Safety angle calculation method of high and low hanging box beams based on laser

Abstract

The embodiment of the present application provides a laser-based method for calculating the safety angle of high and low hanging box girders, which realizes accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; uses Kalman filtering and confidence weights to optimize the fusion of laser ranging and overhead crane attitude data; extracts time series features based on long and short-term memory networks, combines sliding time windows and autoregressive models to predict the trend of box girder inclination changes, and realizes risk classification warning. The method realizes model adaptive updating by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capabilities during the construction of high and low hanging box girders.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method for calculating the safety angle of a high and low hanging box beam based on laser. Background Art

[0002] Safety monitoring has always been an important and challenging issue in the field of high-low hanging box girder construction. Traditional box girder hoisting monitoring mainly relies on manual observation and simple sensor data collection. This method is not only inefficient, but also easily affected by human factors and environmental conditions, making it difficult to timely detect and warn of potential safety hazards. With the development of intelligent sensing technology, although some monitoring solutions based on technologies such as laser ranging and attitude sensing have emerged, these solutions still have obvious limitations.

[0003] Existing monitoring systems often use a single data source or a simple data fusion method, which makes it difficult to fully and accurately reflect the real-time status of the box girder. At the same time, environmental factors have a great impact on laser ranging, and the existing system lacks an effective environmental compensation mechanism, resulting in unstable measurement accuracy. In terms of data processing, simple filtering algorithms are difficult to effectively handle noise interference under complex working conditions, and fixed warning thresholds cannot adapt to dynamically changing construction environments. In addition, existing systems generally lack in-depth mining and utilization of historical data, and cannot achieve continuous optimization and adaptive adjustment of models, which seriously affects the reliability and practicality of the system.

[0004] Therefore, how to effectively integrate multi-source data, establish a dynamic safety assessment mechanism, and improve the system's adaptive capabilities are key issues that need to be addressed. This is not only related to construction safety and efficiency, but also an important topic for promoting the development of intelligent construction. Summary of the invention

[0005] In response to the problems in the prior art, the present application provides a laser-based method for calculating the safety angle of high and low hanging box girders, which can effectively improve the safety monitoring level and early warning capability during the construction of high and low hanging box girders.

[0006] In order to solve at least one of the above problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a laser-based method for calculating the safety angle of a high and low hanging box beam, comprising:

[0008] Construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, the environmental parameter data includes temperature data, humidity data, atmospheric pressure data, and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data, and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0009] A dynamic safety assessment model is constructed based on the real-time state parameters of the box girder, and the time series characteristics of the box girder height data and the overhead crane motion data are respectively extracted using a long short-term memory network. A sliding time window is constructed to calculate the local fluctuation index of the time series characteristics. An autoregressive model is constructed in combination with the local fluctuation index to predict the trend of the box girder inclination angle change. The risk level is calculated based on the trend of the box girder inclination angle change, and a graded alarm signal is triggered when the risk level exceeds a preset threshold.

[0010] The box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data are stored in an optimization database, and the long short-term memory network and the autoregressive model are trained based on the historical data in the optimization database, and the mapping relationship model is updated.

[0011] Furthermore, the adaptive environmental compensation mechanism is established based on the environmental parameter data, the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data, and the light intensity data are extracted, a mapping relationship model between the time-varying characteristics and the ranging error is established, and the environmental compensation coefficient is calculated according to the mapping relationship model, including:

[0012] The temperature data, the humidity data, the atmospheric pressure data, and the light intensity data are respectively subjected to wavelet transformation to extract time-varying features, and statistical features such as mean, standard deviation, kurtosis, and skewness are calculated for each type of data using a sliding time window, and the calculated statistical features are combined to form a feature vector, and the feature vector is mapped to the interval [0,1] using normalization processing to obtain the environmental time-varying features;

[0013] A BP neural network including an input layer, a hidden layer and an output layer is constructed, the environmental time-varying characteristics are input into the input layer of the BP neural network, a mapping relationship between the environmental time-varying characteristics and the laser ranging error is established through the nonlinear transformation of the hidden layer, and an environmental compensation coefficient is determined based on the ranging error value calculated by the output layer, and the environmental compensation coefficient is used to correct the laser ranging data.

[0014] Furthermore, the Kalman filter algorithm is used to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, the confidence weights of the laser ranging data and the overhead crane attitude data are calculated based on the covariance matrix, and the laser ranging data is corrected in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, including:

[0015] Construct a Kalman filter state space model, input the laser ranging data and the crane attitude data as observation quantities into the state space model, calculate the prior state estimate value through the state prediction equation, calculate the new information sequence based on the observation equation, update the state estimate value in combination with the Kalman gain matrix, and obtain the ranging data and attitude data after noise suppression;

[0016] The confidence weights of the ranging data and the attitude data are calculated based on the prediction error covariance matrix in the Kalman filtering process, the confidence weights are linearly combined with the environmental compensation coefficient to obtain a comprehensive correction coefficient, and the comprehensive correction coefficient is used to correct the ranging data after noise suppression to generate box girder height data reflecting the actual height of the box girder.

[0017] Further, the step of correcting the overhead crane attitude data to obtain overhead crane motion data and fusing the box girder height data with the overhead crane motion data to obtain the box girder real-time state parameters includes:

[0018] According to the acceleration, angular velocity and angle information in the attitude data of the overhead crane, the quaternion attitude solution method is used to calculate the motion attitude matrix of the overhead crane, and the attitude matrix is ​​corrected in combination with the confidence weight. The three-dimensional motion trajectory data and motion parameters of the overhead crane are obtained through coordinate transformation, and the motion data of the overhead crane reflecting the actual motion state of the overhead crane are generated;

[0019] A multi-source data fusion model is designed to align the box girder height data and the overhead crane motion data by timestamp, and the fusion weight coefficient is calculated by weighted average method. The aligned data are weighted summed using the weight coefficient to obtain the real-time status parameters of the box girder including the box girder height information and the overhead crane motion information.

[0020] Furthermore, the dynamic safety assessment model is constructed based on the real-time state parameters of the box girder, the time series characteristics of the box girder height data and the overhead crane motion data are respectively extracted using a long short-term memory network, and a sliding time window is constructed to calculate the local fluctuation index of the time series characteristics, including:

[0021] Constructing a long short-term memory network structure, including an input layer, an LSTM layer and a fully connected layer, dividing the box girder height data and the overhead crane motion data into training sequences according to a preset time step and inputting them into the long short-term memory network, extracting the long-term dependency and short-term change features in the data through the synergistic effect of the forget gate, the input gate and the output gate, and outputting the time series feature vectors of the box girder height and the overhead crane motion in the fully connected layer;

[0022] A sliding time window of fixed length is set, and the root mean square value and variance of the time series feature vector are calculated within the time window. The calculation result is used as the characteristic fluctuation intensity index, and the characteristic fluctuation intensity index is smoothed by the exponential weighted average method to obtain a local fluctuation index reflecting the change of the box girder state.

[0023] Furthermore, the autoregressive model is constructed in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, the risk level is calculated based on the trend of the box girder inclination angle change, and a graded alarm signal is triggered when the risk level exceeds a preset threshold, including:

[0024] An autoregressive model is constructed, local fluctuation indicators are input into the model in time series, the autoregressive coefficient is estimated by the least square method, the future predicted value of the box girder inclination angle is calculated based on the autoregressive coefficient and the historical data sequence, and the predicted value is smoothed by a first-order exponential smoothing method to obtain a changing trend curve of the box girder inclination angle;

[0025] The fuzzy comprehensive evaluation method is used to establish risk assessment rules. The slope and acceleration of the box girder inclination trend are used as evaluation factors. The risk score of each evaluation factor is calculated by the membership function. The weighted sum of each score is taken based on the risk evaluation matrix to obtain the comprehensive risk level. The corresponding level of alarm signal is output according to the comparison result between the risk level and the preset threshold.

