Bridge cantilever assembly construction monitoring method, device and equipment and storage medium

Abstract

The invention discloses a bridge cantilever assembly construction monitoring method, device and equipment and a storage medium, and relates to the technical field of construction monitoring, and the method comprises the steps: carrying out the time synchronization of multi-stage monitoring equipment in a corresponding construction region of bridge cantilever assembly construction through a satellite time service method, real-time monitoring data of the construction area is obtained through multi-stage monitoring equipment; extracting point cloud data and image data in the real-time monitoring data by using a first preset neural network to obtain spatial features, and extracting force data and environment data in the real-time monitoring data by using a second preset neural network to obtain time sequence features; and carrying out weighted fusion on the spatial features and the time sequence features, and carrying out real-time monitoring on the construction state of the construction area based on the fused features. Therefore, more comprehensive construction monitoring data can be obtained through multi-stage monitoring equipment, and weighted fusion of spatial features and time sequence features is performed in combination with a double-branch feature extraction network structure, so that the data accuracy is improved, and the construction monitoring effect can be guaranteed.

Description

Technical Field

[0001] This invention relates to the field of construction monitoring technology, and in particular to a method, device, equipment and storage medium for monitoring bridge cantilever assembly construction. Background Technology

[0002] For monitoring and early warning of bridge cantilever assembly in harsh environments such as high-altitude canyons, traditional industry solutions mainly revolve around a combination of "manual inspection + single-point sensor monitoring + static finite element analysis." Specifically, this includes: on-site monitoring primarily through manual inspection, static data acquisition from single-point sensors, offline analysis based on finite element software, and experience-based risk warning mechanisms. However, this approach has significant limitations in bridge cantilever assembly in such harsh environments, failing to meet the demands for dynamic monitoring and real-time early warning. The main drawbacks are insufficient monitoring accuracy and coverage; severe terrain obstruction in canyons (such as steep mountains and dense fog); difficulty for manual inspections to reach critical areas such as the top of the steel beams and the bottom of the supports, resulting in monitoring blind spots; single-point sensors can only cover localized areas, failing to achieve full-area dynamic monitoring; and other issues such as difficulty in capturing dynamic coupling relationships, delayed risk warnings and low decision-making efficiency, weak anti-interference capabilities, and poor adaptability.

[0003] Therefore, improving the accuracy of monitoring during bridge cantilever assembly construction is a problem that needs to be solved in this field. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, device, equipment, and storage medium for monitoring bridge cantilever assembly construction. Through multi-level monitoring equipment, more comprehensive construction monitoring data can be obtained. Furthermore, by combining a dual-branch feature extraction network structure for weighted fusion of spatial and temporal features, data accuracy is improved, ensuring the effectiveness of construction monitoring. The specific solution is as follows:

[0005] Firstly, this application provides a method for monitoring bridge cantilever assembly construction, including:

[0006] The multi-level monitoring equipment in the construction area corresponding to the bridge cantilever assembly construction is synchronized with time using satellite time synchronization method, and real-time monitoring data of the construction area is obtained through the time-synchronized multi-level monitoring equipment; the multi-level monitoring equipment includes lidar, cameras and mechanical sensors.

[0007] The first preset neural network is used to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data to obtain the corresponding spatial features, and the second preset neural network is used to perform feature extraction processing on the force data in the real-time monitoring data and the environmental data corresponding to the construction area to obtain the corresponding temporal features.

[0008] The spatial features and temporal features are weighted and fused, and the construction status of the construction area is monitored in real time based on the fused features.

[0009] Optionally, acquiring real-time monitoring data of the construction area through time-synchronized multi-level monitoring equipment includes:

[0010] Real-time monitoring data of the construction area is obtained through multi-level monitoring equipment synchronized with time and a preset data acquisition frequency; the preset data acquisition frequency includes the data acquisition frequency corresponding to the lidar, the camera and the force sensor respectively;

[0011] The preset data acquisition frequency includes a first data acquisition frequency when the construction area matches preset normal construction conditions, and a second data acquisition frequency when the construction area matches preset abnormal construction conditions.

[0012] Optionally, after acquiring real-time monitoring data of the construction area through time-synchronized multi-level monitoring equipment, the method further includes:

[0013] The image data in the real-time monitoring data is denoised using a preset convolutional neural network to obtain the corresponding denoised image data; the preset convolutional neural network is a network built based on an attention mechanism to repair pixel values ​​through residual connections.

[0014] Accordingly, the feature extraction processing of the point cloud data and image data in the real-time monitoring data using the first preset neural network includes:

[0015] The first preset neural network is used to perform feature extraction processing on the point cloud data and the noise-reduced image data in the real-time monitoring data.

