Construction safety real-time high-precision detection method and system based on environmental characteristics

Abstract

The invention relates to the technical field of video recognition, in particular to a construction safety real-time high-precision detection method and system based on environmental characteristics. Arranging a plurality of sensors in a to-be-detected area to collect field environment data in real time; analyzing the illumination change of the to-be-detected area through the light and shadow analysis model, calculating a light-safety misjudgment index, and identifying a potential light misjudgment area; a visual processing model is utilized to analyze the worker video, human skeleton key points are recognized in real time, personnel behavior safety indexes are obtained, and potential dangerous behaviors are detected; detecting local wind speed and turbulence conditions and vibration and resonance phenomena of a to-be-detected area in real time through a local breeze and vibration detection model, and evaluating a local environment stability index; and the safety level of the to-be-detected area is calculated in real time by using the light-safety misjudgment index, the personnel behavior safety index and the local environment stability index, so that multi-dimensional and real-time construction safety risk assessment is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of video recognition, and particularly to a real-time high-precision detection method and system for construction safety based on environmental features. Background Art

[0002] Working at heights is a high-risk task in construction, and construction workers face various safety threats from environmental factors, equipment operation, and personal behavior. In recent years, the application of image processing technology in construction safety monitoring has gradually received extensive attention, especially in real-time monitoring and dynamic safety detection. By installing camera or sensor devices to capture real-time images and collect data at the construction site, image processing technology can quickly identify potential safety issues in the scene. For example, computer vision algorithms can be used to automatically identify whether construction workers are wearing protective equipment or monitor potential hazards in the construction environment (such as falling objects from height, un-reinforced structures, etc.). Traditional image processing technology can, to a certain extent, identify safety hazards at the construction site by analyzing images.

[0003] However, the application of existing image processing technology in construction safety monitoring still has certain limitations. Traditional image processing technology rarely considers the influence of light and shadow when monitoring construction safety. Light changes can easily lead to visual misjudgment or depth perception errors of workers (such as misjudging steps, missing guardrails). At the same time, existing safety monitoring for working at heights only focuses on the overall wind speed and lacks consideration of the impact of local sudden wind on the body stability of workers.

[0004] Therefore, a real-time high-precision detection method and system for construction safety based on environmental features are proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a real-time high-precision detection method and system for construction safety based on environmental features to achieve high-precision detection of construction safety for working at heights. It includes arranging a variety of sensors in the area to be detected to collect on-site environmental data in real time; analyzing the light change in the area to be detected through a light and shadow analysis model, calculating the light-safety misjudgment index, and identifying potential light misjudgment areas; using a vision processing model to identify human skeletal key points in real time, obtaining the personnel behavior safety index, and detecting potential dangerous behaviors; detecting the local wind speed, turbulence, vibration, and resonance phenomena in the area to be detected through a local breeze and vibration detection model in real time, and evaluating the local environmental stability index; using the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generating a safety detection report, and triggering a safety warning mechanism.

[0006] To achieve the above object, the present invention provides the following technical solutions: A real-time high-precision construction safety detection method based on environmental characteristics, comprising: Deploy a variety of sensors in the area to be detected to collect on-site environmental data in real time, including illumination data, worker videos, wind speed data, humidity data, air flow data, and vibration data; Analyze the illumination changes in the area to be detected through a light and shadow analysis model based on the illumination data and the worker videos, calculate the light-safety misjudgment index, and identify potential light misjudgment areas; Use a vision processing model to analyze the worker videos, real-time identify the key points of the human skeleton, obtain the personnel behavior safety index, and detect potential dangerous behaviors; Real-time detect the local wind speed, turbulence conditions, vibration, and resonance phenomena in the area to be detected through a local breeze and vibration detection model, and evaluate the local environmental stability index; Utilize the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generate a safety detection report, and trigger a safety warning mechanism.

[0007] Preferably, the variety of sensors include: an illumination sensor, a high-definition camera, a temperature and humidity sensor, a wind speed sensor, and an acceleration sensor; The illumination data includes natural light intensity, artificial lighting brightness, illumination direction, shadow distribution, and illumination uniformity parameters; The worker videos include high-definition video sequences of workers operating in the area to be detected; The wind speed data includes the local wind speed vector, the maximum instantaneous wind speed, and the wind direction change trend in the area to be detected; The humidity data includes the relative air humidity and the ground humidity in the area to be detected; The air flow data includes the local air flow direction, small-scale turbulence intensity, and disturbance frequency; The vibration data includes the vibration amplitude, vibration frequency, vibration acceleration, and structural resonance frequency points in the area to be detected.

[0008] Preferably, the light and shadow analysis model includes: an illumination extraction unit, a shadow recognition unit, a dynamic illumination change tracking unit, and a light misjudgment evaluation unit; The illumination extraction unit extracts the pixel brightness histogram and illumination gradient distribution in the image frames of the worker videos after preprocessing; The shadow recognition unit separates the dynamic illumination occlusion shadows and static structure occlusion shadows in the area to be detected and generates a shadow mask map; The dynamic illumination change tracking unit analyzes the illumination mutation trend of consecutive image frames based on the pixel brightness histogram, illumination gradient distribution, and the shadow mask map; The light misjudgment evaluation unit fuses the shadow mask map with the construction path and the operation edge area, evaluates the risk probability of visual illusion caused by light, outputs the light-safety misjudgment index of each area in the area to be detected, and marks the area where the light-safety misjudgment index is greater than the preset misjudgment threshold as the potential light misjudgment area.