[0026] Furthermore, the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data are stored in an optimization database, and the long short-term memory network and the autoregressive model are trained based on the historical data in the optimization database, and the mapping relationship model is updated, including:

[0027] The box girder height data, overhead crane motion data, time series characteristics, local fluctuation index, box girder inclination change trend and risk level data are stored in the optimized database according to the preset data table structure, and a time index and association relationship are established for the stored data. The sliding window mechanism is used to regularly clean up expired data, and the valid sample data in the recent period is retained as the training set;

[0028] Based on the historical data in the training set, the long short-term memory network is incrementally trained, the network weight parameters are updated, and the coefficients of the autoregressive model are re-estimated. At the same time, the BP neural network is retrained according to the historical correspondence between environmental parameters and ranging errors, and the parameters of the environmental compensation mapping model are updated to achieve online optimization and update of the model.

[0029] In a second aspect, the present application provides a laser-based high and low hanging box beam safety angle calculation device, comprising:

[0030] A multi-source fusion module is used to construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, wherein the environmental parameter data includes temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0031] A risk warning module is used to construct a dynamic safety assessment model based on the real-time state parameters of the box girder, use a long short-term memory network to respectively extract the time series characteristics of the box girder height data and the overhead crane motion data, construct a sliding time window to calculate the local fluctuation index of the time series characteristics, and construct an autoregressive model in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, calculate the risk level based on the trend of the box girder inclination angle change, and trigger a graded alarm signal when the risk level exceeds a preset threshold;

[0032] An iterative optimization module is used to store the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into an optimization database, train the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and update the mapping relationship model.

[0033] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the laser-based method for calculating the safety angle of high and low hanging box beams are implemented.

[0034] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the laser-based high and low hanging box beam safety angle calculation method.

[0035] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the laser-based high and low hanging box beam safety angle calculation method.

[0036] It can be seen from the above technical solutions that the present application provides a laser-based method for calculating the safety angle of high and low hanging box girders, which realizes accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; Kalman filtering and confidence weights are used to optimize the fusion of laser ranging and overhead crane attitude data; based on the long short-term memory network, time series features are extracted, and the sliding time window and autoregressive model are combined to predict the trend of box girder inclination angle changes, so as to realize risk classification warning. This method realizes model adaptive updating by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capability during the construction of high and low hanging box girders. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 This is one of the flow charts of the laser-based high and low hanging box beam safety angle calculation method in the embodiment of the present application;

[0039] Figure 2This is a second flow chart of a method for calculating the safety angle of a high and low hanging box beam based on laser in an embodiment of the present application;

[0040] Figure 3 This is a third flow chart of a method for calculating the safety angle of a high and low hanging box beam based on laser in an embodiment of the present application;

[0041] Figure 4 This is a fourth flow chart of a method for calculating the safety angle of a high and low hanging box beam based on laser in an embodiment of the present application;

[0042] Figure 5 This is a fifth flow chart of a method for calculating the safety angle of a high and low hanging box beam based on laser in an embodiment of the present application;

[0043] Figure 6 This is a sixth flow chart of a method for calculating the safety angle of a high and low hanging box beam based on laser in an embodiment of the present application;

[0044] Figure 7 This is the seventh flow chart of the method for calculating the safety angle of the high and low hanging box beams based on laser in the embodiment of the present application;

[0045] Figure 8 This is a structural diagram of a laser-based high and low hanging box beam safety angle calculation device in an embodiment of the present application;

[0046] Fig. 9 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application.

[0047] Reference numerals:

[0048] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0051] Taking into account the problems existing in the prior art, the present application provides a laser-based method for calculating the safety angle of high and low hanging box girders, which realizes accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; Kalman filtering and confidence weights are used to optimize the fusion of laser ranging and overhead crane attitude data; based on the long short-term memory network, time series features are extracted, and the sliding time window and autoregressive model are combined to predict the trend of box girder inclination angle changes, so as to realize risk classification warning. The method realizes model adaptive update by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capability during the construction of high and low hanging box girders.

[0052] In order to effectively improve the safety monitoring level and early warning capability during the construction of high and low hanging box beams, the present application provides an embodiment of a method for calculating the safety angle of high and low hanging box beams based on laser, see Figure 1 The laser-based high and low hanging box beam safety angle calculation method specifically includes the following contents:

[0053] Step S101: construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, the environmental parameter data including temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0054] Optionally, this embodiment proposes a multi-source data fusion solution suitable for high-low hanging box girder construction scenarios, which integrates laser ranging, overhead crane attitude and environmental parameter data to achieve accurate monitoring and status evaluation of the box girder installation process. This solution specifically considers the impact of complex and changeable environmental factors on the measurement accuracy of the construction site.

[0055] This embodiment first constructs a multi-source monitoring data fusion model, deploys a laser ranging sensor array in the box girder hoisting operation area, and installs a posture sensor on the overhead crane to collect motion state data. Environmental monitoring nodes are set up at key locations on the construction site to collect environmental parameters such as temperature, humidity, atmospheric pressure, and light intensity in real time. This multi-dimensional data collection solution ensures all-round monitoring of the box girder installation process.

[0056] In view of the influence of construction environment factors on the accuracy of laser ranging, this embodiment innovatively designs an adaptive environmental compensation mechanism. The time-varying features of environmental parameter data are extracted through wavelet transform to capture the dynamic characteristics of environmental factors such as temperature fluctuations and humidity changes. For example, in hot and rainy weather in summer, the laser ranging results are often significantly affected by temperature gradients and atmospheric refraction. At this time, these influencing factors can be effectively evaluated by extracting the time-varying features of the environment.

[0057] This embodiment establishes a mapping relationship model between environmental time-varying characteristics and ranging errors, and uses a BP neural network structure to achieve adaptive mapping from environmental characteristics to error compensation. The model can automatically calculate the compensation coefficient according to the current environmental conditions and perform real-time correction on the ranging results.

[0058] In terms of data processing, this embodiment applies the Kalman filter algorithm to suppress noise in laser ranging data and overhead crane attitude data. The Kalman filter effectively suppresses measurement noise while maintaining the ability to respond quickly to changes in the box girder state through two stages: prediction and update. Based on the covariance matrix in the filtering process, the confidence weights of different data sources are calculated to achieve dynamic evaluation of data reliability.

[0059] This embodiment combines the environmental compensation coefficient with the data confidence weight to construct a comprehensive correction mechanism. For example, when a laser ranging point is detected to be disturbed by strong sunlight, the system will automatically reduce the weight of the measuring point and increase the environmental compensation strength to ensure the accuracy of the measurement results. In this way, the corrected box girder height data and overhead crane motion data are closer to the actual situation.

[0060] In the data fusion stage, this embodiment adopts a weighted fusion strategy to align and comprehensively process the corrected box girder height data and the overhead crane motion data in time and space, and generate comprehensive parameters reflecting the real-time status of the box girder. These parameters include key indicators such as the spatial position and attitude angle of the box girder, providing a reliable data basis for subsequent safety assessment.

[0061] This embodiment effectively solves the problem of unstable accuracy of traditional single measurement methods in complex construction environments through multi-source data fusion and environmental adaptive compensation. This solution shows good environmental adaptability and measurement stability during the actual box girder hoisting process, providing reliable technical support for the safety monitoring of the box girder installation process.

[0062] By comprehensively using technologies such as data fusion, environmental compensation and noise suppression, this embodiment significantly improves the accuracy and reliability of box girder status monitoring, and provides a more accurate decision-making basis for safety management of the construction site.