[0016] Optionally, after acquiring real-time monitoring data of the construction area through time-synchronized multi-level monitoring equipment, the method further includes:

[0017] The image data in the real-time monitoring data is decomposed into illumination and reflection components using a color constancy model, and then processed using adaptive gamma correction technology to obtain processed image data.

[0018] Accordingly, the feature extraction processing of the point cloud data and image data in the real-time monitoring data using the first preset neural network includes:

[0019] The first preset neural network is used to perform feature extraction processing on the point cloud data and the processed image data in the real-time monitoring data.

[0020] Optionally, the weighted fusion of the spatial features and the temporal features includes:

[0021] Causal correlation calculations are performed on the spatial features and temporal features based on a preset attention weight matrix to obtain the causal correlation degree between the relevant features; the preset attention weight matrix is ​​an attention weight matrix constructed based on the construction status and corresponding causal information related to the bridge cantilever assembly construction.

[0022] Based on the current weather conditions corresponding to the construction area, the current data weights corresponding to the current weather conditions are determined from a preset weight mapping table; the current data weights include the data weights corresponding to the point cloud data and the image data.

[0023] The spatial features and the temporal features are weighted and fused by combining the current data weights and the causal correlation degree.

[0024] Optionally, the real-time monitoring of the construction status of the construction area based on the fused features includes:

[0025] The images of the connecting parts of the bridge cantilever assembly construction-related beam segments corresponding to the fused features are monitored in real time by a preset target detection algorithm in order to determine whether the connection status of the corresponding connecting parts meets the preset stability conditions.

[0026] Based on preset sensor data thresholds, the stress state of the bridge cantilever assembly construction-related beam segments corresponding to the fused features is monitored in real time to determine whether the stress state of the corresponding beam segments meets the preset stress conditions.

[0027] Optionally, after real-time monitoring of the construction status of the construction area based on the fused features, the method further includes:

[0028] Based on digital twin technology, a virtual scene corresponding to the construction area is constructed using data related to the construction status, so that relevant personnel can view the construction status corresponding to the cantilever assembly construction of the bridge through the virtual scene.

[0029] Secondly, this application provides a bridge cantilever assembly construction monitoring device, comprising:

[0030] The multi-level monitoring module is used to synchronize the time of multi-level monitoring equipment in the construction area corresponding to the bridge cantilever assembly construction via satellite time synchronization, and to obtain real-time monitoring data of the construction area through the time-synchronized multi-level monitoring equipment; the multi-level monitoring equipment includes lidar, cameras, and mechanical sensors.

[0031] The feature processing module is used to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data using a first preset neural network to obtain corresponding spatial features, and to perform feature extraction processing on the force data in the real-time monitoring data and the environmental data corresponding to the construction area using a second preset neural network to obtain corresponding temporal features.

[0032] The construction status monitoring module is used to perform weighted fusion of the spatial features and the temporal features, and to monitor the construction status of the construction area in real time based on the fused features.

[0033] Thirdly, this application provides an electronic device, comprising:

[0034] Memory, used to store computer programs;

[0035] A processor is used to execute the computer program to implement the bridge cantilever assembly construction monitoring method described above.

[0036] Fourthly, this application provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the bridge cantilever assembly construction monitoring method described above.

[0037] Therefore, this application can synchronize the time of multi-level monitoring equipment in the construction area corresponding to the cantilever assembly construction of bridges using satellite timing methods, and obtain real-time monitoring data of the construction area through the time-synchronized multi-level monitoring equipment. The multi-level monitoring equipment includes lidar, cameras, and mechanical sensors. Then, a first preset neural network is used to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data to obtain corresponding spatial features, and a second preset neural network is used to perform feature extraction processing on the force data and environmental data corresponding to the construction area in the real-time monitoring data to obtain corresponding temporal features. Subsequently, the spatial features and the temporal features are weighted and fused, and the construction status of the construction area is monitored in real time based on the fused features. In this way, this application can set up multi-level monitoring equipment in the construction area corresponding to the cantilever assembly construction of bridges, and combine lidar, cameras, and mechanical sensors to monitor and collect more comprehensive monitoring data in the construction area. Moreover, it can combine a dual-branch feature extraction network structure to perform weighted fusion of spatial features and temporal features, thereby improving data accuracy and ensuring the effectiveness of subsequent real-time monitoring of the construction status. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0039] Figure 1 This application discloses a flowchart of a bridge cantilever assembly construction monitoring method.

[0040] Figure 2 This application discloses an architecture diagram of a bridge cantilever assembly construction monitoring system.

[0041] Figure 3 This application discloses a flowchart of a specific bridge cantilever assembly construction monitoring method.

