Security situation awareness model training method for hydrological information platform

Abstract

The invention relates to the technical field of hydrological safety perception, in particular to a safety situation perception model training method of a hydrological information platform. The ice flood abnormity perception model comprises a data input layer, a weather data identification layer, a river channel data identification layer, a water source data identification layer, an ice flood abnormity perception layer and a result output layer; the weather data identification layer predicts the weather data set to obtain weather prediction data; the river channel data recognition layer recognizes the river channel data set to obtain river channel feature data; the water source data identification layer identifies the weather prediction data and the water source data set to obtain water source prediction data; and the ice flood abnormity sensing layer identifies the weather prediction data, the river channel characteristic data and the water source prediction data to obtain an ice flood abnormity coefficient. According to the method, the ice flood abnormity perception model is obtained through construction and training, and river ice flood abnormity is accurately identified.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological safety perception, and in particular to a safety situation perception model training method for a hydrological information platform. Background Art

[0002] Ice flood refers to the hydrological phenomenon in which ice blocks the river channel and creates resistance to water flow, causing a significant rise in river water levels.

[0003] As the temperature gradually rises, the temperature of the river surface and the ice layer increases. When the temperature reaches above the freezing point, the ice layer begins to melt, become thinner and more fragile. When the ice layer breaks, ice blocks drift in the water and move in the river channel under the influence of the water flow. Floating ice blocks will accumulate in narrow river channels, bridge piers, river bends or near other obstacles, and even form ice dams, which will hinder the normal flow of the river and cause the water level to rise rapidly.

[0004] Once the ice dam is broken or the water level is too high, the accumulated water will be released instantly, causing ice floods and bringing serious water damage. This kind of flood usually occurs in a short period of time, but it is very destructive and may cause river bank erosion, property loss, farmland flooding, and affect the safety of infrastructure. In addition, ice floods may also have an impact on the ecosystem and change the habitat.

[0005] Therefore, understanding the formation mechanism of ice floods is necessary to help formulate effective flood control measures and reduce their impact on humans and the ecological environment; an effective monitoring and early warning system is an important means of preventing ice floods and can provide timely warnings of possible flood risks.

[0006] In the existing technology, ice floods are mainly identified through the direct impact of temperature changes on river ice in weather data and the breakup of ice. Factors other than ice melting and breakage, such as water source replenishment and riverbed characteristics, are not fully considered, resulting in low accuracy in the identification and prediction of ice dams.

[0007] Therefore, a security situation awareness model training method for hydrological information platform is proposed. Summary of the invention

[0008] The purpose of the present invention is to provide a method for training a safety situation awareness model of a hydrological information platform, including an ice flood anomaly perception model; the ice flood anomaly perception model includes a data input layer, a weather data recognition layer, a river data recognition layer, a water source data recognition layer, an ice flood anomaly perception layer and a result output layer; the weather data recognition layer predicts the weather data set to obtain weather forecast data; the river data recognition layer identifies the river data set to obtain river feature data; the water source data recognition layer identifies the weather forecast data and the water source data set to obtain water source forecast data; the ice flood anomaly perception layer identifies the weather forecast data, the river feature data and the water source forecast data to obtain the ice flood anomaly coefficient. The present invention obtains an ice flood anomaly perception model by constructing and training, and accurately identifies river ice flood anomalies.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] A safety situation awareness model training method for a hydrological information platform, including an ice flood anomaly awareness model;

[0011] The ice flood anomaly perception model includes a data input layer, a weather data recognition layer, a river data recognition layer, a water source data recognition layer, an ice flood anomaly perception layer and a result output layer;

[0012] The data input layer is used to input the collected ice flood data set into the model; the ice flood data set includes a weather data set, a river data set and a water source data set;

[0013] The weather data recognition layer is used to identify the weather data set and predict the weather in the future period according to the weather data set to obtain weather forecast data;

[0014] The river data recognition layer includes a riverbed recognition unit, a water flow recognition unit, a water quality recognition unit and an ice layer recognition unit, which respectively recognize the riverbed data, water flow data, water quality data and ice layer data in the river data set to obtain river channel characteristic data; the river channel characteristic data includes riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data;

[0015] The water source data identification layer identifies and predicts the water source replenishment of the river in the future cycle through weather forecast data and water source data sets to obtain water source prediction data; the water source prediction data includes water source access points, water source kinetic energy and water source potential energy;

[0016] The ice flood anomaly perception layer identifies weather forecast data, river channel characteristic data and water source forecast data to obtain an ice flood anomaly coefficient;

[0017] The result output layer is used to output the ice flood anomaly coefficient.

[0018] The ice flood data set includes a weather data set, a river data set and a water source data set;

[0019] The weather data set includes historical precipitation data, historical temperature data, historical wind direction data and historical wind speed data; the river data set includes riverbed data, water flow data, water quality data and ice layer data; the water source data set includes terrain data, surface water data and groundwater data.

