight, which is conducive to assisting in construction plan formulation, resource allocation, and progress management through standardized modeling and process demonstration.

[0048] To enhance visualization and immersion, the construction scene simulation unit 100 includes a phase division module, a model building module, and a dynamic demonstration module.

[0049] The phase division module is used to divide the entire project into multiple construction phases using project management data and construction process knowledge. Each phase corresponds to a construction status, including structural status, equipment layout, personnel distribution, and risk points.

[0050] The process involves inputting project management data and construction process knowledge through data entry, including the following data: engineering drawings (CAD, BIM), project schedule (Project, Gantt chart), construction organization design, safety specifications and standards, and real-time monitoring data (IoT sensors, video surveillance). By parsing the Project file, process nodes and dependencies are extracted. Clustering algorithms (such as K-means) are used to group processes according to job type and resources. Based on the clustering results and construction logic, several stages are defined (the more detailed the stages, the more refined the construction status). Each stage is named, and construction description text is generated.

[0051] The model building module is used to construct standardized 3D models based on BIM technology. It generates a standard model for each stage output from the stage division module. The model includes: structural components (beams, columns, supports, etc.), equipment layout (tower cranes, hoists, scaffolding, etc.), personnel movement (simulated personnel paths, work areas, etc.), and safety signs and risk point markers. Specifically, it outputs standardized 3D model files (such as FBX and GLTF formats) by inputting CAD drawings and BIM models, and based on the construction content description of each stage and the standard component library (scaffolding, supports, tower cranes, etc.).

[0052] When building a standardized 3D model for each stage, the components for the current stage are extracted from the BIM model, and the components are uniformly processed in terms of naming, material, color, etc. The standard equipment model library is called to automatically arrange the equipment, and the path planning algorithm (such as the A* algorithm) is used to simulate the personnel movement. According to the construction content and specifications, risk areas (such as high-altitude operations, cross-construction, etc.) are highlighted in the model.

[0053] The dynamic demonstration module arranges standardized 3D models of each stage along a timeline to form a dynamic process demonstration diagram. It supports timeline playback (simulating construction progress), stage jumping (viewing specific stage models), multi-view switching (top view, eye view, cross-section, etc.), and animation simulation (such as hoisting, demolition, concrete pouring, etc.). Specifically, it generates timeline controls based on stage time, switches between different stage models according to the timeline, and performs animation simulation of key actions (such as hoisting, moving, demolition). Note that keyframe interpolation algorithms are used to achieve smooth transitions. It supports operations such as pause, play, jump, zoom, and view switching to facilitate interactive control for users.

[0054] Secondly, in order to dynamically identify risk points in the construction process, facilitate early warning in actual projects, and reduce safety accidents, the critical node marking unit 200 includes a node marking module and a hazard quantification and sorting module.

[0055] The node marking module is used to establish a risk factor database, extract the construction content (such as component composition, personnel movement and construction stage) of the current construction stage 3D model in the model building module, input the construction content into the risk factor database, and output the risk factors of all critical nodes in the current stage 3D model.

[0056] The process of establishing a structured, scalable, and queryable risk factor database includes the following steps: setting the construction phase in Revit and exporting the component list for the corresponding phase; extracting component data using Revit API / Dynamo / IfcOpenShell; designing the risk factor database to store risk factor rules and support rule engine calls; receiving construction content to match the risk factor database; matching relevant risk factors from the `risk_factors` table, such as matching factors like "high-altitude operations," "scaffolding erection," and "rebar tying" if the current phase is "main structure construction"; traversing the components displayed in the 3D model using the Revit API (C# or Python); extracting the location coordinates of the 3D model through spatial analysis and coordinate processing; and highlighting the area of ​​critical nodes.

[0057] The hazard quantification and ranking module is used to pre-define the quantification dimensions and scoring rules for each risk factor to establish a risk factor scoring rule table. It sequentially receives multiple risk factors output from the node marking module into the risk factor scoring rule table, outputs weighted scoring values ​​to define hazard values, and uses a bubble sort method to rank multiple risk factors. The risk factor scoring rule table can determine hazard values ​​based on the quantification dimensions of the components of existing high-risk projects, for example:

[0058]

[0059] By setting weighted rule scores based on the quantitative dimensions and example values ​​of risk factors, and using rule scores to represent hazard values, data-driven supervision of critical engineering projects can be achieved.

