Multi-source data fusion tunnel waterlogging disaster intelligent early warning method and device
Through the fusion of multi-source data, a tunnel flood disaster warning model is constructed, which can accurately predict and multi-stage warning of water accumulation in the tunnel, and solve the problems of insufficient data accuracy and insufficient warning timeliness in traditional methods, and improve disaster response capabilities.
Patent Information
- Application Number
- CN202510511937.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-23
- Publication Date
- 2025-05-23
Smart Images
Figure CN120030422A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to multi-source data fusion and intelligent early warning technology, and in particular to a multi-source data fusion intelligent early warning method and device for tunnel waterlogging disasters. Background Art
[0002] With the increase of urban population density and the concentration of economic activities, tunnels, as key nodes connecting urban transportation networks, are directly related to the smooth operation of urban transportation and public safety. Due to their semi-enclosed structures, tunnels are prone to waterlogging, waterlogging and even flooding once they encounter waterlogging disasters, leading to direct economic losses such as traffic interruption, vehicle damage, and equipment failure. They are also more likely to cause casualties and secondary disasters, such as power system paralysis and communication interruption, further expanding the scope of disaster impact.
[0003] Traditional methods usually rely on a single data source or limited sensors to collect hydrological information. The data coverage is narrow and the accuracy is insufficient, making it difficult to fully reflect complex hydrological changes. Secondly, traditional methods lack deep fusion and multi-level modeling of multi-source data, and often only make predictions based on simple historical data or empirical formulas, resulting in low accuracy of prediction results, especially for waterlogging disasters under extreme weather conditions. In addition, the data collection and processing efficiency of traditional methods is low, and real-time monitoring and rapid response cannot be achieved, resulting in delayed warning information and difficulty in meeting the timeliness requirements of modern disaster prevention and mitigation. Finally, traditional warning mechanisms usually adopt a single threshold trigger mode, lack a multi-stage and dynamic warning strategy, and cannot provide differentiated response measures according to different stages of disaster development, thereby reducing the pertinence and practicality of warnings. Summary of the invention
[0004] In order to improve the existing intelligent early warning method and device for tunnel waterlogging disasters, an intelligent early warning method and device for tunnel waterlogging disasters with multi-source data fusion is provided. This method constructs a tunnel water inflow and waterlogging waterlogging prediction model through multi-source sensors and meteorological data, realizes accurate prediction of water flow and waterlogging depth, and combines a multi-stage early warning mechanism to improve disaster prevention response capabilities.
[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: The intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion includes: Divide the tunnel into several areas and obtain image data of each area in the tunnel; Based on the image data, the tunnel water accumulation contour map is obtained, the contour line function is recorded, and the contour line function of each time node is summarized to construct the tunnel water accumulation change model; Obtain tunnel water accumulation contour maps of each area, divide the water accumulation contour line in each area into multiple sections, input each section of the contour line at each time node based on the tunnel water accumulation change model, and obtain the change trend of the function image of each section of the tunnel water accumulation contour line; By comprehensively analyzing the changing trend of the tunnel water accumulation contour function images in each section of the contour line in the same area, the water accumulation change in the tunnel in the current area can be judged, and the probability of waterlogging in the tunnel can be obtained by comprehensively analyzing all areas; Based on the rate of change of the tunnel waterlogging contour function image, a waterlogging depth prediction model is constructed to obtain the water depth in the tunnel in the future time period, classify the waterlogging disaster risk level, and design a multi-stage early warning trigger strategy.
[0006] Preferably, the obtaining of the tunnel water accumulation contour map based on the image data, recording the contour line function, and summarizing the contour line functions at each time node to construct the tunnel water accumulation change model specifically includes: Eliminate lens distortion based on perspective transformation and establish a mapping relationship between pixel coordinates and actual space coordinates of the tunnel; Based on the actual spatial coordinates of the tunnel obtained, the tunnel water accumulation contour nodes are marked according to fixed pixels; Based on the contour nodes, the nodes are connected into continuous contour lines through spatial interpolation method; The contour function of each time node is obtained to form a contour time series, and a tunnel water accumulation change model is constructed.