[0020] The river data identification layer includes a riverbed identification unit, a water flow identification unit, a water quality identification unit and an ice layer identification unit;

[0021] The riverbed identification unit identifies the riverbed data to obtain riverbed characteristic data; the riverbed characteristic data includes riverbed geometric characteristics and riverbed surface characteristics;

[0022] The water flow identification unit identifies the water flow data to obtain water flow characteristic data; the water flow characteristic data includes flow data, flow velocity data, water level data and water temperature data;

[0023] The water quality identification unit identifies the water quality data to obtain water quality characteristic data;

[0024] The ice layer identification unit identifies the ice layer data to obtain ice layer characteristic data.

[0025] The identification process of the water quality identification unit is as follows:

[0026] Select multiple locations in the river as sampling points to obtain river water samples; perform detection and analysis on the river water samples to obtain the composition of the river water as water quality data;

[0027] The water quality data is identified to obtain water quality characteristic data; the water quality characteristic data includes water salinity, water turbidity and water nutrients.

[0028] The ice layer characteristic data includes fixed ice layer characteristic data and moving ice layer characteristic data;

[0029] Fixed ice is an ice layer that is fixed relative to the riverbed; the fixed ice characteristic data include ice thickness, ice texture and fixed ice surface temperature;

[0030] The moving ice layer is an ice layer that moves and drifts relative to the riverbed; the moving ice layer characteristic data includes the ice layer shape, ice layer speed and moving ice layer surface temperature.

[0031] The identification process of the ice layer characteristic data is as follows:

[0032] Acquiring ice layer data, wherein the ice layer data includes an ice layer image, ice layer thickness, and ice layer surface temperature;

[0033] Identifying the ice layer image and dividing it into a fixed ice layer area, a moving ice layer area and an ice-free area;

[0034] Perform image recognition on the fixed ice layer area to obtain the ice layer texture; then obtain the ice layer thickness and the fixed ice layer surface temperature in the fixed ice layer area;

[0035] Image recognition is performed on the moving ice layer area to obtain the ice layer shape and ice layer speed, and at the same time obtain the moving ice layer surface temperature.

[0036] The water source data identification layer includes a terrain identification unit, a precipitation water source identification unit, a ground water source identification unit, a ground water source identification unit and an ice layer water source identification unit;

[0037] The terrain recognition unit recognizes the terrain data in the water source data set to determine the river channel area;

[0038] The precipitation water source identification unit identifies the river area to obtain river distribution data; the precipitation directly falling into the river area is predicted through the river distribution data and weather forecast data to obtain precipitation forecast data;

[0039] The ground water source identification unit identifies and predicts the ground water source flowing into the river channel through the terrain distribution data and the ground water data to obtain the ground water prediction data;

[0040] The groundwater source identification unit identifies and predicts the groundwater source flowing into the river channel through the terrain distribution data and the groundwater data to obtain groundwater prediction data;

[0041] The ice layer water source identification unit identifies the melting of the ice layer according to the weather forecast data and the river channel characteristic data to obtain melt water prediction data;

[0042] Water source prediction data is obtained through the precipitation prediction data, surface water prediction data, groundwater prediction data and melt water prediction data.

[0043] The water source prediction data includes water source access point, water source kinetic energy and water source potential energy;

[0044] The water source access point is the location where the water source flows into the river;

[0045] The water source access point of precipitation water source is the river area of ​​precipitation; the water source access point of ground water source is the location where the ground water source merges into the river; the water source access point of groundwater source is the location where the groundwater source merges into the river; the water source access point of ice layer water source is the location of the ice layer.

[0046] The kinetic energy of the water source is obtained through the velocity of the water source;

[0047] The speed of groundwater sources and underground water sources is measured and predicted at the water source access point; the speed of ice layer water sources is obtained by predicting the speed of the ice layer; the speed of precipitation water sources is predicted by measuring the precipitation speed on the water surface;

[0048] The water source potential energy is obtained by the height difference between the water source and a reference position, and the reference position is obtained according to river distribution data;

[0049] The potential energy of groundwater sources and groundwater sources is obtained from the height difference between the water source access point and the control position; the potential energy of ice layer water sources is obtained from the height difference between the starting position of the ice layer and the control position; the potential energy of precipitation water sources is obtained through the height difference between the center point of the precipitation river area and the control position.

[0050] The training process of the ice flood anomaly perception model is as follows:

[0051] Initialize the model; obtain historical data of river ice flood; the historical data of ice flood includes weather data before ice flood, weather data after ice flood, ice flood water source data and ice flood river data;

[0052] The initialized ice flood anomaly perception model is trained and optimized through the ice flood historical data to obtain the ice flood anomaly perception model.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. The present invention identifies riverbed data through a riverbed identification unit to obtain riverbed characteristic data; identifies water flow data through a water flow identification unit to obtain water flow characteristic data; identifies water quality data through a water quality identification unit to obtain water quality characteristic data; identifies ice layer data through an ice layer identification unit to obtain ice layer characteristic data; accurately identifies river channel characteristics through riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data.