[0060] To further enhance visualization and early warning capabilities, the critical node marking unit 200 also includes a model highlighting module. This module receives risk factors from all critical nodes in the current 3D model from the node marking module, as well as the ranking of multiple risk factors output by the hazard quantification and ranking module. It sets a color mapping relationship where the higher the hazard value, the darker the color. Based on this color mapping relationship, multiple risk factors are highlighted in the 3D model. This highlights the risk factors visually, allowing staff to focus on them through simulated scenarios. The module also displays hazard value labels, and clicking on a node displays a detailed score.

[0061] Then, the multi-scene fusion comparison unit 300 is used to fuse multi-scene monitoring data of critical and major projects in real time to form VR images. The VR images are divided into multiple feature images by the area where the critical and major nodes are located, and the phased visualization demonstration diagrams mapped by the feature images are output in sequence according to the order. The similarity between the feature images and the phased visualization demonstration diagrams is calculated. Real-time data is combined with AI models to dynamically identify risk points in the construction process, provide early warnings, reduce safety accidents, and improve the accuracy and automation level of identification. At the same time, the fusion of multi-source monitoring data and image comparison realizes automatic identification of deviations. Based on historical stage models and early warning data, it can be used for project review, problem tracing and construction optimization.

[0062] Furthermore, the multi-scenario fusion comparison unit 300 includes a multi-source data acquisition module, an actual highlight display module, and a feature matching module;

[0063] The multi-source data acquisition module is used to divide the 3D model of the final stage into multiple monitoring areas. Each area is allocated several cameras so that the monitoring range of multiple cameras covers the entire monitoring area. The installation positions of the cameras in the 3D model are matched with the actual locations of the hazardous engineering projects to obtain actual VR images. The "view frustum" model is used to estimate the camera coverage area. The field of view is usually 45°~90°, and the installation height is recommended to be 3~8 meters (depending on the size of the area). The coverage radius is calculated based on the camera resolution and lens parameters. Visual point placement simulation is performed using tools such as SketchUp to check whether the image stitching is complete. If the stitching is complete, the camera installation positions are output; otherwise, the simulation is repeated.

[0064] The actual highlighting module is used to calculate the actual risk factors of the 3D model risk factors in the VR image through the node marking module, and then highlight the actual risk factors using the model highlighting module. This makes it easier to see the critical nodes in the VR image corresponding to the actual monitoring video in a straightforward way.

[0065] The feature matching module is used to sort actual risk factors based on the hazard quantification and sorting module. It then sequentially uses an image similarity algorithm to compare the features of the actual risk factors with the features of the risk factors in the 3D model, outputting a similarity value. If the similarity value is greater than the matching threshold, a normal signal is output; if the similarity value is not greater than the matching threshold, a construction warning signal is output.

[0066] Image similarity algorithms essentially map image data into a certain feature representation, and then calculate the difference or similarity between these features to obtain a similarity score (0~1 or 0~100%). The higher the similarity score, the more similar the two images are; the lower the score, the greater the difference. This involves extracting high-level semantic features of images through convolutional neural networks (CNNs) and then calculating similarity. Specifically, a dual-branch network with shared weights is used to extract features, and the distance between feature vectors is compared to reflect the similarity.

[0067] And, as Figure 2 The multi-scene fusion comparison unit 300 also includes a personnel movement matching module. This module is used to pre-collect standard personnel features of staff, extract current personnel features from the multi-source data collection module, and compare the matching degree between the current personnel features and the standard personnel features, including the following postures:

[0068] If the output does not match, the warning data display unit 400 will send a warning signal to unauthorized personnel, so as to facilitate timely detection when unauthorized personnel enter the construction site of the dangerous project and improve safety.

[0069] In the second scenario, if the output matches, the dynamic demonstration module extracts the features of the current personnel characteristics in the dynamic process demonstration diagram arranged on the timeline, defining them as standard personnel movement features. Simultaneously, it extracts the current personnel movement features output by the multi-source data acquisition module on the timeline. These personnel movement features include the path and work area features of each personnel characteristic. The matching degree between the standard personnel movement features and the current personnel movement features is compared. If the output matches, a normal signal is issued; if the output does not match, a work warning signal is transmitted to the early warning data display unit 400 to prevent workers from acting disorderly or deviating from the plan on the site of a high-risk project. This avoids situations where other workers or equipment are unaware of the situation and continue working, potentially causing accidents, and also avoids situations where slack work affects project efficiency.

