Reservoir scheduling rule extraction method and system based on multi-source remote sensing satellite

Abstract

The invention provides a reservoir scheduling rule extraction method and system based on a multi-source remote sensing satellite. The method comprises the following steps: respectively acquiring a Sentinel-1 remote sensing image and a GF-3 remote sensing image of a target reservoir and preprocessing the Sentinel-1 remote sensing image and the GF-3 remote sensing image; respectively carrying out mask cutting processing and binarization processing on the preprocessed Sentinel-1 remote sensing image and the preprocessed GF-3 remote sensing image; on the basis of the Sentinel-1 remote sensing image and the GF-3 remote sensing image which are subjected to binarization processing, complete water boundaries are obtained respectively, and then the water surface area is calculated; the water surface area calculated based on the Sentinel-1 remote sensing image and the GF-3 remote sensing image is corrected, and a corrected and uniformized water surface area sequence is obtained; based on the water body boundary and the corrected and uniformized water surface area sequence, reservoir scheduling rule extraction is carried out; the invention constructs a method for judging the long-term reservoir regulation rule in the non-control area drainage basin only through remote sensing image information and runoff hydrological data. Therefore, the data collection process is simplified, and the real-time performance and accuracy of data acquisition are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of reservoir operation rule extraction, and particularly to a method and system for extracting reservoir operation rules based on multi-source remote sensing satellites. Background Art

[0002] The increasing water and energy demands pose great challenges to the development of human society and economy. Reservoir construction is one of the most practical strategies to alleviate water resource and energy shortages. At the same time, flood disasters also have a serious impact on the production and life of human society, restricting the high-quality development of the regional economy. The operation of reservoirs in a region is of great significance for water resource management. In particular, through a series of operation measures such as flood interception, peak shifting, and peak cutting, significant flood control effects have been achieved, while also taking into account water supply and power generation, greatly meeting the needs of activities such as agriculture, industry, and shipping. Scientific reservoir operation is of great significance for maximizing the operation benefits of reservoirs, fully managing and utilizing water resources within the basin, and ecological environment protection.

[0003] However, the actual application scenarios of reservoir operation are extremely complex, involving multiple time scales and multiple flow states, often accompanied by occasional emergencies, and facing many problems. Reservoirs should not only undertake medium- and long-term (seasonal and monthly) operation tasks of managing downstream water supply and optimizing economic benefits, but also undertake short-term (daily and hourly) operation tasks, including managing grid load, water demand, shipping and fish farming, disaster prevention, and emergency operations during floods and droughts. The actual operation requirements often change frequently and often deviate from the operation plan. Although a large number of reservoirs have been built in many river basins in China, only a few large reservoirs are equipped with dedicated personnel to continuously observe and record water storage information. There are difficulties in obtaining hydrological and water regime information in data-deficient areas and a lack of data on the operation rules of non-controlled reservoirs. The remote sensing satellite monitoring technology has the advantages of large scale, high efficiency, and low cost. Therefore, using satellite technology to judge reservoir operation rules, that is, judging the water release and storage processes of reservoirs based on the time series of reservoir outflow and the duration, is of great significance for the optimal operation research of the national reservoir group.

[0004] Currently, according to relevant literature, although there are various schemes for judging reservoir operation rules, there is currently a lack of remote sensing satellites for monitoring flow rate or velocity and related algorithms, making it difficult to determine the inflow and outflow of reservoirs through remote sensing means, and further difficult to extract fine reservoir operation rules and obtain the water storage strategies and operation rules of non-controlled reservoirs, which hinders the application of remote sensing technology in reservoir operation rule services. Summary of the Invention

