AI based facility safety monitoring system

Abstract

The present invention relates to an AI-based facility safety monitoring and subscription-based education linkage service system, and more specifically, to an AI-based facility safety monitoring system that integrates, analyzes, corrects, and visualizes sensor data such as displacement, water level, and temperature of civil engineering and construction facilities, as well as video data, using AI, enables local governments, citizens, engineers, and educational institutions to use it as a subscription service, and allows universities and research institutions to utilize it as educational and research materials and basic design data. The AI-based facility safety monitoring system according to the present invention can solve the problem of false alarms, which is a limitation of existing systems, and improve the accuracy of analysis by integrating and analyzing sensor data as well as video and climate data using AI.

Description

Technology Field

[0001] The present invention relates to an AI-based facility safety monitoring and subscription-based education linkage service system, and more specifically, to an AI-based facility safety monitoring system that integrates, analyzes, corrects, and visualizes sensor data such as displacement, water level, and temperature of civil engineering and construction facilities, as well as video data, using AI, enables local governments, citizens, engineers, and educational institutions to use it as a subscription service, and allows universities and research institutions to utilize it as educational and research materials and basic design data. Background Technology

[0002] Existing safety monitoring systems were limited to simply providing numerical values ​​and graphs measured by sensors.

[0003] As a result, when an anomalous value caused by a sudden spike in sensor values ​​occurred, an alarm would sound immediately, leading to frequent false alarms; conversely, when actual displacement occurred in the structure, it was often mistaken for a problem with the sensor.

[0004] Furthermore, there was a lack of systems capable of directly linking climate change with structural displacement for analysis, or providing visualized information that general citizens and students could intuitively understand. There was also an insufficient system for utilizing real-time field measurement data for educational and research purposes.

[0005] In particular, local governments faced the problem of incurring outsourcing costs for analysis due to the difficulty of analyzing measurement data. In the case of slopes, although management standards should be calculated differently depending on the type of slope (such as soil, rock, or fractured zones), geological layers, and rainfall conditions, the lack of relevant data resulted in standardized management standards, making precise safety management difficult. Prior art literature

[0006] Republic of Korea Registered Patent 10-2268276 Republic of Korea Registered Patent 10-2191008 The problem to be solved

[0007] The present invention aims to solve the aforementioned conventional problems by overcoming the limitations of existing monitoring systems and providing an integrated smart safety monitoring platform that includes AI, video analysis, climate data integration, and educational platform functions.

[0008] In addition, the present invention aims to provide a system that automatically interprets sensor data in comparison with management standards, describes the status of a facility in natural language sentences, and operates adaptively by actively controlling the measurement and camera shooting cycles when anomalies occur.

[0009] Furthermore, the present invention aims to provide a system that quantitatively detects cracks or shape changes using camera images, enhances data reliability by mutually correcting them with sensor data, and intuitively visualizes causes and effects by linking structural displacement with climate data such as rainfall, snowfall, and temperature on the same timeline.

[0010] Furthermore, the present invention aims to provide an educational interface that can be easily understood by non-experts or students by providing detailed explanations, such as how to read the data, and to link actual measurement and AI analysis data so that they can be utilized as basic data for education, research, and design. means of solving the problem

[0011] The AI-based facility safety monitoring system according to the present invention for achieving the above-mentioned purpose comprises: a data collection unit that collects facility-related data including a sensor that detects physical changes in a facility, a camera that captures external changes, and a climate data API that collects external environmental factors; an analysis unit that receives facility-related data collected by the data collection unit and analyzes, corrects, and detects outliers through a machine learning model; a visualization education unit that visualizes the analysis results of the analysis unit by combining them with a 3D model, a timeline, and explanatory text, and generates educational data; and a service provision unit that provides the educational data generated by the visualization education unit in a user-specific mode.

[0012] The analysis unit of the AI-based facility safety monitoring system according to the present invention is characterized by automatically increasing the measurement cycle of the data collection unit and the camera shooting cycle when an anomaly occurs, returning to the original cycle when the situation returns to normal, and recording an event log.

[0013] The analysis unit of the AI-based facility safety monitoring system according to the present invention is characterized by analyzing image data collected from the camera of the data collection unit using a Digital Image Correlation (DIC) technique to quantify displacement, and mutually correcting the quantified displacement data with sensor data.

