A state analysis method and system based on data set

Abstract

The present invention relates to the technical field of levee hidden danger detection, and in particular to a state analysis method and system based on a data set. The method comprises the following steps: using a laser radar to detect the stratigraphic lithology of the levee, obtaining a multidimensional dataset of the levee, using GIS software to perform three-dimensional visualization processing based on the multidimensional dataset, and constructing a stratigraphic analysis map; based on the stratigraphic analysis map, using an unmanned boat and a submersible to monitor and analyze the hydrological data downstream of the levee, and generating a high-definition remote sensing image of the downstream levee; based on the high-definition remote sensing image of the downstream levee, using the sensor upstream of the levee to obtain a real-time multimodal dataset upstream of the levee, and using the real-time multimodal dataset upstream of the levee to perform cloud-based aggregation operations on the levee; based on the cloud-based dataset upstream of the levee and the downstream high-definition remote sensing image, using the levee hydrological artificial intelligence algorithm to construct a comprehensive levee decision support system; the present invention improves the efficiency of levee hidden danger investigation and analysis by processing the levee hidden danger dataset.

Description

Technical Field

[0001] The present invention relates to the technical field of levee hidden danger detection, and in particular to a state analysis method and system based on a data set. Background Art

[0002] Based on the traditional levee detection method, the state analysis method and system based on the data set introduces various physical detection technologies, such as lidar, unmanned sensing equipment, etc., by reading and recording the physical characteristics of the levee, and then using computer technology to process and analyze the data, it finally realizes the rapid and accurate detection and evaluation of levee hidden dangers, thereby improving the safety and stability of the levee. At the same time, with the rapid development of computer technology and image processing technology, the application scope of the state analysis method and system based on the data set has been continuously expanded and improved, becoming an important means of levee detection, greatly improving the efficiency of levee hidden danger detection and improving work safety. Since the traditional levee detection method is mainly manual inspection and survey detection, which has low efficiency, poor accuracy and poor safety, there is an urgent need for an efficient, accurate, safe and reliable levee hidden danger detection method. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a state analysis method and system based on a data set to solve at least one of the above technical problems.

[0004] To achieve the above object, the present invention provides a state analysis method based on a data set, comprising the following steps:

[0005] Step S1: using laser radar to detect the stratigraphic lithology of the levee to obtain a multidimensional dataset of the levee, wherein the multidimensional dataset of the levee includes a stratigraphic lithology dataset, a levee contour dataset, and a terrain height dataset, and using GIS software to perform three-dimensional visualization processing on the multidimensional dataset to construct a stratigraphic analysis map;

[0006] Step S2: Based on the stratigraphic analysis diagram, unmanned boats and submersibles are used to monitor and analyze the hydrological data downstream of the dike, and obtain high-definition remote sensing images of the downstream of the dike;

[0007] Step S3: using the upstream sensor of the dike to obtain the real-time multimodal dataset of the upstream of the dike according to the high-definition remote sensing image of the downstream of the dike, and performing the dike cloud aggregation operation on the real-time multimodal dataset of the upstream of the dike to generate the dike cloud dataset;

[0008] Step S4: Based on the upstream cloud dataset of the levee and the downstream high-definition remote sensing images and using the levee hydrological artificial intelligence algorithm, a comprehensive levee decision support system is constructed to achieve all-round and multi-level monitoring and investigation of the levee's hidden danger status.

[0009] The present invention provides a state analysis method based on a data set. The method uses a stratigraphic lithology data set, a levee contour data set, a terrain height data set and GIS software to perform three-dimensional visualization processing of a multidimensional data set, and constructs a stratigraphic analysis map, thereby improving the accuracy and scope of levee monitoring. The stratigraphic analysis map is used to monitor and analyze the hydrological data downstream of the levee using an unmanned boat and a submersible, and a high-definition remote sensing image of the downstream of the levee is generated. The data are integrated for comprehensive analysis, and decision support is provided. Based on the stratigraphic analysis map and the high-definition remote sensing image of the downstream of the levee, the present invention can perform a cloud-based convergence operation on the levee to generate a cloud-based data set on the levee. A comprehensive decision support system for the levee is constructed using a levee hydrological artificial intelligence algorithm based on the cloud-based data set on the levee, thereby realizing all-round and multi-level monitoring and investigation of the hidden danger status of the levee. Compared with traditional manual investigation methods, the present invention has high security, high monitoring accuracy, high management efficiency, etc. Through a variety of modern technologies such as laser radar technology, unmanned boats, and submersibles, it can provide technical support for innovation in the field of levee safety monitoring, realize automated and digital monitoring, and realize comprehensive, timely, and accurate intelligent analysis of levee hidden danger detection.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: using laser radar to perform multi-dimensional detection of the stratum lithology of the levee to obtain a multi-dimensional data set of the stratum of the levee;

[0012] Step S12: performing data-driven modeling on the levee stratum multidimensional data set based on geostatistical methods to generate a three-dimensional levee physical model;

[0013] Step S13: Performing three-dimensional visualization processing using GIS software based on the three-dimensional levee physical model to construct a visualization view of the levee data;

[0014] Step S14: Use a machine vision recognition algorithm to perform machine vision recognition analysis on the levee data visualization view to construct a stratigraphic analysis diagram.

[0015] The present invention utilizes a levee stratum analysis method, performs multi-dimensional detection of stratum lithology of the levee through laser radar, generates a levee stratum multi-dimensional data set, performs data-driven modeling on the levee stratum multi-dimensional data set using a geostatistical method, generates a levee physical model, and on this basis, utilizes GIS software for three-dimensional visualization processing to construct a levee data visualization view, providing an intuitive visual expression for levee stratum analysis. At the same time, utilizes a machine vision recognition algorithm for machine vision recognition, and generates a stratum analysis diagram by analyzing the levee data visualization view. Compared with the traditional manual analysis method, the present invention greatly shortens the levee detection cycle, reduces the error rate in the manual analysis process, and improves the efficiency and accuracy of stratum analysis through the technical means of data-driven modeling and machine vision recognition algorithms.

[0016] Preferably, step S12 comprises the following steps:

[0017] Step S121: performing data cleaning on the stratum multidimensional data set to generate a dike normalized data set;

[0018] Step S122: using a multivariate statistical analysis method to perform coefficient of variation analysis on the dike normalized data set to generate a dike multivariate analysis data set;

[0019] Step S123: Calculate geological differences using geostatistical spatial interpolation methods based on the dike multivariate analysis data set to generate a stratigraphic data warehouse;

[0020] Step S124: Perform simulation analysis using numerical simulation based on the stratum data warehouse to generate a three-dimensional physical model of the levee;

[0021] The present invention effectively solves the problem of non-standard and incomplete stratigraphic data in the prior art by performing data cleaning on stratigraphic multidimensional datasets and generating standardized levee datasets. Multivariate statistical analysis methods are used to perform analytical calculations on coefficients of variation of standardized levee datasets, thereby improving the accuracy and efficiency of data analysis while ensuring data quality. Geological differences are calculated on multivariate analysis datasets of levees using geostatistical spatial interpolation methods, and the generated stratigraphic data warehouse has higher geological accuracy and reliability. Numerical simulation analysis is performed based on the stratigraphic data warehouse, and the generated three-dimensional physical model of the levee can simulate the levee structure more accurately, thereby improving the flood control and disaster resistance capabilities and safety.