[0070] The early warning data display unit 400 is used to preset the difference threshold, receive feature images with similarity below the difference threshold and issue early warning signals, and dynamically display the reasons for the early warning of critical nodes. Through intelligent push and dynamic display, it improves the efficiency and transparency of the supervision response. The graphical display facilitates communication and collaboration between personnel of different professions and positions.

[0071] The early warning data display unit 400 includes a phased early warning module and a demonstration chart retrieval module;

[0072] The phased early warning module is used to receive construction early warning signals from the feature matching module, identify the construction phase matched by the construction early warning signal, and perform early warning operations according to the construction phase.

[0073] The demo image display module is used to display two warning reason images in a split screen. The two warning reason images include the actual risk factor VR image highlighted by the actual highlight display module in the current stage, and the dynamic process demonstration image output by the dynamic demonstration module before and after the current stage time node. Among them, the dynamic process demonstration image before and after the current stage time node can be a timeline display of the previous stage → current stage → next stage, which helps staff to see the differences.

[0074] It is worth noting that the phased warning module includes a flashing light and a voice broadcaster;

[0075] The number of flashlights is consistent with the number of construction stages divided by the stage division module. Each construction stage is matched with one flashlight. The flashlight generates a pulse signal through a timing circuit (such as a 555 chip) or MCU (microcontroller). The current is transmitted through the LED to emit light. The pulse signal controls the intermittent on-off state to provide early warning. When the stage-based early warning module outputs an engineering early warning signal, it can determine the current construction stage based on the flashlights, and gain a more accurate understanding of the site conditions.

[0076] The voice broadcaster is used to preset reminder voices for each stage, identify the reminder voices that match the stage that issues the construction warning signal, and emit sound through a speaker. On the one hand, it reminds all staff on site of the warning, and on the other hand, it further informs staff about the situation at each stage. The voice player outputs voice signals through electronic synthesis or pre-recorded audio, which are then played through an external speaker.

[0077] The early warning data display unit 400 also includes an early warning feedback module. This module establishes a mapping relationship between early warning signal types and feedback terminals. Early warning signal types include construction early warning signals, unrelated personnel early warning signals, and work early warning signals. The module identifies the early warning signal type and transmits the early warning reason wirelessly to the matching terminal device. For example, when the early warning signal type is a construction early warning signal, the early warning reason is directly sent to the terminal (mobile APP, etc.) of the construction personnel at the current stage, allowing each construction worker to understand the early warning reason in a timely manner, which is beneficial for subsequent adjustments. Similarly, when the early warning is for unrelated personnel, the early warning reason is sent to the terminal of the staff responsible for maintaining the site. When the early warning is for work, the early warning reason is sent to the terminal of the staff whose personnel characteristics and work area characteristics do not match, achieving accurate reminders and improving adjustment efficiency.

[0078] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects, characterized in that: It includes a construction scenario simulation unit (100), a critical node marking unit (200), a multi-scenario fusion comparison unit (300), and an early warning data display unit (400); The construction scenario simulation unit (100) establishes a construction simulation scenario based on the real scene of a major and critical engineering project. The construction simulation scenario includes a standardized 3D model and a phased visualization demonstration diagram. The construction scenario simulation unit (100) includes a phase division module, a model building module, and a dynamic demonstration module. The phase division module is used to divide the entire project into multiple construction phases using project management data and construction process knowledge. Each phase corresponds to a construction state, including structural state, equipment layout, personnel distribution, and risk points. The model building module is used to build a standardized 3D model based on BIM technology and generate a standard model for each phase output by the phase division module. The dynamic demonstration module is used to arrange the standardized 3D models of each phase according to the timeline to form a dynamic process demonstration diagram, which supports timeline playback, phase jump, multi-view switching, and animation simulation. The critical nodes are marked on the 3D model according to the construction stage through the critical node marking unit (200), and the critical nodes are displayed in order of hazard weight. The multi-scene fusion comparison unit (300) is used to fuse multi-scene monitoring data of critical engineering projects in real time to form VR images. The VR images are divided into multiple feature images by the area where the critical node is located, and the phased visualization demonstration map of the feature image is output in sequence according to the order. The similarity between the feature images and the phased visualization demonstration map is calculated. The early warning data display unit (400) is used to preset a difference threshold, receive feature images with similarity lower than the difference threshold, issue construction early warning signals, and dynamically display the reasons for warnings of critical nodes. The early warning data display unit (400) includes a staged early warning module and a demonstration image retrieval module. The staged early warning module is used to receive construction early warning signals, identify the construction stage matched by the construction early warning signal, and perform early warning operations according to the construction stage. The staged early warning module includes a flashlight and a voice broadcaster. The number of flashlights is consistent with the number of multiple construction stages divided by the stage division module, and each construction stage is matched with one flashlight. The voice broadcaster is used to preset reminder voices for each stage, identify the reminder voices matched by the stage that issues the construction early warning signal, and emit sound through a speaker. The demonstration image display module is used to display two warning reason images in a split screen. The two warning reason images include a VR image of the actual risk factor highlighted in the current stage, and a dynamic process demonstration image before and after the current stage time node.

