Reservoir capacity dynamic monitoring method and system based on unmanned aerial vehicle

Abstract

The embodiment of the invention provides a reservoir capacity dynamic monitoring method and system based on an unmanned aerial vehicle, and the method comprises the steps: obtaining a reservoir surface feature data set through the unmanned aerial vehicle, determining the boundary of a water area, scanning the bottom of the reservoir, obtaining the underwater terrain elevation and sediment distribution data, and constructing a three-dimensional terrain model; and determining terrain and deposit change data based on the three-dimensional terrain model dynamic change data of the continuous time sequence. Sedimentation space distribution characteristics are obtained through change rate analysis, and the sediment transportation amount and the sedimentation and deposition space rule are determined in combination with the water flow speed. Key driving factors are screened through correlation analysis, a sedimentation high-risk area is determined according to the key driving factors, and the change of sediments in the area is monitored, so that the change of the reservoir capacity is mastered. The reservoir siltation can be accurately and dynamically monitored, the high-risk area and the change mode are defined, the dynamic database is constructed to update the capacity change in real time, and important technical support is provided for reservoir management and flood control and disaster reduction.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of geographic information monitoring technology, and in particular to a method and system for dynamic monitoring of reservoir capacity based on an unmanned aerial vehicle (UAV). Background Art

[0002] Reservoirs are core hydraulic structures for water resource regulation and flood prevention and disaster reduction. The dynamic changes in their capacity are directly related to regional water supply security, agricultural irrigation efficiency, and the stability of flood control systems. With the intensification of climate change and the increasing intensity of river basin development, reservoir siltation is becoming increasingly prominent. Sedimentation leads to reduced storage capacity and reduced regulation capacity, which not only affects the service life of the projects but also poses a potential threat to downstream flood control safety. Therefore, accurate and efficient monitoring of reservoir topography and the dynamic changes in sedimentation has become a key requirement in water conservancy management.

[0003] Current methods for monitoring reservoir capacity have significant technical limitations. First, traditional manual measurement relies on equipment such as total stations and depth sounders, requiring the deployment of a large number of monitoring points across the reservoir area. This is not only time-consuming and labor-intensive, with limited coverage, but also hindered by hydrological conditions (such as high water levels and turbulent currents during flood season), making it difficult to achieve high-frequency, full-area monitoring. This results in insufficient data integrity and an inability to reflect the spatial heterogeneity of sedimentation. Second, while fixed-point sensors can achieve continuous data collection, their monitoring range is constrained by the density of hardware deployment. This can easily lead to blind spots in reservoirs with complex terrain and uneven sedimentation. Furthermore, sensors are often immersed in water for extended periods, resulting in high maintenance costs and data accuracy that is susceptible to water quality interference. Third, existing technologies often focus on analyzing sedimentation changes at a single point in time, lacking continuous comparison of data across multiple time periods. This makes it difficult to quantify sedimentation rates and long-term trends, and it is even more difficult to integrate driving factors such as rainfall and water flow to reveal the spatial patterns of sedimentation. Fourth, traditional methods lack sedimentation risk models, making it impossible to accurately identify high-risk areas. This results in inadequately targeted dredging projects and significant resource waste.

[0004] In addition, existing three-dimensional terrain modeling technologies mostly rely on a single data source (such as only surface remote sensing or underwater sonar), and the degree of integration between surface and underwater data is low, making it difficult to construct a complete three-dimensional model of the reservoir area, resulting in large errors in the calculation of sedimentation volume; at the same time, the analysis of the sedimentation driving mechanism lacks systematicity, and historical hydrological data and spatial distribution characteristics are not integrated, making it impossible to accurately judge the sediment transport path and deposition law, which restricts the scientific nature of the dynamic prediction of reservoir capacity.

[0005] Therefore, how to break through the technical bottleneck of traditional monitoring methods and achieve high-precision, full-cycle, and intelligent monitoring of reservoir topography and sediment through multi-source data fusion, multi-period dynamic analysis, and driving factor correlation modeling has become an urgent need to improve reservoir management and ensure project safety. Summary of the Invention

[0006] The embodiments of the present invention provide a method and system for dynamic monitoring of reservoir capacity based on drones to address the problem that the existing technology lacks systematic analysis of the sedimentation driving mechanism, fails to integrate historical hydrological data and spatial distribution characteristics, and cannot accurately determine the sediment transport path and deposition pattern, which restricts the scientific nature of dynamic prediction of reservoir capacity.