[0005] The object of the present invention is to overcome the above deficiencies and provide a method and system for extracting reservoir operation rules based on multi-source remote sensing satellites, which do not require precipitation and surface evapotranspiration data of the reservoir location, and can respectively model the inflow and outflow of all small and medium-sized reservoirs in the basin only through the water storage information and runoff estimation obtained from remote sensing images, and improve the calculation efficiency and applicability without loss of accuracy.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is: a method for extracting reservoir operation rules based on multi-source remote sensing satellites, including the following steps: Step S1: Obtain Sentinel-1 remote sensing images and GF-3 remote sensing images of the target reservoir respectively and perform preprocessing; Step S2: Perform mask clipping and binarization processing on the preprocessed Sentinel-1 remote sensing images and GF-3 remote sensing images in Step S1 respectively to obtain binarized raster images; Step S3: Remove other patch information outside the target reservoir from the binarized Sentinel-1 remote sensing images and GF-3 remote sensing images and perform hole filling to respectively obtain complete water body boundaries, and then calculate the water surface area; Step S4: Correct the water surface areas calculated based on Sentinel-1 remote sensing images and GF-3 remote sensing images to obtain a corrected and unified water surface area sequence; Step S5: Extract reservoir operation rules based on the water body boundaries in Step S3 and the corrected and unified water surface area sequence in Step S4.

[0007] Further, Step S2 includes: Outline the maximum inundation boundary vector B of the target reservoir on Tianditu, then construct a buffer boundary B2 of 90 to 500 meters for the maximum inundation boundary vector B, use the buffer boundary B2 as a mask to clip the image, and perform binarization processing using the Otsu method to obtain a binarized raster image, with the water body pixel value being 1 and the non-water body pixel value being 0.

[0008] Further, Step S3 includes: Step S3.1: Further perform mask processing on the binarized Sentinel-1 remote sensing images and GF-3 remote sensing images to remove other patch information outside the target reservoir; Step S3.2: Remove the edge hole areas of the water body in the binarized raster image, then vectorize the raster image and remove the internal holes again to obtain a complete water body boundary; Calculate the water surface area based on the obtained complete water body boundary.

[0009] Further, the step S3.1 includes: calculating the inundation frequency map of the water body, then adjusting the maximum inundation boundary vector B according to the inundation frequency map, and further performing a masking process on the binarization result of step S3 using the adjusted boundary vector B; The step S3.2 includes: filling on the raster image to remove edge holes; then converting the raster image into a polygon vector boundary, deleting small polygons nested inside the polygon vector boundary to remove internal holes, and finally obtaining a complete water body boundary.

[0010] Further, the step S4 includes: selecting the water surface areas calculated from Sentinel-1 remote sensing images and GF-3 remote sensing images at the same or adjacent times for linear regression, and using the regression coefficient to correct the monitoring results of the low-resolution images at other times to obtain a corrected and unified water surface area sequence.

[0011] Further, the step S5 includes: Step S5.1: Obtaining the DEM elevation values corresponding to the water body boundary, and obtaining the water level time series curve based on the DEM elevation values corresponding to the water body boundary; Step S5.2: Based on the water storage - area model and the corrected and unified water surface area sequence, obtaining the water storage time series curve; Step S5.3: Inverting the reservoir outflow according to the target reservoir water volume time series curve and the inflow runoff data to obtain the outflow time series curve; Step S5.4: According to the obtained water level time series curve, water storage time series curve, and outflow time series curve, performing multi-curve correlation analysis and data anomaly processing to generate an envelope curve, analyzing the duration of water storage and release, and realizing the extraction of reservoir operation rules.

[0012] Further, the step S5.1 includes: Obtaining the DEM elevation values corresponding to the water body boundary, registering the water body boundary with the DEM elevation values corresponding to the positions, and obtaining the water level elevation data by taking the median of the water body boundary superimposed with the DEM elevation values corresponding to the positions. Recording the water level elevation data as the water level of the reservoir at different times to obtain the water level time series curve.

[0013] Further, the reservoir outflow in the step S5.3 is: ; where, is the outflow (cubic meters) of the reservoir between the and the day; and are the and the Daily water storage (cubic meters); is the daily inflow of the upstream river channel (cubic meters per second); is 86,400 seconds in a day; Arrange the calculated reservoir outflow in a time series to generate a complete outflow time series curve.

[0014] Furthermore, the multi-curve correlation analysis in step S5.4 includes: aligning the water level time series curve, water storage time series curve, and outflow time series curve in time to form a time series of multi-curve correlation, analyzing the change trends in different stages, and identifying key nodes; Generating the envelope curve includes: (1) In the time series of multi-curve correlation, take each water level and the corresponding maximum water storage to represent the maximum operating range of the reservoir as the upper envelope line; take each water level and the corresponding minimum water storage to represent the minimum operating range of the reservoir as the lower envelope line; (2) Use spline interpolation or polynomial fitting for the discrete points in the upper envelope line or lower envelope line to generate a continuous envelope curve, ensuring that the curve changes smoothly and conforms to the actual situation of reservoir operation.