[0022] Preferably, step S2 comprises the following steps:

[0023] Step S21: Divide the monitoring area according to the formation analysis map and generate a monitoring module table;

[0024] Step S22: extracting monitoring point parameter data according to the monitoring module table and using unmanned boats and submersible-carried detection equipment to generate a downstream hydrological detection data table for the dike;

[0025] Step S23: using the downstream dike regression analysis calculation formula to perform regression calculation on the downstream dike hydrological detection data table to generate a primary downstream dike remote sensing data set;

[0026] Step S24: Based on the primary levee downstream remote sensing data set, a pixel-level iterative classifier method is used to perform pixel classification calculations to obtain a high-definition remote sensing image of the levee downstream.

[0027] The present invention divides the monitoring area based on the stratigraphic analysis map, clearly divides the monitoring area, and helps to improve the monitoring accuracy and efficiency. At the same time, the generation of the monitoring module table facilitates the subsequent extraction and management of the monitoring point parameter data. The monitoring point parameter data extraction operation is performed by using the unmanned boat and the submersible carrying detection equipment, and the hydrological data table can be obtained. This information can be used to analyze the hydrological environment downstream of the levee, which is helpful to predict and evaluate the key parameters such as the water level and flow rate downstream of the levee, and improve the accuracy of monitoring. The regression calculation can be performed by the regression analysis calculation formula of the downstream levee, and the primary levee downstream remote sensing data set can be obtained. The primary levee downstream remote sensing data set is helpful to invert the surface coverage feature information based on the primary levee downstream remote sensing data set, and can provide an important data basis for the subsequent pixel-level iterative classifier. By using this method, the primary levee downstream remote sensing data set can be pixel classified and calculated to obtain a high-definition remote sensing image of the downstream levee. The high-definition remote sensing image of the downstream levee can show the landform and land use status of the downstream levee area, and provide valuable information for subsequent monitoring and management. The present invention establishes a set of efficient and accurate levee downstream monitoring methods, which is helpful to improve the safety management level of the downstream levee and reduce the risk of levee collapse.

[0028] Preferably, the downstream regression analysis calculation formula of the dike in step S23 is specifically:

[0029]

[0030] NDVI = (NIR-RED) / (NIR+RED);

[0031] NDBI=(SWIR-NIR) / (SWIR+NIR);

[0032] Among them, SF is the dike safety factor, T n is the dike terrain adjustment coefficient, n is the size of the fragmented terrain model dataset, H is the water level, HC is the dike top elevation, DC is the vertical distance from the downstream side of the dike top to the riverbed, NDVI is the normalized difference vegetation index, x i is the horizontal coordinate data of the dike terrain detection point from the unmanned boat and the submersible, y i is the vertical coordinate data of the embankment terrain detection point from the unmanned boat and the submersible, NDBI is the normalized building index, NIR is the reflectance value of the near infrared band, RED is the reflectance value of the red band, and SWIR is the reflectance value of the shortwave infrared band.

[0033] The present invention uses a downstream regression analysis calculation formula for the levee. By establishing a safety factor SF, it can effectively evaluate the flood resistance of the levee and reduce the safety risks caused by factors such as water level, levee top elevation, relative height, etc., thereby improving the safety performance of the levee. The levee terrain adjustment coefficient T nThe introduction of NDVI can detect and evaluate under different levee terrain conditions, so as to accurately judge the safety of the levee, adjust relevant parameters and improve the safety of the levee. The application of NDVI can evaluate the vegetation coverage along the levee. By monitoring the changes in the vegetation index, the ecological environment around the levee can be timely grasped, so as to judge the role of vegetation in the stability of the levee and provide a scientific basis for levee management and maintenance. The application of NDBI can evaluate the density and height of buildings from the levee to the river, and then judge which safety risks the buildings around the levee increase or reduce for the levee, providing a useful reference for levee management and planning. The combined use of multi-wavelength reflectance values ​​will better reflect the terrain conditions of the levee.

[0034] Preferably, step S3 comprises the following steps:

[0035] Step S31: deploy sensors upstream of the dike according to the high-definition remote sensing image of the downstream dike, and build a monitoring area upstream of the dike;

[0036] Step S32: monitoring the real-time multimodal data of the upstream of the dike according to the upstream monitoring area of ​​the dike, generating a real-time multimodal data set of the upstream of the dike;

[0037] Step S33: performing sequence calculation on the upstream real-time multimodal dataset of the dike using the upstream sequence weight analysis calculation formula of the dike to generate the upstream real-time multimodal model of the dike;

[0038] Step S34: Performing model homomorphic encryption based on the real-time multimodal model of the upstream of the dike to generate a homomorphic ciphertext data set;

[0039] Step S35: Perform cloud data aggregation processing based on the homomorphic ciphertext data set to generate a levee cloud data set.

[0040] The present invention uses high-definition remote sensing images to perform sensor deployment operations and builds a monitoring area upstream of the levee to increase the coverage area and monitoring efficiency of the upstream monitoring data of the levee, thereby effectively improving the safety of the levee. The upstream monitoring area of ​​the levee is used to monitor multiple signals and generate a real-time multimodal data set. The monitoring of the multimodal data set can improve the single data variable problem under the traditional monitoring method, and more comprehensively and accurately reflect the changes in the levee status, thereby improving the reliability and accuracy of levee safety monitoring. According to the real-time multimodal data set, the upstream sequence weight analysis calculation formula is used to perform sequence calculation to obtain a real-time multimodal model of the upstream levee, which can better reflect the relationship between multi-dimensional data. It has higher prediction accuracy and robustness to improve the accuracy and timeliness of levee safety monitoring. Homomorphic encryption technology is used to encrypt the real-time multimodal model of the upstream of the levee, and a homomorphic ciphertext data set is generated. The confidentiality and security of this data further improve the privacy protection and data security of the levee data, thereby avoiding possible data leakage and abuse. The homomorphic ciphertext data set is used for cloud data aggregation and processing, and a levee cloud data set is generated, so that the cloud can better integrate, analyze and process multimodal data from different sensors, so as to better and more timely monitor, feedback, warn and predict the stability and safety of the levee, improve the safety of the levee project, and thus reduce the pressure brought by manpower and material resources during the flood season.

[0041] Preferably, the calculation formula for the upstream sequence weight analysis of the dike in step S33 is specifically:

[0042]

[0043] Among them, Q is the time series weight analysis coefficient of the upstream of the levee, P is the rainfall, K1 is the regression weight coefficient of the first observation point, t is the time, T1 is the lag error coefficient of the first observation point, K2 is the regression weight coefficient of the second observation point, T2 is the lag error coefficient of the second observation point, K3 is the fixed water capacity coefficient of the upstream land, K4 is the area of ​​the upstream basin of the levee, and A is the area of ​​the downstream basin of the levee.