[0063] Step S102: constructing a dynamic safety assessment model based on the real-time state parameters of the box girder, using a long short-term memory network to respectively extract the time series characteristics of the box girder height data and the overhead crane motion data, constructing a sliding time window to calculate the local fluctuation index of the time series characteristics, and constructing an autoregressive model in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, and calculating the risk level based on the trend of the box girder inclination angle change, and triggering a graded alarm signal when the risk level exceeds a preset threshold;

[0064] Optionally, this embodiment constructs a dynamic safety assessment model based on the real-time state parameters of the box girder, which provides real-time warning for safety risks during the box girder hoisting process. In actual application scenarios, the box girder may experience unstable states such as swinging and tilting during the hoisting process. Timely identification and warning of these states are crucial to construction safety.

[0065] This embodiment uses a long short-term memory network (LSTM) to process the box girder height data and the overhead crane motion data respectively. The special structural design of the LSTM network includes an input gate, a forget gate, and an output gate, which enables it to effectively capture the long-term dependencies of the box girder state changes. For example, during the box girder hoisting process, even if there is a slow tilt trend, the LSTM can promptly detect this potential dangerous trend through long-term state memory.

[0066] In order to extract time series features, this embodiment designs a dynamic sliding time window mechanism. The window length is adjusted according to the actual rhythm of the box girder hoisting construction, and is usually set to 3-5 minutes. In each time window, the statistical features such as the root mean square value and variance of the box girder height and the overhead crane movement are calculated. These features can reflect the local fluctuation of the box girder state.

[0067] Based on the local fluctuation index, this embodiment constructs an autoregressive model to predict the changing trend of the box girder inclination angle. The model takes into account the historical change law of the box girder inclination angle, estimates the autoregressive coefficient through the least squares method, and realizes the prediction of future inclination angle changes. For example, when it is detected that the box girder begins to have a small periodic swing, the model can predict whether this swing will gradually expand.

[0068] In terms of risk level assessment, this embodiment adopts a fuzzy comprehensive evaluation method, taking the slope and acceleration of the box girder inclination change trend as the main evaluation factors. By establishing a risk membership function, different degrees of inclination changes are mapped to corresponding risk levels. For example, when the box girder inclination change rate exceeds the safety threshold, the system will trigger different levels of alarm signals according to the risk level.

[0069] The alarm mechanism of this embodiment is divided into multiple levels, including early warning, general warning and emergency warning, etc. Different levels of alarms correspond to different handling measures. For example, in the early warning state, the operator is prompted to pay attention and observe, and the emergency warning requires the lifting operation to be stopped immediately and safety measures to be taken.

[0070] In practical applications, this embodiment can effectively identify abnormal conditions during the box girder hoisting process. For example, when the box girder swings due to wind, the model can predict possible dangerous situations by analyzing the changing trends of the swing amplitude and frequency, and reserve sufficient response time for the construction personnel.

[0071] This embodiment solves the problems of delayed warning and frequent false alarms in traditional box girder hoisting monitoring by combining LSTM networks, sliding time windows, and autoregressive prediction. This solution can detect abnormal changes in the box girder status in a timely manner, accurately predict potential risks, and significantly improve the safety and controllability of the box girder hoisting process.

[0072] By establishing a dynamic safety assessment system, this embodiment realizes full-process monitoring and intelligent early warning of the box girder hoisting process, provides reliable decision-making support for construction site managers, and effectively reduces the risk of safety accidents during the box girder hoisting process.

[0073] Step S103: storing the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into an optimization database, training the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and updating the mapping relationship model.

[0074] Optionally, this embodiment constructs a self-optimizing data management solution for the box girder hoisting construction process, and realizes continuous optimization and performance improvement of various prediction models by establishing an optimized database storage and utilizing multi-dimensional monitoring data.

[0075] This embodiment uses a distributed database architecture to store box girder height data and overhead crane motion data, and implements fast data retrieval and association analysis through timestamp indexing. During data storage, a multi-level cache mechanism is designed to hierarchically store real-time collected data to ensure fast response capabilities for high-frequency access data.

[0076] For the storage of time series features and local fluctuation indexes, this embodiment designs a special feature vector storage structure to associate and store the time series features extracted by the LSTM network with the local fluctuation indexes calculated by the sliding time window. This storage structure facilitates the rapid construction of training sample sets during subsequent model training and improves the efficiency of model optimization.

[0077] This embodiment uses time series compression storage technology to store the trend data of box girder inclination change, reduces storage space occupancy through differential encoding, and retains the integrity of trend data. Risk level data is stored using an event trigger mechanism to record the key time points of risk level changes and corresponding trigger conditions.

[0078] In terms of data quality management, this embodiment establishes a data reliability assessment mechanism, and by setting data validity rules, high-quality historical data is screened out for model training. For example, when abnormal fluctuations in sensor data are detected during a certain lifting process, the relevant data will be marked as low reliability and given a lower weight during model training.

[0079] For the training optimization of LSTM network, this embodiment adopts incremental learning strategy and regularly updates the model with new high-quality data. During the training process, the optimal network hyperparameters are selected through cross-validation method to improve the prediction accuracy of the model for the state change of the box girder.

[0080] In the process of optimizing the autoregressive model, this embodiment establishes multiple sub-models based on historical data under different working conditions, which are respectively suitable for different construction environments and operating conditions. Through the weight adaptive mechanism, the dynamic fusion of model prediction results is achieved to improve the robustness of the prediction.

[0081] The environmental compensation mapping relationship model is updated using online learning, which continuously optimizes the mapping relationship between environmental parameters and ranging errors based on the environmental impact data accumulated during the actual construction process. When new environmental factors affect the model, the model can adaptively adjust the compensation strategy.

[0082] This embodiment achieves continuous improvement of model performance by establishing a closed-loop data optimization mechanism. For example, during a box girder hoisting process, if the early warning system successfully predicts and prevents a possible safety accident, the relevant status data and early warning process will be used first for subsequent model optimization.

[0083] By optimizing the construction and use of the database, this embodiment solves the problem that the model performance in the traditional box girder hoisting monitoring system is difficult to continuously improve. This solution realizes the adaptive optimization of the monitoring model and significantly improves the prediction accuracy and engineering adaptability of the system.

[0084] This embodiment establishes a continuously evolving box girder hoisting safety monitoring system through a data-driven model optimization strategy, provides more reliable technical support for construction process safety management, and effectively improves the intelligence level of box girder hoisting construction.

[0085] From the above description, it can be seen that the laser-based high and low hanging box girder safety angle calculation method provided in the embodiment of the present application can realize accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; uses Kalman filtering and confidence weights to optimize the fusion of laser ranging and overhead crane attitude data; extracts time series features based on long and short-term memory networks, combines sliding time windows and autoregressive models to predict the trend of box girder inclination changes, and realizes risk classification warning. The method realizes model adaptive updates by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capabilities during the construction of high and low hanging box girders.

[0086] In one embodiment of the laser-based high and low hanging box beam safety angle calculation method of the present application, see Figure 2 , and can also include the following:

[0087] Step S201: extracting time-varying features from the temperature data, the humidity data, the atmospheric pressure data, and the light intensity data respectively through wavelet transform, calculating statistical features such as mean, standard deviation, kurtosis, and skewness for each type of data using a sliding time window, combining the calculated statistical features to form a feature vector, and using normalization processing to map the feature vector to the [0,1] interval to obtain environmental time-varying features;

[0088] Step S202: construct a BP neural network including an input layer, a hidden layer and an output layer, input the environmental time-varying characteristics into the input layer of the BP neural network, establish a mapping relationship between the environmental time-varying characteristics and the laser ranging error through the nonlinear transformation of the hidden layer, determine the environmental compensation coefficient based on the ranging error value calculated by the output layer, and the environmental compensation coefficient is used to correct the laser ranging data.