[0014] The visualization education unit of the AI-based facility safety monitoring system according to the present invention is characterized by visualizing the cause and effect of displacement by linking the 3D structural schematic data of the facility and the climate data collected from the data collection unit to a timeline.

[0015] The visualization education unit of the AI-based facility safety monitoring system according to the present invention is characterized by providing an educational interface that automatically displays detailed explanations of how to read educational data and measuring instruments including sensors, along with visualized educational data.

[0016] The analysis unit of the AI-based facility safety monitoring system according to the present invention is characterized by including an artificial intelligence engine that synthesizes sensor data, image data, and climate data to infer the cause of abnormal events, predicts future risks based on data trends, and performs multi-sensor cross-validation, scenario-based alarms, and automatic report generation.

[0017] The analysis unit of the AI-based facility safety monitoring system according to the present invention is characterized by including an interactive explanation function that responds to a user's natural language question based on analyzed data.

[0018] The analysis unit of the AI-based facility safety monitoring system according to the present invention is characterized by autonomously adjusting and optimizing the data measurement mode of the data collection unit by taking into account the power and network conditions of the site.

[0019] The service provider of the AI-based facility safety monitoring system according to the present invention is characterized by providing customized services to local governments, citizens, engineers, and educational institutions through a subscription service model.

[0020] The analysis unit of the AI-based facility safety monitoring system according to the present invention is characterized by predicting the risk of a slope by learning categorical data, such as the type of slope and geological characteristics, and numerical data, including rainfall intensity, slope, and measured values, through machine learning. Effects of the invention

[0021] The AI-based facility safety monitoring system according to the present invention can solve the problem of false alarms, which is a limitation of existing systems, and improve the accuracy of analysis by integrating and analyzing sensor data as well as video and climate data using AI.

[0022] In addition, the AI-based facility safety monitoring system according to the present invention can induce social participation in safety management by enabling non-expert citizens or students to easily understand the safety status of facilities through 3D visualization, automatic storytelling reports, interactive explanation functions, etc.

[0023] In addition, the AI-based facility safety monitoring system according to the present invention can contribute to industrial development by reducing the budget burden on local governments through a subscription service model, enabling citizens to manage safety in their daily lives at a low cost, providing precise diagnostic services to engineers, and providing actual data to educational and research institutions.

[0024] In addition, the AI-based facility safety monitoring system according to the present invention learns various factors such as slope type, geology, and rainfall using AI to calculate customized management standards for individual slopes, thereby solving problems caused by the application of uniform standards and enabling more precise and preventive slope management. Brief explanation of the drawing

[0025] FIG. 1 is a drawing showing an AI-based facility safety monitoring system according to the present invention. FIGS. 2 and 3 are drawings illustrating the operation process of an AI-based facility safety monitoring system according to the present invention. Specific details for implementing the invention

[0026] Hereinafter, an AI-based facility safety monitoring system according to a preferred embodiment of the present invention will be described in detail with reference to the attached drawings.

[0027] FIGS. 1 to 3 illustrate an AI-based facility safety monitoring system according to the present invention. Referring to FIGS. 1 to 3, the AI-based facility safety monitoring system according to the present invention may be configured to include a data collection unit (100), an analysis unit (200), a visualization education unit (300), and a service provision unit (400).

[0028] The data collection unit (100) collects facility-related data in real time by including a sensor unit (110) that detects physical changes in the facility, a camera (120) that captures changes in the exterior, and a climate data API (130) that collects external environmental factors.

[0029] The analysis unit (200) receives facility-related data collected from the data collection unit (100) and analyzes, corrects, and detects outliers in the data through an AI analysis engine (210) to which AI-based machine learning and deep learning models are applied.

[0030] The visualization education department (300) combines the analysis results of the analysis department (200) with a 3D model, timeline, and explanatory text to visualize them so that users can intuitively understand them and generate customized educational data.

[0031] The service provider (400) provides educational data generated by the visualization education department (300) in a user-specific mode optimized according to the type of user (expert, citizen, student, etc.) and provides continuous safety management services through a subscription service model.

[0032] The analysis unit (200) automatically increases the measurement cycle of the data collection unit (100) and the shooting cycle of the camera (120) when an anomaly occurs to perform intensive monitoring, and returns to the original cycle when the situation returns to normal, and records all related processes in detail in an event log. This prevents the accumulation of unnecessary data and allows for selective response to crisis situations, thereby increasing system operation efficiency.

