A river water level jump monitoring method applicable to artificial rivers

Abstract

The present invention provides a river water level jump monitoring method that can be used for artificial rivers, obtains radar remote sensing data images, establishes a circular buffer zone to extract radar backscatter coefficient data, and establishes a radar backscatter coefficient variation data set; uses GPM precipitation data to interpolate radar backscatter coefficient variation data and obtain radar backscatter coefficient variation and precipitation data sets with continuous time series characteristics; based on the obtained high-time resolution radar backscatter coefficient variation data, interpolated precipitation data, and the established ten-level water level labels, uses the SVM algorithm to establish a river water level jump classification discrimination model; subsequently, variables such as radar backscatter coefficients at any time are obtained and input into the trained model, so that the river water level fluctuation level and early warning information can be obtained in real time. The present invention can solve the problems of limited data sources, low temporal and spatial resolution, and difficulty in discrimination in artificial river water level monitoring and early warning research.

Description

Technical Field

[0001] The invention relates to the research field of river water level monitoring, and in particular to a river water level jump monitoring method that can be used for artificial rivers. Background Art

[0002] River water level, as an important indicator of the hydrological cycle and a basic component of water balance, is the main basis for formulating water resources management and ecological environmental protection measures. It has great application value in warning of heavy rainfall and combating natural disasters such as floods and droughts.

[0003] Frequent natural disasters have become one of the major challenges facing the world today. Disasters such as floods and typhoons continue to affect people's lives and property safety, so disaster monitoring has become urgent. In this context, the rapid progress of satellite remote sensing technology has provided unprecedented support for the monitoring of disaster changes.

[0004] Satellite remote sensing technology, with its all-weather, long-term and high-precision characteristics, has brought revolutionary changes to disaster monitoring. Among them, radar remote sensing has the characteristics of strong penetration of clouds and fog, which greatly makes up for the shortcomings of optical remote sensing images that are easily disturbed and seriously missing data. In addition, river water level is closely related to precipitation. The introduction of precipitation data combined with satellite remote sensing technology can realize the monitoring of surface changes. However, the return cycle of radar remote sensing images is long, and it is difficult to realize the monitoring of actual water level changes and real-time prevention and control and early warning of disasters in 1 to 2 weeks. In addition, with the development of modernization of water conservancy projects, artificial rivers account for an increasing proportion of total rivers. The profile of artificial rivers is vertical. Compared with current scientific research, remote sensing monitoring of water level and runoff based on river width (non-vertical profile) is not applicable.

[0005] Among them, Patent 1: A river runoff monitoring inversion method based on multi-source radar remote sensing technology uses river width data extracted from synthetic aperture radar (SAR) images and relative water depth data extracted from radar altimeter data, and uses the Manning formula to calculate river flow. This method has several shortcomings. First, runoff monitoring based on river width data is not suitable for artificial rivers. Second, due to the limitation of relying on radar altimeter data, this method is only applicable to river data within the area covered by the radar altimeter. Third, because the interpolation method is not used in this patent, which will result in sparse data, it limits the possibility of long-term river flow monitoring.

[0006] Patent 2: A method for monitoring the dynamic flow of water based on GNSS reflection waves. This patent uses the global navigation satellite system (GNSS) reflection wave data to interpolate the short-time series river width data to synthesize a river width data set with long-time series characteristics. The main differences from the river water level jump monitoring method provided by the present invention that can be used for artificial rivers are: first, the mechanism is different. It interpolates the river width data through GNSS reflection waves, and then predicts the water level through the river width; second, the achievable effects are different. The river width extraction method cannot be used for artificial rivers. The method of this patent application does not rely on the river width parameters and can be used for jump monitoring of artificial rivers. In addition, the comparative documents are mainly aimed at the acquisition of water level information in natural rivers. This patent application judges the specific jump level of the water level and achieves different purposes.