[0055] 2. The present invention identifies the terrain where the river is located to obtain the river area and the river influence area; identifies the terrain of the river influence area to obtain terrain distribution data; divides and predicts the water source of the river through the river area and terrain distribution data to obtain precipitation water source, ground water source, groundwater source and ice layer water source; and accurately identifies the water source replenishment situation of the river.

[0056] 3. The present invention identifies weather data sets to obtain weather forecast data; identifies weather forecast data, river channel characteristic data and water source forecast data, and obtains ice flood anomaly coefficients through weather factors, river channel factors and water source factors; and accurately predicts the probability of ice flood anomalies occurring in rivers in future periods. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a structural schematic diagram of the ice flood abnormality perception model of the present invention;

[0058] Figure 2 It is a structural schematic diagram of the river data identification layer of the present invention;

[0059] Figure 3 It is a structural schematic diagram of the water source data identification layer of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] Ice flood refers to the hydrological phenomenon that ice blocks the river channel and creates resistance to water flow, causing a significant rise in river water levels; ice floods may cause river bank erosion, property loss, and farmland flooding, affecting the safety of infrastructure. Therefore, it is necessary to understand the formation mechanism of ice floods in order to formulate effective flood control measures and reduce their impact on humans and the ecological environment. An effective monitoring and early warning system is an important means of preventing ice floods and can provide timely warnings of possible flood risks.

[0062] Therefore, a security situation awareness model training method for hydrological information platform is proposed.

[0063] Embodiment 1

[0064] The present invention proposes a safety situation awareness model training method for a hydrological information platform, including an ice flood anomaly awareness model;

[0065] The ice flood anomaly perception model is constructed based on a deep neural network model. The structure is as follows: Figure 1 As shown, it includes data input layer, weather data recognition layer, river data recognition layer, water source data recognition layer, ice flood anomaly perception layer and result output layer;

[0066] The data input layer is used to input the collected ice flood data set into the model; the ice flood data set includes a weather data set, a river data set and a water source data set;

[0067] The weather data recognition layer is used to identify the weather data set and predict the weather in the future period according to the weather data set to obtain weather forecast data;

[0068] The river data recognition layer includes a riverbed recognition unit, a water flow recognition unit, a water quality recognition unit and an ice layer recognition unit, which respectively recognize the riverbed data, water flow data, water quality data and ice layer data in the river data set to obtain river channel characteristic data; the river channel characteristic data includes riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data;

[0069] The water source data identification layer identifies and predicts the water source replenishment of the river in the future cycle through weather forecast data and water source data sets to obtain water source prediction data; the water source prediction data includes water source access points, water source kinetic energy and water source potential energy;

[0070] The ice flood anomaly perception layer identifies weather forecast data, river channel characteristic data and water source forecast data to obtain an ice flood anomaly coefficient;

[0071] The result output layer is used to output the ice flood anomaly coefficient.

[0072] The present invention collects a weather data set, a river channel data set and a water source data set of a river to obtain an ice flood data set; identifies the weather data set to obtain weather forecast data; identifies riverbed data, water flow data, water quality data and ice layer data in the river channel data set to obtain river channel characteristic data; identifies the weather forecast data and the water source data set to obtain water source forecast data; identifies the weather forecast data, the river channel characteristic data and the water source forecast data to obtain an ice flood anomaly coefficient; and accurately predicts the probability of an ice flood anomaly occurring in a river in a future period.

[0073] The ice flood data set includes a weather data set, a river data set and a water source data set;

[0074] The weather data set includes historical precipitation data, historical temperature data, historical wind direction data and historical wind speed data; the river data set includes riverbed data, water flow data, water quality data and ice layer data; the water source data set includes terrain data, surface water data and groundwater data.

[0075] The present invention collects weather data, river data and water source data of the area where the river is located, obtains weather data set, river data set and water source data, and provides a basis for abnormal perception of ice flood in the river.

[0076] The structure of the river data identification layer is as follows Figure 2 As shown, it includes a riverbed identification unit, a water flow identification unit, a water quality identification unit and an ice layer identification unit;

[0077] The riverbed identification unit identifies the riverbed data to obtain riverbed characteristic data; the riverbed characteristic data includes riverbed geometric characteristics and riverbed surface characteristics;

[0078] The riverbed geometric features include three-dimensional structural data of the riverbed, which are used to reflect width changes, depth changes, slope changes, grooves and depressions of the riverbed.

[0079] The riverbed surface characteristics include the particle size data of the riverbed surface; the particle size on the riverbed refers to particles of different sizes contained on the riverbed surface, which have different effects on water flow, sedimentation and ecosystem.

[0080] Common particle size compositions include: gravel, which is usually larger particles. The presence of gravel affects the flow rate and sedimentation of the river; sand, sandy riverbeds are usually smoother, which is conducive to the smooth distribution of water flow; silt and clay, fine particles, such riverbeds are usually flatter and the water flow rate is slower.