[0033] The analysis unit (200) analyzes image data collected from the camera (120) of the data collection unit (100) using the Digital Image Correlation (DIC) technique to calculate quantitative data regarding cracks, shape changes, and displacements on the surface of the structure, and can mutually correct the calculated displacement data with the sensor data. This compensates for the limitations of sensor data measured at a point level and enables a more reliable safety diagnosis through behavior analysis at the surface level.

[0034] The analysis unit (200) can be configured to include an advanced artificial intelligence engine that synthesizes sensor data, image data, and climate data to infer the cause of abnormal events, predicts future risks based on data trends, and performs multi-sensor cross-validation, scenario-based alarms, and automatic report generation.

[0035] The analysis unit (200) includes an interactive explanation function that responds to a user's natural language question about the displayed data based on the analyzed data. This can help non-expert users who have difficulty interpreting data to understand.

[0036] The analysis unit (200) autonomously adjusts and optimizes the data measurement mode (low power mode, communication optimization mode, etc.) of the data collection unit (100) by taking into account the power and network conditions at the site, thereby enabling stable data collection in any environment.

[0037] The analysis unit (200) learns categorical data such as the type of slope (soil, rock, etc.) and geological characteristics, and numerical data including rainfall intensity, slope, and measured values ​​using machine learning, so that it can accurately predict the risk level according to the unique characteristics of each slope, rather than a single management standard.

[0038] The visualization education department (300) links the 3D structural model data of the facility and the climate data (rainfall, snowfall, temperature, etc.) collected from the data collection department (100) to a timeline to clearly visualize the cause and effect between the occurrence of displacement and the patent weather event.

[0039] The visualization education department (300) can provide an educational interface that automatically displays detailed explanations of ‘how to read educational data’ and detailed specifications and operating principles of instruments including sensors, along with visualized educational data.

[0040] The service provider (400) can provide customized services to local governments, citizens, engineers, and educational institutions through a differentiated subscription service model.

[0041] Hereinafter, the configuration of the AI-based facility safety monitoring system according to the present invention described above will be explained in more detail.

[0042] The data collection unit (100) is equipped with a sensor unit (110) including displacement, inclination, groundwater level, humidity, and temperature sensors to detect changes in the condition of the facility. Additionally, the data collection unit (100) may be configured to include a miniature camera capable of wide-angle, close-up, and night shooting to precisely capture changes in the appearance of the structure. The wide-angle camera can monitor the behavior of the entire structure, the close-up camera monitors the progression of specific cracks, and the night shooting function can ensure continuous monitoring 24 hours a day. Furthermore, the data collection unit (100) can be linked with a climate data API (130) to collect external environmental factors such as rainfall, snowfall, and temperature in the monitoring area, thereby enabling the comprehensive collection of internal and external factors.

[0043] The analysis unit (200) performs the function of removing noise from raw data received from the data collection unit (100) through the data correction unit (260) and correcting drift. In addition, it detects outliers in the raw data through the outlier detection unit (220), automatically controls the data collection cycle when an outlier occurs, and analyzes the image captured by the camera (120) using DIC (Digital Image Correlation) technology to quantify displacement and mutually correct it with the sensor data.

[0044] The analysis unit (200) provides data correction and noise removal functions to correct errors, noise, drift (gradual change in measurements over time) that may be included in the collected raw data, thereby increasing the accuracy of the data.

[0045] The analysis unit (200) provides outlier detection and adaptive operation functions, so that when an abnormal value (outlier) is detected in the data, the measurement and camera shooting cycle is automatically shortened to monitor the situation more precisely. Then, when the situation returns to normal, it returns to the original cycle, and all of this process is recorded as an event.

[0046] The analysis unit (200) is equipped with a DIC image analysis unit (250) to provide an image-based DIC (Digital Image Correlation) analysis function, and analyzes an image captured by a camera using DIC technology to quantitatively detect cracks or shape changes in a structure. The image analysis results can be compared and corrected with sensor data to improve the reliability of the analysis.

[0047] In particular, the analysis unit (200) is equipped with an AI analysis engine (210) and can provide abnormal event cause inference, risk prediction, multi-sensor cross-verification, scenario-based alarm, automatic report and storytelling, interactive descriptor, autonomous optimization, machine learning and deep learning functions.

[0048] The analysis unit (200) can analyze the cause of abnormal phenomena by combining sensor, image, and climate data through an abnormal event cause inference function.

