A full-process intelligent supervision system for power engineering
By deploying various front-end sensing devices and deep learning technology in the power engineering supervision system, real-time data collection and intelligent analysis of construction site data have been achieved, solving the data silo problem, improving the efficiency and quality of supervision work, and ensuring timely response to construction safety and quality issues.
Patent Information
- Application Number
- CN202611246605.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-17
- Publication Date
- 2026-11-10
AI Technical Summary
The existing power engineering supervision system suffers from data silos, making it impossible to achieve cross-system and cross-stage data correlation analysis. This results in information delays during the supervision process, making it difficult to achieve intelligent supervision throughout the entire process and to respond promptly to quality hazards and safety risks at the construction site.
The system adopts a full-process intelligent supervision system for power engineering. It collects data from the construction site in real time by deploying various front-end sensing devices, uses deep learning technology for intelligent analysis, generates early warning information, and realizes full-process management of supervision tasks and multi-party collaborative work through a collaborative management module. It supports multi-dimensional data display and real-time early warning.
It enables intelligent analysis of construction safety, quality, and schedule, improving the efficiency and quality of supervision work, ensuring that relevant personnel receive early warning information in a timely manner, reducing supervision costs, and enhancing the visibility and flexibility of supervision work.
Smart Images

Figure CN123070474A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer technology, and in particular to an intelligent supervision system for the entire process of power engineering. Background Technology
[0002] As a core component of national infrastructure construction, power engineering construction directly impacts socio-economic development and people's well-being through its quality and safety. Power engineering supervision, a crucial system for ensuring construction quality, is responsible for supervising and managing the entire lifecycle of a project, from design review and equipment procurement to construction and final acceptance. Traditional power engineering supervision relies primarily on manual on-site inspections, paper-based documentation, and regular meetings for information collection and transmission. This manual-driven approach has gradually revealed numerous problems when facing large-scale, complex modern power engineering projects, including inefficiency, delayed information transmission, and incomplete regulatory coverage.
[0003] With the rapid development of information technology, information technology has been gradually applied to the field of power engineering supervision. Various power supervision information systems have emerged in the market. These systems typically possess basic functional modules such as contract management, progress tracking, and quality inspection records, achieving a certain degree of digitalization and process-orientation of supervision operations. However, most existing power supervision information systems adopt a monolithic architecture design, with each functional module operating independently, data storage scattered, and interface standards inconsistent between systems. This results in various types of data generated during the supervision process, such as progress data, quality inspection data, safety inspection records, and material acceptance information, being scattered across different modules or even different systems, creating a serious data silo phenomenon.
[0004] Due to data silos, supervisors are unable to comprehensively analyze and assess the overall status of engineering projects within a unified data view. When quality hazards or safety risks arise at the construction site, relevant information often needs to be manually compiled before being reported to management, leading to delayed responses and missed opportunities for optimal intervention. Furthermore, the lack of cross-system and cross-stage data correlation analysis capabilities makes it difficult for existing systems to deeply mine and intelligently warn of the massive amounts of data generated during power engineering supervision. This hinders effective data support for supervisory decisions and severely restricts the intelligent development of the power engineering supervision industry. A comprehensive, integrated, and intelligent power engineering supervision system has become a pressing technical need for the industry. Summary of the Invention
[0005] The purpose of this invention is to provide an intelligent supervision system for the entire process of power engineering, which solves the problems mentioned in the background art.
[0006] This invention is implemented as follows: a full-process intelligent supervision system for power engineering, comprising: The data acquisition module collects visual data, audio data, environmental data, and equipment operation data in real time from various front-end sensing devices deployed at the power engineering construction site, and transmits the collected raw data to the data processing module. The data processing module preprocesses the received raw data, including data cleaning, data format unification, and feature extraction. It stores the preprocessed structured data in the data storage unit and sends the corresponding structured data to the intelligent analysis module according to the preset data analysis task. The intelligent analysis module is based on a pre-trained deep learning model. It analyzes and processes the received structured data, identifies safety hazards, non-compliant operations, and construction quality defects at the construction site, generates corresponding analysis results, and sends the analysis results to the early warning management module. The early warning management module judges the analysis results output by the intelligent analysis module according to the preset early warning rules. When the analysis results meet the early warning conditions, it generates early warning information and determines the recipient and channel of the early warning information according to the early warning level, and sends the early warning information to the corresponding terminal device. The collaborative management module is used to manage the entire process of creating, assigning, executing, and accepting supervision tasks. It supports multi-party collaborative work, generates supervision reports, performs statistical analysis on supervision data, and outputs visualized results.
[0007] In one optional implementation, the data acquisition module includes a video acquisition unit, an audio acquisition unit, an environmental sensing unit, and a device data acquisition unit. The video acquisition unit includes high-definition cameras and infrared cameras deployed in key areas of the construction site, used to acquire real-time video images and nighttime infrared images of the construction site; The audio acquisition unit includes a directional microphone array for acquiring audio signals from the construction site and using beamforming technology to extract sound source information from a specific direction. The environmental sensing unit includes a temperature and humidity sensor, a wind speed sensor, a noise sensor, and an air quality sensor, used to collect environmental parameters at the construction site. The equipment data acquisition unit interfaces with the control system of the construction equipment to obtain the operating status, working parameters and fault codes of the construction equipment in real time.
[0008] In one optional implementation, the data processing module includes a data cleaning unit, a feature extraction unit, and a data indexing unit; The data cleaning unit is used to filter invalid data in the original data, supplement missing data, and correct abnormal data. The feature extraction unit is used to extract feature vectors for intelligent analysis from the cleaned data. For video data, it extracts frame-level features and video-level features; for audio data, it extracts spectral features and semantic features; and for environmental data, it extracts temporal features. The data indexing unit is used to create an index for structured data so that it can be quickly retrieved and queried, and the data is stored in a distributed storage system.