[0081] The water flow identification unit identifies the water flow data to obtain water flow characteristic data; the water flow characteristic data includes flow data, flow velocity data, water level data and water temperature data;

[0082] The water quality identification unit identifies the water quality data to obtain water quality characteristic data;

[0083] The ice layer identification unit identifies the ice layer data to obtain ice layer characteristic data.

[0084] The present invention identifies riverbed data through a riverbed identification unit to obtain riverbed characteristic data; identifies water flow data through a water flow identification unit to obtain water flow characteristic data; identifies water quality data through a water quality identification unit to obtain water quality characteristic data; identifies ice layer data through an ice layer identification unit to obtain ice layer characteristic data; and accurately understands river channel characteristics through riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data.

[0085] The identification process of the water quality identification unit is as follows:

[0086] Select multiple locations in the river as sampling points to obtain river water samples; perform detection and analysis on the river water samples to obtain the composition of the river water as water quality data;

[0087] The water quality data is identified to obtain water quality characteristic data; the water quality characteristic data includes water salinity, water turbidity and water nutrients.

[0088] Among them, the salinity of the water will affect the melting rate of the ice layer. The freezing point of high-salinity water is low, causing the ice layer to melt faster. Suspended matter in the water, such as mud, organic matter, etc., will affect the light transmittance of the water body, thereby affecting the change in water temperature. Turbid water bodies will absorb more heat and may cause faster melting. At the same time, dissolved oxygen and nutrients in the water body can affect the growth of aquatic organisms, and then affect the heat of the water body and the growth of ice, thereby changing the formation and melting process of ice.

[0089] The present invention samples and analyzes river water to obtain the composition of river water as water quality data; identifies the characteristics that affect the melting of ice layers in the water quality data to obtain water quality characteristic data; and accurately identifies the factors that affect the melting of ice layers in rivers.

[0090] The ice layer characteristic data includes fixed ice layer characteristic data and moving ice layer characteristic data;

[0091] The fixed ice layer is an ice layer that is fixed relative to the riverbed; the fixed ice layer characteristic data includes ice layer thickness, ice layer texture and fixed ice layer surface temperature; the moving ice layer is an ice layer that moves and drifts relative to the riverbed; the moving ice layer characteristic data includes ice layer shape, ice layer speed and moving ice layer surface temperature.

[0092] The identification process of the ice layer characteristic data is as follows:

[0093] Acquiring ice layer data, wherein the ice layer data includes an ice layer image and ice layer thickness;

[0094] Identifying the ice layer image and dividing it into a fixed ice layer area, a moving ice layer area and an ice-free area;

[0095] Perform image recognition on the fixed ice layer area to obtain the ice layer texture; then obtain the ice layer thickness and the fixed ice layer surface temperature in the fixed ice layer area;

[0096] Image recognition is performed on the moving ice layer area to obtain the ice layer shape and ice layer speed, and at the same time obtain the moving ice layer surface temperature.

[0097] Among them, the thickness of the ice layer is measured by ultrasonic equipment, and the surface temperature of the ice layer is identified by infrared technology.

[0098] The present invention identifies and divides ice layer data to obtain fixed ice layer areas, moving ice layer areas and ice-free areas; identifies fixed ice layer areas to obtain ice layer texture and fixed ice layer surface temperature; identifies moving ice layer areas to obtain ice layer shape, ice layer speed and moving ice layer surface temperature; and accurately identifies ice layer data in a river.

[0099] The structure of the water source data identification layer is as follows Figure 3As shown, it includes a terrain identification unit, a precipitation water source identification unit, a ground water source identification unit, a ground water source identification unit and an ice layer water source identification unit;

[0100] The terrain recognition unit recognizes the terrain data in the water source data set to determine the river channel area;

[0101] The precipitation water source identification unit identifies the river area to obtain river distribution data, and predicts the precipitation directly falling into the river area through the river distribution data and weather forecast data to obtain precipitation forecast data;

[0102] The ground water source identification unit identifies and predicts the ground water source flowing into the river channel through the terrain distribution data and the ground water data to obtain the ground water prediction data;

[0103] The groundwater source identification unit identifies and predicts the groundwater source flowing into the river through terrain distribution data and groundwater data to obtain groundwater prediction data;

[0104] The ice layer water source identification unit identifies the melting of the ice layer based on the weather forecast data and river channel characteristic data to obtain melt water forecast data;

[0105] Water source prediction data is obtained through the precipitation prediction data, surface water prediction data, groundwater prediction data and melt water prediction data.

[0106] The present invention identifies the terrain where the river is located to obtain river distribution data; obtains precipitation prediction data through river distribution data and weather prediction data; obtains surface water prediction data through terrain distribution data and surface water data, obtains groundwater prediction data through terrain distribution data and groundwater data, identifies the melting of ice layers through weather prediction data and river characteristic data, and obtains meltwater prediction data; and accurately identifies the water source replenishment situation of the river.