[0089] Optionally, this embodiment proposes a method for extracting time-varying features of environmental parameters and compensating errors. Aiming at the problem of the influence of environmental factors on the laser ranging accuracy during the box girder lifting process, an adaptive environmental compensation mechanism is established through wavelet transform and neural network.

[0090] This embodiment first extracts time-varying features from environmental data such as temperature, humidity, atmospheric pressure, and light intensity. Wavelet transform is used to perform multi-scale decomposition of environmental data, and Daubechies wavelet basis functions are selected to obtain feature information of different frequency bands through three-layer wavelet decomposition. This processing method can effectively capture the mutation characteristics and periodic change characteristics of environmental parameters.

[0091] Regarding the setting of the sliding time window, this embodiment adopts a window length of 5 minutes and a window sliding step of 1 minute according to the time scale characteristics of the environmental parameter changes. In each time window, the statistical characteristics such as the mean, standard deviation, kurtosis and skewness of the environmental parameters are calculated. These statistical characteristics can fully reflect the distribution characteristics and change laws of the environmental parameters.

[0092] This embodiment innovatively designs a feature vector combination scheme, which weights and combines the statistical features of each environmental parameter according to their influence on the ranging error. For example, under strong light conditions in summer, the weight of light intensity will be increased accordingly to better compensate for the influence of light on ranging.

[0093] The normalization of the feature vector adopts the improved min-max normalization method, taking into account the seasonal variation characteristics of environmental parameters, and dynamically adjusting the upper and lower bounds of the normalization interval. This processing method ensures the comparability of environmental time-varying characteristics under different seasonal conditions.

[0094] In terms of BP neural network construction, this embodiment designs a three-layer network structure. The number of nodes in the input layer matches the dimension of the time-varying characteristics of the environment. The hidden layer adopts a two-layer structure with 12 and 8 nodes respectively. The output layer outputs the ranging error value as a single node. The hidden layer uses the tanh activation function, which provides good nonlinear mapping capabilities.

[0095] For the training process of the neural network, this embodiment adopts an improved back propagation algorithm, introduces momentum terms and an adaptive learning rate mechanism, and improves the convergence speed and stability of the model. The training data set contains environmental parameters and corresponding ranging error samples under different weather conditions to ensure that the model has good generalization ability.

[0096] The calculation of the environmental compensation coefficient adopts a nonlinear mapping method based on the ranging error. When the predicted ranging error is large, the compensation coefficient will increase accordingly, but an upper limit is set to avoid over-compensation. For example, in rainy conditions, the system will automatically adjust the compensation coefficient according to the humidity change to ensure ranging accuracy.

[0097] This embodiment solves the problem of poor adaptability of traditional distance measurement systems to environmental changes by extracting time-varying environmental features and mapping them using neural networks. This solution can evaluate the impact of environmental factors on distance measurement accuracy in real time and make corresponding compensation adjustments, significantly improving the reliability of box girder height measurement.

[0098] By establishing an environmental adaptive compensation mechanism, this embodiment achieves high-precision ranging capabilities around the clock and in all weather conditions, provides stable and reliable position data support for the box girder hoisting process, and effectively improves the safety and efficiency of the construction process.

[0099] In one embodiment of the laser-based high and low hanging box beam safety angle calculation method of the present application, see Figure 3 , and can also include the following:

[0100] Step S301: construct a Kalman filter state space model, input the laser ranging data and the crane attitude data as observation quantities into the state space model, calculate the prior state estimate value through the state prediction equation, calculate the new information sequence based on the observation equation, update the state estimate value in combination with the Kalman gain matrix, and obtain the ranging data and attitude data after noise suppression;

[0101] Step S302: Calculate the confidence weights of the ranging data and the attitude data based on the prediction error covariance matrix in the Kalman filtering process, linearly combine the confidence weights with the environmental compensation coefficient to obtain a comprehensive correction coefficient, and use the comprehensive correction coefficient to correct the ranging data after noise suppression to generate box girder height data reflecting the actual height of the box girder.

[0102] Optionally, this embodiment proposes a box girder height measurement data processing method based on Kalman filtering, which achieves high-precision correction of measurement data by constructing a state space model and introducing a comprehensive correction mechanism.

[0103] This embodiment first constructs a Kalman filter state space model for the box girder hoisting scenario. The state vector contains the height, velocity and acceleration information of the box girder, and the observation vector contains the laser ranging data and the crane attitude data. The state transfer matrix is ​​designed based on the kinematic characteristics of the box girder, taking into account the acceleration change characteristics during the hoisting process.

[0104] In the state prediction process, this embodiment adopts an improved prediction equation and introduces an adaptive noise covariance estimation mechanism. When the box girder is in a stable hoisting state, the process noise covariance is small, which improves the smoothness of the state prediction; when a sudden disturbance occurs, the process noise covariance automatically increases, improving the model's ability to respond to sudden changes.

[0105] This embodiment designs an observation equation based on multi-source data fusion to optimize the combination of laser ranging data and overhead crane attitude data. The design of the observation matrix takes into account the measurement characteristics of the two types of data, and improves the reliability of the observation data through weighted fusion. For example, when the overhead crane tilts, the weight of the attitude data will increase accordingly.

[0106] The calculation of the new information sequence adopts the residual analysis method, which evaluates the reliability of the measurement data in real time by comparing the difference between the observed value and the predicted value. When the new information sequence is abnormal, the system will automatically adjust the Kalman gain matrix to reduce the impact of the abnormal data.

[0107] This embodiment innovatively designs an adaptive adjustment strategy for the Kalman gain matrix. In the early stage of hoisting, due to inaccurate state estimation, the gain matrix is ​​large and more dependent on observation data; as the filtering process proceeds, the gain matrix gradually decreases and the weight of the state estimation increases accordingly.

[0108] The calculation of confidence weight is based on the eigenvalue analysis of the prediction error covariance matrix. This embodiment adopts the normalization processing method of the matrix diagonal elements to convert the prediction errors of different dimensions into comparable confidence indicators. Smaller prediction errors correspond to higher confidence weights.

[0109] In the construction of the comprehensive correction coefficient, this embodiment designs an adaptive weight allocation mechanism. The combination ratio of the environmental compensation coefficient and the confidence weight will be dynamically adjusted according to the actual working conditions. For example, under severe weather conditions, the weight of the environmental compensation coefficient will be increased accordingly.

[0110] This embodiment solves the problem of insufficient data reliability in traditional box girder height measurement by dual processing of noise suppression and data correction. This solution can effectively suppress measurement noise while taking into account the influence of environmental factors, significantly improving the accuracy of box girder height data.

[0111] By establishing a closed-loop data processing mechanism, this embodiment achieves real-time and accurate compensation of box girder height measurement, providing reliable data support for box girder hoisting and positioning. This solution has shown good anti-interference ability and environmental adaptability in practical applications, effectively improving the accuracy and safety of box girder hoisting construction.

[0112] In one embodiment of the laser-based high and low hanging box beam safety angle calculation method of the present application, see Figure 4 , and can also include the following:

[0113] Step S401: according to the acceleration, angular velocity and angle information in the attitude data of the overhead crane, the quaternion attitude solution method is used to calculate the motion attitude matrix of the overhead crane, the attitude matrix is ​​corrected in combination with the confidence weight, and the three-dimensional motion trajectory data and motion parameters of the overhead crane are obtained through coordinate transformation, so as to generate the motion data of the overhead crane reflecting the actual motion state of the overhead crane;

[0114] Step S402: Design a multi-source data fusion model, align the box girder height data and the overhead crane motion data by timestamp, calculate the fusion weight coefficient using the weighted average method, perform a weighted sum operation on the aligned data using the weight coefficient, and obtain the real-time status parameters of the box girder including the box girder height information and the overhead crane motion information.