[0009] In one optional implementation, the intelligent analysis module includes a safety behavior analysis unit, a quality defect detection unit, a progress analysis unit, and an environmental risk assessment unit. The safety behavior analysis unit analyzes the behavior of construction workers based on a human posture estimation model and a target detection model, identifies unsafe behaviors such as not wearing a safety helmet, not wearing a safety belt, and violating operating procedures, and generates a safety warning signal when an unsafe behavior is identified. The quality defect detection unit analyzes images and videos of construction quality based on image segmentation and defect classification models, identifies quality problems such as concrete pouring defects, rebar binding defects, and welding quality defects, and generates a quality warning signal when a quality defect is identified. The progress analysis unit calculates the deviation between the actual construction progress and the planned construction progress based on construction drawing data and actual construction images, using a comparative analysis algorithm, and generates a progress deviation report. The environmental risk assessment unit, based on data collected by the environmental sensing unit and combined with meteorological forecast data, assesses the weather risks, geological risks, and environmental safety risks at the construction site, and generates an environmental risk assessment report.
[0010] In one optional implementation, the early warning management module includes an early warning rule configuration unit, an early warning level determination unit, and an early warning distribution unit; The warning rule configuration unit supports user-defined warning rules, including warning conditions, warning thresholds, warning levels, and warning durations; The warning level determination unit divides the warning information into four levels: red warning, orange warning, yellow warning, and blue warning, according to the severity of the analysis results and the preset warning level classification standard. The warning distribution unit sends warning information to the corresponding responsible personnel through channels such as instant messaging software, SMS platform, email system and mobile application, according to the level and type of the warning information, and tracks the reception and processing status of the warning information.
[0011] In one optional implementation, the collaborative management module includes a task management unit, a document management unit, a communication and coordination unit, and a report generation unit; The task management unit supports the entire process management of the creation, allocation, execution, and acceptance of supervision tasks, and provides task reminders, task reminders, and task timeout alarms. The document management unit supports uploading, storage, version management, and retrieval of documents such as construction drawings, construction specifications, supervision logs, and acceptance reports; The communication and coordination unit provides online instant messaging functionality, supports various forms of communication such as text, voice, images, and video, and automatically records communication content to form a communication log. The report generation unit automatically generates daily, weekly, monthly, and special reports based on preset report templates, and supports the export and printing of reports.
[0012] In one optional implementation, the intelligent supervision system for the entire power engineering process also includes a three-dimensional visualization module, which is used to construct a three-dimensional digital model of the construction site, overlay the collected real-time data and analysis results onto the three-dimensional model for visualization, and supports multi-angle browsing and hotspot annotation.
[0013] In one optional implementation, the full-process intelligent supervision system for power engineering further includes a mobile inspection module. The mobile inspection module runs on a mobile terminal device, enabling supervisors to conduct on-site inspections through the mobile terminal device, record problems found during inspections in real time, upload photos and videos of the inspection site, and receive early warning information and task notifications.
[0014] In one optional implementation, the intelligent supervision system for the entire power engineering process adopts a cloud computing architecture, including an infrastructure layer, a platform service layer, and an application service layer. The infrastructure layer provides unified management and scheduling of computing resources, storage resources, and network resources; The platform service layer provides data processing services, model training services, and model inference services; The application service layer provides various supervision application functions for users and provides data sharing and service call capabilities to third-party systems through application programming interfaces.
[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: The intelligent supervision system for the entire power engineering process provided in this application achieves comprehensive data collection from the construction site by deploying various front-end sensing devices. It utilizes deep learning technology to achieve intelligent analysis of construction safety, quality, and progress, solving the technical problems of existing technologies that rely on manual inspections, are inefficient, and cannot achieve 24 / 7 real-time monitoring. The system uses an early warning management module to provide real-time warnings of safety hazards and quality defects, automatically selecting the sending channel based on the warning level to ensure relevant personnel receive warning information and take appropriate measures immediately. The collaborative management module enables full-process management of supervision tasks and multi-party collaboration, improving the efficiency and quality of supervision work. The 3D visualization module visually displays construction site data and intelligent analysis results in the form of 3D models, enhancing the visualization of supervision work. The mobile inspection module supports mobile office work for supervisors, improving the flexibility and convenience of supervision work. This invention realizes the intelligence, automation, and visualization of power engineering supervision, improving supervision efficiency and quality, reducing supervision costs, and providing technical support for the safety management of power engineering construction. Attached Figure Description
[0016] Figure 1 This is a schematic diagram of the overall technical architecture of the intelligent supervision system for the entire process of power engineering proposed in this invention; Figure 2 This is a schematic diagram of the core principle framework of the intelligent analysis module based on a deep learning model in this invention. Detailed Implementation
[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0018] In the accompanying drawings, the size of constituent elements, the thickness of layers, or areas may sometimes be exaggerated for clarity. Therefore, any implementation of this disclosure is not necessarily limited to the dimensions shown in the drawings, and the shapes and sizes of the components in the drawings do not reflect true proportions. Furthermore, the drawings schematically illustrate ideal examples, and any implementation of this disclosure is not limited to the shapes or values shown in the drawings.
[0019] Reference Figure 1 , Figure 1 This is a schematic diagram of the overall technical architecture of a full-process intelligent supervision system for power engineering proposed in one embodiment of this application. As shown in the figure, the system includes: a data acquisition module, a data processing module, an intelligent analysis module, an early warning management module, and a collaborative management module.
[0020] The data acquisition module in the full-process intelligent supervision system for power engineering collects visual, audio, environmental, and equipment operation data in real time from various front-end sensing devices deployed at the construction site, and transmits the raw data to the data processing module. In this embodiment, the data acquisition module, as the data source of the entire intelligent supervision system, bears the crucial responsibility of real-time sensing and collection of multi-dimensional information from the construction site. By rationally deploying various front-end sensing devices in key areas of the construction site, it achieves comprehensive and all-round status perception of the entire construction process, providing a data foundation for subsequent intelligent analysis and early warning decisions. The raw data collected by the data acquisition module includes, but is not limited to, video stream data, audio stream data, environmental parameter data, and equipment operation status data. This data is sent to the data processing module in real time through a dedicated data transmission channel for further processing.