[0007] In a first aspect, an embodiment of the present invention provides a method for dynamic monitoring of reservoir capacity based on a drone, comprising: S100, obtaining a reservoir surface feature dataset by using a drone, wherein the reservoir surface feature dataset includes reservoir water surface image data; S200, determining water boundary information based on the reservoir surface characteristic dataset, performing a depth scan of the reservoir bottom based on the water boundary information, obtaining underwater terrain elevation data and sediment distribution data, and constructing a three-dimensional terrain model of the reservoir; S300, determining reservoir topography change data and sediment change data based on the dynamic change data of the three-dimensional topography model of the reservoir in a continuous time series; S400, performing a change rate analysis on the reservoir topography change data and the sediment change data to obtain a spatial distribution characteristic of the reservoir sediment change; S500, determining a sediment transport amount based on the spatial distribution characteristics and the reservoir flow velocity, and determining a spatial pattern of sediment deposition based on the sediment transport amount; S600, determining key driving factors affecting sediment deposition based on the spatial pattern of sediment deposition; the driving factors include rainfall and water flow velocity; S700: Determine a high-risk area for sedimentation based on the key driving factors, and monitor changes in reservoir capacity based on changes in sediment in the high-risk area.

[0008] In a second aspect, an embodiment of the present invention provides a dynamic reservoir capacity monitoring system based on an unmanned aerial vehicle, comprising: A data acquisition module is used to obtain a reservoir surface feature dataset using a multi-source remote sensing device carried by an unmanned aerial vehicle, wherein the reservoir surface feature dataset includes reservoir water surface image data; A terrain modeling module is used to determine water boundary information based on the reservoir surface feature dataset, perform a depth scan of the reservoir bottom based on the water boundary information, obtain underwater terrain elevation data and sediment distribution data, and construct a three-dimensional terrain model of the reservoir; A dynamic analysis module is used to determine reservoir topography change data and sediment change data based on the dynamic change data of the three-dimensional topography model of the reservoir in a continuous time series; A sedimentation rate analysis module is used to analyze the change rate of the reservoir topography change data and the sediment change data to obtain the spatial distribution characteristics of the reservoir sediment change; A sediment transport law analysis module is used to determine the sediment transport amount based on the spatial distribution characteristics and the reservoir water flow velocity, and to determine the spatial law of sediment deposition based on the sediment transport amount; A correlation analysis module is used to determine key driving factors affecting sediment deposition based on the spatial pattern of sediment deposition; the key driving factors include rainfall and water flow velocity; The dynamic monitoring module is used to determine the high-risk area for sedimentation based on the key driving factors, and monitor the change of reservoir capacity based on the change of sediment in the high-risk area for sedimentation.

[0009] The method and system for dynamic reservoir capacity monitoring based on drones provided by the embodiments of the present invention have the following advantages compared with the prior art: (1) The present invention can realize accurate dynamic monitoring of reservoir sedimentation. By combining multi-source remote sensing of drones with underwater sonar detection, the present invention integrates surface images and underwater elevation data to construct a three-dimensional terrain model. Combined with multi-period data collection and difference analysis, the present invention can accurately capture the dynamic changes of reservoir topography and sedimentation, overcoming the problems of incomplete data coverage and limited accuracy of traditional methods, and improving the comprehensiveness and accuracy of monitoring.

[0010] (2) The present invention can clarify the monthly rate, significant areas and spatial distribution characteristics of sedimentation changes through means such as change rate analysis and spatial distribution heterogeneity analysis; combined with water flow velocity simulation and sediment transport estimation, the spatial pattern of sediment deposition can be determined. At the same time, the influence weight of each factor on sedimentation can be determined through driving factor correlation test, making the analysis more scientific and in-depth, and improving the depth and reliability of sedimentation change analysis.

[0011] (3) The present invention uses a high-precision lidar equipped with an unmanned aerial vehicle to collect key data for areas corresponding to driving factors with weights exceeding the threshold. Combined with time trend fitting and outlier detection technology, it can accurately identify the distribution range and change pattern of high-risk areas for siltation, provide precise targets for risk prevention and control, and accurately identify high-risk areas for siltation.