[0042] Figure 4 This is a schematic diagram of a bridge cantilever assembly construction monitoring device disclosed in this application;

[0043] Figure 5 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] See Figure 1 As shown in the figure, an embodiment of the present invention discloses a method for monitoring bridge cantilever assembly construction, including:

[0046] Step S11: Synchronize the time of multi-level monitoring equipment in the construction area corresponding to the bridge cantilever assembly construction using satellite time synchronization method, and obtain real-time monitoring data of the construction area through the time-synchronized multi-level monitoring equipment; the multi-level monitoring equipment includes lidar, cameras and mechanical sensors.

[0047] In this embodiment, to monitor the construction status of the corresponding construction area for bridge cantilever assembly, the monitoring objects may include dynamic changes such as steel beam displacement, support settlement, slope cracks, connector status, and beam segment stress; the monitoring needs to adapt to the real-time evolution of the structural morphology during construction. Specifically, the monitoring equipment can employ a three-dimensional monitoring architecture, utilizing lidar, cameras, and mechanical sensors. Specifically, lidar can be deployed at higher positions on both sides of the bridge cantilever assembly area, such as on the mountainsides, to perform a comprehensive scan of the cantilever assembly area and acquire overall 3D point cloud data of the bridge. Furthermore, for cold environments, temperature-adaptive, cold-resistant lidar can be selected. Additionally, cameras can be deployed around the cantilever assembly supports to capture visual features such as support cracks and slope collapses through optical imaging principles. For cold environments, heating modules can be installed within the cameras, such as automatically activating defrosting functions when temperatures are low. Further, for monitoring methods related to mechanical sensors, fiber optic sensors can be deployed at key bridge sections (such as the box girder section at the mid-support point and the guide beam connection nodes) to measure stress changes. Pressure sensors can also be installed in the hydraulic lines of the cantilever assembly jacks to collect cantilever assembly force data in real time. For cold environments, cold-resistant shielded cables can be used as sensor cables, with insulated sleeves embedded to prevent freezing and cracking.

[0048] Furthermore, when monitoring the construction area corresponding to the cantilever assembly construction of the bridge using multi-level monitoring equipment, time synchronization is required. Specifically, this can be achieved by using satellite time synchronization to provide unified time synchronization for lidar, cameras, and sensors. Then, the data of the construction area is monitored through the time-synchronized multi-level monitoring equipment to obtain the corresponding real-time monitoring data.

[0049] In one specific embodiment, acquiring real-time monitoring data of the construction area through time-synchronized multi-level monitoring devices may include: acquiring real-time monitoring data of the construction area through time-synchronized multi-level monitoring devices and a preset data acquisition frequency; the preset data acquisition frequency includes data acquisition frequencies corresponding to the lidar, the camera, and the force sensor respectively; wherein, the preset data acquisition frequency includes a first data acquisition frequency when the construction area matches preset normal construction conditions, and a second data acquisition frequency when the construction area matches preset abnormal construction conditions. Specifically, to balance monitoring costs and monitoring effectiveness, different data acquisition frequencies can be used under different construction conditions. For example, under normal conditions, the lidar can acquire point cloud data once per second, the camera can acquire one image every 30 seconds, and the sensor can acquire one set of data every 100 milliseconds. When abnormal construction conditions occur, a high-frequency acquisition mode can be triggered to ensure the integrity and accuracy of the monitoring data.

[0050] Step S12: Use a first preset neural network to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data to obtain corresponding spatial features, and use a second preset neural network to perform feature extraction processing on the force data in the real-time monitoring data and the environmental data corresponding to the construction area to obtain corresponding temporal features.

[0051] In this embodiment, real-time monitoring data of the construction area can be acquired through multi-level monitoring equipment via the above steps. Then, the real-time monitoring data can be analyzed. This data specifically includes point cloud data corresponding to the lidar, image data corresponding to the camera, and force data corresponding to the mechanical sensors. Specifically, feature extraction can be performed on the point cloud data and image data using a first preset neural network to obtain corresponding spatial features. Spatial features such as steel beam deformation, support settlement, and slope cracks can be extracted using convolutional layers. Correspondingly, the force data from the sensors can be processed using a second preset neural network to extract temporal features, capture the dynamic change trend during cantilever assembly, and combine it with environmental data corresponding to the construction area for feature stitching to obtain corresponding temporal features. The environmental data here can include specific values ​​of environmental factors such as temperature, humidity, and wind speed. Furthermore, in a specific embodiment, the monitoring data specifically refers to multimodal engineering data collected through lidar (point cloud data), cameras (image data), mechanical sensors (force data), and environmental sensors (temperature, humidity, wind speed, etc.). A dual-branch feature extraction method is used to extract the engineering attributes of different data. Specifically, spatial features of the point cloud data and image data are extracted, focusing on the geometric deformation of the bridge structure (such as steel beam displacement and support cracks); temporal features of the force data and environmental data are extracted, focusing on the mechanical response and environmental impact (such as stress changes and the effects of sudden temperature drops). This dual-branch extraction method yields features that better meet actual engineering needs.