[0021] In this embodiment, the data processing module preprocesses the received raw data, including data cleaning, data format standardization, and feature extraction. The preprocessed structured data is then stored in the data storage unit, and the corresponding structured data is sent to the intelligent analysis module according to a preset data analysis task. Upon receiving the raw data, the data processing module first performs quality screening and data cleaning to remove obvious errors, duplicates, and invalid data. Subsequently, it standardizes the format of raw data from different sources, converting video data, audio data, environmental data, and device data into a standardized data format for internal system use, facilitating subsequent storage and analysis. After data cleaning and format standardization, the feature extraction unit in the data processing module analyzes and processes various types of data, extracting feature vectors for intelligent analysis. For video data, frame-level and video-level features are extracted. Frame-level features include texture, color, and edge features of a single frame image, while video-level features include motion, temporal, and scene features. For audio data, spectral and semantic features are extracted. Spectral features include Mel-frequency characteristics and MFCC features, while semantic features include speech recognition results and abnormal sound recognition results. For environmental data, temporal and statistical features are extracted, including temperature change trends, humidity change curves, and wind speed time-series data. After feature extraction, the structured data is stored in a distributed storage system, and according to the preset data analysis task configuration, the corresponding structured data is sent to the intelligent analysis module for in-depth analysis and processing.
[0022] In this embodiment, the intelligent analysis module, based on a pre-trained deep learning model, analyzes and processes the received structured data, identifies safety hazards, non-compliant operations, and construction quality defects at the construction site, generates corresponding analysis results, and sends these results to the early warning management module. The intelligent analysis module is the core intelligent processing unit of the entire end-to-end intelligent supervision system for power engineering, integrating multiple deep learning models optimized for power engineering construction scenarios. These deep learning models are trained and iterated on massive amounts of labeled data, possessing the ability to perceive and understand complex construction site scenarios. Upon receiving structured data from the data processing module, the intelligent analysis module, according to the data type and analysis task requirements, calls the corresponding deep learning model for analysis and processing, outputting structured analysis results containing information such as hazard type, location coordinates, and severity. The analysis results are sent to the early warning management module in real time via the system's internal message queue, triggering subsequent early warning judgment and distribution processes.
[0023] In this embodiment, the early warning management module judges the analysis results output by the intelligent analysis module according to preset early warning rules. When the analysis results meet the early warning conditions, an early warning message is generated, and the recipient and sending channel of the early warning message are determined according to the early warning level. The early warning management module has a built-in configurable early warning rule engine, which allows supervisors to customize early warning rules according to actual management needs, including parameters such as early warning conditions, early warning thresholds, early warning levels, and early warning duration. When the analysis results output by the intelligent analysis module enter the early warning management module, the early warning rule engine judges the analysis results item by item according to the preset early warning rules. For analysis results that meet the early warning conditions, the early warning generation process is automatically triggered. When an early warning is generated, the system determines the early warning level according to the severity and type of the analysis results, according to the preset early warning level classification standard, and automatically determines the recipient and sending channel based on the early warning level and type, ensuring that the early warning information can be delivered to the relevant responsible personnel in a timely and accurate manner.
[0024] In this embodiment, the collaborative management module is used to manage the entire process of creating, allocating, executing, and accepting supervision tasks. It supports multi-party collaborative work, generates supervision reports, performs statistical analysis on supervision data, and outputs visualized results. The collaborative management module serves as the comprehensive management portal of the full-process intelligent power engineering supervision system, providing supervisors with one-stop functional support for task management, document management, communication and coordination, and report generation. Regarding task management, the collaborative management module supports supervisors in creating various supervision tasks, assigning tasks to specific executors, tracking task execution progress, and conducting acceptance confirmation upon task completion. In terms of document management, the collaborative management module provides electronic management functions for documents such as construction drawings, construction specifications, supervision logs, and acceptance reports, supporting document uploading, storage, version management, and retrieval. For communication and coordination, the collaborative management module provides online instant messaging functionality, supporting various forms of communication such as text, voice, images, and video, and automatically records communication content to form a communication log. Regarding report generation, the collaborative management module automatically generates daily, weekly, monthly, and special reports based on preset report templates, and supports report export and printing.
[0025] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering, in which the data acquisition module includes a video acquisition unit, an audio acquisition unit, an environmental sensing unit, and an equipment data acquisition unit.
[0026] The video acquisition unit includes high-definition cameras and infrared cameras deployed in key areas of the construction site to collect real-time video images and nighttime infrared images. In this embodiment, the video acquisition unit is the core component of the data acquisition module, responsible for 24 / 7 visual monitoring of the construction site. The high-definition cameras use high-resolution image sensors, supporting video acquisition of at least 2 megapixels, and can clearly capture status information such as personnel activities, equipment operation, and material stacking at the construction site. The high-definition cameras have functions such as automatic exposure, automatic white balance, and wide dynamic range, which can adapt to complex lighting conditions at the construction site. The infrared cameras use infrared supplementary lighting technology, automatically switching to infrared imaging mode at night or in low-light environments to ensure that the system can achieve uninterrupted 24-hour visual monitoring. Both the high-definition cameras and infrared cameras in the video acquisition unit support pan-tilt control, allowing monitoring personnel to remotely adjust the camera's shooting angle and zoom level for refined monitoring of key areas. The video stream data acquired by the video acquisition unit is transmitted to the data processing module in real time via the RTSP protocol. During the transmission process, the H.264 or H.265 video encoding standard is used for compression to ensure a balance between network transmission efficiency and image quality.
[0027] The audio acquisition unit includes a directional microphone array for acquiring audio signals from the construction site and using beamforming technology to extract sound source information from specific directions. In this embodiment, the audio acquisition unit uses a multi-microphone array. By leveraging the spatial sampling characteristics of the microphone array and combining it with a beamforming algorithm, it enhances the extraction of sound source information from specific directions. Beamforming technology utilizes the phase difference of the signals received by each element of the microphone array, increasing the sound pressure from the target direction through delay accumulation, while suppressing interference noise from other directions. In the power engineering construction site environment, the audio acquisition unit primarily targets and identifies specific sound sources such as construction machinery operation sounds, safety helmet impact sounds, cries for help, and abnormal noises. The microphone array of the audio acquisition unit uses an omnidirectional pickup mode to cover a 360-degree range, and uses digital signal processing technology to achieve parallel output of multiple beams, each beam corresponding to a specific acquisition direction. The acquired audio data is compressed using AAC or OPUS audio coding standards and then transmitted to the data processing module for subsequent spectral feature extraction and semantic analysis.