[0007] Therefore, this patent uses the change of radar backscatter coefficient, which is sensitive to roughness changes, to capture the rise and fall of river water level. This method is not limited by the width of the river and is particularly suitable for artificial rivers with vertical profiles, and can effectively monitor the jump of water flow. Summary of the invention

[0008] To solve the above problems, the present invention discloses a river water level jump monitoring method that can be used for artificial rivers, so as to provide precise, near real-time monitoring and early water level warning information; the present invention can not only track the formation process of floods, but also give early warning and respond to the occurrence of disasters in a timely manner, effectively reducing the losses caused by disasters; and provide all-weather, highly accurate and timely dynamic feature descriptions for various rivers, so as to better cope with the challenges of frequent natural disasters.

[0009] Technical solution: The present invention provides a method for monitoring river water level jumps that can be used in artificial rivers, comprising:

[0010] Step 1: Obtain radar remote sensing image data at the observation location within the observation date;

[0011] Step 2: Obtain precipitation data based on the Global Precipitation Project (GPM) dataset at the observation location during the observation date, and interpolate it into a 3 km grid;

[0012] Step 3: Obtain wind speed data based on the fifth generation ECMWF atmospheric reanalysis global climate data within the observation date at the observation location;

[0013] Step 4: Obtain the global 30-meter SRTM elevation DEM data for the observation location and observation date;

[0014] Step 5: Preprocess the acquired radar image data, including orbit correction, radiation calibration, filtering, terrain correction and decibel preprocessing steps;

[0015] Step 6: Based on the radar remote sensing image data, a circular buffer with a radius of 500m is established with the observation site as the center, and the radar backscatter coefficient data set corresponding to the observation date within the range is extracted;

[0016] Step 7: Based on the radar backscatter coefficient dataset, generate a radar backscatter coefficient change dataset in combination with the date;

[0017] Step 8: Obtain the measured water level data at the observation site during the observation date, and establish ten water level labels based on the average value, maximum value, minimum value and variance statistical indicators of the measured data, including five levels of rising water and five levels of falling water;

[0018] Step 9: Extract the corresponding radar backscatter coefficient variation data from the acquired time series remote sensing image data, and interpolate the acquired radar backscatter coefficient variation data set in combination with the acquired continuous date precipitation data to obtain the radar backscatter coefficient variation data with continuous dates;

[0019] Step 10: Perform spatiotemporal matching of the water level label information of the corresponding date measured by the hydrological station with the radar backscatter coefficient change data, precipitation data, wind speed data and elevation data obtained based on the interpolated satellite remote sensing image; based on the radar backscatter coefficient change, precipitation data, wind speed data, elevation data, date, longitude, latitude data, and the ten established water level labels, namely, five levels of rising water and five levels of falling water, use the SVM algorithm to establish a river water level jump classification model;

[0020] Step 11: Subsequently, the change in radar backscatter coefficient, precipitation data, wind speed data, elevation data, date, longitude and latitude data at any time are obtained and substituted into the river water level jump classification model to conduct real-time monitoring of river water level jumps and obtain real-time water level warning information.

[0021] Specifically, the GPM precipitation data is processed based on the bilinear interpolation method. The calculation process of bilinear interpolation is as follows:

[0022] In a two-dimensional grid, suppose there are four points with coordinates and corresponding values:

[0023]

[0024] Among them, Q11, Q21, Q12, and Q22 are four known points, which are located at (x1, y1), (x2, y1), (x1, y2), and (x2, y2) on the coordinate axis respectively.

[0025] The calculation formula for bilinear interpolation is as follows:

[0026]

[0027] Among them, (x, y) are the coordinates of the target point to be interpolated, f(x, y) is the interpolation result of the point, and f(Q11), f(Q21), f(Q12), and f(Q22) are the function values ​​of four known points respectively.