[0052] In one specific embodiment, after acquiring real-time monitoring data of the construction area through multi-level monitoring devices synchronized with time, the process may further include: performing noise reduction processing on the image data in the real-time monitoring data using a preset convolutional neural network to obtain corresponding noise-reduced image data; the preset convolutional neural network is a network built based on an attention mechanism for repairing pixel values ​​through residual connections; correspondingly, the step of using a first preset neural network to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data may include: using a first preset neural network to perform feature extraction processing on the point cloud data and the noise-reduced image data in the real-time monitoring data. Specifically, after obtaining the real-time monitoring data of the construction area, to improve the accuracy of the data, noise reduction processing can be performed on the image data therein. A convolutional neural network built based on an attention mechanism for repairing pixel values ​​through residual connections can be used to process the image data to obtain corresponding noise-reduced image data. Correspondingly, after performing noise reduction processing on the image data in the real-time monitoring data, feature extraction can be performed on the noise-reduced image data to improve the accuracy of the data.

[0053] Furthermore, in another specific embodiment, after acquiring real-time monitoring data of the construction area through multi-level monitoring equipment synchronized with time, the process may further include: decomposing the image data in the real-time monitoring data into illumination and reflection components using a color constancy model, and processing the illumination and reflection components using adaptive gamma correction technology to obtain processed image data; correspondingly, the step of using a first preset neural network to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data includes: using a first preset neural network to perform feature extraction processing on the point cloud data and the processed image data in the real-time monitoring data. Specifically, to improve data accuracy, image enhancement can be performed on the image data. This involves decomposing the illumination and reflection components of the image data using a color constancy model, and adjusting the illumination component using adaptive gamma correction to enhance structural details (such as crack textures) in the reflection component. This improves image contrast and ensures the effectiveness of subsequent feature extraction.

[0054] Step S13: Perform weighted fusion of the spatial features and the temporal features, and monitor the construction status of the construction area in real time based on the fused features.

[0055] In this embodiment, the above steps can be used to extract features from real-time monitoring data, and then the extracted spatial features and temporal features can be weighted and fused. It can be understood that the weighted and fused features reflect the overall state of the construction area, and the construction status of the construction area can be monitored in real time based on the fused features.

[0056] In one specific embodiment, the weighted fusion of the spatial features and the temporal features may include: calculating the causal relationship between the spatial features and the temporal features based on a preset attention weight matrix to obtain the causal relationship degree between the relevant features; the preset attention weight matrix is ​​an attention weight matrix constructed based on the construction status related to the bridge cantilever assembly construction and the corresponding causal information; determining the current data weight corresponding to the current weather status from a preset weight mapping table according to the current weather status corresponding to the construction area; the current data weight includes the data weight corresponding to the point cloud data and the image data; and combining the current data weight and the causal relationship degree to perform weighted fusion of the spatial features and the temporal features. Specifically, in the process of weighted fusion of the extracted spatial features and temporal features, the causal relationship between various factors in the construction status related to the bridge cantilever assembly construction can be considered, for example, a sudden drop in temperature → a decrease in material stiffness → an increase in steel beam deflection; that is, the causal relationship degree between the spatial features and the temporal features can be calculated by combining the attention weight matrix to reflect the causal relationship between various construction conditions. Furthermore, different environmental conditions (such as weather, where lidar data may be inaccurate in rainy or snowy weather) have varying impacts on the construction status. For example, in rainy or snowy weather, the weight of lidar data is reduced while the weight of image data is increased. Specifically, the current data weight matching the current weather condition can be determined from a pre-built weight mapping table, i.e., the data weight corresponding to the collected data such as point cloud data and image data can be determined. Then, the spatial features and temporal features can be weighted and fused by combining the current data weight with the previous causal correlation. This can yield more accurate fused features, which can effectively resist environmental interference such as wind, snow, and low light in high-altitude canyons and reflect a more realistic construction status.

[0057] Furthermore, in another specific embodiment, the real-time monitoring of the construction status of the construction area based on the fused features may include: real-time monitoring of the connection images of the bridge cantilever assembly construction-related beam segments corresponding to the fused features using a preset target detection algorithm to determine whether the connection status of the corresponding connection components meets preset stability conditions; and real-time monitoring of the stress state of the bridge cantilever assembly construction-related beam segments corresponding to the fused features based on preset sensor data thresholds to determine whether the stress state of the corresponding beam segments meets preset stress conditions. Specifically, in the process of real-time monitoring of the construction status, image analysis can be used. The target detection algorithm is used to analyze the connection images of the bridge cantilever assembly construction-related beam segments corresponding to the fused features to detect whether the connection status of the connection components meets preset stability conditions and whether there are abnormalities such as loose steel beam connection components or bracket deformation. Furthermore, combined with preset data thresholds, the stress state of the relevant beam segments can be monitored to determine whether the stress state of the beam segments is normal. If a single point of data exceeds the threshold, a relevant alarm can be triggered. Furthermore, if an alarm is triggered, the control system of the cantilever assembly equipment can be linked to directly suspend construction, depending on the severity of the situation, and construction can be resumed after the risk is eliminated.