[0028] The environmental sensing unit includes temperature and humidity sensors, wind speed sensors, noise sensors, and air quality sensors, used to collect environmental parameters at the construction site. In this embodiment, the environmental sensing unit is deployed at multiple key locations on the construction site to monitor the environmental status in real time, providing data support for environmental risk assessment. The temperature and humidity sensors use digital sensor chips, supporting a temperature measurement range of -40 degrees Celsius to 80 degrees Celsius and a humidity measurement range of 0% to 100% relative humidity, with measurement accuracies of ±0.3 degrees Celsius and ±2% relative humidity, respectively. The wind speed sensor uses an ultrasonic measurement principle, supporting a wind speed measurement range of 0 to 60 meters per second, with an initial wind speed below 0.1 meters per second, accurately sensing even slight wind changes at the construction site. The noise sensor uses a sound level meter module conforming to the IEC 61672 standard, supporting a noise measurement range of 30 dB to 130 dB, and possesses A-weighted frequency response characteristics, accurately reflecting the human ear's perception of noise. Air quality sensors can detect multiple air quality indicators such as PM2.5, PM10, CO, CO2, NO2, and SO2, providing data support for environmental monitoring at construction sites. Various sensors in the environmental sensing unit collect data via RS485 bus or LoRa wireless communication and aggregate it to the edge computing gateway of the data acquisition unit. The gateway performs preliminary verification and format conversion on the data before sending it to the data processing module.
[0029] The equipment data acquisition unit interfaces with the control system of the construction equipment to obtain the equipment's operating status, working parameters, and fault codes in real time. In this embodiment, the equipment data acquisition unit connects to the control systems of various construction equipment on the construction site through a standardized data interface, enabling automatic acquisition and real-time transmission of equipment operating data. The equipment data acquisition unit supports multiple industrial communication protocols, including mainstream industrial protocols such as Modbus RTU, Modbus TCP, and OPC UA, and can adapt to the control systems of different types of construction equipment. The acquired equipment data includes, but is not limited to: equipment operating status (start, stop, standby, fault, etc.), working parameters (such as the lifting capacity, luffing angle, and slewing angle of tower cranes, the operating speed and floor position of construction elevators, and the pumping pressure and boom posture of concrete pump trucks), fault codes (fault diagnosis codes generated by the control system when equipment malfunctions), and equipment working time statistics. The equipment data acquisition unit periodically collects equipment data according to a preset sampling frequency. For abnormal states (such as equipment fault reports, equipment parameter exceeding limits, etc.), a real-time reporting mechanism is triggered to ensure that the system can detect abnormal equipment conditions as soon as possible.
[0030] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering, in which the data processing module includes a data cleaning unit, a feature extraction unit, and a data indexing unit.
[0031] The data cleaning unit filters invalid data, supplements missing data, and corrects abnormal data in the original data. In this embodiment, the data cleaning unit is a key step in ensuring data quality, and its processing flow includes three main steps: invalid data filtering, missing data supplementation, and abnormal data correction. Invalid data filtering refers to identifying and removing obviously erroneous or meaningless data from the original data, such as frames with large areas of blue or black screens in video data, long periods of silence in audio data, and values collected by environmental sensors that exceed the physically reasonable range. Missing data supplementation refers to filling in data gaps caused by equipment failure or packet loss. For continuous data sequences, interpolation algorithms are used for supplementation; for discrete data points, the mean of data from adjacent time points is used for filling or missing data is preserved. Abnormal data correction is used to identify and process abnormal measurement values caused by sensor drift, electromagnetic interference, etc. The correction algorithm comprehensively considers multi-dimensional information such as the temporal continuity of the data point with the preceding and following data, the spatial correlation with data from similar sensors, and historical statistical patterns to determine whether it is abnormal data and makes reasonable corrections.
[0032] The feature extraction unit extracts feature vectors for intelligent analysis from the cleaned data. For video data, it extracts frame-level and video-level features; for audio data, it extracts spectral and semantic features; and for environmental data, it extracts temporal features. In this embodiment, the feature extraction unit employs targeted feature extraction algorithms to process different types of data and generate feature vectors suitable for deep learning model input. For video data, the feature extraction unit first performs frame decoding and frame extraction, extracting video frames at an appropriate frame rate (typically 1 to 5 frames per second) for subsequent processing. Frame-level feature extraction uses a pre-trained convolutional neural network model (such as ResNet, VGG, etc.) to extract features from single-frame images, obtaining high-level semantic features such as texture, color, and spatial features for each frame. Video-level feature extraction uses a temporal convolutional network or recurrent neural network model to model the temporal variation of frame-level features, extracting motion features, behavioral pattern features, and scene category features within video segments. For audio data, the feature extraction unit first performs preprocessing operations such as pre-emphasis, framing, and windowing on the original audio signal. Then, it converts the time-domain signal into a frequency-domain representation using short-time Fourier transform or Mel filter bank, extracting spectral features such as Mel spectral features, MFCC features, spectral centroid, and spectral entropy. Simultaneously, for analysis tasks requiring semantic understanding, the audio signal is input into a pre-trained speech recognition model or acoustic event detection model to extract semantic features. For environmental data, the feature extraction unit extracts statistical features (mean, variance, maximum, minimum, peak, etc.) and temporal features (trend, fluctuation amplitude, periodicity, etc.) for time-series environmental data, and category and state features for discrete environmental data.
[0033] The data indexing unit is used to create indexes for structured data to enable rapid retrieval and querying, and to store the data in a distributed storage system. In this embodiment, the data indexing unit establishes a multi-dimensional indexing system for various types of structured data, supporting rapid retrieval and querying based on multiple dimensions such as time, space, event type, and data source. The data indexing unit adopts a distributed indexing architecture, storing index data in shards across multiple index nodes, and achieving efficient index querying through an index routing mechanism. For video data, index items include acquisition time, camera number, video segment start and end times, and key event tags; for audio data, index items include acquisition time, audio segment duration, acoustic event type, and acoustic feature vector; for environmental data, index items include acquisition time, sensor number, environmental parameter type, and numerical range; for device data, index items include acquisition time, device number, device type, and operating status. The data indexing unit works closely with the distributed storage system, automatically creating corresponding index items while data is being written to the storage system, ensuring the real-time performance and accuracy of retrieval and querying.