[0009] Preferably, the visual processing model includes: a human body skeleton key point recognition unit, a posture evaluation unit, and an abnormal behavior detection unit; The human body skeleton key point recognition unit processes the image frames of the worker video through a posture estimation model based on deep learning, extracts the coordinates of human key points, and generates a human skeleton structure diagram; The posture evaluation unit calculates the angle change, center of gravity offset, and movement trajectory curve between skeleton points according to the human skeleton structure diagram, and identifies the human posture information; The abnormal behavior detection unit compares the human posture information with the dangerous behavior database, combines the humidity data to detect potential dangerous behaviors, and outputs a human behavior safety index.

[0010] Preferably, the local breeze and vibration detection model includes: a local wind speed and turbulence condition analysis unit, a vibration and resonance phenomenon analysis unit, and a local environmental stability evaluation unit; The local wind speed and turbulence condition analysis unit identifies local sudden wind events and small-scale turbulence characteristics according to the wind speed data and the air flow data; The vibration and resonance phenomenon analysis unit compares the vibration data with the natural frequency of the construction structure to identify vibration and resonance phenomena; The local environmental stability evaluation unit combines the local sudden wind events, small-scale turbulence characteristics, and the vibration and resonance phenomena, and outputs the local environmental stability index.

[0011] Preferably, the calculation process of the safety level includes: Input the light-safety misjudgment index, the human behavior safety index, and the local environmental stability index into a risk assessment model to generate a quantitative risk level score, including low risk, medium risk, high risk, and extremely high risk, to obtain the safety level of the area to be detected; The risk assessment model dynamically adjusts the weights of each index according to the construction type, operation time, wind force level, and historical safety accident data.

[0012] Preferably, a real-time high-precision construction safety detection system based on environmental characteristics includes: A data acquisition module, which is used to arrange a variety of sensors in the area to be detected to collect on-site environmental data, including light, worker videos, wind speed, humidity, air flow, and vibration; A light and shadow analysis module, which is used to analyze the light change in the area to be detected through a light and shadow analysis model according to the light, calculate the light-safety misjudgment index, and identify potential visual illusion areas; A visual recognition module, which is used to analyze the worker video by using a visual processing model, real-time identify the key points of the human body skeleton, obtain the personnel behavior safety index, and detect potential dangerous behaviors; A local breeze and vibration detection module, which is used to detect the local wind speed and turbulence conditions in the area to be detected and the vibration and resonance phenomena of the platform to be detected in real time through a local breeze and vibration detection model, and evaluate the local environmental stability index; A safety monitoring and early warning module, which is used to use the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generate a safety detection report, and trigger a safety early warning mechanism.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The present invention introduces the concept of "light-safety misjudgment index" in construction safety detection, and jointly models the light data and worker videos through a light and shadow analysis model to identify visual illusion areas caused by uneven light, shadow occlusion, strong light reflection, etc. This model can effectively make up for the neglect of "optical hazards" in traditional safety detection systems, and can effectively reduce accidents such as stepping into the air and misjudging the edge caused by light misjudgment in scenarios such as bridge construction and high-altitude operations, improving the safety detection ability in light and shadow scenarios.

[0014] 2. The present invention adopts the personnel posture recognition and abnormal behavior detection technology based on a visual processing model, abandons the traditional dependence on wearable devices (such as accelerometers, gyroscopes, etc.) and / or RFID tags, and can accurately extract the key points of the human body skeleton and analyze its behavior characteristics only through worker video data. This model greatly improves the adaptability of the system and is applicable to working environments where it is inconvenient to wear devices, such as high altitudes and narrow spaces. At the same time, combined with a deep learning behavior recognition model, dangerous actions such as slipping, excessive bending, and imbalance can be detected in real time, effectively avoiding the risks brought by illegal operations of workers, thereby improving the construction safety detection accuracy in high-altitude operations.

[0015] 3. The present invention establishes a joint modeling mechanism for local micro-airflow and platform vibration at the high-altitude operation construction site through the local micro-wind and vibration detection model based on a high-precision wind speed sensing array and a ground structure vibration sensor. This model can not only monitor micro-scale disturbances such as turbulence, sudden wind, and wind direction deflection in real time, but also give early warning judgments on whether the high-altitude operation construction platform is in a critical resonance state. Compared with the traditional "average wind speed" judgment mode, this method has higher accuracy and faster response, and is particularly suitable for early detection of falling risk hazards in high-altitude operation sensitive scenarios such as high-altitude steel structure erection and bridge support construction, effectively improving the detection ability of construction safety problems caused by local micro-airflow and platform vibration in high-altitude operation scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic flow chart of a method for real-time high-precision detection of construction safety based on environmental characteristics provided by an embodiment of the present invention; Figure 2 It is a schematic structural diagram of a system for real-time high-precision detection of construction safety based on environmental characteristics provided by an embodiment of the present invention; Figure 3 It is a working principle diagram of the light and shadow analysis model provided by an embodiment of the present invention; Figure 4 It is a schematic structural diagram of the visual processing model provided by an embodiment of the present invention; Figure 5 It is a working principle diagram of the local micro-wind and vibration detection model provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] Working at heights is a high-risk task in construction. Construction workers face safety threats from various aspects such as environmental factors, equipment operation, and personal behavior. Traditional image processing technologies for monitoring construction safety rarely consider the influence of light and shadow, and light changes are likely to cause visual misjudgment or depth perception errors of workers (such as misjudging the absence of steps or guardrails). At the same time, existing high-altitude operation safety monitoring only focuses on the overall wind speed and lacks consideration of the impact of local sudden wind force on the body stability of workers.