[0107] The water source prediction data includes water source access point, water source kinetic energy and water source potential energy;

[0108] The water source access point is the location where the water source flows into the river;

[0109] The water source access point of precipitation water source is the river area of ​​precipitation; the water source access point of ground water source is the location where the ground water source merges into the river; the water source access point of groundwater source is the location where the groundwater source merges into the river; the water source access point of ice layer water source is the location of the ice layer.

[0110] The kinetic energy of the water source is obtained through the velocity of the water source;

[0111] The speed of groundwater sources and underground water sources is measured and predicted at the water source access point; the speed of ice layer water sources is obtained by predicting the speed of the ice layer; the speed of precipitation water sources is predicted by measuring the precipitation speed on the water surface;

[0112] The water source potential energy is obtained by the height difference between the water source and a reference position, and the reference position is obtained according to river distribution data;

[0113] The potential energy of groundwater sources and groundwater sources is obtained from the height difference between the water source access point and the control position; the potential energy of ice layer water sources is obtained from the height difference between the starting position of the ice layer and the control position; the potential energy of precipitation water sources is obtained through the height difference between the center point of the precipitation river area and the control position.

[0114] The present invention obtains water source access points of different types of water sources, and identifies the kinetic energy and potential energy of the water source when it flows into a river according to the water source access points, thereby accurately identifying the energy of the water source.

[0115] The training process of the ice flood anomaly perception model is as follows:

[0116] Initialize the model,

[0117] Obtain historical data on river ice floods;

[0118] The ice flood historical data includes weather data before the ice flood, weather data after the ice flood, ice flood water source data and ice flood river data;

[0119] The initialized ice flood anomaly perception model is trained and optimized through the ice flood historical data to obtain the ice flood anomaly perception model.

[0120] The present invention collects data on rivers where ice floods occur to obtain ice flood historical data; and trains the ice flood anomaly perception model through the ice flood historical data to obtain an accurate ice flood anomaly perception model.

[0121] In order to verify the recognition effect of the ice flood anomaly perception model described in the present invention, a control experiment was carried out using data of different dimensions. The present invention constructed three control models in total; Model 1, Model 2, Model 3 and the ice flood anomaly perception model were verified to obtain the recognition accuracy; the data dimensions and recognition accuracy of the control model and the ice flood anomaly perception model are shown in Table 1.

[0122] Table 1. Control experiment data of ice flood anomaly perception model

[0123] Model River data Weather data Water source data Recognition accuracy Model 1 use Not adopted Not adopted 86.96％ Model 2 use use Not adopted 90.72％ Model 3 use Not adopted use 89.55％ Ice flood abnormality perception model use use use 93.26％

[0124] Among them, model one only uses river data to identify and train river ice flood anomalies; model two uses river data and weather data to identify and train ice flood anomalies; model three uses river data and water source data to identify and train ice flood anomalies. It can be seen from the data in Table 1 that the ice flood anomaly perception model constructed by the present invention has the highest recognition accuracy and the best performance in the ice flood anomaly identification process.

[0125] Through the safety situation awareness model training method of a hydrological information platform described in the present invention, an ice flood anomaly perception model can be obtained; through the ice flood anomaly perception model, ice flood anomalies in rivers can be accurately identified and warned, effectively ensuring the safety of river basins.

[0126] Embodiment 2

[0127] River A is a river that flows to the sea. Ice floods have occurred many times in the river, causing losses to both sides of the river. In order to better identify and predict river ice floods, a security situation awareness model training method for a hydrological information platform described in the present invention is adopted.

[0128] The safety situation awareness model training method of the hydrological information platform includes an ice flood anomaly perception model, and the training process of the ice flood anomaly perception model is:

[0129] Initialize the model,

[0130] Get the historical data of ice flood of river A;

[0131] The ice flood historical data include pre-ice flood weather data, post-ice flood weather data, ice flood water source data and ice flood river data; wherein, the pre-ice flood weather data is the weather data before the ice flood occurs in the river, including pre-ice flood precipitation data, pre-ice flood temperature data, pre-ice flood wind direction data and pre-ice flood wind speed data; the post-ice flood weather data is the weather data after the ice flood occurs in the river, including post-ice flood precipitation data, post-ice flood temperature data, post-ice flood wind direction data and post-ice flood wind speed data;

[0132] Ice flood water source data include ice flood topographic data, ice flood surface water data and ice flood groundwater data;

[0133] Ice flood river data include ice flood riverbed data, ice flood water flow data, ice flood water quality data and ice flood ice layer data;

[0134] The initialized ice flood anomaly perception model is trained and optimized through the ice flood historical data to obtain the ice flood anomaly perception model.

[0135] The present invention collects river ice flood data to obtain ice flood historical data, including pre-ice flood weather data, post-ice flood weather data, ice flood water source data and ice flood river data; trains and adjusts an initialized ice flood anomaly perception model through the ice flood historical data; and then identifies the collected ice flood data set through the trained ice flood anomaly perception model to accurately judge the probability of ice flood anomalies occurring in the river.