[0034] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering, in which the intelligent analysis module includes a safety behavior analysis unit, a quality defect detection unit, a progress analysis unit, and an environmental risk assessment unit.
[0035] The safety behavior analysis unit analyzes the behavior of construction workers based on a human posture estimation model and a target detection model, identifying unsafe behaviors such as not wearing a safety helmet, not wearing a safety belt, and violating operating procedures. When an unsafe behavior is detected, a safety warning signal is generated. In this embodiment, the safety behavior analysis unit adopts a multi-model collaborative analysis technical solution to achieve comprehensive intelligent monitoring of the safety behavior of personnel at the construction site. The human posture estimation model uses a real-time posture estimation algorithm based on deep learning (such as OpenPose, HRNet, etc.), which can detect key points of the human skeleton (17 key points including head, neck, shoulders, elbows, hands, hips, knees, and ankles) from video images, and infer the posture and movement of the human body based on the spatial relationship and temporal change pattern between the key points. The target detection model uses a deep learning target detection algorithm based on anchor boxes or without anchor boxes (such as YOLO, Faster R-CNN, etc.), which can detect target objects such as construction workers, safety helmets, and safety belts from video images, and output the target's location bounding box and category label. The safety behavior analysis unit fuses and analyzes the outputs of the human posture estimation model and the target detection model. First, the target detection model identifies construction workers in the video footage and whether they are wearing safety helmets and safety belts. Then, the human posture estimation model analyzes the workers' postures and movements to identify any violations (such as leaning out of guardrails while working at height, inattentiveness while using a mobile phone, or improper hand gestures). When unsafe behavior is detected, the safety behavior analysis unit generates a safety warning signal containing information such as the behavior type, time of occurrence, location, and severity, and sends it to the warning management module for further processing. (Refer to...) Figure 2 , Figure 2 This is a schematic diagram illustrating the core principle framework of the intelligent analysis module based on a deep learning model in this invention. For example... Figure 2 As shown, the deep learning model of the safety behavior analysis unit adopts a dual-branch architecture design. One branch is responsible for the detection and tracking of personnel targets, and the other branch is responsible for the estimation of human posture and behavior analysis. The feature maps of the two branches are fused in the feature fusion layer to output the final analysis result.
[0036] The quality defect detection unit analyzes images and videos related to construction quality based on image segmentation and defect classification models. It identifies quality issues such as concrete pouring defects, rebar tying defects, and welding defects. When a quality defect is detected, a quality warning signal is generated. In this embodiment, the quality defect detection unit has constructed a specialized quality defect detection model library for common quality defect types in power engineering construction. The image segmentation model uses a fully convolutional network architecture (such as U-Net, DeepLabV3, etc.) to perform pixel-level segmentation of construction quality images, accurately identifying the location and boundary range of defect areas. The defect classification model uses a multi-class convolutional neural network architecture to determine the defect type of the segmented defect areas (such as cracks, honeycombing, pitting, exposed rebar, corrosion, welding slag inclusions, incomplete welds, etc.). The workflow of the quality defect detection unit is as follows: First, the input construction quality image is segmented into defect regions using an image segmentation model, and a segmentation mask for the defect regions is output. Then, image blocks of the defect regions are cropped based on the segmentation mask, and these image blocks are input into a defect classification model for type determination. Simultaneously, considering information such as the area ratio of the defect region, defect type, and defect severity, a quality warning signal containing information such as defect location, defect type, defect area, and severity level is generated. For video data, the quality defect detection unit extracts video frames for detection according to a set sampling interval, and merges and confirms defects detected in multiple consecutive frames to reduce false detections and missed detections.
[0037] The progress analysis unit, based on construction drawings and actual construction images, uses comparative analysis algorithms to calculate the deviation between the actual and planned construction progress, generating a progress deviation report. In this embodiment, the progress analysis unit employs an image-based construction progress analysis method. By comparing images of the actual construction scene with construction drawings or BIM models, it quantitatively assesses the deviation between the actual and planned construction progress. The progress analysis unit first uses construction drawings or BIM models as a benchmark reference, generating expected scene images for different construction stages using image rendering technology. Then, it performs feature-level comparative analysis between the actual construction scene images captured by cameras and the expected scene images, using algorithms such as structural similarity measurement, feature matching, and semantic segmentation alignment to calculate the similarity score. Finally, it maps the similarity score to quantitative indicators of construction progress, generating a progress deviation report containing progress deviation values, deviation cause analysis, and suggestions for expediting. The progress analysis unit also supports time-series-based progress trend analysis, predicting subsequent construction progress trends through statistical analysis of historical progress data, providing decision support for project managers.
[0038] The environmental risk assessment unit, based on data collected by the environmental sensing unit and combined with meteorological forecast data, assesses the weather risks, geological risks, and environmental safety risks at the construction site, generating an environmental risk assessment report. In this embodiment, the environmental risk assessment unit comprehensively analyzes multi-source environmental data to achieve a comprehensive assessment of the environmental risks at the construction site. The environmental risk assessment unit receives real-time environmental parameter data such as temperature, humidity, wind speed, noise, and air quality collected by the environmental sensing unit, and simultaneously obtains weather forecast data (including temperature, precipitation, wind force, lightning, etc.) for a future period provided by meteorological forecast services. The environmental risk assessment unit has a built-in environmental risk assessment model, which is built based on expert experience and historical accident data, and can calculate the environmental risk index based on real-time monitoring and forecast values of environmental parameters. When environmental parameters are detected to exceed safety thresholds or severe weather conditions are predicted, the environmental risk assessment unit automatically triggers an early warning mechanism and generates a risk assessment report containing information such as risk type, risk level, impact range, and response recommendations. The types of risks assessed by the environmental risk assessment unit include, but are not limited to: heatstroke risk (calculated based on temperature, humidity, and wind speed), falling object risk (calculated based on wind speed and gust forecasts), electric shock risk (calculated based on lightning activity and air humidity), collapse risk (calculated based on rainfall and soil moisture content), and poisoning and asphyxiation risk (calculated based on the concentration of harmful gases).