[0015] In addition, the present invention also discloses a reservoir operation rule extraction system based on multi-source remote sensing satellites. The reservoir operation rule extraction system based on multi-source remote sensing satellites is used to execute the steps in the reservoir operation rule extraction method based on multi-source remote sensing satellites, including: Remote sensing image acquisition module: It is used to acquire Sentinel-1 remote sensing images and GF-3 remote sensing images of the target reservoir respectively and perform preprocessing; Binarization processing module: It is used to perform mask cropping processing and binarization processing on the preprocessed Sentinel-1 remote sensing images and GF-3 remote sensing images respectively; Water body boundary extraction module: It is used to obtain the complete water body boundary based on the binarized Sentinel-1 remote sensing images and GF-3 remote sensing images respectively, and then calculate the water surface area; Calibration module: It is used to calibrate the water surface areas calculated based on Sentinel-1 remote sensing images and GF-3 remote sensing images to obtain a calibrated and unified water surface area sequence; Reservoir operation rule extraction module: It is used to extract reservoir operation rules based on the water body boundary and the calibrated and unified water surface area sequence.

[0016] Advantages of the present invention: (1) The satellite data used in the present invention is SAR, which can penetrate clouds and rain for all-weather and all-time continuous observation in cloudy and rainy environments. (2) Traditional methods for extracting the outflow rate have too many restrictive conditions. The present invention does not require obtaining precipitation and surface evapotranspiration data of the reservoir location. By only using the water storage information obtained from remote sensing images and runoff estimation, the inflow and outflow of all medium and small reservoirs in the basin can be respectively modeled, and the medium and long-term reservoir operation rules of the non-controlled area basin can be judged. This not only simplifies the data collection process but also improves the real-time and accuracy of data acquisition, greatly improving the calculation efficiency and applicability without sacrificing accuracy.

[0017] (3) Currently, there is a lack of remote sensing satellites and related algorithms for monitoring flow rate or velocity, which to a certain extent limits the accurate acquisition and analysis of these key parameters. The present invention realizes high-frequency monitoring of the reservoir capacity and time-series hydrological parameters of the basin reservoir group through remote sensing images, providing strong technical support for the research in related fields. Brief Description of the Drawings

[0018] Figure 1 It is a schematic flow chart of a method for extracting reservoir operation rules based on multi-source remote sensing satellites; Figure 2 The research area of Example 2 of the present invention; Figure 3 The schematic diagram of the drawn mask of Example 2 of the present invention; Figure 4 The binary effect diagram of Example 2 of the present invention; Figure 5 The characteristic curve of Zhouba Reservoir from 2022 to 2023 in Example 2 of the present invention. Detailed Embodiments

[0019] The present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0020] Example 1: As Figure 1 shown, a method for extracting reservoir operation rules based on multi-source remote sensing satellites includes the following steps: Step S1: Respectively obtain the Sentinel-1 remote sensing image and GF-3 remote sensing image of the target reservoir and perform preprocessing; Step S2: Respectively perform mask cropping and binary processing on the preprocessed Sentinel-1 remote sensing image and GF-3 remote sensing image in Step S1 to obtain binary raster images; Step S3: Remove other patch information outside the target reservoir from the binary processed Sentinel-1 remote sensing image and GF-3 remote sensing image and perform hole filling to respectively obtain complete water body boundaries, and then calculate the water surface area; Step S4: Calibrate the water surface areas calculated based on Sentinel-1 remote sensing images and GF-3 remote sensing images to obtain a calibrated and unified water surface area sequence. Step S5: Extract reservoir operation rules based on the water body boundaries in Step S3 and the calibrated and unified water surface area sequence in Step S4.

[0021] Further, Step S2 includes: On the Tianditu map, outline the maximum inundation boundary vector B of the target reservoir, then construct a buffer boundary B2 of 90 to 500 meters for the maximum inundation boundary vector B, use the buffer boundary B2 as a mask to crop the image, and perform binary processing using the Otsu method to obtain a binary raster image, where the water body pixel value is 1 and the non-water body pixel value is 0.