[0028] Specifically, N radar remote sensing image data of certain dates D1 (year / month / day) are obtained, and a series of remote sensing image preprocessing processes such as radiation correction and atmospheric calibration are required;

[0029] Specifically, the radar backscatter coefficient data obtained include the longitude and latitude of the center coordinates of the observation location {X, Y}, and the radar backscatter coefficient of the observation point , the radar backscatter coefficient is calculated using the following formula:

[0030]

[0031] The system parameters include wavelength , angle of incidence , Polarization , the target parameters include the complex dielectric constant , surface roughness , Volume scattering coefficient in inhomogeneous media .

[0032] Based on this, this patent proposes a water level information extraction method that is different from the traditional method. Usually, the rise and fall of water levels is determined based on the change in river width. This method is mainly applicable to natural wide rivers with non-vertical river profiles. However, the method used in this patent is to capture the rise and fall of river water levels through changes in radar backscatter coefficients. This method is not limited by the width of the river and is particularly suitable for artificial rivers with vertical profiles. It can effectively monitor the jump of water flow;

[0033] Specifically, the historical sequence is formed according to the time D1 (year / month / day) of the acquired radar remote sensing image data as follows:

[0034]

[0035] in, Indicates longitude and latitude, represents the radar backscatter coefficient of the observation point at time t1, and L represents the number of dates in the observation time series.

[0036] Specifically, based on the radar backscatter coefficient dataset, the radar backscatter coefficient change dataset is generated in combination with the date:

[0037] ,

[0038] in, Indicates longitude and latitude, Indicates The change of radar backscatter coefficient,

[0039] Right now

[0040] Where L represents the number of dates in the observed time series. and Both represent the time series number of the radar backscatter coefficient dataset.

[0041] Specifically, the time-space matching of radar backscatter coefficient change data, precipitation data, wind speed data and elevation data is performed, and the precipitation data, wind speed data and elevation data of the corresponding time are found according to the date with radar remote sensing image data, so as to obtain the radar backscatter coefficient change data and its matching precipitation data, wind speed data and elevation data. Specifically, in the radar remote sensing image date D1, the precipitation data, wind speed data and elevation data corresponding to the date D1 are searched, so as to extract the corresponding radar backscatter coefficient change data, precipitation data, wind speed data and elevation data, so as to obtain the radar backscatter coefficient change data, precipitation data, wind speed data and elevation data of N D1 dates.

[0042] Specifically, the measured data of the hydrological station include date D2 and water level. The intersection of date D2 and date D1 is found to obtain the radar backscatter coefficient change data corresponding to date D2, interpolated precipitation data, wind speed data and elevation data. Combined with the longitude and latitude data corresponding to the hydrological station, the date, longitude, latitude, radar backscatter coefficient change, precipitation data, wind speed data and elevation data set after time and space matching is established.

[0043] Specifically, a river water level jump classification model is established through the SVM algorithm: first, according to the average measured water level in the study area, as well as the maximum and minimum water levels, a classification operation is performed, which is divided into five levels of rising water and five levels of falling water, with a total of ten labels; secondly, all inputs and labels are matched for training and modeling, that is, the river water level jump classification model is obtained;

[0044] Specifically, for the dates with LN remote sensing image radar backscatter coefficient changes missing, the missing precipitation data is used. Multiply by the radar backscatter coefficient change of the previous day for discrete or continuous interpolation and complement, and obtain L radar backscatter coefficient change data with consecutive dates D1.

[0045] Specifically, based on the data values ​​of the change in radar backscatter coefficient of L consecutive dates, combined with the continuous time series of date, longitude and latitude, daily average precipitation data, wind speed data and elevation data, and the established river water level jump classification model, continuous, high temporal and spatial resolution, and near real-time water level label information can be acquired.

[0046] Beneficial effects of the present invention:

[0047] 1. The change of radar backscatter coefficient can be used to effectively identify the change of water level in non-natural river channels, and at the same time, continuous and uninterrupted long-term series of water level change information can be obtained with high time resolution; it can solve the problems of limited data source, low time resolution and difficult monitoring in river water level research.