[0019] The present invention provides a real-time high-precision detection method and system for construction safety based on environmental characteristics, which is applicable to high-risk working condition scenarios such as high-altitude operations, bridge construction, and steel structure erection. It can comprehensively sense lighting conditions, personnel behavior, and environmental disturbances, and realize multi-dimensional and real-time construction safety risk assessment and early warning. In order to illustrate that the method of the present invention can play a role in real-time high-precision detection of construction safety for high-altitude operations, the effectiveness of the present invention will be described below from two embodiments.

[0020] Embodiment 1 In the embodiment of the present application, the method proposed by the present invention is used to elaborate on the real-time high-precision detection process of construction safety for high-altitude operations by combining the influences of light and shadow and local sudden wind force, etc. The embodiment of the present application aims at the real-time high-precision detection of construction safety for high-altitude operations in a certain construction site A of a building. The following is based on Figure 1 the content to elaborate on the real-time high-precision detection process of the construction safety for high-altitude operations; among them, Figure 1 is the specific flowchart of the method proposed by the present invention, including: arranging a variety of sensors in the area to be detected to collect on-site environmental data in real time; analyzing the lighting changes in the area to be detected through a light and shadow analysis model, calculating the light-safety misjudgment index, and identifying potential light misjudgment areas; using a vision processing model to identify the key points of the human body skeleton in real time, obtaining the personnel behavior safety index and detecting potential dangerous behaviors; detecting the local wind speed, turbulence situation, vibration and resonance phenomena in the area to be detected in real time through a local breeze and vibration detection model, and evaluating the local environmental stability index; using the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generating a safety detection report and triggering a safety early warning mechanism. Combining Figure 1 and Figure 2 the content in the following description is as follows: A real-time high-precision detection method for construction safety based on environmental characteristics, including: Arranging a variety of sensors in the area to be detected to collect on-site environmental data in real time, including lighting data, worker videos, wind speed data, humidity data, air flow data, and vibration data; The variety of sensors include: a lighting sensor, a high-definition camera, a temperature and humidity sensor, a wind speed sensor, and an acceleration sensor; The lighting data includes natural light intensity, artificial lighting brightness, lighting direction, shadow distribution, and lighting uniformity parameters; The worker videos include high-definition video sequences of the worker operating in the area to be detected; The wind speed data includes the local wind speed vector, the maximum instantaneous wind speed, and the wind direction change trend in the area to be detected; The humidity data includes the relative air humidity and the ground humidity in the area to be detected; The air flow data includes local air flow direction, small-scale turbulence intensity, and perturbation frequency; The vibration data includes the vibration amplitude, vibration frequency, vibration acceleration, and structural resonance frequency points of the area to be detected.

[0021] Specifically, 6 high-resolution industrial cameras are deployed on the operation platform (including the edge area and the entrance and exit channels), with a resolution of not less than 1920×1080, for continuously collecting worker videos; each camera uses a structured light depth camera and can collect RGB images and depth information; the cameras are connected to the local edge computing server through a wired gigabit network; 4 ultrasonic anemometers are respectively deployed at the four corners of the platform, with a collection frequency of 10Hz, for collecting wind speed data and air flow data; 2 multi-channel illuminance sensing modules are set on the top of the platform for collecting illumination data; 3-axis vibration accelerometers are installed at the middle part of the platform and the support nodes, with a collection frequency of 1000Hz, for collecting vibration data; 2 temperature and humidity sensors are respectively arranged in the upper, middle, and lower areas of the platform for collecting humidity data.

[0022] Through the synchronous deployment and collection of multi-type and multi-dimensional sensors (videos, wind speed, illumination, humidity, vibration, etc.), it is possible to realize the comprehensive modeling and real-time perception of the construction site environment. Compared with the traditional monitoring system with a single perception dimension, this method significantly improves the timeliness and spatial coverage of data acquisition, provides a high-resolution and multi-source heterogeneous data basis for subsequent comprehensive perception of illumination conditions, personnel behavior, and environmental perturbations, and realizes multi-dimensional and real-time construction safety risk assessment and early warning, thereby improving the accuracy and response speed of the overall construction safety detection and early warning.

