A water flow dynamic monitoring method based on GNSS reflected wave

Abstract

The application discloses a kind of water flow dynamic monitoring methods based on GNSS reflected wave, it is related to river water flow research field.The method includes: GNSS reflected wave data is carried out data processing, time sequence is formed according to time sequence;Through continuous time series GNSS daily reflectivity, extract time series characteristics;Satellite remote sensing image is preprocessed, obtain several date observation place remote sensing image;Space-time matching is carried out, obtain GNSS reflected wave daily reflectivity and its matched river width information;According to corresponding hydrology, river width information, establish water flow runoff inversion model based on satellite remote sensing image river width;Continuous date GNSS time series variation characteristics is applied to river width value extracted from remote sensing satellite image, is substituted into inversion model and carries out water flow dynamic real-time monitoring and early warning.The above scheme can solve the problem of limited data source, low space-time resolution and difficult to monitor in waterway research, realize near real-time, fine, dynamic water flow monitoring.

Description

Technical Field

[0001] This invention relates to the field of river flow research, and in particular to a method for dynamic monitoring of water flow based on GNSS reflected waves. Background Technology

[0002] The availability of global runoff data is declining year by year, making inversion algorithms that replace measured river flow at hydrological stations extremely important. Obtaining dynamic changes in river flow trends and runoff through satellite remote sensing technology has immense application value for early warning of heavy rainfall and combating natural disasters such as floods and droughts. Satellite remote sensing technology can provide near real-time monitoring and early warning information for natural disasters such as floods and droughts, which is of paramount importance for rapid disaster assessment and emergency response.

[0003] Research on water body extraction based on satellite remote sensing imagery currently utilizes primarily optical and radar imagery as data sources. However, floods are often accompanied by clouds and rain, making it difficult for optical sensors to obtain high-quality, cloud-free images. While radar imagery offers good penetration and is unaffected by weather, the acquired images reflect the radar scattering characteristics of ground features, resulting in low interpretation capabilities and low temporal resolution (16 days), which fails to meet practical application needs. Consequently, research on water body retrieval using remote sensing imagery is largely confined to the post-flood assessment of inundated areas—a stage of coarse-resolution, single-temporal image recognition and interpretation—lacking the ability to provide high-frequency, effective information for high-resolution dynamic changes and early warning of small rivers and waterways.

[0004] GNSS-R technology utilizes polarized waves reflected from the ground by GNSS satellites for ground feature inversion. Unaffected by weather, clouds, or fog, it provides data 24 / 7 and has attracted considerable attention due to its combination of active and passive microwave remote sensing advantages. Its theoretical resolution is 0.5 x 7 km. Current research results are limited to distinguishing between inundated and non-inundated areas, suitable for post-disaster analysis and assessment, but not for the precise real-time monitoring of water flow urgently needed in early warning of natural disasters such as floods and droughts (Zhang S. et al., 2022). Therefore, how to rapidly and accurately acquire high-resolution dynamic information on surface water flow using various satellite remote sensing data is still in the initial exploratory stage and is a key technical problem that water resources, environmental, and other departments urgently need to solve.

[0005] Zhang,S.,Ma,Z.,Liu,Q.,Hu,S.,Feng,Y.,Zhao,H.,Guo,Q.,POBIinterpolationgorithmforCYGNSSnearrealtimeflooddetectionresearch:acases tudyofextremeprecipitationeventsinHenan,Chinain2021,AdvancesinSpaceResearch(2022),doi:https: / / doi.org / 10.1016 / j.asr.2022.11.016 Summary of the Invention

[0006] Purpose of the invention: The present invention aims to provide a method for dynamic monitoring of water flow using GNSS reflected waves, enabling it to provide detailed, near real-time monitoring and early warning information on water flow.

[0007] Technical solution: This embodiment of the invention provides a method for dynamic monitoring of water flow based on GNSS reflected waves, including:

[0008] Acquire GNSS surface reflection data for the observation location within the observation date;

[0009] Acquire satellite remote sensing image data of the observation location within the observation date;

[0010] The acquired GNSS reflected wave data is processed to extract the latitude and longitude information of the observation location and the daily average reflectance value, and a time series is formed according to the time sequence.

[0011] Temporal features are extracted from the daily average reflectance of continuous time-series GNSS.

[0012] Remote sensing image preprocessing is performed on the acquired satellite remote sensing images to obtain remote sensing images of the observation locations on several dates;

[0013] Spatiotemporal matching was performed between remote sensing images of observation locations acquired on several dates and GNSS data;

[0014] Based on the water level and runoff information measured by the hydrological station on the corresponding date, and the river width information from satellite remote sensing images, a water flow and runoff inversion model based on the river width from satellite remote sensing images is established.