[0007] Preferably, the step of obtaining the tunnel waterlogging contour map of each area, dividing the waterlogging contour line in each area into multiple sections, inputting each section of the contour line at each time node based on the tunnel waterlogging change model, and obtaining the change trend of the function image of each section of the tunnel waterlogging contour line specifically includes: Obtain tunnel water accumulation contour maps in different areas, and divide the tunnel water accumulation contour line in each area into multiple contour lines according to the curvature extreme value points; Based on the contour line functions of different time nodes of the same contour line, the difference and overall change are calculated; Based on the obtained contour line changes, obtain the area changes and shape changes of which the changes exceed 50%; A time series is constructed based on the changes in each time node, and the change trend of the tunnel water accumulation contour function image of each contour line is analyzed through the regression model.
[0008] Preferably, based on the acquired contour line changes, acquiring the area changes and shape changes whose changes exceed 50% specifically includes: Based on the calculated change area, calculate and obtain the part of the area change exceeding 50% in each contour line segment at different time periods, and obtain the area change; Based on the changed contour line, the curvature, aspect ratio, and boundary smoothness of each position are calculated to obtain the contour line shape parameters; The shape change is obtained based on the contour line shape parameter change of each position node.
[0009] Preferably, the step of judging the change of water accumulation in the tunnel in the current area by comprehensively analyzing the change trend of the tunnel water accumulation contour line function image in each section of the contour line in the same area, and comprehensively obtaining the probability of waterlogging in the tunnel in all areas specifically includes: Statistically calculate the changing trend of the contour function image of each section of the tunnel in the area. If the changing trend is consistent, the changing trend of the area can be determined; Based on the inconsistent change trends, the change trends of the adjacent areas of the area are obtained, and different weight values are assigned based on the area of the adjacent areas; Based on the distribution of change trends in the region and the weight change trends of adjacent regions, the change trend of the region is comprehensively judged; Based on the changing trend of the tunnel water accumulation contour function image in each area, the probability of waterlogging in the tunnel is calculated through the physical driving model.
[0010] Preferably, the method of constructing a waterlogging depth prediction model based on the tunnel waterlogging contour line function image change rate, obtaining the waterlogging depth in the tunnel in the future time period, and classifying the waterlogging disaster risk level, and designing a multi-stage early warning trigger strategy specifically includes: Based on the waterlogging situation obtained by calculation, the function image change data of the tunnel waterlogging contour line in the historical waterlogging data is obtained, and the data set is divided; The divided data is input into the LSTM model for training, and the trained tunnel flooding prediction model is output; Based on the real-time acquired tunnel water accumulation contour line function image change data, it is input into the waterlogging water accumulation depth prediction model to obtain the water accumulation depth in the tunnel in the future time period; Based on the acquired data of water accumulation depth, water accumulation growth rate and duration in the tunnel in the future time period, four levels of danger are divided; Based on the dangerous situations at various levels, multi-stage dangerous situation warning responses are triggered.
[0011] Furthermore, an intelligent early warning device for tunnel waterlogging disasters based on multi-source data fusion is proposed, including: Data acquisition and area management module: The data acquisition and area management module is mainly used to realize the grid partition management of tunnel space, integrating multiple cameras to synchronously collect RGB-D images and temperature and humidity sensor data; Image processing and coordinate mapping module: The image processing and coordinate mapping module is mainly used to establish a geometric mapping relationship between the pixel coordinate system and the tunnel BIM model through adaptive perspective transformation to eliminate lens distortion; Contour line completion module: The contour line completion module is mainly used to interpolate and generate continuous contour line functions; Segment trend analysis module: The segment trend analysis module is mainly used to segment the contour line into characteristic segments according to the extreme points of curvature, and calculate and obtain the morphological evolution trend of each segment; Regional risk assessment module: The regional risk assessment module is mainly used to integrate the trend consistency analysis of the internal contour segments of the region and the spatial weights of the adjacent regions to calculate the probability of waterlogging disasters caused by the spread of accumulated water; Deep learning prediction module: The deep learning prediction module is mainly used to build an LSTM hybrid network model to predict the evolution of water depth based on the historical profile function time series; Danger level classification module: The hazard level classification module is mainly used to define three-dimensional assessment indicators and realize dynamic risk classification; Multi-level early warning response module: The multi-level early warning response module is mainly used to design a hierarchical early warning strategy that links threshold triggering with trend prediction, and supports the generation of linkage control instructions for the drainage system; Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.