[0022] Further, Step S3 includes: Step S3.1: Further perform masking on the binary-processed Sentinel-1 remote sensing image and GF-3 remote sensing image to remove other patch information outside the target reservoir. Step S3.2: Remove the edge hole areas of the water body in the binary raster image, then vectorize the raster image and remove the internal holes again to obtain a complete water body boundary. Based on the obtained complete water body boundary, calculate the water surface area.

[0023] Even further, Step S3.1 includes: Calculate the inundation frequency map of the water body, then adjust the maximum inundation boundary vector B according to the inundation frequency map, and use the adjusted boundary vector B to further perform masking on the binary result of Step S3. Step S3.2 includes: Fill the raster image to remove the edge holes; then convert the raster image into a polygon vector boundary, delete the small polygons nested inside the polygon vector boundary to remove the internal holes, and finally obtain a complete water body boundary.

[0024] Further, Step S4 includes: Select the water surface areas calculated from Sentinel-1 remote sensing images and GF-3 remote sensing images at the same time or adjacent times for linear regression, and use the regression coefficient to calibrate the monitoring results of the low-resolution images at other times to obtain a calibrated and unified water surface area sequence.

[0025] Further, Step S5 includes: Step S5.1: Obtain the DEM elevation values corresponding to the water body boundary positions, and obtain the water level time series curve based on the DEM elevation values corresponding to the water body boundary positions. Step S5.2: Based on the water storage - area model and the corrected and unified water surface area sequence, obtain the water storage time - series curve; Step S5.3: Invert the reservoir outflow according to the target reservoir water volume time - series curve and the inflow runoff data to obtain the outflow time - series curve; Step S5.4: According to the obtained water level time - series curve, water storage time - series curve, and outflow time - series curve, conduct multi - curve correlation analysis and data anomaly processing, generate the envelope curve, analyze the duration of water storage and release, and realize the extraction of reservoir operation rules.

[0026] Furthermore, the step S5.1 includes: Obtain the DEM elevation value at the corresponding position of the water body boundary, register the water body boundary with the DEM elevation value at the corresponding position, and obtain the water level elevation data by taking the median of the water body boundary superimposed with the DEM elevation value at the corresponding position. Record the water level elevation data as the water level of the reservoir at different times to obtain the water level time - series curve.

[0027] Furthermore, the reservoir outflow in the step S5.3 is: ; where, is the outflow of the reservoir (cubic meters) between the and the day; and are the water storage volumes (cubic meters) of the reservoir on the and the day respectively; is the daily inflow of the upstream river (cubic meters per second); is 86400 seconds in a day. Arrange the calculated reservoir outflows in a time series to generate a complete outflow time - series curve.

[0028] Furthermore, the multi - curve correlation analysis in the step S5.4 includes: Align the water level time - series curve, water storage time - series curve, and outflow time - series curve in time to form a time series of multi - curve correlation, analyze the change trends in different stages, and identify key nodes; Generating the envelope curve includes: (1) In the time series of multi - curve correlation, take each water level and the corresponding maximum water storage volume, which represents the maximum operation range of the reservoir, as the upper envelope line; take each water level and the corresponding minimum water storage volume, which represents the minimum operation range of the reservoir, as the lower envelope line; (2) Use spline interpolation or polynomial fitting for the discrete points in the upper envelope line or the lower envelope line to generate a continuous envelope curve, ensuring that the curve changes smoothly and conforms to the actual operation of the reservoir.

[0029] Embodiment 2: like Figure 1 As shown, this embodiment provides a reservoir scheduling rule extraction method based on multi-source remote sensing satellites, comprising the following steps: Step S1: respectively obtain the Sentinel-1 remote sensing image and GF-3 remote sensing image of the target reservoir and perform preprocessing; In this example, the Mintuo River Basin is selected as the research area. Figure 2 shown.

[0030] In this embodiment, the processing operations of Sentinel-1 remote sensing images mainly include orbit correction, image stitching, regional cropping, thermal noise removal, radiation calibration, filtering, terrain correction, and decibel processing. The orbit information contained in the Sentinel-1 metadata file is not very accurate, and the status information of the satellite orbit needs to be updated by SNAP networking for orbit correction. It is necessary to determine the rough positions of the five reservoirs in the Mintuo River Basin based on longitude and latitude to determine the image stitching and cropping process. The terrain correction part is performed using a 5-meter DEM of the Mintuo River Basin. The above operations can all be batch processed using Python by calling snappy gpt through the command line combined with the SNAP flowchart, which greatly improves efficiency and saves time costs.

