CN-122022464-A - Construction safety risk monitoring and early warning method and system based on multi-source perception

CN122022464ACN 122022464 ACN122022464 ACN 122022464ACN-122022464-A

Abstract

The invention discloses a construction safety risk monitoring and early warning method and system based on multi-source perception, which relate to the technical field of construction site safety and comprise the steps of constructing a construction site monitoring information network according to a multi-source sensor, establishing a three-dimensional construction monitoring space of the construction site, and constructing a construction safety monitoring model according to the plurality of safety characteristic parameters and the two-way long-short-term neural memory network, correcting the risk prediction result through a safety correction factor, and adjusting the corrected risk prediction result by combining the action track of constructors. According to the invention, the construction site data is collected through the distributed multi-source sensor, after the construction site data is processed, the processed construction site data is calculated through the two-way long-short-term memory network, the predicted data is corrected in real time by combining a safety correction factor mechanism, the position and working condition change of personnel are fused, a four-stage grading early warning matrix is constructed, the monitoring-predicting-correcting-responding closed loop is realized, and the risk early warning responding speed is improved.

Inventors

LIU DONGSHUANG
WANG CHAO
SU JUN
WU ZHITING
WANG JIULIN
ZHU YANGCHAO

Assignees

重庆沙坪坝交通实业有限公司

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (10)