[0044] The present invention utilizes a weight analysis calculation formula for the upstream sequence of the levee, which fully considers that P is the rainfall, K1 is the regression weight coefficient of the first observation point, t is the time, T1 is the lag error coefficient of the first observation point, K2 is the regression weight coefficient of the second observation point, T2 is the lag error coefficient of the second observation point, K3 is the fixed water capacity coefficient of the upstream land, K4 is the area of ​​the upstream basin of the levee, and A is the area of ​​the upstream basin of the levee. The formula includes a variety of factors that affect the stability of the levee, involving the action and resultant force of multiple factors such as water, soil, and rock base. By calculating the action and resultant force of water and soil through K3*exp(-K4A), the stability of the levee can be evaluated more comprehensively and accurately, which is helpful to guide engineering design and inspection and improve the design quality of the levee. It represents the contribution of the second observation point to the weight coefficient, where the sine function with the regression weight coefficient K2 and the lag error coefficient T2 is multiplied by the factors related to the fixed water capacity K3 of the upstream land and the upstream basin area K4A of the dam. By analyzing the role of the embankment slope and the embankment soil type, the influence of the embankment slope and the embankment soil type on the embankment stability is considered. It can comprehensively analyze the influence of these factors on the embankment stability, and provide a useful reference for optimizing engineering design and strengthening protective measures.

[0045] Preferably, step S4 comprises the following steps:

[0046] Step S41: extracting stratigraphic features using an image processing algorithm based on the upstream cloud data set and the downstream high-definition remote sensing image of the dike to generate a stratigraphic feature matrix;

[0047] Step S42: Based on the stratum feature matrix, the levee hydrological artificial intelligence algorithm is used to perform deep learning on the levee, and a flow chart of the levee hydrological data model is generated;

[0048] Step S43: constructing a physical sequence process based on the dike hydrological data model flow chart using the dike hazard coefficient analysis formula to generate a dike hidden danger detection model;

[0049] Step S44: Based on the levee hidden danger detection model, the feature engineering method is used to fuse the triple data to generate a comprehensive levee decision support framework;

[0050] Step S45: Based on the levee comprehensive decision support framework, levee data is integrated to build a levee comprehensive decision support system.

[0051] The present invention uses an image processing algorithm to extract stratum features and generate a stratum feature matrix, which can accurately and efficiently obtain information on strata around the levee, thereby establishing a foundation for subsequent levee hidden danger detection and improving the accuracy and efficiency of detection. Through levee deep learning, model training can be combined with data-driven ideas, which greatly improves the detection accuracy and makes full use of the upstream cloud data set and downstream high-definition remote sensing images of the levee, making the detection data more accurate and reliable. According to the levee hazard coefficient analysis formula, by processing and analyzing the collected data, the levee hidden dangers can be more comprehensively grasped, thereby providing more reliable data support for subsequent detection. Through triple data fusion, data from different fields can be used to improve the credibility and accuracy of the data, thereby more comprehensively evaluating the safety of the levee. At the same time, the feature engineering method can also improve the accuracy of the model and mine more meaningful data features. The comprehensive decision support system for the levee can abstract data fusion and key parameters, making the detection results easier to understand and support decisions, thereby more accurately evaluating the safety of the levee and carrying out effective safety management.

[0052] Preferably, the levee hazard factor analysis formula in step S43 is specifically:

[0053]

[0054] Where FS is the safety factor of the levee, W is the weight of unit soil water in the levee hydrological data model, β is the slope of the levee, C is the unit weight of soil, γ is the unit weight of water, D is the water depth, Ks is the slope of the rock base, s is the sub-surface pressure coefficient, i is the levee top extension coefficient, Tz is the wave period, Do is the levee soil type and strength coefficient, sinβ represents the force of water on the levee surface, Ccosβ represents the gravity of the soil in the vertical direction, (1-sinβ) represents the resultant force of water and soil above the levee, and Kssinβ represents the downward force of the rock base.

[0055] The present invention provides a formula for analyzing and calculating the hazard factor of a levee, which fully considers that W is the weight of unit soil water content, β is the slope of the levee, C is the unit weight of soil, γ is the unit weight of water, D is the water depth, Ks is the slope of the rock base, s is the sub-surface pressure coefficient, c is the levee top extension coefficient, Tz is the wave period, Do is the levee soil type and strength coefficient, sinβ represents the force of water on the levee surface, Ccosβ represents the gravity of the soil in the vertical direction, (1-sinβ)γD represents the combined force of water and soil above the levee, and Kssinβ represents the downward force of the rock base , the forces of water and soil on the embankment surface are calculated by W(sinβ+Ccosβ), including the weight of water and the gravity of soil in the vertical direction, (1-sinβ)γD is used to calculate the resultant force of water and soil above the embankment, including the pressure of water on the embankment surface and the resistance of soil in the embankment body, and Kssinβ takes into account the downward force of the rock base, which can more accurately evaluate the stability of the embankment. The sub-surface pressure coefficient and the embankment crest extension coefficient are calculated by s*i*(Tz / Do), and the influence of wave period and embankment soil type and strength coefficient are considered, which can more comprehensively evaluate the stability and safety of the embankment.

[0056] In one embodiment of the present specification, a state analysis system based on a data set is provided, comprising:

[0057] at least one processor;

[0058] a memory communicatively connected to at least one processor; wherein,

[0059] The memory stores a computer program that can be executed by at least one processor. The computer program is executed by the at least one processor so that the at least one processor can perform the data set-based state analysis method as described above.

[0060] The present invention provides a hidden danger prevention and investigation and analysis system based on physical detection. The system uses physical detection technology to investigate and analyze hidden dangers of levees, which can improve the safety and stability of levees, avoid disasters such as floods, and protect people's lives and property. At the same time, the system can investigate levees over a large area in a short period of time, greatly improving the investigation efficiency, and can automatically perform the investigation and analysis process, reducing the cost of manual operation and personnel risks, and improving resource utilization efficiency. In addition, the system can monitor and warn of the hidden dangers of levees in real time, take timely measures to repair and reinforce them, and effectively prevent the occurrence of disasters caused by hidden dangers of levees. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A schematic diagram of the steps of a state analysis method based on a data set according to the present invention;

[0062] Figure 2 Detailed implementation flow chart of step S1;

[0063] Figure 3 Detailed implementation flow chart of step S2. DETAILED DESCRIPTION

[0064] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0065] The present application example provides a state analysis method and system based on a data set. The execution subjects of the dike hidden danger inspection and analysis method and system include but are not limited to: mechanical equipment, data processing platform, cloud server nodes, network transmission equipment, etc. equipped with the system can be regarded as general computing nodes of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0066] See also Figures 1 to 3 The present invention provides a state analysis method based on a data set, the method comprising the following steps:

[0067] Step S1: using laser radar to detect the stratigraphic lithology of the levee to obtain a multidimensional dataset of the levee, including a stratigraphic lithology dataset, a levee contour dataset, and a terrain height dataset; using GIS software to perform three-dimensional visualization processing on the multidimensional dataset to construct a stratigraphic analysis map;

[0068] Step S2: Based on the stratigraphic analysis diagram, unmanned boats and submersibles are used to monitor and analyze the hydrological data downstream of the dike, and obtain high-definition remote sensing images of the downstream of the dike;

[0069] Step S3: using the upstream sensor of the dike to obtain the real-time multimodal dataset of the upstream of the dike according to the high-definition remote sensing image of the downstream of the dike, and performing the dike cloud aggregation operation on the real-time multimodal dataset of the upstream of the dike to generate the dike cloud dataset;

[0070] Step S4: Based on the upstream cloud dataset of the levee and the downstream high-definition remote sensing images and using the levee hydrological artificial intelligence algorithm, a comprehensive levee decision support system is constructed to achieve all-round and multi-level monitoring and investigation of the levee's hidden danger status.

[0071] The present invention provides a state analysis method based on a data set. The method uses a stratigraphic lithology data set, a levee contour data set, a terrain height data set and GIS software to perform three-dimensional visualization processing of a multidimensional data set, and constructs a stratigraphic analysis map, thereby improving the accuracy and scope of levee monitoring. The stratigraphic analysis map is used to monitor and analyze the hydrological data downstream of the levee using an unmanned boat and a submersible, and a high-definition remote sensing image of the downstream of the levee is generated. The data are integrated for comprehensive analysis, and decision support is provided. Based on the stratigraphic analysis map and the high-definition remote sensing image of the downstream of the levee, the present invention can perform a cloud-based convergence operation on the levee to generate a cloud-based data set on the levee. A comprehensive decision support system for the levee is constructed using a levee hydrological artificial intelligence algorithm based on the cloud-based data set on the levee, thereby realizing all-round and multi-level monitoring and investigation of the hidden danger status of the levee. Compared with traditional manual investigation methods, the present invention has high security, high monitoring accuracy, high management efficiency, etc. Through a variety of modern technologies such as laser radar technology, unmanned boats, and submersibles, it can provide technical support for innovation in the field of levee safety monitoring, realize automated and digital monitoring, and realize comprehensive, timely, and accurate intelligent analysis of levee hidden danger detection.

[0072] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a state analysis method based on a data set of the present invention. In this example, the steps of the state analysis method based on a data set include:

[0073] Step S1: using laser radar to detect the stratigraphic lithology of the levee to obtain a multidimensional dataset of the levee, including a stratigraphic lithology dataset, a levee contour dataset, and a terrain height dataset; using GIS software to perform three-dimensional visualization processing on the multidimensional dataset to construct a stratigraphic analysis map;

[0074] In an embodiment of the present invention, for example, a laser radar device is used to emit a laser beam for scanning, and stratigraphic lithology detection is performed on the levee to obtain a multidimensional data set of the levee, including a stratigraphic lithology data set, levee contour data and terrain height data. The original data obtained from the laser radar device is preprocessed, including data denoising, data filtering and data registration. Through these processes, accurate original data is obtained. On the basis of the original data, combined with the results of field investigations and literature materials such as geological maps and stratigraphic columns, computer algorithms and manual interpretation methods are used to extract stratigraphic lithology information to form a stratigraphic data set. Multidimensional data sets such as the stratigraphic lithology data set, the levee contour data set and the terrain height data set are integrated, and data format conversion and data correction are performed. GIS software is used to perform three-dimensional visualization processing on the levee multidimensional data set, and multidimensional data sets such as the stratigraphic lithology data set, the levee contour data set and the terrain height data set are spatially superimposed to form stratigraphic analysis maps of different colors, different heights and different thicknesses, thereby generating a stratigraphic analysis map of the levee.

[0075] Step S2: Based on the stratigraphic analysis diagram, unmanned boats and submersibles are used to monitor and analyze the hydrological data downstream of the dike, and obtain high-definition remote sensing images of the downstream of the dike;

[0076] In an embodiment of the present invention, for example, surveying instruments and other equipment are used to perform stratigraphic analysis in an area that needs to be monitored, and various data related to the geological structure of the area are obtained to generate a monitoring module table. Based on the monitoring module table, unmanned boats and submersibles are used to perform remote control by calling the control system of the remote sensing cloud platform to perform ferrying. The control system of the remote sensing cloud platform is called to provide a carrier detection device to extract monitoring point parameter data, generate a hydrological detection data table downstream of the dike, and use a computer to perform image processing on the data based on the hydrological detection data table downstream of the dike to generate a primary remote sensing data set downstream of the dike. Based on the primary remote sensing data set downstream of the dike, a pixel-level iterative classifier method is used to perform pixel classification calculations to generate a high-definition remote sensing image downstream of the dike.

[0077] Step S3: using the upstream sensor of the dike to obtain the real-time multimodal dataset of the upstream of the dike according to the high-definition remote sensing image of the downstream of the dike, and performing the dike cloud aggregation operation on the real-time multimodal dataset of the upstream of the dike to generate the dike cloud dataset;

[0078] In an embodiment of the present invention, for example, sensors are deployed upstream of the dike. The sensors may include but are not limited to meteorological sensors, hydrological sensors, geological sensors and other types. A monitoring area upstream of the dike is established. Real-time multimodal data of the upstream of the dike is obtained through the dike sensors to generate a real-time multimodal dataset of the upstream of the dike. According to the high-definition remote sensing image of the downstream of the dike, the real-time multimodal dataset of the upstream of the dike is obtained by using the upstream sensors of the dike. A real-time multimodal model of the upstream of the dike is generated by using deep learning and other technologies. The established real-time multimodal model is homomorphically encrypted, and the original dataset is converted into a homomorphic ciphertext dataset to ensure the confidentiality and security of the data. The homomorphic ciphertext dataset is subjected to cloud data aggregation processing, including but not limited to cloud computing technology and big data technology, to integrate data from different sources, and to perform data processing and analysis to generate a dike cloud dataset.

[0079] Step S4: Based on the upstream cloud dataset of the levee and the downstream high-definition remote sensing images and using the levee hydrological artificial intelligence algorithm, a comprehensive levee decision support system is constructed to achieve all-round and multi-level monitoring and investigation of the levee's hidden danger status.

[0080] In an embodiment of the present invention, for example, the data collected by various sensors are uploaded to the cloud in real time using perception data acquisition technology. For high-definition remote sensing images downstream of the levee, the characteristics of the levee's surrounding environment, including various vegetation, roads, buildings, etc., are extracted through image processing technology. At the same time, remote sensing technology is used to scan the internal conditions of the levee to understand the internal conditions of the levee. Then, the levee hydrological artificial intelligence algorithm is used to analyze and process the data upstream and downstream of the levee, and characteristic information related to the levee is extracted. Based on the levee characteristic information, a levee safety assessment is performed to achieve early warning and investigation of levee hidden dangers. Finally, a comprehensive levee decision support system is constructed to achieve all-round and multi-level monitoring and investigation of the levee's hidden danger status, timely discover and warn of hidden dangers in the levee, and provide effective decision support for levee managers.