[0012] Compared with the prior art, the advantages of the present invention are: Real-time tracking and prediction capabilities for dynamic changes of water accumulation in tunnels. By analyzing the contour lines of water accumulation at each time node, the temporal and spatial change trends of water accumulation can be fully reflected, providing early warning for the occurrence of disasters. In addition, the tunnel water accumulation change model and contour function image change trend analysis can accurately evaluate the changes in water accumulation in different areas, and comprehensively obtain the overall probability of waterlogging disasters based on the risk status of each area. This method not only improves the accuracy of prediction, but also builds a waterlogging depth prediction model based on the change rate of water accumulation depth, further refines the disaster level classification, and improves the timeliness and pertinence of warnings. By designing a multi-stage warning trigger strategy, it can effectively implement response measures at different stages, provide a scientific basis for tunnel management departments, and thus reduce the risks and losses caused by tunnel waterlogging disasters. This method has strong real-time, accuracy and operability, and can provide important support for tunnel safety management. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 A schematic diagram of the method proposed by the present invention; Figure 2 A schematic diagram of constructing a tunnel water accumulation change model proposed by the present invention; Figure 3A schematic diagram for obtaining the changing trend proposed by the present invention; Figure 4 This is a schematic diagram of the contour line variation proposed by the present invention; Figure 5 This is a schematic diagram of the waterlogging probability calculation proposed by the present invention; Figure 6 This is a schematic diagram of water accumulation depth prediction and disaster warning proposed by the present invention; Figure 7 This is a schematic diagram of the electronic device in this solution; Figure 8 This is a schematic diagram of the computer-readable storage medium structure in this solution. DETAILED DESCRIPTION
[0014] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.
[0015] The intelligent early warning device for tunnel waterlogging disasters based on multi-source data fusion includes: Data acquisition and area management module: The data acquisition and area management module is mainly used to realize the grid partition management of tunnel space, integrating multiple cameras to synchronously collect RGB-D images and temperature and humidity sensor data; Image processing and coordinate mapping module: The image processing and coordinate mapping module is mainly used to establish a geometric mapping relationship between the pixel coordinate system and the tunnel BIM model through adaptive perspective transformation to eliminate lens distortion; Contour line completion module: The contour line completion module is mainly used to interpolate and generate continuous contour line functions; Segment trend analysis module: The segment trend analysis module is mainly used to segment the contour line into characteristic segments according to the extreme points of curvature, and calculate and obtain the morphological evolution trend of each segment; Regional risk assessment module: The regional risk assessment module is mainly used to integrate the trend consistency analysis of the internal contour segments of the region and the spatial weights of the adjacent regions to calculate the probability of waterlogging disasters caused by the spread of accumulated water; Deep learning prediction module: The deep learning prediction module is mainly used to build an LSTM hybrid network model to predict the evolution of water depth based on the historical profile function time series; Danger level classification module: The hazard level classification module is mainly used to define three-dimensional assessment indicators and realize dynamic risk classification; Multi-level early warning response module: The multi-level early warning response module is mainly used to design a hierarchical early warning strategy that links threshold triggering with trend prediction, and supports the generation of linkage control instructions for the drainage system; Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.
[0016] See also Figure 1 As shown in the figure, the intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion includes: Step 1: Divide the tunnel into several areas and obtain image data of each area in the tunnel; Step 2: Obtain the tunnel water accumulation contour map based on the image data, record the contour line function, and summarize the contour line functions at each time node to build a tunnel water accumulation change model; Step 3: Obtain the tunnel water accumulation contour map of each area, divide the water accumulation contour line in each area into multiple segments, input each contour line of each time node based on the tunnel water accumulation change model, and obtain the change trend of the function image of each tunnel water accumulation contour line; Step 4: By comprehensively analyzing the changing trend of the tunnel water accumulation contour function images in each section of the contour line in the same area, the change of water accumulation in the tunnel in the current area is determined, and the probability of waterlogging in the tunnel is obtained by comprehensively analyzing all areas; Step 5: Based on the rate of change of the tunnel waterlogging contour function image, a waterlogging depth prediction model is constructed to obtain the water depth in the tunnel in the future time period, classify the waterlogging disaster risk level, and design a multi-stage early warning trigger strategy.
[0017] See also Figure 2 As shown in the figure, the tunnel water accumulation contour map is obtained based on the image data, the contour line function is recorded, and the contour line function of each time node is summarized to construct the tunnel water accumulation change model, which specifically includes: Eliminate lens distortion based on perspective transformation and establish a mapping relationship between pixel coordinates and actual space coordinates of the tunnel; Based on the actual spatial coordinates of the tunnel obtained, the tunnel water accumulation contour nodes are marked according to fixed pixels; Based on the contour nodes, the nodes are connected into continuous contour lines through spatial interpolation method; The contour function of each time node is obtained to form a contour time series, and a tunnel water accumulation change model is constructed.