[0023] Preferably, as Figure 3 shown, the illumination change of the area to be detected is analyzed through the light and shadow analysis model according to the illumination data and the worker videos, the light-safety misjudgment index is calculated, and potential light misjudgment areas are identified; The light and shadow analysis model includes: an illumination extraction unit, a shadow recognition unit, a dynamic illumination change tracking unit, and a light misjudgment evaluation unit; The illumination extraction unit extracts the pixel brightness histogram and illumination gradient distribution in the image frames of the worker videos after preprocessing; The shadow recognition unit separates the dynamic illumination occlusion shadows and static structure occlusion shadows in the area to be detected and generates a shadow mask map; The dynamic illumination change tracking unit analyzes the illumination mutation trend of consecutive image frames according to the pixel brightness histogram, illumination gradient distribution, and the shadow mask map; The light misjudgment evaluation unit fuses the shadow mask map with the construction path and the operation edge area, evaluates the risk probability of visual illusion caused by light, outputs the light-safety misjudgment index of each area in the area to be detected, and marks the area where the light-safety misjudgment index is greater than the preset misjudgment threshold as the potential light misjudgment area.

[0024] Specifically, the shadow recognition unit is based on a trained deep convolutional neural network model (such as ShadowNet or U-Net), inputs the video image frame, the pixel brightness histogram and the light gradient distribution obtained by the light extraction unit, and performs pixel-level shadow area segmentation; Combined with the time series analysis method, it is judged whether the shadow area moves with time or remains unchanged all the time, so as to separate the dynamic light occlusion shadow and the static structure occlusion shadow; Output a shadow mask map, in which the dynamic light occlusion shadow area and the static structure occlusion shadow area are respectively marked for use by the subsequent light change tracking module.

[0025] The dynamic light change tracking unit compares the video image frames frame by frame, compares the brightness histogram and the light gradient map in multiple consecutive frames of images, and calculates the light change amplitude in the time dimension; Perform inter-frame difference on the shadow mask map, and analyze the change trend of the shadow area, position and shape with time; Introduce the optical flow estimation technology to judge whether the light source (such as the sun or the construction lamp) has a relative displacement or a sudden change in brightness; Based on the above information, output a set of time series feature vectors describing the degree of regional light change, reflecting whether there are potential visually misleading changes in this area.

[0026] The light misjudgment evaluation unit fuses and compares the shadow mask map with the predefined construction path map and the operation area boundary map, and judges whether the shadow or strong reflection covers the key operation path or the edge of the safety guardrail. If it covers, there is a visual illusion risk in this area; Referring to the human eye visual model (including contrast sensitivity, edge blur recognition ability, etc.), evaluate the probability of misjudgment that may occur to construction workers, and output the light-safety misjudgment index of each area to be detected; if the light-safety misjudgment index of a certain area exceeds the preset misjudgment threshold (such as 0.8), then mark it as a potential light misjudgment area for subsequent risk synthesis evaluation and early warning system triggering.

[0027] Table 1 is the light misjudgment area recognition efficiency table.

[0028] Table 1 Light misjudgment area recognition efficiency table ; This embodiment proposes a light and shadow analysis model based on image brightness, shadow recognition, and time series analysis, which can intelligently detect potential visual risks caused by changes in lighting conditions, such as misjudgment problems of steps and holes caused by shadow occlusion. This model makes up for the shortcoming of the insufficient recognition ability of traditional monitoring systems in complex lighting scenarios, realizes the quantitative evaluation of the risk of safety misjudgment induced by light, and improves the practicality and intelligence level under real high-altitude operation construction conditions.

[0029] Preferably, as Figure 4 shown, use a visual processing model to analyze the worker video, real-time identify the key points of the human body skeleton, obtain the personnel behavior safety index, and detect potential dangerous behaviors; The visual processing model includes: a human body skeleton key point recognition unit, a posture evaluation unit, and an abnormal behavior detection unit; The human body skeleton key point recognition unit processes the image frames of the worker video through a deep learning-based pose estimation model, extracts the coordinates of the human key points, and generates a human skeleton structure diagram; The posture evaluation unit calculates the angular changes, center of gravity offsets, and motion trajectory curves between the skeleton points according to the human skeleton structure diagram to identify the personnel posture information; The abnormal behavior detection unit compares the personnel posture information with the dangerous behavior database, combines the humidity data to detect potential dangerous behaviors, and outputs the personnel behavior safety index.

[0030] Specifically, the human body skeleton key point recognition unit includes: a personnel target detection layer, a skeleton key point detection layer, a key point coordinate post-processing layer, and a perspective fusion recognition layer; The personnel target detection layer uses a human target detection model to detect the worker main body in each frame of the image. Preferably, lightweight or high-precision detection networks such as YOLOv7 and Faster R-CNN are used to identify the worker target in the image and obtain its bounding box coordinates; the detected bounding box area is cropped to obtain an independent image area for each human body; The skeleton key point detection layer performs skeleton key point recognition on the extracted human body image area using a deep learning pose estimation model; the deep learning pose estimation model can select advanced network structures such as HRNet, OpenPose, ViTPose, and DETR-Pose to realize the coordinate prediction of the skeleton key points of each worker (25 points in total: head, shoulders, elbows, wrists, hips, knees, ankles, feet), and output the confidence score corresponding to each key point. If the confidence of a certain key point is lower than the set confidence threshold (such as 0.7), the system regards it as unavailable data and performs missing compensation or ignoring processing in the post-processing link.