[0058] In another specific embodiment, after real-time monitoring of the construction status of the construction area based on the fused features, it may further include: constructing a virtual scene corresponding to the construction area using data related to the construction status based on digital twin technology, so that relevant personnel can view the construction status corresponding to the bridge cantilever assembly construction through the virtual scene. Specifically, to facilitate monitoring and management, a virtual scene corresponding to the construction area can be constructed based on digital twin technology, mapping information such as the location and stress distribution of relevant steel beams in real time, allowing for panoramic viewing. Further, in a specific embodiment, a database can be maintained to record historical optimal construction handling solutions. A case-based reasoning algorithm can be used to match the current working condition and generate appropriate handling suggestions (e.g., reducing the assembly speed by 20% and adding temporary supports when the guide beam deflection exceeds 15%), and this can be linked with the cantilever assembly equipment control system, automatically sending a pause command when a red alert is triggered. This achieves a closed-loop engineering application of monitoring data, improves construction management efficiency, and shortens risk response time.

[0059] Therefore, this application can collect more comprehensive monitoring data in the construction area by setting up multi-level monitoring equipment in the corresponding construction area of ​​bridge cantilever assembly construction, combined with lidar, cameras and mechanical sensors. It can also improve the accuracy of data by combining spatial and temporal features with a dual-branch feature extraction network structure. Furthermore, by combining data detection and digital twin technology, the construction status can be displayed intuitively, thus improving the construction monitoring effect.

[0060] like Figure 2 As shown in the figure, this application discloses a system architecture diagram corresponding to a bridge cantilever assembly construction monitoring method, involving components such as lidar, cameras, sensors, edge computing terminals, 5G data transmission modules, cloud servers, construction management and control platforms, and equipment control systems; the following embodiments will be combined with this system architecture, and as shown in the figure. Figure 3 The flowchart shown illustrates a specific method for monitoring bridge cantilever assembly construction, which includes:

[0061] In this embodiment, the multi-level monitoring equipment used may include cold-resistant lidar, cold-resistant high-definition cameras, fiber optic grating sensors, and environmental sensors, and can be divided into a three-dimensional monitoring network architecture of "air-ground-structure". Among them, the high-altitude monitoring layer can be selected to deploy 3 sets of industrial-grade cold-resistant lidar (operating temperature -20℃~70℃, ranging accuracy ±2mm) on the mountains on both sides of the canyon to perform 360° scanning of the steel beam cantilever assembly area and obtain the overall three-dimensional point cloud data of the steel beam. The lidar system uses pulsed laser ranging to emit laser beams towards the target and calculates the distance based on the beam reflection time. It can collect 1 million data points per second, achieving full-area coverage monitoring of steel beam displacement. The ground monitoring layer can deploy multiple cold-resistant high-definition cameras around the cantilever assembly support, using fisheye lenses to cover the support settlement area and slope protection zone. The cameras have built-in heating modules that automatically activate defrosting at -20°C. They capture visual features such as support cracks and slope collapses through optical imaging, and the images are transmitted in real-time to the edge computing terminal. Furthermore, the structural monitoring layer can embed fiber optic grating sensors (corresponding to fixed strain measurement ranges) at key sections of the steel beam (such as the box section at the mid-support point and the guide beam connection node), sensing stress changes through changes in the fiber optic grating's wavelength. Pressure sensors are also installed in the hydraulic lines of the cantilever assembly jacks to collect cantilever assembly force data in real time. The sensor cables are cold-resistant shielded and embedded in insulated conduits to prevent freezing and cracking. Environmental sensors can collect environmental data such as temperature, humidity, and wind speed. Furthermore, the edge computing terminal can collect real-time monitoring data from multi-level monitoring devices, construct a ring data transmission network through industrial Ethernet, and use 5G wireless links to ensure continuous transmission even in environments with obstructed views, such as canyons and mountainous terrain. Moreover, the multimodal data acquisition process can utilize a BeiDou time synchronization module to provide unified time synchronization for lidar, cameras, and sensors, ensuring data consistency.

[0062] Furthermore, visual data, i.e., image data acquired by the camera, needs to undergo anti-interference processing, such as noise reduction and enhancement. Specifically, in the noise reduction process, a U-Net (a convolutional neural network-based model) image denoising model is pre-constructed based on an attention mechanism. The input is raw image data containing wind and snow noise. The encoder extracts image features, and a spatial attention module is introduced in the decoder stage to focus on identifying high-frequency noise regions such as snowflakes and fog. Pixel values ​​are repaired through residual connections. This model, trained on a large dataset of high-altitude construction images, can improve the peak signal-to-noise ratio and further enhance the accuracy of steel beam edge recognition. Correspondingly, in the image enhancement process, for low-light environments, the Retinex (color constancy) theory can be used to decompose the illumination and reflection components of the image. Adaptive gamma correction is used to adjust the illumination component, and structural details (such as crack textures) in the reflection component are enhanced, improving image contrast and meeting the needs of subsequent feature extraction.