[0081] In the embodiment of the present invention, reference Figure 2 The above is a flowchart of the detailed implementation steps of step S1. In one embodiment of the present specification, the detailed implementation steps of step S1 include:

[0082] Step S11: using laser radar to perform multi-dimensional detection of the stratum lithology of the levee to obtain a multi-dimensional data set of the stratum of the levee;

[0083] Step S12: performing data-driven modeling on the levee stratum multidimensional data set based on geostatistical methods to generate a three-dimensional levee physical model;

[0084] Step S13: Performing three-dimensional visualization processing using GIS software based on the three-dimensional levee physical model to construct a visualization view of the levee data;

[0085] Step S14: Use a machine vision recognition algorithm to perform machine vision recognition analysis on the levee data visualization view to construct a stratigraphic analysis map.

[0086] The present invention provides a levee stratum analysis method, which uses laser radar to perform multi-dimensional detection of stratum lithology of the levee to generate a multi-dimensional data set of levee strata, uses a geological statistical method to perform data-driven modeling on the multi-dimensional data set of levee strata, and generates a three-dimensional physical model of the levee. On this basis, GIS software is used to perform three-dimensional visualization processing to construct a visualization view of the levee data, which provides an intuitive visual expression for the levee stratum analysis. At the same time, machine vision recognition is performed using a machine vision recognition algorithm, and a stratum analysis diagram is generated by analyzing the visualization view of the levee data. Compared with the traditional manual analysis method, the present invention greatly shortens the levee detection cycle, reduces the error rate in the manual analysis process, and improves the efficiency and accuracy of stratum analysis through the technical means of data-driven modeling and machine vision recognition algorithms.

[0087] In an embodiment of the present invention, a laser radar is used to emit a laser beam to scan the levee to obtain levee stratum data, a data acquisition module is used to acquire levee stratum data, the acquired data is uploaded to a data processing module, the data processing module is used to screen, process and organize the acquired data based on an algorithm to form a levee stratum multidimensional data set, a data import module is used to receive the levee stratum multidimensional data set, the data is transmitted to a geological data modeling module, the geological data modeling module performs data mining and modeling based on a geostatistical method to generate a three-dimensional levee physical model, the model output module outputs the generated three-dimensional levee physical model, the data processing module performs secondary processing and processing on the three-dimensional levee physical model based on GIS software to form a levee data visualization view, the data import module is used to receive the generated levee data visualization view, the data is transmitted to an image decomposition module, the image decomposition module performs image decomposition on the levee data visualization view based on image processing technology to form multiple sub-images, and then the feature extraction module performs feature extraction on each sub-image based on a machine vision recognition algorithm to obtain a feature vector of the sub-image, and the analysis output module performs cluster analysis on all the obtained sub-image feature vectors to construct a stratum analysis diagram.

[0088] In one embodiment of this specification, the specific steps of step S12 are:

[0089] Step S121: performing data cleaning on the stratum multidimensional data set to generate a dike normalized data set;

[0090] Step S122: using a multivariate statistical analysis method to perform coefficient of variation analysis on the dike normalized data set to generate a dike multivariate analysis data set;

[0091] Step S123: Calculate geological differences using geostatistical spatial interpolation methods based on the dike multivariate analysis data set to generate a stratigraphic data warehouse;

[0092] Step S124: Perform simulation analysis using numerical simulation based on the stratum data warehouse to generate a three-dimensional physical model of the levee;

[0093] The present invention effectively solves the problem of non-standard and incomplete stratigraphic data in the prior art by performing data cleaning on stratigraphic multidimensional datasets and generating standardized levee datasets. Multivariate statistical analysis methods are used to perform analytical calculations on coefficients of variation of standardized levee datasets, thereby improving the accuracy and efficiency of data analysis while ensuring data quality. Geological differences are calculated on multivariate analysis datasets of levees using geostatistical spatial interpolation methods, and the generated stratigraphic data warehouse has higher geological accuracy and reliability. Numerical simulation analysis is performed based on the stratigraphic data warehouse, and the generated three-dimensional physical model of the levee can simulate the levee structure more accurately, thereby improving the flood control and disaster resistance capabilities and safety.

[0094] In an embodiment of the present invention, for example, a data acquisition module is used to collect multidimensional data of levee strata, and the collected data is transmitted to a data cleaning module. The data cleaning module performs data preprocessing and cleaning operations to obtain a standardized levee data set. A data import module is used to receive the levee standardized data set, and the data is transmitted to a multivariate statistical analysis module. The multivariate statistical analysis module is used to analyze the levee standardized data set based on a coefficient of variation analytical calculation method to generate a levee multivariate analysis data set. The analysis output module outputs the generated levee multivariate analysis data set, and the data is transmitted to a geostatistical spatial interpolation module. The geostatistical spatial interpolation module is used to interpolate the levee multivariate analysis data set based on a spatial interpolation method to obtain a stratum data warehouse. The warehouse generation module stores the generated stratum data warehouse in a database, and a model construction module is used to construct a three-dimensional physical model of the levee based on the stratum data warehouse and modeling principles.

[0095] In the embodiment of the present invention, reference Figure 3 The above is a flowchart of the detailed implementation steps of step S2. In one embodiment of the present specification, the detailed implementation steps of step S2 include:

[0096] Step S21: Divide the monitoring area according to the formation analysis map and generate a monitoring module table;

[0097] Step S22: extracting monitoring point parameter data according to the monitoring module table and using unmanned boats and submersible-carried detection equipment to generate a downstream hydrological detection data table for the dike;

[0098] Step S23: using the downstream dike regression analysis calculation formula to perform regression calculation on the downstream dike hydrological detection data table to generate a primary downstream dike remote sensing data set;

[0099] Step S24: Based on the primary levee downstream remote sensing data set, a pixel-level iterative classifier method is used to perform pixel classification calculations to obtain a high-definition remote sensing image of the levee downstream.

[0100] The present invention divides the monitoring area based on the stratigraphic analysis map, clearly divides the monitoring area, and helps to improve the monitoring accuracy and efficiency. At the same time, the generation of the monitoring module table facilitates the subsequent extraction and management of the monitoring point parameter data. The monitoring point parameter data extraction operation is performed by using the unmanned boat and the submersible carrying detection equipment, and the hydrological data table can be obtained. This information can be used to analyze the hydrological environment downstream of the levee, which is helpful to predict and evaluate the key parameters such as the water level and flow rate downstream of the levee, and improve the accuracy of monitoring. The regression calculation can be performed by the regression analysis calculation formula of the downstream levee, and the primary levee downstream remote sensing data set can be obtained. The primary levee downstream remote sensing data set is helpful to invert the surface coverage feature information based on the primary levee downstream remote sensing data set, and can provide an important data basis for the subsequent pixel-level iterative classifier. By using this method, the primary levee downstream remote sensing data set can be pixel classified and calculated to obtain a high-definition remote sensing image of the downstream levee. The high-definition remote sensing image of the downstream levee can show the landform and land use status of the downstream levee area, and provide valuable information for subsequent monitoring and management. The present invention establishes a set of efficient and accurate levee downstream monitoring methods, which is helpful to improve the safety management level of the downstream levee and reduce the risk of levee collapse.