[0031] Considering the differences in GF-3 imaging quality, visual inspection methods were used to screen out all high-quality images within the time range of the target reservoir area. The processing operations of GF-3 remote sensing images mainly include radiometric calibration, atmospheric correction, RPC orthorectification and image cropping.

[0032] like Figure 4 As shown, step S2: mask cropping and binarization are performed on the Sentinel-1 remote sensing image and the GF-3 remote sensing image preprocessed in step S1 to obtain a binary raster image; In this embodiment, the SDWI is first calculated for the preprocessed Sentinel-1 SAR image.

[0033] The calculation equation of SDWI index is:

[0034] Among them, VV represents the VV polarization band, and VH represents the VH polarization band.

[0035] like Figure 3The figure shows a schematic diagram of the drawn mask in an embodiment of the present invention. The base map is the TianDiTu image of the Maoergai Reservoir area, and the accuracy can reach the sub-meter level. The areas frequently flooded on the image are shown as light-colored tidal flat areas, and the maximum inundation boundary of the reservoir is drawn based on these areas. After the mask is drawn, buffer processing is performed on the drawn vector, and the buffer is set to 90 meters. Then, the obtained buffer vector boundary is used to perform mask clipping on the Sentinel-1 and GF-3 images, and then the Otsu method is used for binarization to further exclude the interference of irrelevant pixels and improve the accuracy of the binarization algorithm.

[0036] Step S3: Remove other patch information outside the target reservoir from the binarized Sentinel-1 remote sensing image and GF-3 remote sensing image and perform hole filling to respectively obtain complete water body boundaries, and then calculate the water surface area; After the image in step S2 is binarized, except for the target reservoir, all are unnecessary patch information. Therefore, it is necessary to further perform mask processing on the binarization result in step S2 to remove other patch information. Then, a hole filling method is adopted to remove the vacant areas of the water body, obtain complete water body boundaries, and then calculate the water surface area to realize the automatic extraction of lake and reservoir water bodies in the basin from multi-source remote sensing data.

[0037] The obtained binary image is used to calculate the inundation frequency map of the reservoir. The drawn vector boundary is adjusted according to the boundary inundation situation of the water body in the inundation frequency map to obtain a more accurate maximum inundation boundary of the reservoir. The maximum inundation boundary is used to further perform mask processing on the binarization result. Then, the water body holes are filled to obtain complete water body boundaries.

[0038] Step S4: Calibrate the water surface areas calculated based on the Sentinel-1 remote sensing image and GF-3 remote sensing image to obtain a calibrated and unified water surface area sequence; Select the water surface areas calculated from the Sentinel-1 remote sensing image and GF-3 remote sensing image at the same time or adjacent times for linear regression, and use the regression coefficient to calibrate the monitoring results of the low-resolution image at other times to obtain a calibrated and unified water surface area sequence.

[0039] Step S5: Extract reservoir operation rules based on the water body boundary in step S3 and the calibrated and unified water surface area sequence in step S4.

[0040] Step S5 includes: Step S5.1. Obtain the DEM elevation value at the corresponding position of the water body boundary, and obtain the water level time series curve based on the DEM elevation value at the corresponding position of the water body boundary; Step S5.2. Based on the water storage - area model and the calibrated and unified water surface area sequence, obtain the water storage time series curve; Step S5.3. Invert the reservoir discharge based on the target reservoir water volume time series curve and the inflow runoff data to obtain the discharge time series curve; Step S5.4. According to the obtained water level time series curve, water storage volume time series curve, and discharge time series curve, conduct multi-curve correlation analysis and data anomaly processing, generate an envelope curve, analyze the duration of water storage and release, and realize the extraction of reservoir operation rules.