[0136] The ice flood abnormality perception model is as follows: Figure 1 As shown, it includes data input layer, weather data recognition layer, river data recognition layer, water source data recognition layer, ice flood anomaly perception layer and result output layer;

[0137] The data input layer is used to input the collected ice flood data set into the model; the ice flood data set includes a weather data set, a river data set and a water source data set;

[0138] The weather data recognition layer is used to identify the weather data set and predict the weather in the future period according to the weather data set; obtain weather forecast data; obtain the weather forecast data by identifying the weather data before the ice flood and the weather data after the ice flood;

[0139] The river data recognition layer includes a riverbed recognition unit, a water flow recognition unit, a water quality recognition unit and an ice layer recognition unit, which respectively recognize the riverbed data, water flow data, water quality data and ice layer data in the river data set to obtain river channel characteristic data; the river channel characteristic data includes riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data;

[0140] The water source data identification layer identifies and predicts the water source replenishment of the river in the future cycle through weather forecast data and water source data sets to obtain water source prediction data; the water source prediction data includes water source access points, water source kinetic energy and water source potential energy;

[0141] The ice flood anomaly perception layer identifies weather forecast data, river channel characteristic data and water source forecast data to obtain an ice flood anomaly coefficient;

[0142] The result output layer is used to output the ice flood anomaly coefficient.

[0143] The present invention collects a weather data set, a river channel data set and a water source data set of a river to obtain an ice flood data set; identifies the weather data set to obtain weather forecast data; identifies riverbed data, water flow data, water quality data and ice layer data in the river channel data set to obtain river channel characteristic data; identifies the weather forecast data and the water source data set to obtain water source forecast data; identifies the weather forecast data, the river channel characteristic data and the water source forecast data to obtain an ice flood anomaly coefficient; and accurately predicts the probability of an ice flood anomaly occurring in a river in a future period.

[0144] Among them, the ice flood data set includes the collected weather data set, river data set and water source data set; the weather data set includes historical precipitation data, historical temperature data, historical wind direction data and historical wind speed data; the river data set includes riverbed data, water flow data, water quality data and ice layer data; the water source data set includes terrain data, surface water data and groundwater data.

[0145] The river data identification layer includes a riverbed identification unit, a water flow identification unit, a water quality identification unit and an ice layer identification unit;

[0146] The riverbed identification unit identifies the riverbed data to obtain riverbed characteristic data; the riverbed characteristic data includes riverbed geometric characteristics and riverbed surface characteristics;

[0147] The water flow identification unit identifies the water flow data to obtain water flow characteristic data; the water flow characteristic data includes flow data, flow velocity data, water level data and water temperature data;

[0148] The water quality identification unit identifies the water quality data to obtain water quality characteristic data; the identification process of the water quality identification unit is: selecting multiple locations in the river as sampling points to obtain river water samples; detecting and analyzing the river water samples to obtain the components of the river water as water quality data; identifying the water quality data to obtain water quality characteristic data; the water quality characteristic data includes water salinity, water turbidity and water nutrients.

[0149] The ice layer identification unit is used to identify ice layer data to obtain ice layer characteristic data. The ice layer characteristic data includes fixed ice layer characteristic data and moving ice layer characteristic data; the fixed ice layer is an ice layer that is fixed relative to the riverbed; the fixed ice layer characteristic data includes ice layer thickness, ice layer texture and fixed ice layer surface temperature; the moving ice layer is an ice layer that moves and drifts relative to the riverbed; the moving ice layer characteristic data includes ice layer shape, ice layer speed and moving ice layer surface temperature.

[0150] The identification process of the ice layer characteristic data is as follows:

[0151] Acquiring ice layer data, wherein the ice layer data includes an ice layer image and ice layer thickness;

[0152] Identifying the ice layer image and dividing it into a fixed ice layer area, a moving ice layer area and an ice-free area;

[0153] Perform image recognition on the fixed ice layer area to obtain the ice layer texture; then obtain the ice layer thickness and the fixed ice layer surface temperature in the fixed ice layer area;

[0154] Image recognition is performed on the moving ice layer area to obtain the ice layer shape and ice layer speed, and at the same time obtain the moving ice layer surface temperature.

[0155] The riverbed data is identified by the riverbed identification unit to obtain riverbed characteristic data; the water flow data is identified by the water flow identification unit to obtain water flow characteristic data; the water quality data is identified by the water quality identification unit to obtain water quality characteristic data; the ice layer data is identified by the ice layer identification unit to obtain ice layer characteristic data; the river channel characteristics are accurately understood through the riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data.