[0031] After the post - processing layer of the key - point coordinates obtains the preliminary recognition result, it performs post - processing on the key - point coordinates; maps the two - dimensional points in the image coordinate system to the three - dimensional space to obtain the spatial bone - point coordinates in the actual high - altitude operation environment; smooths the positions of the bone points between frames by methods such as Kalman filtering and exponentially weighted moving average; calculates the connection vectors and angular relationships between adjacent key points to construct a complete human skeleton structure diagram for action evaluation.

[0032] The perspective - fusion recognition layer fuses the multi - camera perspective image information based on image stitching, feature fusion, or spatial reconstruction methods to enhance the recognition accuracy.

[0033] The posture - evaluation unit constructs bone vectors between each pair of key points in the human skeleton structure diagram, calculates the included angle between the bones using the cosine - angle formula, and calculates the angle change between the bone points. The center - of - gravity deviation calculates the position of the human body's center of mass using the coordinates of all body key points, and then combines the trajectory of the center - of - gravity point in the time series to determine whether there is a significant deviation or imbalance tendency. Model the trajectories of the bone key - point coordinates in several consecutive frames, construct the human motion curve by methods such as polynomial fitting or spline interpolation, identify dynamic behaviors such as bending, jumping, and sudden squatting, and obtain the motion - trajectory curve. Based on the above - mentioned angle changes, center - of - gravity deviation, and the degree of change in the motion curve, construct a calculation function for the personnel behavior safety index to quantify the stability of the personnel's posture; the formula for the personnel behavior safety index is: ; Among them, is the personnel behavior safety index; is the adjustment weight of the angle - change term; is the total number of bone key points; is the th bone key point's frame angle; is the frame interval; is the reference angle - fluctuation threshold of this bone key point within the normal range; is the adjustment weight of the center - of - gravity deviation term; is the frame's human - body center - of - gravity coordinate; is the reference center - of - gravity coordinate of the human body when standing normally; is the maximum acceptable center - of - gravity deviation distance; is the th key point's speed sequence; is the adjustment weight of the motion - curve change - degree term; is the th standard deviation of the speed sequence of the bone key point; is the reference standard value for the fluctuation of the steady motion speed; Classify the personnel behavior safety index and detect potential dangerous behaviors: If : The posture is stable and there are no potential dangerous behaviors; If : There may be slight posture instability and slight potential dangerous behaviors; If : The posture is unstable and there are serious potential dangerous behaviors.

[0034] The abnormal behavior detection unit compares the skeletal key point angle features output by the posture evaluation unit, the dynamic time warping (DTW) of the center of gravity trajectory, or the human behavior recognition model based on the graph convolutional network (GCN) with the built-in dangerous behavior database of the system, and jointly analyzes the humidity data and the posture recognition result; When it is recognized that the person shows signs of foot sliding and imbalance, and at the same time the environmental humidity is higher than 90%, or the dew point is close to the ground temperature, the system increases the confidence level of the risk level judgment of this behavior; Establish a dual judgment mechanism (posture matching + humidity anomaly) for suspected abnormal behaviors, and output the behavior results, including the abnormal type, start timestamp, and confidence score, and finally output the personnel behavior safety index.

[0035] Table 2 gives a comparison table of the recognition accuracy of personnel behaviors.

[0036] Table 2 Comparison Table of Personnel Behavior Recognition Accuracy ; The embodiment of this application proposes that the visual processing model adopts a deep learning posture recognition network and an abnormal behavior recognition module, which can, without relying on wearable devices, perform real-time tracking and analysis on the dynamic postures of construction workers. By recognizing dangerous behaviors such as "slip", "climb", and "irregular bending", it realizes high-efficiency and low-invasive human factor safety risk monitoring. Compared with traditional manual inspections and wearable devices, this method is more adaptable to complex construction environments and has higher recognition accuracy and coverage.

[0037] Preferably, as Figure 5 shown, the local micro-wind and vibration detection model is used to detect the local wind speed, turbulence situation, vibration, and resonance phenomenon in the area to be detected in real time, and evaluate the local environmental stability index; The local micro-wind and vibration detection model includes: a local wind speed and turbulence situation analysis unit, a vibration and resonance phenomenon analysis unit, and a local environmental stability evaluation unit; The local wind speed and turbulence situation analysis unit identifies local sudden wind force events and small-scale turbulence characteristics according to the wind speed data and the air flow data; The vibration and resonance phenomenon analysis unit compares the vibration data with the natural frequency of the construction structure to identify vibration and resonance phenomena; The local environmental stability evaluation unit combines the local sudden wind force event, small-scale turbulence characteristics and the vibration and resonance phenomena to output the local environmental stability index.

[0038] Specifically, the local wind speed and turbulence condition analysis unit determines whether a local sudden wind force event occurs by calculating the short-term (within 3 seconds) wind speed change amplitude and wind direction deviation angle; calculates the spectral density function of the local air flow through Fourier transform, and identifies the air flow disturbance with high-frequency components (such as > 3 Hz) in the frequency distribution, and determines it as small-scale turbulence characteristics.