[0041] First, obtain the water level time series curve by superimposing the water body boundary on the DEM elevation value at the corresponding position of the water body boundary. Then, select the reservoir area within a certain period of time and fit the water storage volume - area model based on the measured water storage volume. Finally, invert the reservoir water storage volume according to all the area results to obtain the water storage volume curve. Taking a typical reservoir as an example, the fitted parameters are as follows:

[0042] Among them, S is the reservoir water storage volume (100 million cubic meters), A is the area of the corresponding reservoir extracted by remote sensing (million square meters), and α and β are the regression constants respectively. For reservoirs in the basin with a similar slope to the Maoergai Reservoir, the same parameters can be selected to calculate the water storage volume. After calculating the reservoir water storage volume, obtain the discharge curve from the obtained inflow data and the previously calculated water storage volume. Including: After calculating the reservoir water storage volume according to the water storage volume - area model for the selected GF-1 remote sensing image area and Sentinel-1 remote sensing image area, list the results with adjacent time as a group. Let the nth group have the times of the th day and th day respectively, corresponding to the water storage volume calculated from the images. The discharge of the reservoir on the th and the th day is shown in the following formula.

[0043]

[0044] Among them, is the discharge of the reservoir (cubic meters) between the th and the th day; and are the water storage volumes (cubic meters) of the reservoir on the th and the th day respectively, and the difference between the two is calculated to obtain the change in reservoir capacity; is the daily inflow of the upstream river channel (cubic meters per second). is 86,400 seconds in a day. At the th and the th days in the time series, the discharge can be calculated according to the formula by calculating the change in reservoir capacity and the cumulative inflow. Arrange the calculation results in a time series to generate a complete discharge time series curve.

[0045] Based on the obtained water level time series curve, water storage volume time series curve, and outflow time series curve, the reservoir operation envelope curve is obtained, and the duration of water storage and release in the reservoir is obtained to derive the reservoir operation rules.

[0046] 1. Multi-curve correlation analysis: Align the water level, water storage volume, and outflow curves in time to form a time series with multi-curve correlation, analyze the changing trends in different stages, and identify key nodes (such as rising water level, increasing or decreasing reservoir capacity, etc.).

[0047] 2. Data anomaly processing: Use methods such as interpolation, moving average, or ARIMA to identify and repair abnormal points, remove the interference of extreme values, and ensure that the curves are smooth and realistic.

[0048] 3. Generation of envelope curve: (1) Generate upper and lower envelope lines according to the changing rules of water level and water storage volume. In the time series, take each water level and the corresponding maximum water storage volume to represent the maximum operation range of the reservoir as the upper envelope line; take each water level and the corresponding minimum water storage volume to represent the minimum operation range of the reservoir as the lower envelope line.

[0049] (2) Use spline interpolation or polynomial fitting for discrete points to generate a continuous envelope curve, ensuring that the curve changes smoothly and conforms to the actual operation of the reservoir.

[0050] 4. Analysis of the duration of water storage and release: (1) Time period division. According to the extreme value intervals of the water level, water storage volume, and outflow curves, divide the water storage period, water release period, and static water period. Water storage period: The water level and water storage volume rise simultaneously, and the outflow significantly decreases. Water release period: The water level and water storage volume decrease simultaneously, and the outflow significantly increases. Static water period: The water level and water storage volume are basically stable, and the outflow is close to the inflow.

[0051] (2) Duration calculation. According to the length of the extreme value intervals of the water level, water storage volume, and outflow curves, count the time lengths of the water storage period, water release period, and static water period, thereby providing important support for the reservoir operation optimization, flood control and disaster reduction, water resources management, and long-term operation planning.

[0052] 5. Extraction of operation rules Extract the rules of water storage and release based on the functions of the reservoir (such as flood control, power generation, irrigation). Stop water release when the water level is below the dead water level, and only maintain the ecological flow when necessary; start flood discharge when the water level is above the flood limit water level to control the risk of the outflow superimposed with the flood peak; conduct routine operation when the water level is between the dead water level and the normal storage water level, and determine the water storage and release strategy according to the downstream demand. Generally speaking, the reservoir mainly releases water during the dry season of the basin to ensure downstream water use demand and avoid the water level being too low to affect the ecological environment; mainly stores water during the wet season to give priority to storing rain and flood resources.

[0053]

[0054] Among them, is the outflow, is the reservoir water level. is the dead water level of the reservoir, is the normal storage level, is the flood control limit level, is the flow rate at which water discharge stops when the water level is below the dead water level, generally 0, and only maintains the ecological flow rate when necessary. is the water discharge flow rate determined according to downstream demands. is the water discharge volume when flood discharge is initiated when the water level is above the flood control limit level.