[0156] The water source data identification layer includes a terrain identification unit, a precipitation water source identification unit, a ground water source identification unit, a ground water source identification unit and an ice layer water source identification unit;

[0157] The terrain recognition unit recognizes the terrain data in the water source data set to determine the river channel area;

[0158] The precipitation water source identification unit identifies the river area to obtain river distribution data, and predicts the precipitation directly falling into the river area through the river distribution data and weather forecast data to obtain precipitation forecast data;

[0159] The ground water source identification unit identifies and predicts the ground water source flowing into the river channel through the terrain distribution data and the ground water data to obtain the ground water prediction data;

[0160] The groundwater source identification unit identifies and predicts the groundwater source flowing into the river through terrain distribution data and groundwater data to obtain groundwater prediction data;

[0161] The ice layer water source identification unit identifies the melting of the ice layer based on the weather forecast data and river channel characteristic data to obtain melt water forecast data;

[0162] Water source prediction data is obtained through the precipitation prediction data, surface water prediction data, groundwater prediction data and melt water prediction data.

[0163] The water source prediction data includes water source access point, water source kinetic energy and water source potential energy;

[0164] The water source access point is the location where the water source flows into the river;

[0165] The water source access point of precipitation water source is the river area of ​​precipitation; the water source access point of ground water source is the location where the ground water source merges into the river; the water source access point of groundwater source is the location where the groundwater source merges into the river; the water source access point of ice layer water source is the location of the ice layer.

[0166] The kinetic energy of the water source is obtained through the velocity of the water source;

[0167] The speed of groundwater sources and underground water sources is measured and predicted at the water source access point; the speed of ice layer water sources is obtained by predicting the speed of the ice layer; the speed of precipitation water sources is predicted by measuring the precipitation speed on the water surface;

[0168] The water source potential energy is obtained by the height difference between the water source and a reference position, and the reference position is obtained according to river distribution data;

[0169] The potential energy of groundwater sources and groundwater sources is obtained from the height difference between the water source access point and the control position; the potential energy of ice layer water sources is obtained from the height difference between the starting position of the ice layer and the control position; the potential energy of precipitation water sources is obtained through the height difference between the center point of the precipitation river area and the control position.

[0170] Taking the ground water source of River A as an example, the access point of the ground water source is the intersection point where the ground water source is injected into River A. The velocity of the ground water source is obtained by measuring the water flow velocity at the intersection point. The potential energy of the ground water source is obtained by the height difference between the intersection point and the reference position. The reference position is the position of the estuary of River A. The ground water data of River A are shown in Table 2.

[0171] Table 2 Surface water data of River A

[0172] aboveground water source Water source speed Water flow Height difference Above ground water source 1 1.35m / s <![CDATA[87.61m 3 / s]]> 56.74m Above ground water source 2 2.82m / s <![CDATA[127.61m 3 / s]]> 84.08m Above ground water source 3 1.74m / s <![CDATA[105.61m 3 / s]]> 36.67m

[0173] The surface water data of the river A is used to predict the surface water data of the future period to obtain the surface water prediction data.

[0174] The present invention identifies the terrain where the river is located to obtain river distribution data; obtains precipitation prediction data through river distribution data and weather prediction data; obtains surface water prediction data through terrain distribution data and surface water data, obtains groundwater prediction data through terrain distribution data and groundwater data, identifies the melting of ice layers through weather prediction data and river characteristic data, and obtains meltwater prediction data; and accurately identifies the water source replenishment situation of the river.

[0175] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for training a safety situation awareness model of a hydrological information platform, including an ice flood anomaly perception model, characterized in that: The ice flood anomaly perception model includes a data input layer, a weather data recognition layer, a river data recognition layer, a water source data recognition layer, an ice flood anomaly perception layer and a result output layer; The data input layer is used to input the collected ice flood data set into the model; the ice flood data set includes a weather data set, a river data set and a water source data set; The weather data recognition layer is used to recognize the weather data set and predict the weather in the future period according to the weather data set to obtain weather prediction data; The river data recognition layer includes a riverbed recognition unit, a water flow recognition unit, a water quality recognition unit and an ice layer recognition unit, which respectively recognize the riverbed data, water flow data, water quality data and ice layer data in the river data set to obtain river channel characteristic data; the river channel characteristic data includes riverbed characteristic data, water flow characteristic data, water quality characteristic data and ice layer characteristic data; The water source data identification layer identifies and predicts the water source replenishment of the river in the future cycle through weather forecast data and water source data sets to obtain water source prediction data; the water source prediction data includes water source access points, water source kinetic energy and water source potential energy; The ice flood anomaly perception layer identifies weather forecast data, river channel characteristic data and water source forecast data to obtain an ice flood anomaly coefficient; The result output layer is used to output the ice flood anomaly coefficient.

2. The method for training a security situation awareness model for a hydrological information platform according to claim 1 is characterized in that: The ice flood data set includes a weather data set, a river data set and a water source data set; The weather data set includes historical precipitation data, historical temperature data, historical wind direction data and historical wind speed data; the river data set includes riverbed data, water flow data, water quality data and ice layer data; the water source data set includes terrain data, surface water data and groundwater data.