[0115] Optionally, this embodiment proposes a box girder state parameter calculation method based on quaternion attitude solution and multi-source data fusion, and realizes accurate monitoring of the box girder hoisting process through overhead crane motion state analysis and data fusion optimization.

[0116] In the process of solving the attitude of the overhead crane, this embodiment adopts the quaternion representation method to avoid the universal joint deadlock problem in the Euler angle representation. By constructing the quaternion differential equation, the measurement data of the accelerometer and the gyroscope are converted into attitude changes, thereby realizing the continuous tracking of the spatial attitude of the overhead crane.

[0117] In the calculation of the attitude matrix, this embodiment designs a prediction update mechanism based on Kalman filtering. The prediction process uses angular velocity information to infer attitude changes, and the update process combines acceleration information for correction. When the sky car is in a low-speed uniform motion state, the weight of the gyroscope data is higher; when acceleration and deceleration occur, the weight of the accelerometer data increases accordingly.

[0118] This embodiment innovatively introduces a confidence correction mechanism for the attitude matrix. Based on the confidence weights calculated in the previous steps, each component of the attitude matrix is ​​adaptively adjusted. For example, when severe vibration is detected, the system will reduce the credibility of the attitude data in the corresponding period to avoid vibration interference affecting the accuracy of attitude solution.

[0119] In the coordinate transformation process, this embodiment establishes a multi-level coordinate system transformation chain. Starting from the local coordinate system of the overhead crane, after the transformation of the intermediate coordinate system, the motion parameters in the global coordinate system are finally obtained. This hierarchical transformation method is convenient for handling the nonlinear characteristics of the overhead crane during motion.

[0120] In order to extract the three-dimensional motion trajectory, this embodiment designs a trajectory smoothing algorithm. By setting an adaptive sliding window, the original trajectory data is locally weighted averaged, which not only retains the motion trend information but also suppresses the influence of high-frequency jitter.

[0121] This embodiment adopts an improved timestamp alignment method in the multi-source data fusion link. Considering that the sampling frequencies of the box girder height data and the overhead crane motion data may be different, a dynamic interpolation algorithm is designed to ensure accurate alignment of the data in the time dimension.

[0122] The calculation of the fusion weight coefficient adopts an adaptive weighting scheme. This embodiment dynamically adjusts the weight of each data source by analyzing the time-varying characteristics of the data. When the reliability of a data source decreases, the system will automatically reduce its weight contribution in the fusion process.

[0123] In the process of generating the real-time state parameters of the box girder, this embodiment comprehensively considers the coupling relationship between the change in the height of the box girder and the movement of the overhead crane. For example, when the overhead crane is accelerating, the system will focus on the swing response of the box girder; when the overhead crane is moving at a constant speed, it will pay more attention to the stability monitoring of the height of the box girder.

[0124] This embodiment solves the problem of incomplete data and unstable accuracy in traditional box girder status monitoring through the collaborative processing of posture solution and data fusion. This solution realizes high-precision monitoring of the position and posture of the box girder during the hoisting process, providing a reliable decision-making basis for hoisting control.

[0125] By establishing an efficient data processing and fusion mechanism, this embodiment significantly improves the calculation accuracy and reliability of the box girder state parameters. This solution shows excellent real-time performance and stability in practical applications, and effectively supports the intelligent control and safety management of the box girder hoisting process.

[0126] In one embodiment of the laser-based high and low hanging box beam safety angle calculation method of the present application, see Figure 5 , and can also include the following:

[0127] Step S501: constructing a long short-term memory network structure, including an input layer, an LSTM layer and a fully connected layer, dividing the box girder height data and the overhead crane motion data into training sequences according to a preset time step and inputting them into the long short-term memory network, extracting the long-term dependency and short-term change features in the data through the synergistic effect of the forget gate, the input gate and the output gate, and outputting the time series feature vectors of the box girder height and the overhead crane motion in the fully connected layer;

[0128] Step S502: Set a sliding time window of fixed length, calculate the root mean square value and variance of the time series feature vector within the time window, use the calculation result as the characteristic fluctuation intensity index, smooth the characteristic fluctuation intensity index through the exponential weighted average method, and obtain a local fluctuation index reflecting the change of the box girder state.

[0129] Optionally, this embodiment proposes a box girder state time series feature extraction method based on a long short-term memory network, which realizes accurate monitoring of box girder state changes by combining deep learning and statistical analysis.

[0130] This embodiment designs an LSTM network structure for the box girder hoisting scenario, and the input layer receives the time series of the box girder height data and the overhead crane motion data. The time step is set based on the box girder motion characteristics, and 30 seconds is selected as the basic unit, which can retain sufficient historical information and respond to state changes in a timely manner.

[0131] In the design of the LSTM layer, this embodiment adopts a two-layer structure, each layer contains 128 neurons. The first layer of LSTM mainly processes short-term dependencies and captures the instantaneous change characteristics of the box girder and the overhead crane; the second layer of LSTM focuses on long-term dependencies and extracts the periodic pattern and trend characteristics of the box girder movement.

[0132] This embodiment introduces an adaptive threshold mechanism in the design of the forget gate. When a sudden change in the box girder state is detected, the forget gate will reduce the weight of historical information, allowing the network to respond to new state changes more quickly. For example, during the box girder start-up and braking process, the system will pay more attention to the current state information.

[0133] The control strategy of the input gate adopts a dynamic weight allocation method. This embodiment automatically adjusts the importance of new input information according to the time-varying characteristics of the data. When the box girder is in a stable operation state, the input gate tends to maintain a stable information flow; when abnormal fluctuations occur, the input gate will increase the weight of the new information.

[0134] In the design of the output gate, this embodiment implements a selective information output mechanism. By analyzing the changing characteristics of the cell state, the system can identify and output characteristic information closely related to the box beam state and filter out irrelevant interference information.

[0135] The fully connected layer adopts a multi-layer structure, and maps the high-dimensional features output by the LSTM layer to a suitable feature space through dimensionality reduction processing. In this embodiment, a dropout mechanism is added to the fully connected layer to improve the generalization ability of the model and avoid overfitting problems.

[0136] In the processing of the time series feature vector, this embodiment innovatively designs a multi-scale analysis method. By setting observation windows of different lengths, the system can simultaneously capture the rapid changes and slow evolution characteristics of the box girder state.

[0137] In the calculation process of the characteristic fluctuation intensity index, this embodiment adopts an adaptive window mechanism. The window length is dynamically adjusted according to the motion state of the box girder. When there is a drastic change, the window length is shortened to increase the response speed, and in the stable stage, the window length is extended to enhance the anti-interference ability.

[0138] The exponential weighted average smoothing process uses a dynamic attenuation factor. This embodiment analyzes the time-varying characteristics of feature fluctuations and adaptively adjusts the size of the attenuation factor. When the fluctuation is severe, a smaller attenuation factor is used to retain more real-time features; when the fluctuation is gentle, a larger attenuation factor is used to enhance the smoothing effect.

[0139] This embodiment solves the problem of insufficient extraction of time series features in traditional box girder state monitoring methods by combining deep learning and statistical analysis. This solution can effectively identify subtle changes in the box girder state and provide a reliable basis for early warning of abnormal conditions.

[0140] By establishing an efficient feature extraction and analysis mechanism, this embodiment realizes real-time and accurate monitoring of the box girder status. In practical applications, this solution demonstrates excellent time series feature capture and fluctuation warning capabilities, effectively improving the safety and controllability of the box girder hoisting process.