[0018] Specifically, the perspective transformation can be represented by a 3x3 matrix. Given a point in the camera coordinate system, the perspective transformation matrix is used to map the pixel coordinates in the image coordinate system. The formula is:
[0019] in, is a perspective transformation matrix, which is related to factors such as the internal and external parameters of the camera and the distortion in the image. , is the three-dimensional point coordinate in the camera coordinate system, , is the transformed two-dimensional pixel coordinate, corresponding to the position in the image plane; Using the internal and external parameters of the camera and the perspective transformation matrix, the mapping relationship between the pixel coordinates in the image and the actual spatial coordinates of the tunnel is established; Use spatial interpolation methods to connect contour nodes into smooth continuous contour lines. Assuming there are several contour nodes in space, a continuous contour line is generated by interpolation methods, where t is the interpolation parameter, which usually ranges from [0,1] or a discrete time step.
[0020] See also Figure 3 As shown, tunnel water accumulation contour maps of various areas are obtained, and the water accumulation contour lines in each area are divided into multiple sections. Based on the tunnel water accumulation change model, each section of the contour line at each time node is input to obtain the change trend of the function image of each section of the tunnel water accumulation contour line. Specifically, the following are included: Obtain tunnel water accumulation contour maps in different areas, and divide the tunnel water accumulation contour line in each area into multiple contour lines according to the curvature extreme value points; Based on the contour line functions of different time nodes of the same contour line, the difference and overall change are calculated; Based on the obtained contour line changes, obtain the area changes and shape changes of which the changes exceed 50%; A time series is constructed based on the changes in each time node, and the change trend of the tunnel water accumulation contour function image of each contour line is analyzed through the regression model.
[0021] It is understandable that the area calculation of the contour line may be inaccurate due to irregular or missing boundaries, especially when the contour changes dramatically. The shape change calculation of the contour line may be affected by local noise or curvature mutation, resulting in large fluctuations in the shape change. Therefore, a more accurate area calculation method (such as Monte Carlo integration or grid point method) is used to reduce errors. At the same time, when calculating the shape change, a smoothing algorithm or local change analysis (such as Gaussian weighted average) can be used to reduce the impact of local mutations. For area changes exceeding 50%, a dynamic threshold is set to adjust the judgment criteria according to the actual change range of the area.
[0022] If the contour changes are very complex, a simple regression model may not be able to capture the trend, resulting in inaccurate trend analysis. It is necessary to select a suitable regression model based on the characteristics of the data, such as nonlinear regression, support vector machine regression, neural network regression, etc. For situations with complex changing trends, deep learning methods (such as convolutional neural networks) can also be considered as a solution. Perform feature engineering on time series data to extract effective features (such as the rate of change of contours, changes in curvature, changes in area, etc.) to provide stronger prediction capabilities for the regression model.
[0023] See also Figure 4As shown, based on the obtained contour line changes, obtaining the area changes and shape changes where the changes exceed 50% specifically includes: Based on the calculated change area, calculate and obtain the part of the area change exceeding 50% in each contour line segment at different time periods, and obtain the area change; Based on the changed contour line, the curvature, aspect ratio, and boundary smoothness of each position are calculated to obtain the contour line shape parameters; The shape change is obtained based on the contour line shape parameter change of each position node.
[0024] Specifically, curvature describes the degree of curvature of a curve at a certain point. For a plane curve , the curvature is calculated as:
[0025] in, is the curvature at time t, and is the velocity component of the curve at time t (i.e., the first-order derivative), representing the tangent vector of the curve, and is the acceleration component (second-order derivative) of the curve at time t, indicating the curvature of the curve; Curvature Take the derivative and get the speed of curvature change. By solving To find the extreme points of curvature (i.e. the turning points of the curve), these points represent where the shape of the contour changes, that is, where the bending trend of the curve changes; Given the area at two different time points, the area change can be calculated using the following formula:
[0026] in, is the percentage change in area, , The time points and The area value when ; For each contour line, perform similar calculations to find the parts with area changes exceeding 50%; The aspect ratio of the contour line is an important parameter for describing the contour shape. It is the ratio of the length to the width of the circumscribed rectangle of the contour. The aspect ratio reflects the degree of stretching of the contour. When the aspect ratio is close to 1, the contour is close to a circle; when the aspect ratio is much greater than 1, the contour is closer to a long strip; Boundary smoothness measures the smoothness of the contour boundary, which can be achieved by calculating the average curvature of the contour line or the volatility of the contour. A common definition method is to calculate the geometric curvature of the contour, the formula is:
[0027] Where S is the geometric curvature of the profile, is the curvature at each point, is the total number of points. A smaller smoothness value indicates that the contour line is smoother, while a larger value indicates that the contour boundary has obvious curvature and irregularities.