[0048] 2. Usually, the rise and fall of water level is determined based on the change of river width. This method is mainly applicable to natural wide river channels with non-vertical river channel profiles; however, the method adopted by this patent is to capture the rise and fall of river water level through the change of radar backscatter coefficient. This method is not limited by the width of the river, and is particularly suitable for artificial rivers with vertical profiles, and can effectively monitor the jump of water flow.

[0049] 3. This patent chooses to establish a circular buffer to extract radar backscatter coefficients. Compared with the traditional square buffer, the use of a circular buffer can maintain the spatial consistency of the analysis area, so that the backscatter characteristics of a given area can be more unified and standardized. The circular buffer can provide an area with higher spatial uniformity for data analysis; due to the geometric characteristics of the circle, it can provide an equidistant boundary, and the distance extending from the center point in all directions is equal, which helps to maintain the uniformity of the extracted backscatter coefficients and improve the reliability of the analysis results. The traditional square buffer, although it can also provide a clearly defined area around the target, because the corner area is farther from the center to the edge, this may introduce additional variability during the analysis, resulting in more unnecessary background information being included in the analysis, thereby affecting the calculation and interpretation of the backscatter coefficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A schematic flow chart of a method for monitoring river water level jumps in an artificial river provided in an embodiment of the present invention;

[0051] Figure 2 This is a water level distribution map of the northern Jiangsu irrigation canal area predicted by the present invention based on the river water level jump classification model.

[0052] Figure 3 This is a regional water level distribution map of the northern Jiangsu irrigation canal predicted based on the background technology. DETAILED DESCRIPTION

[0053] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the accompanying drawings, and the words "inner" and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0054] See also Figure 1 , which is a flow chart of a method for monitoring river water level jumps in artificial rivers provided in this embodiment, and is described in detail in combination with specific steps.

[0055] Step S101, obtaining radar remote sensing image data of northern Jiangsu in 2022;

[0056] Step S102, obtaining precipitation data for northern Jiangsu in 2022 based on the Global Satellite Precipitation Project (GPM) dataset, and interpolating it into a 3km grid;

[0057] In the specific implementation, the calculation process of bilinear interpolation is as follows:

[0058]

[0059] Among them, Q11, Q21, Q12, and Q22 are four known points, which are located at (x1, y1), (x2, y1), (x1, y2), and (x2, y2) on the coordinate axis respectively.

[0060] The calculation formula for bilinear interpolation is as follows:

[0061]

[0062] Among them, (x, y) are the coordinates of the target point to be interpolated, f(x, y) is the interpolation result of the point, and f(Q11), f(Q21), f(Q12), and f(Q22) are the function values ​​of four known points respectively.

[0063] Step S103, obtaining wind speed data for northern Jiangsu in 2022 based on the fifth-generation ECMWF atmospheric reanalysis global climate data;

[0064] Step S104, obtaining the global 30-meter SRTM elevation DEM data for northern Jiangsu in 2022;

[0065] Step S105, performing data preprocessing on the acquired radar image data, including preprocessing processes such as radiation correction, geometric correction, and image enhancement;

[0066] In the specific implementation, 161 radar remote sensing image data of northern Jiangsu in 2022 were obtained, and then a series of remote sensing image preprocessing processes such as radiation correction and atmospheric calibration were carried out;

[0067] Step S106, based on the radar remote sensing image data, a circular buffer zone with a radius of 500m is established with 13 hydrological stations in northern Jiangsu as the center, and radar backscatter coefficient data in each buffer zone throughout 2022 is extracted;

[0068] The radar backscatter coefficient data obtained include the longitude and latitude of the hydrological station coordinates , and the radar backscatter coefficient of the observation point , the radar backscatter coefficient can be calculated using the following formula:

[0069]

[0070] The system parameters include wavelength , angle of incidence , Polarization , the target parameters include the complex dielectric constant , surface roughness , Volume scattering coefficient in inhomogeneous media .