[0055] In this embodiment, as Figure 5 shown are the time series curves of the water level, water storage volume, and outflow of the Zhouba Reservoir from 2022 to 2023, where (a) is the time series curve of the water level and water storage volume of the Zhouba Reservoir, and (b) is the time series curve of the outflow of the Zhouba Reservoir. During this period, the highest water level of the Zhouba Reservoir was 384.21 m, and the lowest water level was 373.34 m; the maximum water storage volume was 162 million cubic meters, and the minimum was 75 million cubic meters; the maximum outflow was 291031076 cubic meters between August 16, 2023 and August 23, 2023. When the water level and water storage volume curves show a downward trend, the reservoir is in the water discharge state, and when the curves show an upward trend, the reservoir is in the water storage state.

[0056] By simulating the operation of the reservoir under different scheduling schemes, the impact of various schemes on the outflow can be evaluated, thereby selecting the optimal scheduling strategy.

[0057] Example 3: A reservoir scheduling rule extraction system based on multi-source remote sensing satellites, the reservoir scheduling rule extraction system based on multi-source remote sensing satellites is used to execute the steps in the reservoir scheduling rule extraction method based on multi-source remote sensing satellites, including: Remote sensing image acquisition module: It is used to respectively acquire the Sentinel-1 remote sensing image and GF-3 remote sensing image of the target reservoir and perform preprocessing; Binary processing module: It is used to perform mask cropping processing and binary processing on the preprocessed Sentinel-1 remote sensing image and GF-3 remote sensing image respectively; Water body boundary extraction module: It is used to respectively obtain the complete water body boundary based on the binary processed Sentinel-1 remote sensing image and GF-3 remote sensing image, and then calculate the water surface area; Calibration module: It is used to calibrate the water surface area calculated based on Sentinel-1 remote sensing images and GF-3 remote sensing images, and obtain the calibrated and unified water surface area sequence; Reservoir operation rule extraction module: It is used to extract reservoir operation rules based on the water body boundary and the calibrated and unified water surface area sequence.

[0058] The above embodiments are only the preferred technical solutions of the present invention, and should not be regarded as limitations on the present invention. The protection scope of the present invention should be the technical solutions recorded in the claims, including the equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, the equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A reservoir dispatching rule extraction method based on multi-source remote sensing satellites, characterized in that: The steps include: Step S1, respectively obtaining Sentinel-1 remote sensing images and GF-3 remote sensing images of the target reservoir and performing preprocessing; Step S2, performing mask cropping and binarization processing on the Sentinel-1 remote sensing image and the GF-3 remote sensing image preprocessed in step S1 to obtain a binary raster image; Step S3, removing other image patch information outside the target reservoir from the binarized Sentinel-1 remote sensing image and the GF-3 remote sensing image and performing hole filling to obtain complete water body boundaries, and then calculating the water surface area; Step S4, correcting the water surface area calculated based on the Sentinel-1 remote sensing image and the GF-3 remote sensing image to obtain a corrected and consistent water surface area sequence; Step S5: extracting reservoir operation rules based on the water body boundary in step S3 and the corrected and unified water surface area sequence in step S4.

2. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 1 is characterized in that: The step S2 comprises: The maximum inundation boundary vector B of the target reservoir is drawn on the Tiandi Map, and then a buffer boundary B2 of 90 to 500 meters is constructed for the maximum inundation boundary vector B. The buffer boundary B2 is used as a mask to crop the image and binarize it using the Otsu method to obtain a binary raster image with water body pixel values ​​of 1 and non-water body pixel values ​​of 0.

3. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 1 is characterized in that: The step S3 comprises: Step S3.1, further masking the binarized Sentinel-1 remote sensing image and GF-3 remote sensing image to remove other image information outside the target reservoir; Step S3.2, removing the edge hole area of ​​the water body in the binary raster image, then vectorizing the raster image and removing the internal holes again to obtain the complete water body boundary; Calculate the water surface area based on the obtained complete water body boundary.

4. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 3 is characterized in that: The step S3.1 comprises: calculating a flooding frequency map of the water body, then adjusting a maximum flooding boundary vector B according to the flooding frequency map, and further performing mask processing on the binarization result of step S3 using the adjusted boundary vector B; The step S3.2 includes: filling in the raster image to remove edge holes; then converting the raster image into a polygonal vector boundary, deleting small polygons nested inside the polygonal vector boundary, removing internal holes, and finally obtaining a complete water body boundary.

5. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 1 is characterized in that: The step S4 comprises: selecting the water surface areas calculated by the Sentinel-1 remote sensing images and the GF-3 remote sensing images at the same or adjacent moments for linear regression, using the regression coefficients to correct the monitoring results of the low-resolution images at other moments, and obtaining a corrected and unified water surface area sequence.

6. The method for extracting reservoir operation rules based on multi-source remote sensing satellites according to claim 1 is characterized in that: The step S5 comprises: Step S5.1: obtaining the DEM elevation value of the position corresponding to the water body boundary, and obtaining the water level time series curve based on the DEM elevation value of the position corresponding to the water body boundary; Step S5.2: Based on the water storage capacity-area model and the corrected and consistent water surface area sequence, obtain a water storage capacity time series curve; Step S5.3: Invert the reservoir outflow according to the target reservoir water volume time series curve and the inflow runoff data to obtain the outflow time series curve; Step S5.4: Based on the acquired water level time series curve, water storage time series curve, and outflow time series curve, multi-curve correlation analysis and data anomaly processing are performed to generate an envelope curve, analyze the duration of water storage and release, and realize the extraction of reservoir scheduling rules.

7. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 6 is characterized in that: The step S5.1 comprises: The DEM elevation value of the corresponding position of the water body boundary is obtained, and the water body boundary is aligned with the DEM elevation value of the corresponding position. The water level elevation data is obtained by superimposing the DEM elevation value of the corresponding position on the water body boundary. The water level elevation data is recorded as the water level of the reservoir at different times to obtain the water level time series curve.

8. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 6 is characterized in that: In step S5.3, the reservoir outflow is: ； in, For the and The outflow of the reservoir between days (cubic meters); and Reservoir No. and Daily water storage capacity (cubic meters); is the daily inflow of the upstream river (cubic meters per second); A day is 86400 seconds; The calculated reservoir outflow is arranged in time series to generate a complete outflow time series curve.

9. The method for extracting reservoir dispatching rules based on multi-source remote sensing satellites according to claim 6 is characterized in that: The multi-curve correlation analysis in step S5.4 includes: aligning the water level time series curve, the water storage time series curve, and the outflow time series curve in time to form a time series of multi-curve correlation, analyzing the change trends in different stages, and identifying key nodes; Generating an envelope curve includes: (1) In the time series associated with multiple curves, take each water level and the corresponding maximum water storage to represent the maximum operating range of the reservoir as the upper envelope; take each water level and the corresponding minimum water storage to represent the minimum operating range of the reservoir as the lower envelope; (2) Use spline interpolation or polynomial fitting to generate a continuous envelope curve for discrete points in the upper envelope or lower envelope to ensure that the curve changes smoothly and conforms to the actual operation of the reservoir.

10. A system for extracting water reservoir operation rules based on multi-source remote sensing satellites, the system for extracting water reservoir operation rules based on multi-source remote sensing satellites being used to execute the steps in the method for extracting water reservoir operation rules based on multi-source remote sensing satellites according to any one of claims 1 to 9, characterized in that: include: Remote sensing image acquisition module: It is used to obtain Sentinel-1 remote sensing images and GF-3 remote sensing images of the target reservoir and perform preprocessing; Binarization processing module: It is used to perform mask cropping and binarization processing on the pre-processed Sentinel-1 remote sensing image and GF-3 remote sensing image respectively; Water body boundary extraction module: It is used to obtain the complete water body boundary based on the binarized Sentinel-1 remote sensing image and GF-3 remote sensing image, and then calculate the water surface area; Correction module: It is used to correct the water surface area calculated based on Sentinel-1 remote sensing images and GF-3 remote sensing images, and obtain the corrected and consistent water surface area sequence; Reservoir operation rule extraction module: It is used to extract reservoir operation rules based on water body boundaries and corrected and unified water surface area sequences.

Images