[0039] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering, in which the early warning management module includes an early warning rule configuration unit, an early warning level determination unit, and an early warning distribution unit.
[0040] The early warning rule configuration unit supports user-defined early warning rules, including early warning conditions, early warning thresholds, early warning levels, and early warning durations. In this embodiment, the early warning rule configuration unit provides a visual rule configuration interface, allowing supervisors to flexibly define various early warning rules according to project management needs. The basic components of an early warning rule include: triggering conditions (the type and numerical range of analysis results output by the intelligent analysis module), early warning thresholds (the numerical boundaries that trigger an early warning), early warning levels (the severity level of the early warning information), early warning durations (the effective display period of the early warning information in the system), and the associated intelligent analysis unit (the data source that triggers the rule). The early warning rule configuration unit supports the configuration of complex rule logic, allowing users to combine multiple triggering conditions using logical operators such as AND, OR, and NOT to achieve accurate early warnings for complex scenarios. For example, the rule "When the safety behavior analysis unit detects that a safety helmet is not being worn and the temperature exceeds 35 degrees Celsius, generate an orange early warning" can be configured, which includes an AND combination of two triggering conditions. The early warning rule configuration unit also supports version management and historical change records for early warning rules, ensuring the traceability of rule adjustments.
[0041] The warning level determination unit classifies warning information into four levels—red, orange, yellow, and blue—based on the severity of the analysis results and according to preset warning level classification standards. In this embodiment, the warning level determination unit adopts rule-based judgment logic to determine the warning level based on attributes such as the type, severity, and scope of impact of the analysis results. A red warning corresponds to particularly serious safety hazards or quality defects, requiring immediate cessation of construction and rectification. Construction can only resume after rectification is completed and accepted. An orange warning corresponds to serious safety hazards or quality defects, requiring rectification within a specified time. A yellow warning corresponds to general safety hazards or quality defects, requiring attention and prompt rectification. A blue warning corresponds to minor violations or potential risks, requiring reminders and attention. While determining the warning level, the warning level determination unit also generates a warning message template containing warning details, suggested handling measures, and guidance for responsible personnel, providing standardized support for subsequent warning distribution and handling.
[0042] The early warning distribution unit sends early warning information to relevant responsible personnel through channels such as instant messaging software, SMS platforms, email systems, and mobile applications, based on the level and type of the early warning information, and tracks the reception and processing status of the early warning information. In this embodiment, the early warning distribution unit implements a multi-channel, multi-level early warning information distribution mechanism. Different distribution strategies are adopted for different levels of early warning information: red and orange warnings are sent to all relevant responsible personnel simultaneously through all available channels (instant messaging software, SMS, telephone, email, and mobile application push notifications) to ensure that the early warning information is delivered as soon as possible; yellow warnings are sent to relevant responsible personnel through instant messaging software and mobile application push notifications; blue warnings are only sent through mobile application push notifications. The early warning distribution unit interfaces with external systems such as instant messaging software (such as DingTalk and WeChat Work), SMS platforms, and email systems through API interfaces to achieve automated push notifications of early warning information. The early warning distribution unit also tracks the sending status (including sent, delivered, read, unread, etc.) and processing status (including pending, processing, resolved, closed, etc.) of each early warning information, and displays the processing progress of early warning information in real time on the management interface, providing managers with a basis for supervision and assessment.
[0043] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering, in which the collaborative management module includes a task management unit, a document management unit, a communication and coordination unit, and a report generation unit.
[0044] The task management unit supports the entire process management of supervision tasks, including creation, allocation, execution, and acceptance, and provides task reminders, task expediting, and task timeout alarms. In this embodiment, the task management unit realizes closed-loop management of supervision tasks, covering the entire lifecycle from task creation to acceptance. The task creation function allows supervisors to create supervision tasks according to preset task templates or custom methods. Task attributes include task name, task description, task type, priority, planned start time, planned completion time, executors, associated construction area or construction process, etc. The task allocation function allows the creation of tasks to be assigned to specific executors, and the assigned tasks are automatically pushed to the executor's task to-do list. The task execution function allows executors to update task progress, upload task execution records (such as inspection photos, test reports, etc.), and submit task completion applications. The task acceptance function allows acceptance personnel to review and accept the completed tasks. Once the acceptance is passed, the task is closed; if the acceptance fails, it is returned to the executor for reprocessing. The task reminder function automatically sends reminder messages to the personnel when the task deadline is approaching; the task expediting function allows managers to manually send expediting messages to the personnel; the task overtime alarm function automatically triggers an alarm when the task has not been accepted after the planned completion time, notifying relevant managers to intervene.
[0045] The document management unit supports the uploading, storage, version management, and retrieval of documents such as construction drawings, construction specifications, supervision logs, and acceptance reports. In this embodiment, the document management unit provides electronic document management capabilities for the full-process intelligent supervision system for power engineering. The document management unit supports uploading documents in various formats, including PDF, Word, Excel, images, and CAD drawings, meeting the diverse storage needs of engineering data. Document storage is organized using a distributed file system, supporting categorized storage according to project, section, unit project, and sub-project dimensions. The version management function automatically records the document's modification history, saves all versions of the document, and supports version comparison and historical version backtracking. The retrieval function supports combined searches based on document name, document type, upload time, uploader, and associated construction location, helping supervisors quickly locate the required documents. The document management unit also integrates with the task management module, supporting the associated management of documents as attachments or deliverables for tasks.