[0028] See also Figure 5 As shown in the figure, by integrating the changing trend of the tunnel water accumulation contour function image in each section of the contour line in the same area, the change of water accumulation in the tunnel in the current area is judged, and the probability of waterlogging in the tunnel is obtained by integrating all areas, including: Statistically calculate the changing trend of the contour function image of each section of the tunnel in the area. If the changing trend is consistent, the changing trend of the area can be determined; Based on the inconsistent change trends, the change trends of the adjacent areas of the area are obtained, and different weight values are assigned based on the area of the adjacent areas; Based on the distribution of change trends in the region and the weight change trends of adjacent regions, the change trend of the region is comprehensively judged; Based on the changing trend of the tunnel water accumulation contour function image in each area, the probability of waterlogging in the tunnel is calculated through the physical driving model.
[0029] Specifically, the water accumulation contour function of each tunnel section can be obtained by fitting the contour data of different time nodes. The change trend of each contour section can be expressed as the change of time series, and the change amount of the same contour section is calculated, and the change is measured by using the mean square error (MSE); Combining the change trends of multiple adjacent areas, the weighted average method can be used to comprehensively judge the change trend of the area. The formula is:
[0030] in, The trend of water accumulation in this area is shown in Figure 2. The proportion of this area is expanded. is the weight ratio of the region, is the additional weight of each adjacent region, is the changing trend of adjacent regions, and n is the total number of adjacent regions.
[0031] By aggregating the change trends of different segments in the region, the overall change trend of the region can be obtained. The standard deviation can be used to represent the degree of discreteness of the change. The formula is:
[0032] in, is the degree of discreteness, is the mean value of the change trend in the region, is the changing trend of each tunnel section, is the number of segments in the region; The occurrence of urban flooding is related to physical factors such as waterlogging area, waterlogging depth, and rainfall. It can be calculated using a hydrological model combined with the changing trend of the waterlogging contour line. The input of the model includes the change in waterlogging area, water flow rate in the tunnel, precipitation amount and precipitation intensity, terrain changes, etc.
[0033] See also Figure 6 As shown in the figure, based on the change rate of the tunnel water accumulation contour function image, a waterlogging depth prediction model is constructed to obtain the water accumulation depth in the tunnel in the future time period, and the waterlogging disaster risk level is divided. The multi-stage early warning trigger strategy is designed, including: Based on the waterlogging situation obtained by calculation, the function image change data of the tunnel waterlogging contour line in the historical waterlogging data is obtained, and the data set is divided; The divided data is input into the LSTM model for training, and the trained tunnel flooding prediction model is output; Based on the real-time acquired tunnel water accumulation contour line function image change data, it is input into the waterlogging water accumulation depth prediction model to obtain the water accumulation depth in the tunnel in the future time period; Based on the acquired data of water accumulation depth, water accumulation growth rate and duration in the tunnel in the future time period, four levels of danger are divided; Based on the dangerous situations at various levels, multi-stage dangerous situation warning responses are triggered.