[0049] The analysis unit (200) can calculate the probability of future risk occurrence based on data trends through a risk prediction function.

[0050] The analysis unit (200) can determine whether there is a sensor error by cross-verifying adjacent sensor and image information through a multi-sensor cross-verification function.

[0051] The analysis unit (200) can operate a scenario-based (Level 1~3) alarm system that considers complex situations rather than simple thresholds through a scenario-based alarm function.

[0052] The analysis unit (200) may further be equipped with a report generation unit (250) to provide automatic report generation and storytelling functions, and may automatically generate a storytelling report in the form of natural language sentences by synthesizing measurement data, climate conditions, and events that occurred in the video.

[0053] The analysis unit (200) can respond in real time to natural language questions from citizens or students about measurement data through an interactive descriptor function.

[0054] The analysis unit (200) can autonomously adjust the data measurement mode by considering the power and network conditions at the site through an autonomous optimization function.

[0055] The analysis unit (200) can predict risk by learning categorical data such as slope type and geological characteristics and numerical data such as rainfall intensity and measured values ​​through machine learning and deep learning functions, and can analyze unstructured data such as DIC images using deep learning (CNN, Transformer, etc.) to extract crack and displacement patterns and correct them with sensor data.

[0056] The visualization education department (300) is equipped with a 3D visualization department (310) to visualize facilities as 3D models and display the direction and magnitude of displacement in the form of arrows, and overlays actual filmed footage to help users intuitively perceive the situation without going to the site.

[0057] In addition, the visualization education unit (300) is equipped with a timeline combination unit (320) to display climate events (rain, snow, etc.) and sensor data linked on the same timeline, thereby making it easy to identify the correlation between the two.

[0058] In addition, the visualization education department (300) is equipped with an educational interface department (330) to automatically provide detailed explanations on how to read data or the principles of measuring instruments, and can support an education mode that can be linked to classes at elementary, middle, high schools, and universities. The education mode, which is designed for elementary, middle, and high school students and non-experts, explains the meaning of data and measurement principles in an easy-to-understand manner to encourage citizens' participation in science, and can improve the quality of education and research by providing actual measurement and AI analysis data to universities and research institutions.

[0059] The service provider (400) is equipped with a user mode management unit (410) to provide differentiated interfaces and information access rights, such as expert mode, citizen mode, student and researcher mode, according to user characteristics.

[0060] In addition, the service provider (400) is equipped with a subscription management unit (420) and can provide customized safety management services through a subscription-based service model targeting local governments, citizens, engineers, and educational institutions.

[0061] The service provider (400) can provide local government subscription type, citizen subscription type, engineer subscription type, and educational institution subscription type services.

[0062] The local government subscription service prevents major accidents by immediately installing sensors on aging facilities where abnormal signs are detected during regular or detailed safety inspections for continuous management, and enables budget savings through the automation of analysis reports.

[0063] The citizen subscription service allows users to directly monitor everyday hazards such as construction sites and aging retaining walls around their homes through an easy-to-install simple sensor package, and to check data via an app and web.

[0064] The engineer subscription service provides specialized services such as 3D scanners, stability analysis, and measurement diagnosis when an urgent safety diagnosis is requested, and enables quick and accurate decision-making by utilizing analysis data.

[0065] The subscription service for educational institutions enables universities and research institutes to bridge the gap between theory and practice by utilizing real-time measurement data and AI analysis results of actual facilities, which are difficult to access, as foundational data for classes, research, and design.

[0066] According to a preferred embodiment of the present invention, the AI-based facility safety monitoring system collects data by synchronizing it with sensors and cameras installed in the facility and through a climate API. The collected data is transmitted to an analysis unit (200), where it is interpreted by AI, anomalies are verified, and cause inference and future risk prediction are performed. At the same time, video data is mutually corrected with sensor data through DIC analysis to increase reliability. All analyzed information is visualized by combining it with 3D models, climate event timelines, video overlays, etc. Finally, a storytelling-style report synthesizing data, graphs, video, and AI analysis results is automatically generated and provided as a subscription service optimized for each user group.

[0067] The AI-based facility safety monitoring system according to the present invention described above has been explained with reference to the attached drawings, but this is merely illustrative, and those skilled in the art will understand that various modifications and equivalent alternative embodiments are possible therefrom.