[0071] In the specific implementation, the historical sequence is formed according to the time sequence of the acquired radar remote sensing image data as follows:

[0072]

[0073] in, Indicates The central longitude and latitude of the hydrological station, Indicates Hydrological station Radar backscatter coefficient of the image.

[0074] Step S107, generating a radar backscatter coefficient variation dataset based on the radar backscatter coefficient dataset and in combination with the date;

[0075]

[0076] in, represents the central longitude and latitude of the th hydrological station, Indicates The first hydrological station The change of radar backscatter coefficient,

[0077] Right now

[0078] in, and Both represent the time series number of the radar backscatter coefficient dataset.

[0079] Step S108, obtaining measured water level data of 13 hydrological stations in northern Jiangsu, and establishing ten water level labels according to the average value, maximum value and minimum value of the measured data, namely, five levels of rising water and five levels of falling water;

[0080] Step S109, based on the acquired radar backscatter coefficient variation data set, combined with the precipitation data with continuous time series characteristics, the acquired radar backscatter coefficient variation data set is interpolated to obtain radar backscatter coefficient variation data with continuous time series characteristics;

[0081] In the specific implementation, for the dates when the change in the backscatter coefficient of the remote sensing image radar of each hydrological station is missing, the missing precipitation data is used:

[0082]

[0083] By multiplying the change in radar backscatter coefficient of the previous day and performing discrete or continuous interpolation, 4745 radar backscatter coefficient change data with continuous time series characteristics can be obtained.

[0084] Step S110, the measured water level label information of the hydrological station is matched in time and space with the radar backscatter coefficient change data, interpolated precipitation data, wind speed data and elevation data obtained based on satellite remote sensing images; combined with the matched date, longitude and latitude data, based on the established ten water level labels, namely, five levels of rising water and five levels of falling water, a river water level jump classification model is established using the SVM algorithm; first, according to the average value of the measured water level in the study area, as well as the maximum and minimum water levels, a classification operation is performed, which is divided into five levels of rising water and five levels of falling water, that is, a total of ten labels; secondly, all inputs and labels are matched for training and modeling, that is, a river water level jump classification model is obtained;

[0085] In the specific implementation, the time-space matching of radar backscatter coefficient change data, precipitation data, wind speed data and elevation data is carried out. According to the date of radar remote sensing image data, the precipitation data, wind speed data and elevation data of the corresponding time are found to obtain the radar backscatter coefficient change data and its matching precipitation data, wind speed data and elevation data. Taking the radar remote sensing image of the first hydrological station on January 1, 2022 as an example, the precipitation data, wind speed data and elevation data on January 1, 2022 are searched to extract the corresponding radar backscatter coefficient change data. Based on the data, precipitation data, wind speed data and elevation data, 4745 radar backscatter coefficient change data, precipitation data, wind speed data and elevation data sets with continuous time series characteristics can be obtained; the measured data of the hydrological station include date and water level, and the time and space matching of the measured data of the hydrological station, radar backscatter coefficient change data, precipitation data, wind speed data and elevation data is carried out, and the date, longitude, latitude, radar backscatter coefficient change data, precipitation data, wind speed data and elevation data sets after time and space matching are established in combination with the longitude and latitude data corresponding to the hydrological station.

[0086] Step S111, obtaining real-time water level information based on the established river water level jump classification model, see Figure 2 , which is a water level distribution map of the Subei Irrigation Main Canal area predicted by the river water level jump classification model provided by the present invention. Taking the Subei Irrigation Main Canal as the research area, the date, longitude, latitude, radar backscatter coefficient change, precipitation data, wind speed data and elevation data set of the Subei Irrigation Main Canal river area on June 13, 2023 are obtained, and can be substituted into the river water level jump classification model obtained based on step S110 to perform real-time monitoring of river water level jumps and obtain real-time water level jump information, such as Figure 2 shown.