[0034] Specifically, historical tunnel waterlogging contour function image data is collected, time series features are extracted and converted into standardized numerical data sets, which are divided into training set, validation set and test set in a ratio of 7:2:1 to retain time continuity. A multi-layer LSTM network is then constructed to output the predicted waterlogging depth for the next hour. The Adam optimizer and MAE loss function are used in the training phase, combined with the validation set early stopping strategy to prevent overfitting, and the final test set is used to evaluate the model accuracy (RMSE ≤ 5cm). During deployment, the contour data collected in real time is input into the model after the same preprocessing, and the waterlogging depth curve for future time periods is dynamically generated through the sliding window mechanism, and the prediction results are output in combination with the confidence interval; The water depth is specifically the measured or predicted water height at the lowest point of the tunnel. The water growth rate is specifically the depth change per unit time, reflecting the speed of the dangerous situation deteriorating. The duration is specifically the length of time that the water exceeds the threshold. In some preferred embodiments, the relationship between the dangerous situation level and various indicators is shown in the following table:
[0035] Based on the compound condition trigger mechanism, any indicator exceeding the standard will trigger the warning: for example, even if h=0.4m (yellow level), if v=0.4m / h (orange level), it will be handled as orange warning. Correction at night / peak traffic period: During low visibility or high traffic flow period, the warning level will automatically increase by one level (e.g. yellow→orange); Based on the blue warning, the response actions include: System self-check: Automatically start the drainage system self-check program to check for pump station failures. Information broadcast: The LED screen at the tunnel entrance displays "Slight water accumulation, drive carefully". Data recording: Record the location and time of water accumulation for subsequent analysis; Based on the yellow warning, the response actions include: Drainage enhancement: Start the backup pump and increase the drainage capacity to 80%. Traffic control: Speed limit 30km / h, close the rightmost flooded lane. Manual inspection: Send personnel to the scene with portable pumping equipment; Based on the orange warning, the response actions include: Full-power drainage: Activate all drainage pipe pumps and request external support drainage vehicles. Traffic closure: No trucks and buses are allowed to pass, and small cars are allowed to pass in one lane. Structural monitoring: Activate tunnel deformation sensors and report structural safety data every 10 minutes; Based on the red warning, the response actions include: Emergency tunnel closure: lowering the waterproof gate and evacuating stranded vehicles and personnel. Inter-departmental linkage: notifying fire and medical units to stand by, and remote traffic diversion by traffic police. Public warning: push "tunnel closure" alerts via SMS, broadcast, and navigation apps.
[0036] Furthermore, the method according to the embodiment of the present application can also be performed by Figure 7 The electronic device architecture shown in FIG. Figure 7 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store the intelligent early warning method and device for tunnel waterlogging disasters with multi-source data fusion provided in the present application. The electronic device 500 may also include a user interface 508. Of course, Figure 7 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 7 One or more components of an electronic device are shown.
[0037] Figure 8 Schematic diagram of a computer-readable storage medium structure provided by an embodiment of the present application. Figure 8As shown, a computer-readable storage medium 600 according to an embodiment of the present application is shown. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by the processor, the intelligent early warning method and device for tunnel waterlogging disasters based on multi-source data fusion according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.
[0038] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0039] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0040] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. An intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion, characterized in that: include: Divide the tunnel into several areas and obtain image data of each area in the tunnel; Based on the image data, the tunnel water accumulation contour map is obtained, the contour line function is recorded, and the contour line function of each time node is summarized to construct the tunnel water accumulation change model; Obtain tunnel water accumulation contour maps of each area, divide the water accumulation contour line in each area into multiple sections, input each section of the contour line at each time node based on the tunnel water accumulation change model, and obtain the change trend of the function image of each section of the tunnel water accumulation contour line; By comprehensively analyzing the changing trend of the tunnel water accumulation contour function images in each section of the contour line in the same area, the water accumulation change in the tunnel in the current area can be judged, and the probability of waterlogging in the tunnel can be obtained by comprehensively analyzing all areas; Based on the rate of change of the tunnel waterlogging contour function image, a waterlogging depth prediction model is constructed to obtain the water depth in the tunnel in the future time period, classify the waterlogging disaster risk level, and design a multi-stage early warning trigger strategy.
2. The intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion according to claim 1 is characterized in that: The method of obtaining a tunnel water accumulation contour map based on image data, recording contour line functions, and summarizing contour line functions at various time nodes to construct a tunnel water accumulation change model specifically includes: Eliminate lens distortion based on perspective transformation and establish a mapping relationship between pixel coordinates and actual space coordinates of the tunnel; Based on the actual spatial coordinates of the tunnel obtained, the tunnel water accumulation contour nodes are marked according to fixed pixels; Based on the contour nodes, the nodes are connected into continuous contour lines through spatial interpolation method; The contour function of each time node is obtained to form a contour time series, and a tunnel water accumulation change model is constructed.