[0063] To integrate multi-source data, it is necessary to perform spatiotemporal alignment of LiDAR point clouds, camera images, sensor data, and environmental parameters (temperature, wind speed). Point cloud data is converted into a 3D tensor through voxel meshing, and spatially correlated with the 2D feature maps of camera images through coordinate mapping. Temporal features of sensor time-series data are extracted through a Long Short-Term Memory (LSTM) network and concatenated with real-time values ​​of environmental parameters. A weighted fusion algorithm can be used to dynamically allocate weights based on data reliability (e.g., the weight of LiDAR data is set to 0.7 when there is no obstruction, and reduced to 0.4 in rainy or snowy weather, while the weight of image data is increased accordingly). Furthermore, the feature extraction process involves a dual-branch deep learning architecture. The spatial feature branch can use a CNN (Convolutional Neural Network) to process image and point cloud data, extracting spatial features such as steel beam deformation, support settlement, and slope cracks through convolutional layers, outputting a 128-dimensional spatial feature vector. The temporal feature branch can use a bidirectional LSM network to process the stress and displacement time-series data of sensors, capturing the dynamic changes during cantilever assembly, outputting a 64-dimensional temporal feature vector. Furthermore, this embodiment introduces a causal attention mechanism, which calculates the causal correlation between spatial features and temporal features (such as the causal chain weight of "sudden temperature drop → decrease in material stiffness → increase in steel beam deflection") through an attention weight matrix. Finally, the spatial feature vector and the temporal feature vector can be fused into a 256-dimensional comprehensive feature vector, i.e., the fused feature.

[0064] Furthermore, based on the fused comprehensive feature vector, pre-trained prediction models (Gradient Boosting Trees) can be used to predict data for a future period, such as numerical predictions of steel beam deflection and support settlement. Bayesian networks can also be used to calculate the probability of occurrence of "steel beam instability," "support collapse," and "slope sliding," and combined with a risk matrix to output red / yellow / blue three-level warnings. When a warning is triggered, SHAP value analysis can be used to identify key influencing factors (such as wind speed > 12m / s contributing 65% to the lateral displacement of the steel beam) and clarify the root cause of the risk.

[0065] Correspondingly, industrial-grade edge servers can be deployed at the construction site to perform target detection on image data, detecting anomalies such as loose steel beam connectors and deformed supports; real-time filtering and outlier removal of sensor data, while controlling response latency; if an anomaly is detected in a data point, an alarm can be triggered, and the relevant data can be uploaded to the cloud server. Furthermore, a management platform for the cloud server client can be set up. This platform can construct a virtual scene of bridge cantilever assembly based on digital twin technology, mapping the steel beam position, stress distribution, and warning areas in real time, supporting a 360° viewing angle. It can also pre-store optimal handling solutions related to historical construction states (e.g., "when the guide beam deflection exceeds 15%, reduce the cantilever assembly speed by 20% and add temporary supports"), matching the current working condition with case-based reasoning algorithms to generate handling suggestions; and it can also link with the cantilever assembly equipment control system, automatically sending a pause command when a red warning is triggered, waiting for the risk to be eliminated before resuming cantilever assembly construction.

[0066] Therefore, this application can reduce the error in bridge deformation monitoring by setting up multi-level monitoring equipment in the corresponding construction area of ​​bridge cantilever assembly construction, and by combining lidar, cameras and mechanical sensors to monitor and collect more comprehensive monitoring data in the construction area. In addition, by combining dynamic weighted fusion technology of multimodal data, it can also combine digital twin technology to monitor the construction status in real time, predict the construction status and issue early warnings for abnormal situations, which can adapt to complex construction scenarios and handle abnormal construction situations in a timely manner.

[0067] like Figure 4 As shown in the figure, this application discloses a bridge cantilever assembly construction monitoring device, including:

[0068] The multi-level monitoring module 11 is used to synchronize the time of multi-level monitoring equipment in the construction area corresponding to the bridge cantilever assembly construction through satellite time synchronization, and to obtain real-time monitoring data of the construction area through the time-synchronized multi-level monitoring equipment; the multi-level monitoring equipment includes lidar, cameras and mechanical sensors.

[0069] The feature processing module 12 is used to perform feature extraction processing on the point cloud data and image data in the real-time monitoring data using a first preset neural network to obtain corresponding spatial features, and to perform feature extraction processing on the force data in the real-time monitoring data and the environmental data corresponding to the construction area using a second preset neural network to obtain corresponding temporal features.