[0015] By applying the obtained daily average reflectance variation rate of GNSS for consecutive dates to the river width values ​​of certain dates extracted from remote sensing satellite images, the interpolated river width values ​​of satellite remote sensing images with consecutive dates can be obtained.

[0016] Based on the river width values ​​obtained from satellite remote sensing images of consecutive dates, the data can be substituted into a water flow runoff inversion model based on river width to conduct real-time monitoring and early warning of water flow dynamics, thereby obtaining real-time runoff and water level information.

[0017] Specifically, the acquired GNSS reflected wave data includes the latitude and longitude coordinates {X,Y} of the center of a 500-meter resolution grid, and the surface reflectance Γ represented by this grid. x,y The surface reflectance can be calculated using the following formula:

[0018]

[0019] Where σ is the peak value of the radar cross section, R t R is the distance between the surface reflection point and the GNSS satellite. r This represents the distance between the ground reflection point and the GNSS receiver.

[0020] Specifically, a historical sequence is formed according to the obtained GNSS reflected wave time D1 (year / month / day). Where X and Y represent latitude and longitude, t1 represents the surface reflectance at the surface reflection point, and L represents the number of dates in the observation time series.

[0021] Specifically, based on the historical GNSS reflectance sequence, the temporal variation characteristics of GNSS are extracted, and the daily average reflectance variation rate is calculated. Where X and Y represent latitude and longitude, t1 represents the rate of change of the daily average reflectance of the surface at the surface reflection point, and L represents the number of dates in the observation time series.

[0022] Specifically, N high-resolution satellite remote sensing images (<1 meter) for certain dates (D2, year / month / day) need to undergo a series of remote sensing image preprocessing processes such as radiometric correction and atmospheric calibration.

[0023] Specifically, in the spatiotemporal matching of GNSS and remote sensing images, the GNSS data corresponding to the date of the remote sensing image is found. Then, based on the latitude and longitude provided by the GNSS data, the points in the remote sensing image of the corresponding date and their corresponding river widths are found. This yields the daily average reflectance of the GNSS reflected waves and the matching river width information. That is, in the GNSS data date D1, the GNSS data corresponding to the satellite remote sensing image date D2 is found. Based on the date D2, the latitude and longitude of the GNSS are determined. Then, the corresponding latitude and longitude points in the remote sensing image of date D2 are matched to obtain the daily average reflectance of GNSS and the river width for N dates D2.

[0024] Specifically, the measured data from the hydrological station includes date D3, water level, and runoff information. The intersection of date D3 and date D2 is found to obtain the river width information from the satellite remote sensing image corresponding to date D3.

[0025] Specifically, a water flow runoff inversion model based on river width from remote sensing images is established using machine learning and other methods.

[0026] Specifically, for the dates when the river width is missing in LN remote sensing images, the missing daily average reflectance change rate is used. By multiplying the river width by the previous day's width and then performing discrete or continuous interpolation, we can obtain L river width values ​​from satellite remote sensing images with consecutive dates D1.

[0027] Specifically, based on the river width values ​​from satellite remote sensing images for L consecutive dates, and the established water flow runoff inversion model based on the river width from remote sensing images, continuous, high spatiotemporal resolution, and near real-time runoff and water level information can be obtained, greatly solving the dilemma of limited remote sensing image data sources in water flow monitoring and early warning.

[0028] Beneficial effects:

[0029] Compared with the prior art, the present invention has the following significant advantages:

[0030] (1) It is not affected by weather factors such as clouds and fog. It extracts the temporal characteristics of GNSS satellite reflection waves and matches and interpolates them with remote sensing images. It can be used to obtain continuous and uninterrupted long-term water flow change information in water runoff inversion models, with high temporal resolution.

[0031] (2) It is applicable to satellite remote sensing images of various spatial resolutions, and can ultimately obtain sub-meter level monitoring results. Therefore, the proposed technical solution systematically solves the problems of limited data sources, low spatiotemporal resolution and difficulty in monitoring in waterway research, and realizes near real-time, fine and dynamic water flow monitoring.

[0032] (3) The required data preprocessing is relatively simple, the number of manually set empirical parameters is small, the requirements for the experience and professional background of the implementer are not high, it is applicable to various geographical areas and scenarios, and the method has high feasibility, robustness and accuracy of prediction results. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating the dynamic water flow monitoring method based on GNSS reflected waves provided in this embodiment of the invention.

[0034] Figure 2 Spatiotemporal matching diagram of GNSS reflectivity and Sentinel-1 remote sensing imagery;

[0035] Figure 3This is a schematic diagram comparing measured runoff and inverted runoff at hydrological stations in the research area, provided for the purposes of this invention. Detailed Implementation

[0036] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0037] See Figure 1 This is a flowchart illustrating the dynamic water flow monitoring method based on GNSS reflected waves provided by the present invention. It includes specific steps, which will be described in detail below.