[0141] In one embodiment of the laser-based high and low hanging box beam safety angle calculation method of the present application, see Figure 6 , and can also include the following:

[0142] Step S601: construct an autoregressive model, input the local fluctuation index into the model in time series, estimate the autoregressive coefficient by using the least square method, calculate the future predicted value of the box girder inclination angle based on the autoregressive coefficient and the historical data sequence, smooth the predicted value by a first-order exponential smoothing method, and obtain a changing trend curve of the box girder inclination angle;

[0143] Step S602: A fuzzy comprehensive evaluation method is used to establish risk assessment rules, and the slope and acceleration of the box girder inclination angle change trend are used as evaluation factors. The risk score of each evaluation factor is calculated through a membership function, and the weighted sum of each score is performed based on the risk evaluation matrix to obtain a comprehensive risk level. According to the comparison result between the risk level and the preset threshold, an alarm signal of the corresponding level is output.

[0144] Optionally, this embodiment proposes a box girder inclination angle prediction and risk assessment method based on an autoregressive model and fuzzy comprehensive evaluation, which realizes intelligent early warning of the box girder hoisting process by combining trend analysis and risk quantification.

[0145] In the construction of the autoregressive model, this embodiment adopts the sliding time window method to select the optimal order. By analyzing the autocorrelation characteristics of the local volatility index, the dynamic range of the model order is determined, which not only ensures the timeliness of the prediction, but also avoids the overfitting problem.

[0146] In the process of estimating the autoregressive coefficient, this embodiment introduces the weighted least square method to assign different weights to the data at different time points. The weight of recent data is higher, and the weight of long-term data gradually decays. This processing method can better reflect the latest change trend of the box girder inclination angle.

[0147] This embodiment innovatively designs an adaptive smoothing mechanism for the predicted value. By analyzing the fluctuation characteristics of the predicted sequence, the exponential smoothing coefficient is dynamically adjusted. When the predicted value fluctuates violently, the smoothing coefficient is reduced to retain important trend information; when the predicted sequence is relatively stable, the smoothing coefficient is increased to improve the smoothness of the curve.

[0148] In the process of generating the trend curve, this embodiment adopts a segmented fitting strategy. The system automatically identifies the key time points of the box girder inclination change, and adopts different fitting parameters at different stages, thereby improving the accuracy of trend prediction. For example, in the box girder starting and braking stages, a shorter prediction step and faster response parameters are selected.

[0149] In the design of risk assessment rules, this embodiment constructs a multi-level fuzzy evaluation system. The slope index reflects the speed of change of the box girder inclination, and the acceleration index reflects the severity of the change. The combination of the two can fully characterize the unstable state of the box girder.

[0150] The design of the membership function adopts a trapezoidal and Gaussian mixture form. According to the safety requirements of box girder hoisting, this embodiment divides the risk level into multiple levels, and each level corresponds to a different membership function. Through the sophisticated function design, a smooth transition of the risk state is achieved.

[0151] In the construction of the risk evaluation matrix, this embodiment introduces expert experience knowledge. Based on a large number of hoisting case analyses, the weight relationship between the evaluation factors is established. When the box girder is in different movement stages, the system will automatically adjust the weight ratio of each factor.

[0152] This embodiment designs a multi-level alarm mechanism. Different risk levels correspond to different forms of alarm signals, from prompts, warnings to emergency alarms, and the intensity of the alarm is gradually increased. The system will dynamically adjust the trigger threshold of the alarm signal according to the changing trend of the risk level.

[0153] By establishing a closed-loop mechanism for prediction and evaluation, this embodiment solves the problem of untimely risk warning and inaccurate evaluation in traditional box girder hoisting. This solution can identify potential safety hazards in advance and reserve sufficient response time for operators.

[0154] This embodiment achieves full monitoring and risk warning of the box girder hoisting process through intelligent prediction and precise evaluation. This solution has demonstrated excellent prediction accuracy and risk identification capabilities in practical applications, effectively improving the safety and controllability of box girder hoisting operations.

[0155] In one embodiment of the laser-based high and low hanging box beam safety angle calculation method of the present application, see Figure 7 , and can also include the following:

[0156] Step S701: storing the box girder height data, overhead crane motion data, time series characteristics, local fluctuation index, box girder inclination change trend and risk level data into the optimization database according to the preset data table structure, establishing a time index and association relationship for the stored data, using a sliding window mechanism to regularly clean up expired data, and retaining valid sample data in the most recent period as a training set;

[0157] Step S702: Incrementally train the long short-term memory network based on the historical data in the training set, update the network weight parameters, re-estimate the coefficients of the autoregressive model, and retrain the BP neural network according to the historical correspondence between the environmental parameters and the ranging error, update the parameters of the environmental compensation mapping model, and realize online optimization and update of the model.

[0158] Optionally, this embodiment proposes a model adaptive updating method based on an optimized database, which realizes continuous optimization of the box girder monitoring system by combining data management and incremental learning.

[0159] This embodiment adopts a hierarchical storage architecture in database design, and stores different types of data in corresponding data tables. Box girder height data and overhead crane motion data are used as the basic data layer, time series characteristics and local fluctuation indicators constitute the feature data layer, and box girder inclination trend and risk level form the decision data layer.

[0160] In the process of building the data index, this embodiment designs a multidimensional index structure. The primary index is established based on the timestamp, and the auxiliary index is created based on the data type and the association relationship. This index mechanism significantly improves the efficiency of data retrieval and supports fast multi-condition query operations.

[0161] This embodiment innovatively introduces a data association management mechanism. By establishing a mapping relationship between data, association analysis of data at different levels is achieved. For example, the original monitoring data corresponding to a risk warning event can be quickly traced back, which is convenient for problem analysis and experience summary.

[0162] In terms of data cleaning, this embodiment adopts a dual evaluation mechanism based on time and importance. The basic data retention period is determined through a sliding window, while retaining historical data with special significance, such as abnormal event records and key status data.

[0163] This embodiment realizes dynamic evaluation of data quality during the training set construction process. By analyzing the integrity, consistency and validity of the data, high-quality sample data is screened out. In the case of missing or abnormal data, the system will use interpolation or smoothing to ensure the reliability of the training data.

[0164] In the incremental training of the LSTM network, this embodiment designs a progressive learning strategy. The newly added data is first trained on a small batch, and the training scope is expanded after the model performance is verified to be improved. This method not only ensures the stability of the model, but also can absorb new data features in a timely manner.

[0165] The parameters of the autoregressive model are updated using a sliding re-estimation method. This embodiment determines whether to accept parameter updates by comparing the performance of the new and old models on the validation set. When the new model can significantly improve the prediction effect, the system will adopt the new parameter configuration.

[0166] In the retraining process of the BP neural network, this embodiment realizes the adaptive modeling of environmental factors. By analyzing the dynamic relationship between environmental parameters and ranging errors, the mapping ability of the compensation model is continuously optimized. For example, under different weather conditions, the system can accurately compensate for the impact of environmental factors on ranging accuracy.

[0167] This embodiment solves the problem that traditional fixed models are difficult to adapt to complex and changing environments through online learning and model optimization. This solution can continuously absorb new monitoring experience and continuously improve the system's prediction accuracy and risk identification capabilities.

[0168] By establishing a complete data management and model update mechanism, this embodiment realizes the continuous evolution of the box girder monitoring system. This solution has demonstrated strong environmental adaptability and performance optimization potential in practical applications, effectively improving the intelligence level and reliability of the box girder hoisting process.

[0169] In order to effectively improve the safety monitoring level and early warning capability during the construction of high and low hanging box girders, the present application provides an embodiment of a laser-based high and low hanging box girder safety angle calculation device for realizing all or part of the contents of the laser-based high and low hanging box girder safety angle calculation method, see Figure 8 The laser-based high and low hanging box beam safety angle calculation device specifically includes the following contents:

[0170] The multi-source fusion module 10 is used to construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, wherein the environmental parameter data includes temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0171] The risk warning module 20 is used to construct a dynamic safety assessment model based on the real-time state parameters of the box girder, use a long short-term memory network to respectively extract the time series characteristics of the box girder height data and the overhead crane motion data, construct a sliding time window to calculate the local fluctuation index of the time series characteristics, and construct an autoregressive model in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, calculate the risk level based on the trend of the box girder inclination angle change, and trigger a graded alarm signal when the risk level exceeds a preset threshold;

[0172] The iterative optimization module 30 is used to store the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into the optimization database, train the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and update the mapping relationship model.