[0039] The vibration and resonance phenomenon analysis unit applies the fast Fourier transform (FFT) to the vibration sequence collected by the sensor, extracts the main frequency components, including the vibration frequency peak, amplitude and harmonic number, and generates a structural vibration spectrogram; queries the natural frequency of the construction component from the structural design parameter library and compares it with the current main vibration frequency. If the preset conditions are met, it is determined that the structure has a resonance risk.

[0040] The local environmental stability evaluation unit inputs the local sudden wind force event frequency, turbulence intensity, vibration and resonance phenomena into the local environmental stability index calculation function: ; where, is the local environmental stability index; is the weight of the local sudden wind force event frequency term; is the local sudden wind force event frequency; is the evaluation time window; is the weight of the turbulence intensity term; is the turbulence intensity; is the weight of the vibration and resonance phenomenon term; is the vibration and resonance phenomenon, is the presence of resonance risk, is the absence of resonance risk.

[0041] Table 3 gives the correlation analysis table of platform disturbance and resonance risk.

[0042] Table 3 Correlation Analysis Table of Platform Disturbance and Resonance Risk ; Analyze wind speed, turbulence, and structural resonance risks through a local breeze and vibration detection model. This model can capture the micro-perturbation changes of the local wind field in real time at high altitudes or marginal areas, and predict the structural stability risks through the vibration spectrum and structural resonance identification mechanism. This fine-grained wind-vibration coupling monitoring method is significantly better than the coarse-grained monitoring of conventional overall anemometers, and can perceive in advance the risks of personnel falling or abnormal structural sway caused by local environmental instability, provide technical support for high-risk operation scenarios, and enhance the dynamic prevention and control ability of on-site safety.

[0043] Preferably, use the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generate a safety detection report, and trigger a safety warning mechanism; The calculation process of the safety level includes: Input the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index into a risk assessment model to generate a quantitative risk level score, including low risk, medium risk, high risk, and extremely high risk, to obtain the safety level of the area to be detected; The risk assessment model dynamically adjusts the weights of each index according to the construction type, operation time, wind force level, and historical safety accident data.

[0044] Specifically, the formula for the risk assessment model to generate a quantitative risk level score is: ; Among them, is the quantitative risk level score; is the weight of the light-safety misjudgment index; is the light-safety misjudgment index; is the weight of the personnel behavior safety index; is the personnel behavior safety index; is the weight of the local environmental stability index; is the local environmental stability index; Divide the risk degree according to the quantitative risk level score: : Low risk; : Medium risk; : High risk; : Extremely high risk.

[0045] The risk assessment model dynamically adjusts the weights of each index according to the construction type, operation time, wind force level, and historical safety accident data. The dynamic adjustment process is: Light-safety misjudgment index weight: Dynamically adjusted according to lighting conditions and working hours (such as day or night). When it is night or the light is poor, the risk of lighting misjudgment is higher, and the weight will be appropriately increased.

[0046] Weight of personnel behavior safety index: The weight of the personnel behavior safety index is adjusted according to the construction type and historical accident data. For example, if historical accidents occur frequently in a certain type of operation, the risk assessment of personnel behavior will increase the weight.

[0047] Weight of local environmental stability index: The stability of the local environment is affected by factors such as wind speed and turbulence. When the wind level is high or the environment is unstable, the weight of environmental stability will increase.

[0048] In this embodiment, by comprehensively considering the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index, a real-time safety level assessment of the construction area is realized by using a dynamically adjusted risk assessment model, and a comprehensive risk level under multi-dimensional coordination is output, and multi-channel warning means are triggered according to the risk level. Compared with traditional qualitative analysis or single-factor threshold judgment, this method comprehensively considers the impacts of light misjudgment, personnel behavior, and local sudden wind disturbance on the construction safety of high-altitude operations, and has significant advantages in terms of response accuracy, coverage comprehensiveness, and intelligent level, thus improving the accuracy and real-time performance of the safety detection for high-altitude operation construction.

[0049] The present invention realizes comprehensive, real-time monitoring and high-precision safety risk assessment of the construction site for high-altitude operations by integrating a variety of environmental sensors and intelligent analysis models. First, by deploying light sensors, video cameras, wind speed sensors, humidity sensors, air flow sensors, and vibration sensors, multi-dimensional environmental data of the construction area is collected in real time. Through the processing and analysis of the light and shadow analysis model, visual processing model, and local breeze and vibration detection model, these data can accurately identify potential risk factors such as light changes, worker behavior, wind speed turbulence, and structural vibration. Specifically, the light and shadow analysis model analyzes the light changes in the construction area through light data and video images, calculates the light-safety misjudgment index, and identifies areas with insufficient light or misjudgment to ensure reducing the risk of misjudgment in an environment with complex light changes; the visual processing model calculates the personnel behavior safety index and identifies potential dangerous behaviors, such as slipping, losing balance, or not wearing protective equipment, by real-time identifying the key points of the human body skeleton to ensure the safety of the operators; the local breeze and vibration detection model monitors the local environmental stability of the construction area in real time through wind speed, turbulence, and vibration data, and identifies potential safety hazards such as wind impact and structural vibration. After comprehensive evaluation of these data, a smart risk assessment model calculates the safety level of the construction area. Based on the evaluation results, the system can generate a safety detection report in real time, and when potential high-risk situations are detected, automatically trigger a safety warning mechanism to timely remind construction workers and managers to take necessary safety measures, reducing the possibility of accidents, thus greatly improving the safety detection level of the construction site.