3. The intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion according to claim 1 is characterized in that: The method of obtaining the tunnel water accumulation contour map of each area, dividing the water accumulation contour line in each area into multiple sections, inputting each section of the contour line at each time node based on the tunnel water accumulation change model, and obtaining the change trend of the function image of each section of the tunnel water accumulation contour line specifically includes: Obtain tunnel water accumulation contour maps in different areas, and divide the tunnel water accumulation contour line in each area into multiple contour lines according to the curvature extreme value points; Based on the contour line functions of different time nodes of the same contour line, the difference and overall change are calculated; Based on the obtained contour line changes, obtain the area changes and shape changes of which the changes exceed 50%; A time series is constructed based on the changes in each time node, and the change trend of the tunnel water accumulation contour function image of each contour line is analyzed through the regression model.
4. The intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion according to claim 3 is characterized in that: The step of obtaining the area change and shape change whose change exceeds 50% based on the obtained contour line change specifically includes: Based on the calculated change area, calculate and obtain the part of the area change exceeding 50% in each contour line segment at different time periods, and obtain the area change; Based on the changed contour line, the curvature, aspect ratio, and boundary smoothness of each position are calculated to obtain the contour line shape parameters; The shape change is obtained based on the contour line shape parameter change of each position node.
5. The intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion according to claim 1 is characterized in that: The method of judging the change of water accumulation in the tunnel in the current area by integrating the change trend of the tunnel water accumulation contour line function image in each section of the contour line in the same area, and obtaining the probability of waterlogging in the tunnel by integrating all areas specifically includes: Statistically calculate the changing trend of the contour function image of each section of the tunnel in the area. If the changing trend is consistent, the changing trend of the area can be determined; Based on the inconsistent change trends, the change trends of the adjacent areas of the area are obtained, and different weight values are assigned based on the area of the adjacent areas; Based on the distribution of change trends in the region and the weight change trends of adjacent regions, the change trend of the region is comprehensively judged; Based on the changing trend of the tunnel water accumulation contour function image in each area, the probability of waterlogging in the tunnel is calculated through the physical driving model.
6. The intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion according to claim 1 is characterized in that: The method of constructing a waterlogging depth prediction model based on the tunnel waterlogging contour line function image change rate, obtaining the waterlogging depth in the tunnel in the future time period, and classifying the waterlogging disaster risk level, and designing a multi-stage early warning trigger strategy specifically includes: Based on the waterlogging situation obtained by calculation, the function image change data of the tunnel waterlogging contour line in the historical waterlogging data is obtained, and the data set is divided; The divided data is input into the LSTM model for training, and the trained tunnel flooding prediction model is output; Based on the real-time acquired tunnel water accumulation contour line function image change data, it is input into the waterlogging water accumulation depth prediction model to obtain the water accumulation depth in the tunnel in the future time period; Based on the acquired data of water accumulation depth, water accumulation growth rate and duration in the tunnel in the future time period, four levels of danger are divided; Based on the dangerous situations at various levels, multi-stage dangerous situation warning responses are triggered.
7. An intelligent early warning device for tunnel waterlogging disasters in combination with multi-source data fusion, used to implement the intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion as claimed in any one of claims 1 to 6, characterized in that: include: Data acquisition and area management module: The data acquisition and area management module is mainly used to realize the grid partition management of tunnel space, integrating multiple cameras to synchronously collect RGB-D images and temperature and humidity sensor data; Image processing and coordinate mapping module: The image processing and coordinate mapping module is mainly used to establish a geometric mapping relationship between the pixel coordinate system and the tunnel BIM model through adaptive perspective transformation to eliminate lens distortion; Contour line completion module: The contour line completion module is mainly used to interpolate and generate continuous contour line functions; Segment trend analysis module: The segment trend analysis module is mainly used to segment the contour line into characteristic segments according to the extreme points of curvature, and calculate and obtain the morphological evolution trend of each segment; Regional risk assessment module: The regional risk assessment module is mainly used to integrate the trend consistency analysis of the internal contour segments of the region and the spatial weights of the adjacent regions to calculate the probability of waterlogging disasters caused by the spread of accumulated water; Deep learning prediction module: The deep learning prediction module is mainly used to build an LSTM hybrid network model to predict the evolution of water depth based on the historical profile function time series; Danger level classification module: The hazard level classification module is mainly used to define three-dimensional assessment indicators and realize dynamic risk classification; Multi-level early warning response module: The multi-level early warning response module is mainly used to design a hierarchical early warning strategy that links threshold triggering with trend prediction, and supports the generation of linkage control instructions for the drainage system; Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.
8. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion as described in any one of claims 1-6.
9. A computer-readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by a processor, the intelligent early warning method for tunnel waterlogging disasters based on multi-source data fusion according to any one of claims 1 to 6 is implemented.