[0173] From the above description, it can be seen that the laser-based high and low hanging box girder safety angle calculation device provided in the embodiment of the present application can realize accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; uses Kalman filtering and confidence weights to optimize the fusion of laser ranging and overhead crane attitude data; extracts time series features based on long and short-term memory networks, combines sliding time windows and autoregressive models to predict the trend of box girder inclination changes, and realizes risk classification warning. The method realizes model adaptive updating by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capabilities during the construction of high and low hanging box girders.

[0174] From the hardware level, in order to effectively improve the safety monitoring level and early warning capability during the construction of high and low hanging box girders, the present application provides an embodiment of an electronic device for implementing all or part of the contents of the laser-based high and low hanging box girder safety angle calculation method, and the electronic device specifically includes the following contents:

[0175] Processor, memory, communication interface and bus; wherein the processor, memory and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between the laser-based high and low hanging box beam safety angle calculation device and the core business system, user terminal and related database and other related equipment; the logic controller can be a desktop computer, a tablet computer and a mobile terminal, etc., but the present embodiment is not limited thereto. In the present embodiment, the logic controller can be implemented with reference to the embodiment of the laser-based high and low hanging box beam safety angle calculation method and the laser-based high and low hanging box beam safety angle calculation device in the embodiment, and the contents thereof are incorporated herein and the repeated parts are not repeated.

[0176] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0177] In practical applications, part of the laser-based high and low hanging box beam safety angle calculation method can be executed on the electronic device side as described above, or all operations can be completed in the client device. The specific selection can be based on the processing power of the client device and the limitations of the user's usage scenario. This application does not limit this. If all operations are completed in the client device, the client device may also include a processor.

[0178] The client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and other implementation scenarios may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0179] Fig. 9 FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Fig. 9As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Fig. 9 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0180] In one embodiment, the laser-based high and low hanging box beam safety angle calculation method function can be integrated into the central processor 9100. The central processor 9100 can be configured to perform the following control:

[0181] Step S101: construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, the environmental parameter data including temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0182] Step S102: constructing a dynamic safety assessment model based on the real-time state parameters of the box girder, using a long short-term memory network to respectively extract the time series characteristics of the box girder height data and the overhead crane motion data, constructing a sliding time window to calculate the local fluctuation index of the time series characteristics, and constructing an autoregressive model in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, and calculating the risk level based on the trend of the box girder inclination angle change, and triggering a graded alarm signal when the risk level exceeds a preset threshold;

[0183] Step S103: storing the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into an optimization database, training the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and updating the mapping relationship model.

[0184] From the above description, it can be seen that the electronic device provided in the embodiment of the present application realizes accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; uses Kalman filtering and confidence weights to optimize the fusion of laser ranging and overhead crane attitude data; extracts time series features based on long short-term memory networks, combines sliding time windows and autoregressive models to predict the trend of box girder inclination changes, and realizes risk classification warning. The method realizes model adaptive updating by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capabilities during the construction of high and low hanging box girders.

[0185] In another embodiment, the laser-based high and low hanging box beam safety angle calculation device can be configured separately from the central processing unit 9100. For example, the laser-based high and low hanging box beam safety angle calculation device can be configured as a chip connected to the central processing unit 9100, and the function of the laser-based high and low hanging box beam safety angle calculation method can be realized through the control of the central processing unit.

[0186] like Fig. 9 As shown, the electronic device 9600 may also include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Fig. 9 In addition, the electronic device 9600 may also include Fig. 9 For components not shown, reference may be made to the prior art.

[0187] like Fig. 9 As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0188] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0189] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0190] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0191] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0192] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0193] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0194] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps in the method for calculating the safety angle of a high and low hanging box beam based on laser, where the execution subject is a server or a client in the above embodiments. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, all the steps in the method for calculating the safety angle of a high and low hanging box beam based on laser, where the execution subject is a server or a client in the above embodiments are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0195] Step S101: construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, the environmental parameter data including temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0196] Step S102: constructing a dynamic safety assessment model based on the real-time state parameters of the box girder, using a long short-term memory network to respectively extract the time series characteristics of the box girder height data and the overhead crane motion data, constructing a sliding time window to calculate the local fluctuation index of the time series characteristics, and constructing an autoregressive model in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, and calculating the risk level based on the trend of the box girder inclination angle change, and triggering a graded alarm signal when the risk level exceeds a preset threshold;

[0197] Step S103: storing the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into an optimization database, training the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and updating the mapping relationship model.

[0198] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application realizes accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; uses Kalman filtering and confidence weights to optimize the fusion of laser ranging and overhead crane attitude data; extracts time series features based on long and short-term memory networks, combines sliding time windows and autoregressive models to predict the trend of box girder inclination changes, and realizes risk classification warning. The method realizes model adaptive updating by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capabilities during the construction of high and low hanging box girders.

[0199] The embodiments of the present application also provide a computer program product capable of implementing all the steps in the laser-based high and low hanging box beam safety angle calculation method in the above-mentioned embodiment, where the execution subject is a server or a client. When the computer program / instruction is executed by a processor, the steps of the laser-based high and low hanging box beam safety angle calculation method are implemented. For example, the computer program / instruction implements the following steps:

[0200] Step S101: construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, the environmental parameter data including temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder;

[0201] Step S102: constructing a dynamic safety assessment model based on the real-time state parameters of the box girder, using a long short-term memory network to respectively extract the time series characteristics of the box girder height data and the overhead crane motion data, constructing a sliding time window to calculate the local fluctuation index of the time series characteristics, and constructing an autoregressive model in combination with the local fluctuation index to predict the trend of the box girder inclination angle change, and calculating the risk level based on the trend of the box girder inclination angle change, and triggering a graded alarm signal when the risk level exceeds a preset threshold;

[0202] Step S103: storing the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into an optimization database, training the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and updating the mapping relationship model.

[0203] From the above description, it can be seen that the computer program product provided by the embodiment of the present application realizes accurate monitoring and risk warning of the box girder hoisting process through multi-source data fusion and dynamic safety assessment. The method establishes a mapping relationship between environmental parameters and ranging errors, and adopts an adaptive environmental compensation mechanism to improve measurement accuracy; uses Kalman filtering and confidence weights to optimize the fusion of laser ranging and overhead crane attitude data; extracts time series features based on long short-term memory networks, combines sliding time windows and autoregressive models to predict the trend of box girder inclination changes, and realizes risk classification warning. The method realizes model adaptive updating by continuously optimizing the database, which effectively improves the safety monitoring level and early warning capabilities during the construction of high and low hanging box girders.

[0204] It should be understood by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0205] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0206] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0207] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0208] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A laser-based method for calculating the safety angle of high and low hanging box beams, characterized in that: The method comprises: Construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, the environmental parameter data includes temperature data, humidity data, atmospheric pressure data, and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data, and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder; Construct a long short-term memory network structure, including an input layer, an LSTM layer and a fully connected layer, divide the box girder height data and the overhead crane motion data into training sequences according to a preset time step and input them into the long short-term memory network, extract the long-term dependency and short-term change characteristics in the data through the synergistic effect of the forget gate, the input gate and the output gate, output the time series feature vector of the box girder height and the overhead crane motion in the fully connected layer, set a sliding time window of fixed length, calculate the root mean square value and variance of the time series feature vector within the time window, use the calculation result as the characteristic fluctuation intensity index, smooth the characteristic fluctuation intensity index by an exponential weighted average method, obtain a local fluctuation index reflecting the change of the box girder state, construct an autoregressive model, input the local fluctuation index into the model in time series, estimate the autoregressive coefficient by the least squares method, calculate the future predicted value of the box girder inclination angle based on the autoregressive coefficient and the historical data sequence, smooth the predicted value by a first exponential smoothing method, obtain a curve of the box girder inclination angle change trend, calculate the risk level based on the box girder inclination angle change trend, and trigger a graded alarm signal when the risk level exceeds a preset threshold; The box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data are stored in an optimization database, and the long short-term memory network and the autoregressive model are trained based on the historical data in the optimization database, and the mapping relationship model is updated.