[0050] Embodiment 2 In Embodiment 1, the method proposed by the present invention successfully realizes real-time high-precision detection of the construction safety of high-altitude operations by combining the influences of light and shadow and local sudden wind force. To further verify the effectiveness of the present invention, in the embodiments of this application, real-time high-precision detection of the construction safety of high-altitude operations at a certain construction site B of a building is also carried out.

[0051] A real-time high-precision detection system for construction safety based on environmental characteristics, as Figure 2 shown, includes: A data acquisition module, configured to arrange a variety of sensors in the area to be detected to collect on-site environmental data, including light, worker video, wind speed, humidity, air flow, and vibration; The variety of sensors include: a light sensor, a high-definition camera, a temperature and humidity sensor, a wind speed sensor, and an acceleration sensor; The light data includes natural light intensity, artificial lighting brightness, light direction, shadow distribution, and light uniformity parameters; The worker video includes a high-definition video sequence of the worker operating in the area to be detected; The wind speed data includes the local wind speed vector, the maximum instantaneous wind speed, and the wind direction change trend in the area to be detected; The humidity data includes the relative air humidity and the ground humidity in the area to be detected; The air flow data includes the local air flow direction, the small-scale turbulence intensity, and the perturbation frequency; The vibration data includes the vibration amplitude, vibration frequency, vibration acceleration, and structural resonance frequency points in the area to be detected.

[0052] Preferably, a light and shadow analysis module is used to analyze the light change in the area to be detected through a light and shadow analysis model according to the illumination, calculate the light-safety misjudgment index, and identify potential visual illusion areas; The light and shadow analysis model includes: a light extraction unit, a shadow recognition unit, a dynamic light change tracking unit, and a light misjudgment evaluation unit; The light extraction unit extracts the pixel brightness histogram and the illumination gradient distribution in the image frames of the worker video after preprocessing; The shadow recognition unit separates the dynamic light occlusion shadows and static structure occlusion shadows in the area to be detected and generates a shadow mask map; The dynamic light change tracking unit analyzes the illumination mutation trend of consecutive image frames according to the pixel brightness histogram, the illumination gradient distribution, and the shadow mask map; The light misjudgment evaluation unit fuses the shadow mask map with the construction path and the operation edge area, evaluates the risk probability of visual illusion caused by illumination, outputs the light-safety misjudgment index of each area in the area to be detected, and marks the area where the light-safety misjudgment index is greater than the preset misjudgment threshold as the potential light misjudgment area.

[0053] Preferably, a visual recognition module is used to analyze the worker video by using a visual processing model, real-time identify the key points of the human body skeleton, obtain the personnel behavior safety index, and detect potential dangerous behaviors; The visual processing model includes: a human body skeleton key point recognition unit, a posture evaluation unit, and an abnormal behavior detection unit; The human body skeleton key point recognition unit processes the image frames of the worker video through a deep learning-based pose estimation model, extracts the coordinates of the human body key points, and generates a human body skeleton structure diagram; The posture evaluation unit calculates the angle change, center of gravity offset, and motion trajectory curve between the skeleton points according to the human body skeleton structure diagram to identify the personnel posture information; The abnormal behavior detection unit compares the personnel posture information with the dangerous behavior database, combines the humidity data to detect potential dangerous behaviors, and outputs the personnel behavior safety index.

[0054] Preferably, the local breeze and vibration detection module is used to detect the local wind speed and turbulence condition of the area to be detected and the vibration and resonance phenomena of the platform to be detected in real time through the local breeze and vibration detection model, and evaluate the local environmental stability index; The local breeze and vibration detection model includes: a local wind speed and turbulence condition analysis unit, a vibration and resonance phenomenon analysis unit, and a local environmental stability evaluation unit; The local wind speed and turbulence condition analysis unit identifies local sudden wind force events and small-scale turbulence characteristics according to the wind speed data and the air flow data; The vibration and resonance phenomenon analysis unit compares the vibration data with the natural frequency of the construction structure to identify vibration and resonance phenomena; The local environmental stability evaluation unit combines the local sudden wind force events, small-scale turbulence characteristics, and the vibration and resonance phenomena, and outputs the local environmental stability index.

[0055] The safety monitoring and early warning module is used to use the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generate a safety detection report, and trigger a safety early warning mechanism.

[0056] The calculation process of the safety level includes: Input the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index into the risk assessment model to generate a quantitative risk level score, including low risk, medium risk, high risk, and extremely high risk, to obtain the safety level of the area to be detected; The risk assessment model dynamically adjusts the weights of each index according to the construction type, operation time, wind force level, and historical safety accident data.

[0057] Table 4 gives the overall performance comparison table of the system of the present invention.