[0101] In the example of the present invention, for example, the monitoring area is divided into several monitoring modules using the stratigraphic analysis map downstream of the dike, and a monitoring module table is generated. According to the monitoring module table, the monitoring points are regularly collected using unmanned boats and submersibles to carry detection equipment, and the water level, flow rate, water temperature and other parameters of the monitoring points are obtained through image processing technology, and the data are stored in the hydrological detection data table downstream of the dike. Based on the regression analysis calculation formula of the dike downstream and the hydrological detection data table downstream of the dike, regression calculation is performed to obtain a primary remote sensing data set downstream of the dike, wherein the primary remote sensing data set downstream of the dike includes water The depth, sediment concentration, water transparency, water temperature, color and other parameter data are collected and classified according to the primary remote sensing data set of the downstream of the dike. The pixel-level iterative classifier method is used to perform pixel classification calculations, and image preprocessing is performed, such as edge extraction, mean filtering and other operations to distinguish water, land, sediment, etc., and finally a high-definition remote sensing image of the downstream of the dike is obtained. Feature extraction algorithms, such as principal component analysis and wavelet transform, are used to obtain the feature information of the image. The image is further classified through classification algorithms, such as support vector machine and K nearest neighbor, to obtain high-quality remote sensing images of the downstream of the dike.

[0102] In the embodiment of the present invention, the downstream regression analysis calculation formula of the dike in step S23 is specifically:

[0103]

[0104] NDVI = (NIR-RED) / (NIR+RED);

[0105] NDBI=(SWIR-NIR) / (SWIR+NIR);

[0106] Among them, SF is the dike safety factor, T n is the dike terrain adjustment coefficient, n is the size of the fragmented terrain model dataset, H is the water level, HC is the dike top elevation, DC is the vertical distance from the downstream side of the dike top to the riverbed, NDVI is the normalized difference vegetation index, x i is the horizontal coordinate data of the dike terrain detection point from the unmanned boat and the submersible, y i is the vertical coordinate data of the embankment terrain detection point from the unmanned boat and the submersible, NDBI is the normalized building index, NIR is the reflectance value of the near infrared band, RED is the reflectance value of the red band, and SWIR is the reflectance value of the shortwave infrared band.

[0107] The present invention provides a downstream regression analysis calculation formula for the levee. By establishing a safety factor SF, the flood resistance capacity of the levee can be effectively evaluated, and the safety risks caused by factors such as water level, levee top elevation, and relative height can be reduced, thereby improving the safety performance of the levee. The levee terrain adjustment coefficient T n The introduction of NDVI can detect and evaluate under different levee terrain conditions, so as to accurately judge the safety of the levee, adjust relevant parameters and improve the safety of the levee. The application of NDVI can evaluate the vegetation coverage along the levee. By monitoring the changes in the vegetation index, the ecological environment around the levee can be timely grasped, so as to judge the role of vegetation in the stability of the levee and provide a scientific basis for levee management and maintenance. The application of NDBI can evaluate the density and height of buildings from the levee to the river, and then judge which safety risks the buildings around the levee increase or reduce for the levee, providing a useful reference for levee management and planning. The combined use of multi-wavelength reflectance values ​​will better reflect the terrain conditions of the levee.

[0108] In one embodiment of the present specification, step S3 includes the following steps:

[0109] Step S31: deploy sensors upstream of the dike according to the high-definition remote sensing image of the downstream dike, and build a monitoring area upstream of the dike;

[0110] Step S32: monitoring the real-time multimodal data of the upstream of the dike according to the upstream monitoring area of ​​the dike, generating a real-time multimodal data set of the upstream of the dike;

[0111] Step S33: performing sequence calculation on the upstream real-time multimodal dataset of the dike using the upstream sequence weight analysis calculation formula of the dike to generate the upstream real-time multimodal model of the dike;

[0112] Step S34: Performing model homomorphic encryption based on the real-time multimodal model of the upstream of the dike to generate a homomorphic ciphertext data set;

[0113] Step S35: Perform cloud data aggregation processing based on the homomorphic ciphertext data set to generate a levee cloud data set.

[0114] In the example of the present invention, the position and range of the upstream monitoring area of ​​the levee are determined according to the high-definition remote sensing image of the downstream of the levee, and a sensor deployment plan for the upstream monitoring area of ​​the levee is designed, including the type, quantity, and position of sensors. According to the sensor deployment plan, the sensors are wired, installed, and debugged to ensure that the sensors can work normally. A data acquisition, transmission, and processing system for the upstream monitoring area of ​​the levee is built, and the data collected by the sensors are processed and stored. Real-time data acquisition is performed on each sensor in the upstream monitoring area of ​​the levee. The data collected by multiple sensors are integrated and processed to generate a real-time multimodal data set for the upstream of the levee. The real-time data set is preprocessed, such as data cleaning, denoising, filtering, etc., to improve the data quality and accuracy. The results are obtained by extracting features from the data to obtain feature vectors. The sequence weight analysis calculation formula is used to perform sequence calculation on the feature vectors to generate sequence weights. According to the sequence weights and feature vectors, machine learning and other methods are used to build models to generate a real-time multimodal model of the upstream of the levee. According to the existing real-time multimodal model of the upstream of the levee, the model is encrypted using a homomorphic encryption algorithm. The encrypted model is applied to the real-time data set to generate a homomorphic ciphertext data set, which is uploaded to the cloud server. On the cloud server, the homomorphic ciphertext data set is decrypted and aggregated to obtain the levee cloud data set. The cloud data set is analyzed and processed to extract useful information, such as water level, flow velocity, rainfall and other indicators upstream of the levee, and the data is visualized and displayed.

[0115] In one example of this specification, the calculation formula for the upstream sequence weight analysis of the dike in step S33 is specifically:

[0116]

[0117] Among them, Q is the time series weight analysis coefficient of the upstream of the levee, P is the rainfall, K1 is the regression weight coefficient of the first observation point, t is the time, T1 is the lag error coefficient of the first observation point, K2 is the regression weight coefficient of the second observation point, T2 is the lag error coefficient of the second observation point, K3 is the fixed water capacity coefficient of the upstream land, K4 is the area of ​​the upstream basin of the levee, and A is the area of ​​the downstream basin of the levee.