[0038] Step S101: Obtain GNSS surface reflection wave data for the observation location within the observation period.

[0039] In practice, the observation location and date are determined, and the raw data from GNSS satellites for the corresponding date and location are obtained.

[0040] Step S102: Obtain satellite remote sensing image data of the observation location within the observation date.

[0041] In practice, the raw image data from high-resolution remote sensing satellites for the corresponding dates and locations are acquired.

[0042] Step S103: Process the acquired GNSS reflected wave data, extract the latitude and longitude information of the observation location and the daily average reflectance value, and form a time series according to the time sequence.

[0043] In practice, the acquired GNSS reflected wave data includes the latitude and longitude coordinates {X,Y} of the center of a 500-meter resolution grid, and the surface reflectivity Γ represented by this grid. x,y The surface reflectance can be calculated using the following formula:

[0044]

[0045] Where σ is the peak value of the radar cross section, R t R is the distance between the surface reflection point and the GNSS satellite. r This represents the distance between the ground reflection point and the GNSS receiver.

[0046] In practice, a historical sequence is formed according to the obtained GNSS reflected wave time D1 (year / month / day). Where X and Y represent latitude and longitude, The surface reflectance at the surface reflection point at time t1, and L represents the number of dates in the observation time series.

[0047] Step S104, using the continuous time-series GNSS daily average reflectance, refer to... Figure 2 (Left) Shows the reflectivity values ​​of GNSS reflected waves, from which temporal features are extracted.

[0048] In practice, based on the historical GNSS reflectance sequence, the temporal variation characteristics of GNSS are extracted, and the daily average reflectance variation rate is calculated. Where X and Y represent latitude and longitude, The daily average reflectance change rate of the surface at the surface reflection point at time t1, where L represents the number of dates in the observation time series.

[0049] Step S105: Perform remote sensing image preprocessing on the acquired satellite remote sensing images to obtain remote sensing images of observation locations on several dates.

[0050] In practice, N satellite remote sensing image data points for certain dates (D2, year / month / day) need to undergo a series of remote sensing image preprocessing processes, such as radiometric correction and atmospheric calibration.

[0051] Step S106: Spatiotemporal matching of remote sensing images of observation locations on several dates with GNSS data, that is, finding the GNSS data corresponding to the date with remote sensing images, and then using the latitude and longitude provided by the GNSS data to find the points in the remote sensing images of the corresponding date and their corresponding river widths, so as to obtain the daily average reflectance of GNSS reflected waves and the matching river width information.

[0052] In practice, within the GNSS data for date D1, the GNSS data corresponding to the satellite remote sensing image for date D2 is located. The latitude and longitude of the GNSS data are determined based on date D2, and then the corresponding latitude and longitude points in the remote sensing image for date D2 are matched. (See [reference needed]). Figure 2 and Figure 3 ,according to Figure 2 (Left) Latitude and longitude coordinates corresponding to the point where reflectivity is located, to find... Figure 2 (Right) By taking the latitude and longitude position corresponding to the satellite remote sensing image and the corresponding river width, the average daily reflectance of GNSS and the river width for N D2 dates can be obtained.

[0053] Step S107: Based on the water level and runoff information measured by the hydrological station on the corresponding date, and the river width information from satellite remote sensing images, establish a water flow and runoff inversion model based on the river width from satellite remote sensing images.

[0054] In practice, the measured data from hydrological stations include date D3, water level, and runoff information. The intersection of date D3 and date D2 is found to obtain the river width information from the corresponding satellite remote sensing imagery on date D3. (See reference...) Figure 3 The hydrological stations in the embodiment include: Chaohu Sluice Gate, Yuxi Sluice Gate, and Tongcheng Sluice Gate.

[0055] In practice, the river width information from satellite remote sensing images is used as input, and the measured runoff is used as output. Through machine learning methods for training and modeling, a water flow runoff inversion model based on the river width from remote sensing images is obtained.

[0056] Step S108: The obtained daily average reflectance change rate of GNSS for consecutive dates is applied to the river width values ​​of certain dates extracted from remote sensing satellite images to obtain interpolated river width values ​​of satellite remote sensing images with consecutive dates.

[0057] In practice, for the dates when the river width is missing in LN remote sensing images, the missing daily average reflectance change rate is used. By multiplying the river width by the previous day's value and performing discrete or continuous interpolation, L river width values ​​from satellite remote sensing images with consecutive dates D1 can be obtained.

[0058] In step S109, the river width values ​​from the satellite remote sensing images for consecutive dates can be substituted into the water flow runoff inversion model based on river width to perform real-time monitoring and early warning of water flow dynamics, thereby obtaining real-time runoff and water level information.