1. The construction safety risk monitoring and early warning method based on multi-source perception is characterized by comprising the following steps of: Traversing the construction site environment and the construction flow, arranging the multi-source sensors which are related to each other, and constructing a construction site monitoring information network; The method comprises the steps of carrying out multidimensional information acquisition on constructors and construction environments according to a construction site monitoring information network, obtaining a multidimensional information data set, processing and fusing, and establishing a construction site three-dimensional construction monitoring space according to the multidimensional information data set; Traversing the three-dimensional construction monitoring space of the construction site to perform feature extraction, generating a plurality of safety feature parameters, and constructing a construction safety monitoring model based on the plurality of safety feature parameters; monitoring and calculating the site environment of constructors through a construction safety monitoring model to obtain risk prediction data; dynamically analyzing according to the multidimensional information data set, generating a safety correction factor, and dynamically feeding back risk prediction data according to the safety correction factor to generate a construction environment change trend graph; And (3) carrying out real-time monitoring on constructors according to the construction environment change trend graph and the risk prediction data, quantifying risk levels, and establishing a multi-level early warning mechanism according to the risk levels and the positions of the constructors.
2. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 1, wherein the traversing of the construction site environment and the construction flow is realized, the interrelated multi-source sensors are arranged, and a construction site monitoring information network is constructed, and the method specifically comprises the following steps: traversing the construction site environment and the construction flow to perform environment safety analysis, and determining the layout positions of the multi-source sensors; Selecting a type of a multi-source sensor according to the construction site environment and the construction flow, and arranging and correlating the multi-source sensors according to the arrangement positions to generate an equipment correlation chain; connecting a plurality of different equipment association chains, carrying out multidimensional analysis on the equipment association chains, and dividing a plurality of construction site monitoring subnetworks according to analysis results; and connecting a plurality of construction site monitoring subnetworks to each other for data correction, and constructing a construction site monitoring information network.
3. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 1, wherein the method is characterized in that multi-dimensional information acquisition is carried out on constructors and construction environments according to a construction site monitoring information network, a multi-dimensional information data set is obtained, the multi-dimensional information data set is processed and fused, and a construction site three-dimensional construction monitoring space is built according to the multi-dimensional information data set, and specifically comprises the following steps: Carrying out multidimensional analysis on constructors and construction environments according to a construction site monitoring information network, and determining types of a plurality of multidimensional sensing data; processing abnormal values and missing data in the multidimensional sensing data, and unifying data dimension and format of the multidimensional data to generate a clean data set; Mapping a plurality of construction site monitoring subnetworks according to the types of the cleaning data set and the multidimensional sensing data to obtain a plurality of monitoring space diagrams; And carrying out regression analysis on the types of the multidimensional sensing data and the clean data sets according to the plurality of monitoring space diagrams to generate a monitoring space risk data set, and constructing a three-dimensional construction monitoring space of the construction site according to real-time analysis of the monitoring space risk data set.
4. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 3, wherein the traversing the three-dimensional construction monitoring space of the construction site performs feature extraction to generate a plurality of safety feature parameters, and specifically comprises the following steps: Extracting a data acquisition time sequence according to the types of the multidimensional sensing data, traversing the three-dimensional construction monitoring space of the construction site according to the data acquisition time sequence, performing risk analysis, and confirming the construction risk types; Traversing the three-dimensional construction monitoring space of the construction site according to the data acquisition time sequence, analyzing the environmental equipment, and confirming the equipment state of the construction site; and carrying out standardized processing on the construction risk type and the data in the equipment state of the construction site to generate a plurality of safety characteristic parameters.
5. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 4, wherein the construction safety monitoring model is constructed based on a plurality of safety feature parameters, and specifically comprises the following steps: carrying out standardization processing and formatting on the safety feature parameters, and sequencing the safety feature parameters according to the data acquisition time sequence to obtain a safety feature parameter sequence table; Carrying out multidimensional fusion on the security feature parameter sequence table, the construction risk type and the site environment equipment state, and constructing a risk state association rule matrix; the method comprises the steps of calling the states of site environment equipment around constructors, and determining a training data set according to a risk state association rule matrix and site three-dimensional construction monitoring space division; introducing the behavior state of constructors, training and calibrating according to the data acquisition time sequence based on the equipment state of the construction site and the training data set, extracting time sequence association characteristics in the training data set according to the two-way long-short-term memory network, and capturing long-short-term dependency relationship; And fusing the time sequence correlation characteristics with the long-short-period dependency relationship, dynamically distributing weights according to the data acquisition time sequence, and carrying out iterative training calibration through a two-way long-short-period memory network according to the time sequence data and the training data set of the equipment state of the building site to construct a construction safety monitoring model.
6. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 5, wherein the construction safety monitoring model monitors and calculates the site environment where constructors are located to obtain risk prediction data, and the method specifically comprises the following steps: Extracting real-time coordinates of constructors according to the multidimensional information data set, and drawing moving tracks of the constructors according to the three-dimensional construction monitoring space; screening constructor state data and construction environment data of the movement track of constructors, marking the constructor state data and the construction environment data as constructor peripheral data, and carrying out standardized processing on the constructor peripheral data to generate a standardized real-time data set; And calculating the standardized real-time data set according to the construction safety monitoring model to obtain risk prediction data.
7. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 6, wherein the dynamic analysis is performed according to a multi-dimensional information data set to generate a safety correction factor, and the method specifically comprises the following steps: traversing the multidimensional information data set, obtaining constructor behavior dimension, construction environment change dimension and equipment running state dimension, which are called as three core dimensions, and performing time sequence dynamic analysis to generate constructor behavior risk factors, environment mutation factors and equipment influence factors; carrying out weighted summation calculation on constructor behavior risk factors, environment mutation factors and equipment influence factors to obtain a safety correction factor; and calculating the security risk contribution values of the three core dimensions according to the ratio of the security correction factors to the constructor behavior risk factors, the environment mutation factors and the equipment influence factors, and establishing a security correction factor updating mechanism consistent with the multi-source sensor updating sampling frequency.
8. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 7, wherein the method is characterized in that risk prediction data are dynamically fed back according to a safety correction factor to generate a construction environment change trend graph, and specifically comprises the following steps: Traversing time sequence records of the safety correction factors, the risk prediction data and the multidimensional information data set, obtaining a core object needing dynamic correction feedback, and performing time sequence dynamic analysis; Global correction is carried out on the risk level quantized value of the core object by using the security correction factors, and the security risk contribution values of the core risk type occurrence probability and the three core dimensions are corrected independently according to the corresponding dimension factors to generate corrected risk data; Extracting data acquisition time stamps, arranging according to time ascending order, dividing fixed time windows, calculating the mean value, peak value and change rate of risk data in each window, and generating a time sequence of time-corrected risk data; The method comprises the steps of determining a core visual dimension of a construction environment change trend graph, synchronously associating time sequence data of three core dimensions, performing visual conversion on the time sequence data of the three core dimensions, generating a multi-curve fusion construction environment change trend graph, and synchronously establishing a trend graph dynamic updating mechanism consistent with the update sampling frequency of a multi-source sensor.
9. The construction safety risk monitoring and early warning method based on multi-source perception according to claim 8, wherein the method is characterized in that the constructor is monitored in real time according to the construction environment change trend chart and the risk prediction data, the risk level is quantified, and a multi-stage early warning mechanism is established according to the risk level and the constructor position, and specifically comprises the following steps: traversing time sequence curve data and risk prediction data in a construction environment change trend graph, acquiring action tracks and expected tasks of constructors, and performing multidimensional time sequence association analysis; correlating the real-time position and the action track of constructors with the safety risk contribution values of three core dimensions, analyzing the influence weight of the expected task of constructors on the risk level, and pre-judging the risk change direction by combining the risk evolution slope of the construction environment change trend graph; defining a risk level threshold based on the quantization range of the corrected risk prediction data, updating sampling frequency according to a multi-source sensor, continuously collecting constructor state data and risk data, comparing the latest curve and historical data of a construction environment change trend graph, recalculating a risk level quantization value, and synchronously updating a risk level label of constructors; Cross-verifying the real-time quantified risk level and the security risk contribution values of three core dimensions in the construction environment change trend graph, and setting a threshold value; And carrying out multi-element time sequence association analysis according to the risk level quantized value, the real-time position of constructors and the risk evolution data of the construction environment change trend graph, constructing a two-dimensional early warning trigger matrix, and synchronously establishing an early warning dynamic adjustment mechanism.
10. A construction safety risk monitoring and early warning system based on multi-source perception for implementing a construction safety risk monitoring and early warning method based on multi-source perception as described in the above claims 1 to 9, which is characterized in that the system comprises: the multi-source information acquisition and data processing module is used for analyzing the construction site environment and the construction flow, laying multi-source sensors to construct a monitoring network, acquiring personnel, environment and equipment multi-type data, and outputting a high-quality clean fusion data set through cleaning, standardization and multi-source fusion processing to provide reliable data support for the subsequent modules and ensure the accuracy and consistency of the data; the monitoring space establishing and feature extracting module is used for fusing the data mapping monitoring space diagram, constructing a three-dimensional construction monitoring space through regression analysis, traversing the space to extract safety feature parameters, realizing site space visualization, defining construction risk types and equipment states, generating targeted safety feature parameters and supporting model training; The prediction model training and risk prediction module is used for training a two-way long-short-term memory network model, predicting risk data, generating safety correction factor correction data and drawing a trend chart, wherein the model has accurate risk prediction capability, the corrected data is more practical, and the trend chart intuitively presents risk dynamics, so that risk prediction efficiency is improved; The real-time monitoring quantification and risk early warning module is used for associating trend graphs with risk data, monitoring personnel states in real time, quantifying risk levels, establishing a multi-level early warning mechanism, accurately quantifying four-level risk, differentiating early warning response in time, forming closed-loop management, and effectively guaranteeing safety of constructors.