[0118] The present invention provides a weight analysis calculation formula for the upstream sequence of the levee, which fully considers that P is the rainfall, K1 is the regression weight coefficient of the first observation point, t is the time, T1 is the lag error coefficient of the first observation point, K2 is the regression weight coefficient of the second observation point, T2 is the lag error coefficient of the second observation point, K3 is the fixed water capacity coefficient of the upstream land, K4 is the area of ​​the upstream basin of the levee, and A is the area of ​​the upstream basin of the levee. The formula includes a variety of factors that affect the stability of the levee, involving the action and resultant force of multiple factors such as water, soil, and rock base. By calculating the action and resultant force of water and soil through K3*exp(-K4*A), the stability of the levee can be evaluated more comprehensively and accurately, which is helpful to guide engineering design and inspection and improve the design quality of the levee. It represents the contribution of the second observation point to the weight coefficient, where the sine function with the regression weight coefficient K2 and the lag error coefficient T2 is multiplied by the factors related to the fixed water capacity K3 of the upstream land and the upstream basin area K4*A of the dam. By analyzing the role of the embankment slope and the embankment soil type, the influence of the embankment slope and the embankment soil type on the embankment stability is considered. It can comprehensively analyze the influence of these factors on the embankment stability, and provide a useful reference for optimizing engineering design and strengthening protective measures.

[0119] In one embodiment of the present specification, step S4 includes the following steps:

[0120] Step S41: extracting stratigraphic features using an image processing algorithm based on the upstream cloud data set and the downstream high-definition remote sensing image of the dike to generate a stratigraphic feature matrix;

[0121] Step S42: Based on the stratum feature matrix, the levee hydrological artificial intelligence algorithm is used to perform deep learning on the levee, and a flow chart of the levee hydrological data model is generated;

[0122] Step S43: constructing a physical sequence process based on the dike hydrological data model flow chart using the dike hazard coefficient analysis formula to generate a dike hidden danger detection model;

[0123] Step S44: Based on the levee hidden danger detection model, the feature engineering method is used to fuse the triple data to generate a comprehensive levee decision support framework;

[0124] Step S45: Based on the levee comprehensive decision support framework, levee data is integrated to build a levee comprehensive decision support system.

[0125] The present invention uses an image processing algorithm to extract stratum features and generate a stratum feature matrix, which can accurately and efficiently obtain information on strata around the levee, thereby establishing a foundation for subsequent levee hidden danger detection and improving the accuracy and efficiency of detection. Through levee deep learning, model training can be combined with data-driven ideas, which greatly improves the detection accuracy and makes full use of the upstream cloud data set and downstream high-definition remote sensing images of the levee, making the detection data more accurate and reliable. According to the levee hazard coefficient analysis formula, by processing and analyzing the collected data, the levee hidden dangers can be more comprehensively grasped, thereby providing more reliable data support for subsequent detection. Through triple data fusion, data from different fields can be used to improve the credibility and accuracy of the data, thereby more comprehensively evaluating the safety of the levee. At the same time, the feature engineering method can also improve the accuracy of the model and mine more meaningful data features. The comprehensive decision support system for the levee can abstract data fusion and key parameters, making the detection results easier to understand and support decisions, thereby more accurately evaluating the safety of the levee and carrying out effective safety management.

[0126] In the example of the present invention, various environmental data obtained by multiple sensors at different locations are used, and the data is pre-processed by an image processing algorithm, the image data is converted into a digital matrix, and the formation feature information is extracted, which is integrated with the environmental data to generate a formation feature matrix, and the formation feature matrix is ​​input into the artificial intelligence algorithm, processed and a hydrological data model is established, and the levee hydrological artificial intelligence algorithm is used to perform deep learning of the levee, and a flow chart of the levee hydrological data model is generated. Based on the flow chart of the levee hydrological data model, a physical sequence process is constructed using a levee hazard coefficient analysis formula to generate a levee hidden danger detection model, and based on the results output by the levee hidden danger detection model and other related data, a feature engineering method is used to perform triple data fusion to generate a levee comprehensive decision support framework, and the results output by the levee comprehensive decision support framework are fused with other data sources, such as manual inspection data, Internet of Things data, historical monitoring data, etc. The data is stored in a database, and the data is visualized through an application to realize the construction of a comprehensive levee decision support system.

[0127] In one example of this specification, the levee hazard factor analysis calculation formula in step S43 is specifically:

[0128]

[0129] Where FS is the safety factor of the levee, W is the weight of unit soil water in the levee hydrological data model, β is the slope of the levee, C is the unit weight of soil, γ is the unit weight of water, D is the water depth, Ks is the slope of the rock base, s is the sub-surface pressure coefficient, i is the levee top extension coefficient, Tz is the wave period, Do is the levee soil type and strength coefficient, sinβ represents the force of water on the levee surface, Ccosβ represents the gravity of the soil in the vertical direction, (1-sinβ) represents the resultant force of water and soil above the levee, and Kssinβ represents the downward force of the rock base.

[0130] The present invention provides a formula for analyzing and calculating the hazard factor of a levee, which fully considers that W is the weight of unit soil water content, β is the slope of the levee, C is the unit weight of soil, γ is the unit weight of water, D is the water depth, Ks is the slope of the rock base, s is the sub-surface pressure coefficient, c is the levee top extension coefficient, Tz is the wave period, Do is the levee soil type and strength coefficient, sinβ represents the force of water on the levee surface, Ccosβ represents the gravity of the soil in the vertical direction, (1-sinβ)γD represents the combined force of water and soil above the levee, and Kssinβ represents the downward force of the rock base , the forces of water and soil on the embankment surface are calculated by W(sinβ+Ccosβ), including the weight of water and the gravity of soil in the vertical direction, (1-sinβ)γD is used to calculate the resultant force of water and soil above the embankment, including the pressure of water on the embankment surface and the resistance of soil in the embankment body, and Kssinβ takes into account the downward force of the rock base, which can more accurately evaluate the stability of the embankment. The sub-surface pressure coefficient and the embankment crest extension coefficient are calculated by s*i*(Tz / Do), and the influence of wave period and embankment soil type and strength coefficient are considered, which can more comprehensively evaluate the stability and safety of the embankment.

[0131] In one embodiment of the present specification, a state analysis system based on a data set is provided, comprising:

[0132] processor;

[0133] a memory communicatively connected to at least one processor; wherein,

[0134] The memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor.

[0135] At least one processor executes so that the at least one processor can execute any of the above methods for detecting and analyzing embankment hazards.

[0136] The present invention provides a state analysis system based on a data set. The system uses physical detection technology to check and analyze hidden dangers of levees, which can improve the safety and stability of levees and avoid disasters such as floods. At the same time, the system can check levees over a large area in a short period of time, greatly improving the inspection efficiency, and can automatically perform the inspection and analysis process, reducing the cost of manual operation and personnel risks, and improving resource utilization efficiency. In addition, the system can monitor and warn of the hidden dangers of levees in real time, take measures to repair and reinforce them in time, and effectively prevent the occurrence of disasters caused by hidden dangers of levees.