2. The supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects according to claim 1, characterized in that: The critical node marking unit (200) includes a node marking module and a hazard quantification and sorting module; The node marking module is used to establish a risk factor database, extract the construction content of the 3D model of the current construction stage from the model building module, input the construction content into the risk factor database, and output the risk factors of all critical nodes in the 3D model of the current stage. The risk quantification and ranking module is used to preset the quantification dimension and scoring rules of each risk factor to establish a risk factor scoring rule table. It sequentially receives multiple risk factors output by the node marking module into the risk factor scoring rule table, outputs weighted scoring values ​​to define risk values, and uses bubble sort to sort multiple risk factors.

3. The supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects according to claim 2, characterized in that: The critical node marking unit (200) also includes a model highlighting module. The model highlighting module is used to receive the risk factors of all critical nodes in the current stage of the 3D model from the node marking module, and to receive the sorting of multiple risk factors output by the hazard quantification and sorting module. It sets a color mapping relationship where the higher the hazard value, the darker the color, and highlights multiple risk factors in the 3D model according to the color mapping relationship.

4. The supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects according to claim 3, characterized in that: The multi-scene fusion comparison unit (300) includes a multi-source data acquisition module, an actual highlight display module, and a feature matching module; The multi-source data acquisition module is used to divide the 3D model of the final stage into multiple monitoring areas, and allocate several cameras to each area so that the monitoring range of multiple cameras covers the entire monitoring area. The module also matches the camera installation positions in the 3D model with the actual locations of the critical engineering projects to obtain actual VR images. The actual highlighting display module is used to calculate the actual risk factors of the 3D model risk factors in the VR image through the node marking module, and to highlight the actual risk factors using the model highlighting display module. The feature matching module is used to sort the actual risk factors based on the hazard quantification and sorting module, and then use an image similarity algorithm to compare the features of the actual risk factors with the features of the risk factors in the 3D model, and output a similarity value. If the similarity value is greater than the matching threshold, a normal signal is output; if the similarity value is not greater than the matching threshold, a construction warning signal is output.

5. The supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects according to claim 4, characterized in that: The multi-scene fusion comparison unit (300) further includes a personnel movement matching module. The personnel movement matching module is used to pre-collect standard personnel features of staff, extract current personnel features from the multi-source data acquisition module, and compare the matching degree between the current personnel features and the standard personnel features, including the following postures: If the output does not match, the warning data display unit (400) will send a warning signal to irrelevant personnel. Second, if the output matches, the current personnel features are extracted from the dynamic flow diagram on the timeline based on the dynamic demonstration module, and defined as standard personnel movement features. At the same time, the current personnel movement features output by the multi-source data acquisition module on the timeline are extracted. The personnel movement features include the path and work area features of each personnel feature. The matching degree between the standard personnel movement features and the current personnel movement features is compared. If the output matches, a normal signal is issued. If the output does not match, a work warning signal is issued to the warning data display unit (400).

6. The supervision information platform for intelligent identification and process-oriented management of critical and major engineering projects according to claim 1, characterized in that: The early warning data display unit (400) also includes an early warning feedback module. The early warning feedback module is used to establish a mapping relationship between early warning signal types and feedback terminals. The early warning signal types include construction early warning signals, unrelated personnel early warning signals and work early warning signals. The module identifies the early warning signal type and feeds back the cause of the early warning to the matching terminal device via wireless communication.

Images