[0046] The communication and coordination unit provides online instant messaging functionality, supporting various communication formats such as text, voice, images, and video, and automatically records communication content to form a communication log. In this embodiment, the communication and coordination unit provides a platform for real-time communication and collaboration among all project participants. The unit establishes communication groups according to the project's organizational structure, including project-level groups, section-level groups, professional-level groups, and topical discussion groups for specific matters. Group members can send text messages, voice messages, images, files, videos, etc., for communication and discussion. Messages are automatically synchronized to the clients of all group members after being sent. The unit automatically records all communication content in the groups, forming a searchable communication log. The log includes information such as the sender, sending time, message content, and recipient, and supports searching by time, personnel, keywords, and other criteria. The unit also supports the @mention function, allowing users to specifically remind specific individuals to pay attention to messages; the message pinning function, allowing important notifications to be displayed at the top; and the read receipt function, allowing the sender to view the read status of messages.
[0047] The report generation unit automatically generates daily, weekly, monthly, and special reports based on preset report templates, and supports report export and printing. In this embodiment, the report generation unit realizes the automated generation and distribution of supervision reports. The report generation unit has built-in multiple report templates covering various business areas of supervision work, including daily safety supervision reports, weekly quality supervision reports, monthly progress supervision reports, and special environmental monitoring reports. The report templates define the data sources, statistical scope, display format, and distribution rules of the reports. The report generation unit automatically extracts relevant data from the system according to preset generation cycles (daily, weekly, monthly) or manual triggering, performs statistical analysis according to the report templates, and generates standardized supervision reports. The generated reports support exporting to common formats such as Word, PDF, and Excel for easy archiving and secondary editing. The report generation unit also supports automatic report distribution, automatically sending reports to relevant personnel according to a preset distribution list.
[0048] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering. In this system, the system further includes a three-dimensional visualization module for constructing a three-dimensional digital model of the construction site, overlaying the collected real-time data and analysis results onto the three-dimensional model for visualization, and supporting multi-angle browsing and hotspot annotation.
[0049] In this embodiment, the 3D visualization module uses BIM technology or 3D reconstruction technology to construct a 3D digital model of the construction site. The geometric data of the 3D digital model comes from the BIM model of the construction drawings or 3D point cloud data reconstructed through methods such as UAV oblique photogrammetry and LiDAR scanning. The 3D visualization module overlays and displays real-time data (such as video streams and sensor data) collected by the data acquisition module and analysis results (such as hazard markings, quality defect markings, and progress status) output by the intelligent analysis module on the corresponding locations of the 3D model in the form of labels, markers, and heat maps. Supervisors can browse the overall layout and local details of the construction site from any angle through the 3D visualization module, and annotate and mark specific locations using the hotspot marking function. The 3D visualization module also supports advanced functions such as sectional views, measurement tools, and collision detection, providing visual analysis tools for construction organization design review and construction scheme demonstration.
[0050] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering. In this system, the system further includes a mobile inspection module, which runs on a mobile terminal device and supports supervisors to conduct on-site inspections through the mobile terminal device, record problems found during inspections in real time, upload photos and videos of the inspection site, and receive early warning information and task notifications.
[0051] In this embodiment, the mobile inspection module is a functional extension of the full-process intelligent supervision system for power engineering on mobile terminals, allowing supervisors to continue their supervision work even after leaving fixed monitoring points. The mobile inspection module runs on smartphones or tablets and provides functions such as viewing inspection tasks, filling out inspection records, collecting on-site photos and videos, receiving early warning information, and receiving task notifications. After arriving at the inspection site, supervisors use the mobile inspection module to scan a QR code or locate their current position, and the system automatically retrieves the inspection task list for that location. Supervisors then inspect each item on the inspection checklist, recording any discovered hazards or problems by taking photos or videos with the mobile terminal's camera and filling in descriptions, severity, and other attributes. Once the inspection record is edited on the mobile terminal, it is automatically synchronized to the system backend and enters the subsequent intelligent analysis and early warning process along with real-time data collected by the data acquisition module. The mobile inspection module also integrates instant messaging functionality, allowing supervisors to communicate and coordinate with other project members in real time during the inspection process.
[0052] In one embodiment, this application also provides a full-process intelligent supervision system for power engineering, in which the system adopts a cloud computing architecture, including an infrastructure layer, a platform service layer, and an application service layer.
[0053] The infrastructure layer provides unified management and scheduling of computing, storage, and network resources. In this embodiment, the infrastructure layer employs cloud computing virtualization technology to virtualize the underlying physical computing devices (such as servers, storage devices, and network devices) into logical resource pools, enabling on-demand resource allocation and elastic scaling. The infrastructure layer includes a computing resource pool (providing CPU and GPU computing resources), a storage resource pool (providing block storage, object storage, and file storage), and a network resource pool (providing virtual private networks, load balancers, and public network bandwidth). The infrastructure layer achieves unified scheduling and management of resources through a cloud platform management system (such as OpenStack or Kubernetes), supporting automatic scaling up and down based on business load to ensure sufficient processing capacity during peak business periods and save resource costs during off-peak periods.
[0054] The platform service layer provides data processing services, model training services, and model inference services. In this embodiment, the platform service layer is built on top of the infrastructure layer, providing general technical platform capabilities for upper-layer applications. The data processing service provides ETL (Extract, Transform, Load) processing capabilities for large-scale data, supporting batch and real-time processing of both structured and unstructured data. The model training service provides a training environment for deep learning models, supporting distributed training, GPU scheduling, training task management, and other capabilities to help developers quickly complete model training iterations. The model inference service provides deployment and inference capabilities for trained deep learning models, supporting online and batch inference, and optimizing inference performance through techniques such as model quantization, model pruning, and model distillation to reduce inference latency and computational resource consumption.
[0055] The application service layer provides various supervision application functions for users and offers data sharing and service access capabilities to third-party systems through application programming interfaces (APIs). In this embodiment, the application service layer is the functional layer for direct user interaction, providing various business functions of the full-process intelligent power engineering supervision system. The functional modules of the application service layer include the aforementioned data acquisition, data processing, intelligent analysis, early warning management, collaborative management, 3D visualization, and mobile inspection modules. These modules are built using a microservice architecture, with each module deployed and expanded independently. Communication and coordination between modules are achieved through a service registration and discovery mechanism. The application service layer also provides standardized interfaces such as RESTful API and GraphQL API, allowing third-party systems (such as owner project management platforms, construction project transaction platforms, and government regulatory platforms) to call system functions or obtain system data, achieving data interconnection and business collaboration.