[0059] In practical implementation, based on the river width values ​​from satellite remote sensing images of L consecutive dates and the established water flow runoff retrieval model based on the river width from remote sensing images, continuous, high spatiotemporal resolution, and near real-time runoff and water level information can be obtained. (See [reference needed]). Figure 3 The chart shows a comparison between the actual runoff values ​​monitored by three hydrological stations and the inverted runoff values. The chart indicates that the inverted runoff results are largely consistent with the measured runoff, but with significant fluctuations. Investigation revealed that the Tongcheng Sluice Gate was opened for flood discharge in September 2021, which is the main reason for the runoff trough. Furthermore, summer is the season with frequent heavy rainfall, and increased precipitation is also a significant factor contributing to runoff fluctuations.

[0060] In practice, this method can be applied to remote sensing images of any spatial resolution, greatly solving the dilemma of limited data sources for remote sensing images in practical applications.

Claims

1. A method for dynamic monitoring of water flow based on GNSS reflected waves, characterized in that, include: Acquire GNSS surface reflection data for the observation location within the observation date; Acquire satellite remote sensing image data of the observation location within the observation date; The acquired GNSS surface reflection wave data are processed to extract the latitude and longitude information of the observation location and the daily average reflectance, and a time series is formed according to the time sequence. By extracting the temporal features from the continuous time series GNSS daily average reflectance, the daily average reflectance change rate is obtained; Remote sensing image preprocessing is performed on the acquired satellite remote sensing image data to obtain remote sensing images of observation locations on several dates; Spatiotemporal matching was performed between the remote sensing images of the observation locations on several dates and GNSS data; Based on the water level and runoff information measured by the hydrological station on the corresponding date, and the river width information from satellite remote sensing images, a water flow and runoff inversion model based on the river width from satellite remote sensing images is established. The obtained daily average reflectance change rate is applied to the river width value of the date extracted from the remote sensing satellite image to obtain the interpolated river width value of the satellite remote sensing image with continuous dates. Based on the river width values ​​obtained from satellite remote sensing images for consecutive dates, the water flow runoff inversion model based on the river width of the satellite remote sensing images is substituted to obtain real-time runoff and water level information, enabling real-time monitoring and early warning of water flow dynamics. When acquiring river width values ​​from satellite remote sensing images for the consecutive dates, for LN dates where river width is missing from remote sensing images, the missing daily average reflectance change rate is used. Multiply by the river width of the previous day and perform discrete or continuous interpolation to complete the data; where X and Y represent latitude and longitude, and L represents the number of dates in the observation time series.

2. The method for dynamic monitoring of water flow based on GNSS reflected waves according to claim 1, characterized in that, The GNSS surface reflection data includes the latitude and longitude coordinates {X,Y} of the center of a 500-meter resolution grid, and the surface reflectance Γ of this grid. x,y The surface reflectance can be calculated using the following formula: Where σ is the peak value of the radar cross section, R t R is the distance between the surface reflection point and the GNSS satellite. r This represents the distance between the ground reflection point and the GNSS receiver.

3. The method for dynamic monitoring of water flow based on GNSS reflected waves according to claim 1, characterized in that, The time series is formed into a historical sequence according to the obtained GNSS reflection wave time D1. Where X and Y represent latitude and longitude, t1 represents the surface reflectance at the surface reflection point, and L represents the number of dates in the observation time series.

4. The method for dynamic monitoring of water flow based on GNSS reflected waves according to claim 3, characterized in that, The daily average reflectance change rate is calculated by extracting the GNSS temporal variation characteristics based on the historical GNSS reflectance sequence. Where X and Y represent latitude and longitude, t1 represents the rate of change of the daily average reflectance of the surface at the surface reflection point, and L represents the number of dates in the observation time series.

5. The method for dynamic monitoring of water flow based on GNSS reflected waves according to claim 1, characterized in that, The preprocessing involves acquiring satellite high-resolution remote sensing image data for several dates (D2) and performing radiometric correction and atmospheric calibration.

6. The method for dynamic monitoring of water flow based on GNSS reflected waves according to claim 1, characterized in that, The spatiotemporal matching involves finding the GNSS data corresponding to the date D1 when remote sensing imagery is available, and then using the latitude and longitude provided by the GNSS data to find the point in the remote sensing image for the corresponding date D2 and its corresponding river width, thereby obtaining the daily average reflectance of the GNSS reflected wave and its matched river width information.

7. The method for dynamic monitoring of water flow based on GNSS reflected waves according to claim 6, characterized in that, The inversion model is established using machine learning methods; by using the measured data date D3, water level, and runoff information from the hydrological station, the intersection of date D3 and date D2 from the satellite remote sensing image is found, and the river width information of the satellite remote sensing image corresponding to date D3 is obtained.

Images