[0087] See also Figure 3 , which is a predicted water level distribution map of the Subei irrigation canal area based on the background technology, and is described in detail below in conjunction with specific steps.

[0088] Step S101, obtain river width data of 13 hydrological stations in northern Jiangsu in 2022, obtain precipitation data of 13 hydrological stations in northern Jiangsu in 2022, obtain water level data of 13 hydrological stations in northern Jiangsu in 2022, establish ten water level labels according to the average value and maximum and minimum values ​​of the measured data, i.e., five levels of rising water and five levels of falling water, perform spatiotemporal matching of river width data and water level labels, establish a data set of date, longitude, latitude, river width and precipitation after spatiotemporal matching, perform modeling based on the SVM method, and obtain a river water level jump classification model based on river width;

[0089] Step S102, obtaining river width data of the river channel area of ​​the Subei Irrigation Main Canal on June 13, 2023;

[0090] Step S103, based on the river water level jump classification model based on river width in step S101, obtain the water level information of the river channel area of ​​the Subei Irrigation Main Canal on June 13, 2023 (such as Figure 3 as shown).

[0091] This embodiment effectively reduces the losses caused by disasters and provides all-weather, highly accurate and timely dynamic feature descriptions for various rivers to better cope with the challenges of frequent natural disasters.

[0092] The technical means disclosed in the scheme of the present invention are not limited to the technical means disclosed in the above-mentioned implementation mode, but also include technical schemes composed of any combination of the above technical features.

Claims

1. A method for monitoring river water level jumps that can be used in artificial rivers, characterized in that: The specific steps include: Step 1: Obtain radar remote sensing image data at the observation location within the observation date; Step 2: Obtain precipitation data based on the Global Precipitation Project (GPM) dataset at the observation location during the observation date, and interpolate it into a 3 km grid; Step 3: Obtain wind speed data based on the fifth generation ECMWF atmospheric reanalysis global climate data within the observation date at the observation location; Step 4: Obtain the global 30-meter SRTM elevation DEM data within the observation date of the observation location; Step 5: Preprocess the acquired radar remote sensing image data, including orbit correction, radiation calibration, filtering, terrain correction and decibel preprocessing steps; Step 6: Based on the radar remote sensing image data, a circular buffer with a radius of 500m is established with the observation site as the center, and the radar backscatter coefficient data set corresponding to the observation date within the range is extracted; the radar backscatter coefficient data obtained in step 6 includes the longitude and latitude of the center coordinates of the observation site , and the radar backscatter coefficient of the observation point , the radar backscatter coefficient is calculated using the following formula: The system parameters include wavelength , angle of incidence , Polarization , the target parameters include the complex dielectric constant , Surface roughness , Volume scattering coefficient in inhomogeneous media ; The historical sequence is formed according to the time D1 sequence of the acquired radar remote sensing image data as follows: ， in, Indicates longitude and latitude, represents the radar backscatter coefficient of the observation point at time t1, L represents the number of dates in the observation time series; D1 includes year / month / day; Step 7: Based on the radar backscatter coefficient data set, the radar backscatter coefficient variation data set is generated in combination with the date; in said step 7, based on the radar backscatter coefficient data set, the radar backscatter coefficient variation data set is generated in combination with the date: ， in, Indicates longitude and latitude, Indicates The change of radar backscatter coefficient, Right now ; Where L represents the number of dates in the observed time series, and All represent the time series number of radar backscatter coefficient data set; Step 8: Obtain the measured water level data at the observation site during the observation date, and establish ten water level labels based on the average, maximum, minimum and variance statistical indicators of the measured data, including five levels of rising water and five levels of falling water; Step 9: Extract radar backscatter coefficient variation data from the acquired time series remote sensing image data, and interpolate the acquired radar backscatter coefficient variation data set in combination with the acquired continuous date precipitation data to obtain radar backscatter coefficient variation data with continuous dates; Step 10: Perform spatiotemporal matching of the water level label information of the corresponding date measured at the hydrological station with the radar backscatter coefficient change data and precipitation data, wind speed data and elevation data obtained based on the interpolated satellite remote sensing image; based on the radar backscatter coefficient change, precipitation data, wind speed data, elevation data, date, longitude, latitude data, and the ten water level labels established, use the SVM algorithm to establish a river water level jump classification model; Step 11: Subsequently, the change in radar backscatter coefficient, precipitation data, wind speed data, elevation data, date, longitude and latitude data at any time are obtained and substituted into the river water level jump classification model to conduct real-time monitoring of river water level jumps and obtain real-time water level warning information.