[0056] The intelligent supervision system for the entire power engineering process proposed in this application has the following significant advantages: This application achieves comprehensive data collection from the construction site by deploying multiple front-end sensing devices and utilizes deep learning technology to achieve intelligent analysis of construction safety, quality, and progress. This solves the technical problems of existing technologies, such as reliance on manual inspections, low efficiency, and difficulty in achieving 24 / 7 real-time monitoring. The system uses an early warning management module to provide real-time warnings of safety hazards and quality defects, automatically selecting the sending channel based on the warning level to ensure relevant personnel receive warning information and take appropriate action immediately. The collaborative management module enables full-process management of supervision tasks and multi-party collaboration, improving the efficiency and quality of supervision work. The 3D visualization module visually displays construction site data and intelligent analysis results in the form of 3D models, enhancing the visualization of supervision work. The mobile inspection module supports mobile on-site work for supervisors, improving the flexibility and convenience of supervision. This invention realizes intelligent, automated, and visualized power engineering supervision, improving supervision efficiency and quality, reducing supervision costs, and providing technical support for the safety management of power engineering construction.
[0057] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0058] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0059] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0060] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0062] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.
[0063] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.
[0064] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A full-process intelligent supervision system for power engineering, characterized in that, include: The data acquisition module is used to collect visual data, audio data, environmental data, and equipment operation data in real time from various front-end sensing devices deployed at the power engineering construction site, and transmit the raw data to the data processing module. The data processing module is used to preprocess the received raw data, including data cleaning, data format unification and feature extraction. The preprocessed structured data is stored in the data storage unit, and the corresponding structured data is sent to the intelligent analysis module according to the preset data analysis task. The intelligent analysis module is used to analyze and process the received structured data based on a pre-trained deep learning model, identify safety hazards, non-compliant operations and construction quality defects at the construction site, generate corresponding analysis results, and send the analysis results to the early warning management module. The early warning management module is used to judge the analysis results output by the intelligent analysis module according to the preset early warning rules. When the analysis results meet the early warning conditions, an early warning message is generated. The early warning message is sent to the corresponding terminal device according to the early warning level. The collaborative management module is used to manage the entire process of creating, assigning, executing, and accepting supervision tasks. It supports multi-party collaborative work, generates supervision reports, performs statistical analysis on supervision data, and outputs visualized results.
2. The intelligent supervision system for the entire process of power engineering according to claim 1, characterized in that, The data acquisition module includes a video acquisition unit, an audio acquisition unit, an environmental sensing unit, and a device data acquisition unit. The video acquisition unit includes high-definition cameras and infrared cameras deployed in key areas of the construction site to acquire real-time video images and nighttime infrared images of the construction site. The audio acquisition unit includes a directional microphone array to acquire audio signals from the construction site and uses beamforming technology to extract sound source information from specific directions. The environmental sensing unit includes temperature and humidity sensors, wind speed sensors, noise sensors, and air quality sensors to acquire environmental parameters of the construction site. The device data acquisition unit interfaces with the control system of the construction equipment to obtain the operating status, working parameters, and fault codes of the construction equipment in real time.
3. The intelligent supervision system for the entire process of power engineering according to claim 1, characterized in that, The data processing module includes a data cleaning unit, a feature extraction unit, and a data indexing unit. The data cleaning unit filters invalid data from the original data, supplements missing data, and corrects abnormal data. The feature extraction unit extracts feature vectors for intelligent analysis from the cleaned data, extracting frame-level and video-level features for video data, spectral and semantic features for audio data, and temporal features for environmental data. The data indexing unit indexes the structured data for rapid retrieval and querying, and stores the data in a distributed storage system.
4. The intelligent supervision system for the entire process of power engineering according to claim 1, characterized in that, The intelligent analysis module includes a safety behavior analysis unit, a quality defect detection unit, a progress analysis unit, and an environmental risk assessment unit. The safety behavior analysis unit, based on a human posture estimation model and a target detection model, analyzes the behavior of construction workers, identifying unsafe behaviors such as not wearing safety helmets, not wearing safety belts, and violating operating procedures. When an unsafe behavior is detected, a safety warning signal is generated. The quality defect detection unit, based on an image segmentation model and a defect classification model, analyzes images and videos of construction quality, identifying quality problems such as concrete pouring defects, rebar tying defects, and welding quality defects. When a quality defect is detected, a quality warning signal is generated. The progress analysis unit, based on construction drawings and actual construction images, uses a comparative analysis algorithm to calculate the deviation between the actual construction progress and the planned construction progress, generating a progress deviation report. The environmental risk assessment unit, based on data collected by the environmental sensing unit and combined with meteorological forecast data, assesses the weather risks, geological risks, and environmental safety risks at the construction site, generating an environmental risk assessment report.
5. The intelligent supervision system for the entire process of power engineering according to claim 1, characterized in that, The early warning management module includes an early warning rule configuration unit, an early warning level determination unit, and an early warning distribution unit; the early warning rule configuration unit supports user-defined early warning rules, including early warning conditions, early warning thresholds, early warning levels, and early warning durations; The warning level determination unit classifies warning information into four levels—red, orange, yellow, and blue—based on the severity of the analysis results and according to a preset warning level classification standard. The warning distribution unit sends warning information to the corresponding responsible personnel through channels such as instant messaging software, SMS platforms, email systems, and mobile applications, based on the level and type of the warning information, and tracks the reception and processing status of the warning information.
6. The intelligent supervision system for the entire process of power engineering according to claim 1, characterized in that, It also includes a 3D visualization module, which is used to build a 3D digital model of the construction site, overlaying the collected real-time data and analysis results onto the 3D model for visualization, and supporting multi-angle browsing and hotspot annotation.
7. The intelligent supervision system for the entire process of power engineering according to claim 1, characterized in that, It also includes a mobile inspection module, which runs on a mobile terminal device and allows supervisors to conduct on-site inspections through the mobile terminal device, record problems found during inspections in real time, upload photos and videos of the inspection site, and receive early warning information and task notifications.