[0070] The construction status monitoring module 13 is used to perform weighted fusion of the spatial features and the temporal features, and to monitor the construction status of the construction area in real time based on the fused features.

[0071] Therefore, this application can improve data accuracy and ensure the real-time monitoring effect of the construction status by setting up multi-level monitoring equipment in the construction area corresponding to the bridge cantilever assembly construction, and by combining lidar, cameras and mechanical sensors to monitor and collect more comprehensive monitoring data in the construction area. Furthermore, it can combine a dual-branch feature extraction network structure to perform weighted fusion of spatial and temporal features, thereby ensuring the data accuracy and the real-time monitoring effect of the construction status.

[0072] In one specific embodiment, the multi-level monitoring module 11 may include:

[0073] A multi-level monitoring unit is used to acquire real-time monitoring data of the construction area through time-synchronized multi-level monitoring equipment and a preset data acquisition frequency; the preset data acquisition frequency includes data acquisition frequencies corresponding to the lidar, the camera and the mechanical sensor respectively; wherein, the preset data acquisition frequency includes a first data acquisition frequency when the construction area matches preset normal construction conditions, and a second data acquisition frequency when the construction area matches preset abnormal construction conditions.

[0074] In one specific embodiment, the device may further include:

[0075] The image denoising module is used to denoise the image data in the real-time monitoring data through a preset convolutional neural network to obtain the corresponding denoised image data; the preset convolutional neural network is a network built based on an attention mechanism to repair pixel values ​​through residual connections.

[0076] Accordingly, the feature processing module 12 may include:

[0077] The first feature processing unit is used to perform feature extraction processing on the point cloud data and the noise-reduced image data in the real-time monitoring data using a first preset neural network.

[0078] In another specific embodiment, the device may further include:

[0079] The image adjustment unit is used to decompose the image data in the real-time monitoring data into illumination and reflection components using a color constancy model, and to process the illumination and reflection components using adaptive gamma correction technology to obtain processed image data.

[0080] Accordingly, the feature processing module 12 may include:

[0081] The second feature processing unit is used to perform feature extraction processing on the point cloud data and the processed image data in the real-time monitoring data using the first preset neural network.

[0082] In one specific embodiment, the construction status monitoring module 13 may include:

[0083] The causal correlation calculation unit is used to perform causal correlation calculation on the spatial features and the temporal features based on a preset attention weight matrix to obtain the causal correlation degree between the relevant features; the preset attention weight matrix is ​​an attention weight matrix constructed based on the construction status and corresponding causal information related to the bridge cantilever assembly construction.

[0084] The data weight determination unit is used to determine the current data weight corresponding to the current weather state from a preset weight mapping table based on the current weather state corresponding to the construction area; the current data weight includes the data weight corresponding to the point cloud data and the image data;

[0085] The weighted fusion unit is used to perform weighted fusion of the spatial features and the temporal features by combining the current data weights and the causal correlation degree.

[0086] In another specific embodiment, the construction status monitoring module 13 may include:

[0087] The connector image monitoring unit is used to monitor the connector images of the bridge cantilever assembly construction related beam segments corresponding to the fused features in real time through a preset target detection algorithm, so as to determine whether the connection status of the corresponding connector meets the preset stability conditions.

[0088] The stress monitoring unit is used to monitor the stress state of the bridge cantilever assembly construction-related beam segments corresponding to the fused features in real time based on preset sensor data thresholds, so as to determine whether the stress state of the corresponding beam segment meets the preset stress conditions.

[0089] In one specific embodiment, the device may further include:

[0090] The virtual scene construction module is used to construct a virtual scene corresponding to the construction area based on digital twin technology and data related to the construction status, so that relevant personnel can view the construction status corresponding to the bridge cantilever assembly construction through the virtual scene.

[0091] Furthermore, embodiments of this application also disclose an electronic device, Figure 5 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0092] Figure 5 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the bridge cantilever assembly construction monitoring method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0093] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0094] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0095] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the bridge cantilever assembly construction monitoring method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs capable of performing other specific tasks.

[0096] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned bridge cantilever assembly construction monitoring method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0097] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0098] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0099] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0100] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0101] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for monitoring cantilever assembly construction of a bridge, characterized by, The method comprises: Time synchronization is performed on a plurality of monitoring devices in a construction area of a bridge cantilever assembly construction by a satellite time service method, and real-time monitoring data of the construction area is obtained by the plurality of monitoring devices after time synchronization; the plurality of monitoring devices comprise a laser radar, a camera and a mechanical sensor; Feature extraction processing is performed on point cloud data and image data in the real-time monitoring data by a first preset neural network to obtain corresponding spatial features, and feature extraction processing is performed on force data in the real-time monitoring data and corresponding environmental data of the construction area by a second preset neural network to obtain corresponding time sequence features; The spatial features and the time sequence features are weightedly fused, and a construction state of the construction area is monitored in real time based on the fused features.