2. A method for monitoring river water level jumps in artificial rivers according to claim 1, characterized in that: In the step 1, the observation location and observation date are determined, and the original data of the radar remote sensing satellite corresponding to the date and location are obtained.

3. A method for monitoring river water level jumps in artificial rivers according to claim 1, characterized in that: In step 2, the GPM precipitation data is processed based on the bilinear interpolation method, and the calculation process of the bilinear interpolation is as follows: In a two-dimensional grid, suppose there are four points with coordinates and corresponding values: Among them, Q11, Q21, Q12, and Q22 are four known points, which are located at (x1, y1), (x2, y1), (x1, y2), and (x2, y2) on the coordinate axis respectively; The calculation formula for bilinear interpolation is as follows: Among them, (x, y) are the coordinates of the target point to be interpolated, f(x, y) is the interpolation result of the point, and f(Q11), f(Q21), f(Q12), and f(Q22) are the function values ​​of four known points respectively.

4. A method for monitoring river water level jumps in artificial rivers according to claim 1, characterized in that: Step 9: For the dates on which the LN remote sensing image radar backscatter coefficient changes are missing, the precipitation data corresponding to the missing dates are used. The rate of change is multiplied by the radar backscatter coefficient change of the previous day to perform discrete or continuous interpolation to obtain L radar backscatter coefficient change data with consecutive dates D1.

5. A method for monitoring river water level jumps in artificial rivers according to claim 1, characterized in that: The step 10: according to the date with radar remote sensing image data, find the precipitation data, wind speed data and elevation data of the corresponding time, obtain radar backscatter coefficient change data and its matching precipitation data, wind speed data and elevation data, specifically: in the radar remote sensing image date D1, find the precipitation data, wind speed data and elevation data corresponding to the date D1, so as to extract the corresponding radar backscatter coefficient change data, precipitation data, wind speed data and elevation data, and obtain the radar backscatter coefficient change data, precipitation data, wind speed data and elevation data of N D1 dates; the measured data of the hydrological station includes the date D2 and the water level, find the intersection of the D2 date and the D1 date, obtain the radar backscatter coefficient change data corresponding to the D2 date, the interpolated precipitation data, wind speed data and elevation data, and combine the longitude and latitude data corresponding to the hydrological station to establish the date, longitude, latitude, radar backscatter coefficient change, precipitation data, wind speed data and elevation data set after time and space matching.

6. A method for monitoring river water level jumps in artificial rivers according to claim 1, characterized in that: In step 10, a river water level jump classification model is established by using the SVM algorithm. First, classification and labeling operations are performed based on the maximum value, average value and variance statistical indicators of the measured water level in the study area; secondly, all inputs and labels are matched for training and modeling to obtain a river water level jump classification model.

7. A method for monitoring river water level jumps in artificial rivers according to claim 1, characterized in that: In the step 11, based on the radar backscatter coefficient change data of L consecutive dates, combined with the continuous time series of date, longitude and latitude, daily average precipitation data, wind speed data and elevation data, and the established river water level jump classification model, continuous, high temporal and spatial resolution and near real-time water level jump monitoring and early warning are carried out.

Images