Description

Construction safety risk monitoring and early warning method and system based on multi-source perception Technical Field The invention relates to the technical field of construction site safety, in particular to a construction safety risk monitoring and early warning method and system based on multi-source perception. Background The construction safety is to build a responsibility system required by regulations for preventing casualties and property loss by technical, management and educational means, and is characterized by building a responsibility system required by regulations, implementing dynamic risk management, integrating entity protection and intelligent monitoring technologies, covering the whole process from start-up to completion and aiming at realizing the fundamental transition from post-treatment to pre-prevention. Sensor technology is gradually introduced in current building construction safety monitoring, and data of personnel positions, environment parameters (such as gas concentration and wind speed) and equipment states (such as tower crane inclination) are tried to be collected, but the data are mostly distributed in a single dimension, so that basic risk identification can be realized, but the dynamic change of construction environment and multi-risk coupling scene are difficult to adapt, and the fine safety management and control requirements of complex sites are difficult to meet. Disclosure of Invention In order to solve the technical problems, the technical scheme provides a construction safety risk monitoring and early warning method and system based on multi-source perception, and the technical scheme solves the problems that the current construction safety monitoring proposed in the background technology gradually introduces a sensor technology, attempts to collect personnel position, environmental parameters (such as gas concentration and wind speed) and equipment state (such as tower crane inclination) data, but is mainly distributed in a single dimension, basic risk recognition can be realized, dynamic change of construction environment and multi-risk coupling scene are difficult to adapt, and the refined safety management and control requirement of complex sites is difficult to meet. In order to achieve the above purpose, the invention adopts the following technical scheme: A construction safety risk monitoring and early warning method based on multi-source perception comprises the following steps: Traversing the construction site environment and the construction flow, arranging the multi-source sensors which are related to each other, and constructing a construction site monitoring information network; The method comprises the steps of carrying out multidimensional information acquisition on constructors and construction environments according to a construction site monitoring information network, obtaining a multidimensional information data set, processing and fusing, and establishing a construction site three-dimensional construction monitoring space according to the multidimensional information data set; Traversing the three-dimensional construction monitoring space of the construction site to perform feature extraction, generating a plurality of safety feature parameters, and constructing a construction safety monitoring model based on the plurality of safety feature parameters; monitoring and calculating the site environment of constructors through a construction safety monitoring model to obtain risk prediction data; dynamically analyzing according to the multidimensional information data set, generating a safety correction factor, and dynamically feeding back risk prediction data according to the safety correction factor to generate a construction environment change trend graph; And (3) carrying out real-time monitoring on constructors according to the construction environment change trend graph and the risk prediction data, quantifying risk levels, and establishing a multi-level early warning mechanism according to the risk levels and the positions of the constructors. Preferably, the traversing of the construction site environment and the construction flow lays out the multi-source sensors related to each other, and constructs a construction site monitoring information network, which specifically comprises: traversing the construction site environment and the construction flow to perform environment safety analysis, and determining the layout positions of the multi-source sensors; Selecting a type of a multi-source sensor according to the construction site environment and the construction flow, and arranging and correlating the multi-source sensors according to the arrangement positions to generate an equipment correlation chain; connecting a plurality of different equipment association chains, carrying out multidimensional analysis on the equipment association chains, and dividing a plurality of construction site monitoring subnetworks according to analysis resu