2. The laser-based high and low hanging box beam safety angle calculation method according to claim 1 is characterized in that: The method of establishing an adaptive environmental compensation mechanism based on the environmental parameter data, extracting the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data, and the light intensity data, establishing a mapping relationship model between the time-varying characteristics and the ranging error, and calculating the environmental compensation coefficient according to the mapping relationship model includes: The temperature data, the humidity data, the atmospheric pressure data, and the light intensity data are respectively subjected to wavelet transformation to extract time-varying features, and statistical features such as mean, standard deviation, kurtosis, and skewness are calculated for each type of data using a sliding time window, and the calculated statistical features are combined to form a feature vector, and the feature vector is mapped to the interval [0,1] using normalization processing to obtain the environmental time-varying features; A BP neural network including an input layer, a hidden layer and an output layer is constructed, the environmental time-varying characteristics are input into the input layer of the BP neural network, a mapping relationship between the environmental time-varying characteristics and the laser ranging error is established through the nonlinear transformation of the hidden layer, and an environmental compensation coefficient is determined based on the ranging error value calculated by the output layer, and the environmental compensation coefficient is used to correct the laser ranging data.

3. The laser-based high and low hanging box beam safety angle calculation method according to claim 1 is characterized in that: The method comprises: using a Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculating the confidence weights of the laser ranging data and the overhead crane attitude data based on a covariance matrix, and correcting the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, including: Construct a Kalman filter state space model, input the laser ranging data and the crane attitude data as observation quantities into the state space model, calculate the prior state estimate value through the state prediction equation, calculate the new information sequence based on the observation equation, update the state estimate value in combination with the Kalman gain matrix, and obtain the ranging data and attitude data after noise suppression; The confidence weights of the ranging data and the attitude data are calculated based on the prediction error covariance matrix in the Kalman filtering process, the confidence weights are linearly combined with the environmental compensation coefficient to obtain a comprehensive correction coefficient, and the comprehensive correction coefficient is used to correct the ranging data after noise suppression to generate box girder height data reflecting the actual height of the box girder.

4. The laser-based high and low hanging box beam safety angle calculation method according to claim 1 is characterized in that: The step of correcting the crane attitude data to obtain crane motion data and fusing the box girder height data with the crane motion data to obtain the box girder real-time state parameters includes: According to the acceleration, angular velocity and angle information in the attitude data of the overhead crane, the quaternion attitude solution method is used to calculate the motion attitude matrix of the overhead crane, and the attitude matrix is ​​corrected in combination with the confidence weight. The three-dimensional motion trajectory data and motion parameters of the overhead crane are obtained through coordinate transformation, and the motion data of the overhead crane reflecting the actual motion state of the overhead crane are generated; A multi-source data fusion model is designed to align the box girder height data and the overhead crane motion data by timestamp, and the fusion weight coefficient is calculated by weighted average method. The aligned data are weighted summed using the weight coefficient to obtain the real-time status parameters of the box girder including the box girder height information and the overhead crane motion information.

5. The laser-based high and low hanging box beam safety angle calculation method according to claim 1 is characterized in that: When the risk level exceeds a preset threshold, triggering a graded alarm signal includes: The fuzzy comprehensive evaluation method is used to establish risk assessment rules. The slope and acceleration of the box girder inclination trend are used as evaluation factors. The risk score of each evaluation factor is calculated by the membership function. The weighted sum of each score is taken based on the risk evaluation matrix to obtain the comprehensive risk level. The corresponding level of alarm signal is output according to the comparison result between the risk level and the preset threshold.

6. The laser-based high and low hanging box beam safety angle calculation method according to claim 1 is characterized in that: The step of storing the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data in an optimization database, training the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and updating the mapping relationship model includes: The box girder height data, overhead crane motion data, time series characteristics, local fluctuation index, box girder inclination change trend and risk level data are stored in the optimized database according to the preset data table structure, and a time index and association relationship are established for the stored data. The sliding window mechanism is used to regularly clean up expired data, and the valid sample data in the recent period is retained as the training set; Based on the historical data in the training set, the long short-term memory network is incrementally trained, the network weight parameters are updated, and the coefficients of the autoregressive model are re-estimated. At the same time, the BP neural network is retrained according to the historical correspondence between environmental parameters and ranging errors, and the parameters of the environmental compensation mapping model are updated to achieve online optimization and update of the model.

7. A laser-based high and low hanging box beam safety angle calculation device, characterized in that: The device comprises: A multi-source fusion module is used to construct a multi-source monitoring data fusion model, collect laser ranging data, overhead crane attitude data and environmental parameter data, wherein the environmental parameter data includes temperature data, humidity data, atmospheric pressure data and light intensity data; establish an adaptive environmental compensation mechanism based on the environmental parameter data, extract the time-varying characteristics of the temperature data, the humidity data, the atmospheric pressure data and the light intensity data, establish a mapping relationship model between the time-varying characteristics and the ranging error, and calculate the environmental compensation coefficient according to the mapping relationship model; use the Kalman filter algorithm to suppress noise and estimate the state of the laser ranging data and the overhead crane attitude data, calculate the confidence weights of the laser ranging data and the overhead crane attitude data based on the covariance matrix, correct the laser ranging data in combination with the confidence weight and the environmental compensation coefficient to obtain the box girder height data, correct the overhead crane attitude data to obtain the overhead crane motion data, and fuse the box girder height data with the overhead crane motion data to obtain the real-time state parameters of the box girder; The risk warning module is used to construct a long short-term memory network structure, which includes an input layer, an LSTM layer and a fully connected layer. The box girder height data and the overhead crane motion data are divided into training sequences according to a preset time step and input into the long short-term memory network. The long-term dependency and short-term change characteristics in the data are extracted through the synergy of the forget gate, the input gate and the output gate. The time series feature vectors of the box girder height and the overhead crane motion are output in the fully connected layer, and a sliding time window of a fixed length is set. The root mean square value and variance of the time series feature vector are calculated within the time window, and the calculation results are used as feature fluctuations. Strength index, smoothing the characteristic fluctuation strength index by exponential weighted average method to obtain a local fluctuation index reflecting the change of the box girder state, constructing an autoregressive model, inputting the local fluctuation index into the model in time series, estimating the autoregressive coefficient by least square method, calculating the future predicted value of the box girder inclination angle based on the autoregressive coefficient and the historical data sequence, smoothing the predicted value by a single exponential smoothing method to obtain a curve of the box girder inclination angle change trend, calculating the risk level based on the box girder inclination angle change trend, and triggering a graded alarm signal when the risk level exceeds a preset threshold; An iterative optimization module is used to store the box girder height data, the overhead crane motion data, the time series characteristics, the local fluctuation index, the box girder inclination change trend, and the risk level data into an optimization database, train the long short-term memory network and the autoregressive model based on the historical data in the optimization database, and update the mapping relationship model.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the laser-based high and low hanging box beam safety angle calculation method described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the laser-based high and low hanging box beam safety angle calculation method described in any one of claims 1 to 6 are implemented.

Images