3. The method for training a security situation awareness model for a hydrological information platform according to claim 1 is characterized in that: The river data identification layer includes a riverbed identification unit, a water flow identification unit, a water quality identification unit and an ice layer identification unit; The riverbed identification unit identifies the riverbed data to obtain riverbed characteristic data; the riverbed characteristic data includes riverbed geometric characteristics and riverbed surface characteristics; The water flow identification unit identifies the water flow data to obtain water flow characteristic data; the water flow characteristic data includes flow data, flow velocity data, water level data and water temperature data; The water quality identification unit identifies the water quality data to obtain water quality characteristic data; The ice layer identification unit identifies the ice layer data to obtain ice layer characteristic data.

4. The method for training a security situation awareness model for a hydrological information platform according to claim 3 is characterized in that: The identification process of the water quality identification unit is as follows: Select multiple locations in the river as sampling points to obtain river water samples; perform detection and analysis on the river water samples to obtain the composition of the river water as water quality data; The water quality data is identified to obtain water quality characteristic data; the water quality characteristic data includes water salinity, water turbidity and water nutrients.

5. The method for training a security situation awareness model for a hydrological information platform according to claim 3 is characterized in that: The ice layer characteristic data includes fixed ice layer characteristic data and moving ice layer characteristic data; Fixed ice is an ice layer that is fixed relative to the riverbed; the fixed ice characteristic data include ice thickness, ice texture and fixed ice surface temperature; The moving ice layer is an ice layer that moves and drifts relative to the riverbed; the moving ice layer characteristic data includes the ice layer shape, ice layer speed and moving ice layer surface temperature.

6. The method for training a security situation awareness model for a hydrological information platform according to claim 5 is characterized in that: The identification process of the ice layer characteristic data is as follows: Acquiring ice layer data, wherein the ice layer data includes an ice layer image, ice layer thickness, and ice layer surface temperature; Identifying the ice layer image and dividing it into a fixed ice layer area, a moving ice layer area and an ice-free area; Perform image recognition on the fixed ice layer area to obtain the ice layer texture; Then, the ice thickness and the surface temperature of the fixed ice layer in the fixed ice layer area are obtained; Image recognition is performed on the moving ice layer area to obtain the ice layer shape and ice layer speed, and at the same time obtain the moving ice layer surface temperature.

7. The method for training a security situation awareness model for a hydrological information platform according to claim 1 is characterized in that: The water source data identification layer includes a terrain identification unit, a precipitation water source identification unit, a ground water source identification unit, a ground water source identification unit and an ice layer water source identification unit; The terrain recognition unit recognizes the terrain data in the water source data set to determine the river channel area; The precipitation water source identification unit identifies the river area to obtain river distribution data; the precipitation directly falling into the river area is predicted through the river distribution data and weather forecast data to obtain precipitation forecast data; The ground water source identification unit identifies and predicts the ground water source flowing into the river channel through the terrain distribution data and the ground water data to obtain the ground water prediction data; The groundwater source identification unit identifies and predicts the groundwater source flowing into the river channel through the terrain distribution data and the groundwater data to obtain groundwater prediction data; The ice layer water source identification unit identifies the melting of the ice layer based on the weather forecast data and the river channel characteristic data to obtain melt water forecast data; Water source prediction data is obtained through the precipitation prediction data, surface water prediction data, groundwater prediction data and melt water prediction data.

8. The method for training a security situation awareness model for a hydrological information platform according to claim 7, characterized in that: The water source prediction data includes water source access point, water source kinetic energy and water source potential energy; The water source access point is the location where the water source flows into the river; The water source access point of precipitation water source is the river area of ​​precipitation; the water source access point of ground water source is the location where the ground water source merges into the river; the water source access point of groundwater source is the location where the groundwater source merges into the river; the water source access point of ice layer water source is the location of the ice layer.

9. The method for training a security situation awareness model for a hydrological information platform according to claim 1, characterized in that: The kinetic energy of the water source is obtained through the velocity of the water source; The speed of groundwater sources and underground water sources is measured and predicted at the water source access point; the speed of ice layer water sources is obtained by predicting the speed of the ice layer; the speed of precipitation water sources is predicted by measuring the precipitation speed on the water surface; The water source potential energy is obtained by the height difference between the water source and a reference position, and the reference position is obtained according to river channel distribution data; The potential energy of groundwater sources and groundwater sources is obtained from the height difference between the water source access point and the control position; the potential energy of ice layer water sources is obtained from the height difference between the starting position of the ice layer and the control position; the potential energy of precipitation water sources is obtained through the height difference between the center point of the precipitation river area and the control position.

10. The method for training a security situation awareness model for a hydrological information platform according to claim 1, characterized in that: The training process of the ice flood anomaly perception model is as follows: Initialize the model, Acquire historical data of river ice floods; the historical data of ice floods include weather data before and after the ice floods, water source data of ice floods, and river channel data of ice floods; The initialized ice flood anomaly perception model is trained and optimized through the ice flood historical data to obtain the ice flood anomaly perception model.

Images