[0137] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0138] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0139] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A state analysis method based on a data set, characterized in that: The following steps are involved: Step S1: using laser radar to detect the stratigraphic lithology of the levee to obtain a multidimensional dataset of the levee, wherein the multidimensional dataset of the levee includes a stratigraphic lithology dataset, a levee contour dataset, and a terrain height dataset, and using GIS software to perform three-dimensional visualization processing on the multidimensional dataset to construct a stratigraphic analysis map; Step S2: Based on the stratigraphic analysis diagram, unmanned boats and submersibles are used to monitor and analyze the hydrological data downstream of the dike, and obtain high-definition remote sensing images of the downstream of the dike; Step S3: According to the high-definition remote sensing image of the downstream dike, the upstream dike sensor is used to obtain the upstream dike real-time multimodal dataset, and the upstream dike real-time multimodal dataset is aggregated on the dike cloud to generate the dike cloud dataset; the specific steps of step S3 are: Step S31: deploy sensors upstream of the dike according to the high-definition remote sensing image of the downstream dike, and build a monitoring area upstream of the dike; Step S32: monitoring the real-time multimodal data of the upstream of the dike according to the upstream monitoring area of ​​the dike, generating a real-time multimodal data set of the upstream of the dike; Step S33: Perform sequence calculation on the upstream real-time multimodal dataset of the dike using the upstream sequence weight analysis calculation formula to generate the upstream real-time multimodal model of the dike; the upstream time series weight analysis calculation formula of the dike in step S33 is specifically: Where Q is the time series weight analysis coefficient of the upstream of the levee, P is rainfall, K1 is the regression weight coefficient of the first observation point, t is time, T1 is the lag error coefficient of the first observation point, K2 is the regression weight coefficient of the second observation point, T2 is the lag error coefficient of the second observation point, K3 is the fixed water capacity coefficient of the upstream land, K4 is the area of ​​the upstream basin of the levee, and A is the area of ​​the downstream basin of the levee; Step S34: Performing model homomorphic encryption based on the real-time multimodal model of the upstream of the dike to generate a homomorphic ciphertext data set; Step S35: performing cloud data aggregation processing based on the homomorphic ciphertext data set to generate a levee cloud data set; Step S4: Based on the upstream cloud dataset of the levee and the downstream high-definition remote sensing images and using the levee hydrological artificial intelligence algorithm, a comprehensive levee decision support system is constructed to achieve all-round and multi-level monitoring and investigation of the levee's hidden danger status.

2. The state analysis method based on data set according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: using laser radar to perform multi-dimensional detection of the stratum lithology of the levee to obtain a multi-dimensional data set of the stratum of the levee; Step S12: performing data-driven modeling on the levee stratum multidimensional data set based on geostatistical methods to generate a three-dimensional levee physical model; Step S13: Performing three-dimensional visualization processing using GIS software based on the three-dimensional levee physical model to construct a visualization view of the levee data; Step S14: Use a machine vision recognition algorithm to perform machine vision recognition analysis on the levee data visualization view to construct a stratigraphic analysis diagram.

3. The state analysis method based on data set according to claim 2, characterized in that: The specific steps of step S12 are: Step S121: performing data cleaning on the stratum multidimensional data set to generate a dike normalized data set; Step S122: using a multivariate statistical analysis method to perform coefficient of variation analysis on the dike normalized data set to generate a dike multivariate analysis data set; Step S123: Calculate geological differences using geostatistical spatial interpolation methods based on the dike multivariate analysis data set to generate a stratigraphic data warehouse; Step S124: Perform simulation analysis using numerical simulation based on the stratum data warehouse to generate a three-dimensional physical model of the levee.

4. The state analysis method based on data set according to claim 2, characterized in that: The specific steps of step S2 are: Step S21: Divide the monitoring area according to the formation analysis map and generate a monitoring module table; Step S22: extracting monitoring point parameter data according to the monitoring module table and using unmanned boats and submersible-carried detection equipment to generate a downstream hydrological detection data table for the dike; Step S23: using the downstream dike regression analysis calculation formula to perform regression calculation on the downstream dike hydrological detection data table to generate a primary downstream dike remote sensing data set; Step S24: Based on the primary levee downstream remote sensing data set, a pixel-level iterative classifier method is used to perform pixel classification calculations to obtain a high-definition remote sensing image of the levee downstream.

5. The state analysis method based on data set according to claim 4, characterized in that: The downstream regression analysis calculation formula of the dike in step S23 is specifically: NDVI=(NIR-RED) / (NIR+RED); NDBI=(SWIR-NIR) / (SWIR+NIR); Among them, SF is the levee safety factor, T n is the dike terrain adjustment coefficient, n is the size of the fragmented terrain model dataset, H is the water level, HC is the dike top elevation, DC is the vertical distance from the downstream side of the dike top to the riverbed, NDVI is the normalized difference vegetation index, x i is the horizontal coordinate data of the dike terrain detection point from the unmanned boat and the submersible, y i is the vertical coordinate data of the embankment terrain detection point from the unmanned boat and the submersible, NDBI is the normalized building index, NIR is the reflectance value of the near infrared band, RED is the reflectance value of the red band, and SWIR is the reflectance value of the shortwave infrared band.

6. The state analysis method based on data set according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: extracting stratigraphic features using an image processing algorithm based on the upstream cloud data set and the downstream high-definition remote sensing image of the dike to generate a stratigraphic feature matrix; Step S42: Based on the stratum feature matrix, the levee hydrological artificial intelligence algorithm is used to perform deep learning on the levee to generate a levee hydrological data model; Step S43: constructing a physical sequence process based on the dike hydrological data model using the dike hazard coefficient analysis formula to generate a dike hidden danger detection model; Step S44: Based on the levee hidden danger detection model, the feature engineering method is used to fuse the triple data to generate a comprehensive levee decision support framework; Step S45: Based on the levee comprehensive decision support framework, levee data is integrated to build a levee comprehensive decision support system.

7. The state analysis method based on data set according to claim 6, characterized in that: The specific formula for analyzing the embankment hazard coefficient in step S43 is: Where FS is the safety factor of the levee, W is the weight of unit soil water in the levee hydrological data model, β is the slope of the levee, C is the unit weight of soil, γ is the unit weight of water, D is the water depth, Ks is the slope of the rock base, s is the sub-surface pressure coefficient, i is the levee top extension coefficient, Tz is the wave period, Do is the levee soil type and strength coefficient, sinβ represents the force of water on the levee surface, Ccosβ represents the gravity of the soil in the vertical direction, (1-sinβ)γD represents the resultant force of water and soil above the levee, and Kssinβ represents the downward force of the rock base.

8. A state analysis system based on a data set, characterized in that: include: at least one processor; a memory communicatively coupled to the at least one processor; The memory stores a computer program executable by at least one processor, and the computer program is executed by at least one processor so as to enable the at least one processor to perform the data set-based state analysis method according to any one of claims 1 to 7.

Images