[0058] Table 4 Overall performance comparison table of the system of the present invention ; Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirits of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A real-time high-precision detection method for construction safety based on environmental characteristics, characterized in that Including: Deploy a variety of sensors in the area to be detected to collect on-site environmental data in real time, including light data, worker videos, wind speed data, humidity data, air flow data, and vibration data; Analyze the light changes in the area to be detected through a light and shadow analysis model based on the light data and the worker videos, calculate the light-safety misjudgment index, and identify potential light misjudgment areas; Use a visual processing model to analyze the worker videos, identify human skeleton key points in real time, obtain the personnel behavior safety index, and detect potential dangerous behaviors; Use a local breeze and vibration detection model to detect the local wind speed, turbulence conditions, vibration, and resonance phenomena in the area to be detected in real time, and evaluate the local environmental stability index; Use the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to calculate the safety level of the area to be detected in real time, generate a safety detection report, and trigger a safety warning mechanism.

2. The real-time high-precision detection method for construction safety based on environmental features according to claim 1, characterized in that The variety of sensors include: a light sensor, a high-definition camera, a temperature and humidity sensor, a wind speed sensor, and an acceleration sensor; The light data includes natural light intensity, artificial lighting brightness, light direction, shadow distribution, and light uniformity parameters; the worker videos include high-definition video sequences of workers operating in the area to be detected; the wind speed data includes the local wind speed vector, the maximum instantaneous wind speed, and the wind direction change trend in the area to be detected; the humidity data includes the relative air humidity and the ground humidity in the area to be detected; the air flow data includes the local air flow direction, small-scale turbulence intensity, and disturbance frequency; the vibration data includes the vibration amplitude, vibration frequency, vibration acceleration, and structural resonance frequency points in the area to be detected.

3. A real-time high-precision detection method for construction safety based on environmental characteristics according to claim 1, characterized in that The light and shadow analysis model includes: a light extraction unit, a shadow recognition unit, a dynamic light change tracking unit, and a light misjudgment evaluation unit; The light extraction unit extracts the pixel brightness histogram and light gradient distribution in the image frames of the worker videos after preprocessing; the shadow recognition unit separates the dynamic light occlusion shadows and static structure occlusion shadows in the area to be detected and generates a shadow mask map; the dynamic light change tracking unit analyzes the light mutation trend of consecutive image frames based on the pixel brightness histogram, light gradient distribution, and the shadow mask map; the light misjudgment evaluation unit fuses the shadow mask map with the construction path and the operation edge area, evaluates the risk probability of visual illusion caused by light, outputs the light-safety misjudgment index of each area in the area to be detected, and marks the area where the light-safety misjudgment index is greater than the preset misjudgment threshold as the potential light misjudgment area.

4. A real-time high-precision construction safety detection method based on environmental characteristics according to claim 1, characterized in that, The visual processing model includes: a human skeleton key point recognition unit, a posture evaluation unit, and an abnormal behavior detection unit; The human body bone key point recognition unit processes the image frames of the worker video through a pose estimation model based on deep learning, extracts the coordinates of human key points, and generates a human skeleton structure diagram; the pose evaluation unit calculates the angular changes, center of gravity offsets, and movement trajectory curves between bone points according to the human skeleton structure diagram to identify the personnel pose information; the abnormal behavior detection unit compares the personnel pose information with the dangerous behavior database, combines the humidity data to detect potential dangerous behaviors, and outputs a personnel behavior safety index.

5. The real-time high-precision detection method for construction safety based on environmental features according to claim 1, characterized in that, The local breeze and vibration detection model includes: a local wind speed and turbulence situation analysis unit, a vibration and resonance phenomenon analysis unit, and a local environmental stability evaluation unit; The local wind speed and turbulence situation analysis unit identifies local sudden wind force events and small-scale turbulence characteristics according to the wind speed data and the air flow data; the vibration and resonance phenomenon analysis unit compares the vibration data with the natural frequency of the construction structure to identify vibration and resonance phenomena; the local environmental stability evaluation unit combines the local sudden wind force events, small-scale turbulence characteristics, and the vibration and resonance phenomena to output the local environmental stability index.

6. The real-time high-precision construction safety detection method based on environmental features according to claim 1, characterized in that, The calculation process of the safety level includes: Input the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index into a risk assessment model to generate a quantitative risk level score, including low risk, medium risk, high risk, and extremely high risk, to obtain the safety level of the area to be detected; the risk assessment model dynamically adjusts the weights of each index according to the construction type, working hours, wind force level, and historical safety accident data.

7. A real-time high-precision construction safety detection system based on environmental characteristics, characterized in that, Including: A data acquisition module for arranging various sensors in the area to be detected to collect on-site environmental data, including light, worker video, wind speed, humidity, air flow, and vibration; A light and shadow analysis module for analyzing the light change in the area to be detected through a light and shadow analysis model according to the light, calculating the light-safety misjudgment index, and identifying potential visual illusion areas; A visual recognition module for analyzing the worker video using a visual processing model to real-time identify the key points of the human body bones, obtain the personnel behavior safety index, and detect potential dangerous behaviors; A local breeze and vibration detection module for real-time detecting the local wind speed and turbulence situation in the area to be detected and the vibration and resonance phenomena of the platform to be detected through a local breeze and vibration detection model, and evaluating the local environmental stability index; A safety monitoring and warning module for using the light-safety misjudgment index, the personnel behavior safety index, and the local environmental stability index to real-time calculate the safety level of the area to be detected, generate a safety detection report, and trigger a safety warning mechanism.

Images