[0068] Therefore, the true scope of technical protection of the present invention should be determined solely by the technical concept of the appended claims. Explanation of the symbols

[0070] 100 : Data Collection Unit 200 : Analysis Department 300 : Visualization Ministry of Education 400 : Service Provider

Claims

Claim 1 An AI-based facility safety monitoring system comprises: a data collection unit that collects facility-related data including sensors that detect physical changes in a facility, cameras that capture external changes, and a climate data API that collects external environmental factors; an analysis unit that receives facility-related data collected by the data collection unit and analyzes, corrects, and detects outliers through a machine learning model; a visualization and education unit that visualizes the analysis results of the analysis unit by combining them with a 3D model, timeline, and explanatory text, and generates educational data; and a service provision unit that provides the educational data generated by the visualization and education unit in user-specific modes. The data collection unit includes a sensor unit that includes displacement, inclination, groundwater level, humidity, and temperature sensors to detect changes in the state of a facility, and a camera capable of wide-angle, close-up, and night shooting to capture external changes in a structure. The analysis unit removes noise from raw data received from the data collection unit through a data correction unit and corrects drift, detects outliers in the raw data through an outlier detection unit, automatically controls the data collection cycle when an abnormal situation occurs, analyzes images captured by the camera to quantify displacement and mutually corrects with sensor data, and multi-sensor mutual The system determines whether there is a sensor error by cross-verifying adjacent sensor and image information through a cross-verification function, and the analysis unit autonomously adjusts the data measurement mode, including the low-power mode and communication optimization mode of the data collection unit, by considering the power and network conditions of the site; the visualization education unit provides an educational interface that automatically displays detailed explanations of measuring instruments, including sensors, and how to read the educational data along with the visualized educational data; the visualization education unit is equipped with a timeline linkage unit to link and display climate events, including rainfall and snowfall, and sensor data on the same timeline, thereby enabling the identification of the correlation between the two; and the service provider unit [enables] local governments, citizens, engineers,An AI-based facility safety monitoring system characterized by providing customized services to educational institutions through a subscription service model, wherein the service provider offers local government subscription, citizen subscription, engineer subscription, and educational institution subscription services; wherein the local government subscription service prevents major accidents by installing and managing sensors on aging facilities where abnormal signs are detected during regular or precise safety inspections, and enables budget savings through the automation of analysis reports; wherein the citizen subscription service monitors risk factors in daily life, including construction sites and aging retaining walls near homes, through a simplified sensor package and allows data to be checked via an app and the web; wherein the engineer subscription service provides professional services, including 3D scanners, stability analysis, and measurement diagnosis, upon request for an emergency safety inspection, and enables decision-making using analysis data; and wherein the educational institution subscription service reduces the gap between theory and practice by utilizing real-time measurement data of actual facilities and AI analysis results, which are difficult for universities or research institutions to access, as basic data for classes, research, and design. Claim 2 An AI-based facility safety monitoring system according to claim 1, characterized in that the analysis unit automatically increases the measurement cycle of the data collection unit and the shooting cycle of the camera when an anomaly occurs, returns to the original cycle when the situation returns to normal, and records an event log. Claim 3 The AI-based facility safety monitoring system according to claim 1, wherein the analysis unit analyzes image data collected from the camera of the data collection unit using a Digital Image Correlation (DIC) technique to quantify displacement and mutually corrects the quantified displacement data with sensor data. Claim 4 An AI-based facility safety monitoring system according to claim 1, characterized in that the visualization education unit visualizes the causes and results of displacement by linking 3D structural schematic data of the facility and climate data collected from the data collection unit to a timeline. Claim 5 delete Claim 6 An AI-based facility safety monitoring system according to claim 1, wherein the analysis unit comprises an artificial intelligence engine that synthesizes sensor data, image data, and climate data to infer the cause of abnormal events, predicts future risks based on data trends, and performs multi-sensor cross-validation, scenario-based alarms, and automatic report generation. Claim 7 An AI-based facility safety monitoring system according to claim 1, wherein the analysis unit includes an interactive explanation function that responds to a user's natural language question based on analyzed data. Claim 8 delete Claim 9 delete Claim 10 An AI-based facility safety monitoring system according to claim 1, wherein the analysis unit learns categorical data such as slope type and geological characteristics and numerical data including rainfall intensity, slope, and measured values ​​through machine learning to predict the risk of the slope.

Images