2. The bridge cantilever construction monitoring method according to claim 1, characterized by, The real-time monitoring data of the construction area is obtained by the plurality of monitoring devices after time synchronization, comprising: The real-time monitoring data of the construction area is obtained by the plurality of monitoring devices after time synchronization and a preset data acquisition frequency; the preset data acquisition frequency comprises data acquisition frequencies corresponding to the laser radar, the camera and the mechanical sensor respectively; The preset data acquisition frequency comprises a first data acquisition frequency when the construction area matches a preset normal construction condition, and a second data acquisition frequency when the construction area matches a preset abnormal construction condition.

3. The bridge cantilever construction monitoring method according to claim 1, characterized by, After the real-time monitoring data of the construction area is obtained by the plurality of monitoring devices after time synchronization, the method further comprises: Noise reduction processing is performed on image data in the real-time monitoring data by a preset convolutional neural network to obtain corresponding denoised image data; the preset convolutional neural network is a network for repairing pixel values by residual connection based on an attention mechanism; Correspondingly, the feature extraction processing on the point cloud data and the image data in the real-time monitoring data by the first preset neural network comprises: The feature extraction processing on the point cloud data and the denoised image data in the real-time monitoring data by the first preset neural network.

4. The bridge cantilever construction monitoring method according to claim 1, characterized by, After the real-time monitoring data of the construction area is obtained by the plurality of monitoring devices after time synchronization, the method further comprises: An illumination component and a reflection component of the image data in the real-time monitoring data are decomposed by a color constancy model, and the illumination component and the reflection component are processed by an adaptive gamma correction technique to obtain processed image data; Correspondingly, the feature extraction processing on the point cloud data and the image data in the real-time monitoring data by the first preset neural network comprises: The feature extraction processing on the point cloud data and the processed image data in the real-time monitoring data by the first preset neural network.

5. The bridge cantilever construction monitoring method according to claim 1, wherein The weighted fusion of the spatial features and the time sequence features comprises: The spatial features and the time sequence features are calculated based on a preset attention weight matrix to obtain a causal correlation degree between the related features; the preset attention weight matrix is an attention weight matrix constructed based on a construction state related to the bridge cantilever assembly construction and corresponding causal information; A current data weight corresponding to the current weather state of the construction area is determined from a preset weight mapping table according to the current weather state; the current data weight includes data weights corresponding to the point cloud data and the image data; The spatial features and the time sequence features are weighted and fused in combination with the current data weight and the causal correlation degree.

6. The bridge cantilever construction monitoring method according to claim 1, wherein The construction state of the construction area is monitored in real time based on the fused features, including: The connection state of the corresponding connecting piece is determined by monitoring the image of the connecting piece of the bridge cantilever assembly construction related beam segment corresponding to the fused features in real time through a preset target detection algorithm to determine whether the corresponding connecting piece meets a preset stability condition; The stress state of the corresponding beam segment is determined by monitoring the stress state of the bridge cantilever assembly construction related beam segment corresponding to the fused features in real time based on a preset sensor data threshold to determine whether the stress state of the corresponding beam segment meets a preset stress condition.

7. The method of monitoring the cantilever construction of a bridge according to any one of claims 1 to 6, wherein After the construction state of the construction area is monitored in real time based on the fused features, it further includes: A virtual scene corresponding to the construction area is constructed based on the data related to the construction state by using a digital twin technology, so that the relevant staff can view the construction state of the bridge cantilever assembly construction through the virtual scene.

8. A bridge cantilever assembly construction monitoring device, characterized by, It includes: A multi-level monitoring module for time synchronizing a plurality of monitoring devices in a construction area corresponding to the bridge cantilever assembly construction through a satellite time service method, and obtaining real-time monitoring data of the construction area through the time-synchronized multi-level monitoring devices; the multi-level monitoring devices include a laser radar, a camera and a mechanical sensor; A feature processing module for performing feature extraction processing on point cloud data and image data in the real-time monitoring data by using a first preset neural network to obtain corresponding spatial features, and performing feature extraction processing on force data in the real-time monitoring data and environment data corresponding to the construction area by using a second preset neural network to obtain corresponding time sequence features; A construction state monitoring module for weighted fusion of the spatial features and the time sequence features, and real-time monitoring of the construction state of the construction area based on the fused features.

9. An electronic device, comprising: It includes: A memory for saving a computer program; A processor for executing the computer program to implement the bridge cantilever assembly construction monitoring method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A memory for saving a computer program; the computer program is executed by a processor to implement the bridge cantilever assembly construction monitoring method of any one of claims 1 to 7.

Images