[0012] (4) The present invention constructs a dynamic monitoring database, associates the classification results with the real-time collected data, and can judge the change status of the reservoir capacity due to siltation in real time, generate dynamic monitoring update data and real-time reports, and provide timely and reliable basis for reservoir dredging planning, water resources optimization allocation, flood control decision-making, etc., which helps to ensure the safety of reservoir projects and the sustainable use of regional water resources, and provides strong technical support for reservoir management and flood control and disaster reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0014] Figure 1 A flowchart of a method for dynamic monitoring of reservoir capacity based on drones provided in one embodiment of the present invention; Figure 2 This is a structural block diagram of a reservoir capacity dynamic monitoring system based on drones provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0016] Figure 1 This is a flowchart of a method for dynamic monitoring of reservoir capacity based on drones according to an embodiment of the present invention, referring to Figure 1 , the method comprising: S100, obtaining a reservoir surface feature dataset using a drone, wherein the reservoir surface feature dataset includes reservoir water surface image data; specifically, the dataset includes: Specifically, in this embodiment, a drone equipped with multi-source remote sensing imaging equipment conducts a comprehensive scan of the reservoir surface and surrounding terrain within a certain radius (e.g., a 1-kilometer radius). This multi-source remote sensing imaging equipment may include visible light cameras, infrared cameras, and lidar. The visible light camera captures image data of the reservoir surface, clearly displaying information such as its surface status and boundaries. The lidar acquires elevation data of the surrounding terrain, accurately measuring topographic relief. During the scan, the drone flies along a pre-set route to ensure comprehensive coverage and no areas are missed.

[0017] As a preferred implementation, the S100 specifically includes: S110: Using the multi-source remote sensing imaging equipment carried by the drone, a full-coverage scan is performed of the reservoir surface and surrounding terrain within a preset range, acquiring reservoir surface image data and surrounding terrain elevation data. The preset surrounding terrain range can be selected from a land area 1-3 kilometers outside the reservoir water boundary, with the specific range adjusted based on the reservoir size to ensure coverage of the catchment area and terrain transition zones that may affect the reservoir. The drone flies along a preset route (such as a grid or serpentine pattern) at a speed of approximately 50 kilometers per hour, covering a reservoir area of ​​approximately 10 square kilometers. The image resolution reaches 0.1 meters per pixel, and the elevation data accuracy is controlled within 0.05 meters. This ensures that there are no blind spots in the scan, and adjacent scan swaths have a 10%-20% overlap rate to ensure accurate data splicing.

[0018] S120, denoising and correcting the reservoir surface image data and the surrounding terrain elevation data to obtain a first reservoir surface image dataset and a first terrain elevation dataset; spatially registering and fusing the first reservoir surface image dataset and the first terrain elevation dataset, extracting terrain features from the fused data, and determining preliminary characteristic distribution information of the reservoir surface; During image processing, a deep learning-based image enhancement algorithm was used to remove noise and correct the image data. Geometric correction methods were also used to eliminate image distortion caused by the drone's flight attitude. For example, if the original image contained shadow areas caused by uneven lighting, the effects of these shadows were significantly reduced after processing. The clarity of the first image dataset of the reservoir surface was improved by approximately 30%, providing more reliable data support for subsequent feature extraction.

[0019] Spatial registration is achieved through geographic information system (GIS) tools (such as ArcGIS), aligning the coordinate system of the water surface imagery and terrain elevation data (e.g., WGS84), with a registration error within 0.2 meters. Data fusion integrates the two types of data into a unified dataset, linking water surface features with the surrounding terrain (for example, clearly showing the relationship between the slope changes along the reservoir bank and the connection with the water surface). Automated algorithms (such as edge detection and threshold segmentation) are used to extract key features from the fused data, including abnormal terrain such as cracks, subsidence zones, and scarps along the reservoir bank, as well as the boundary between the water surface and the land, to form preliminary feature distribution information.

[0020] S130, performing data filling on the preliminary characteristic distribution information to obtain a complete reservoir surface characteristic data set, performing three-dimensional modeling based on the reservoir surface characteristic data set to obtain a three-dimensional distribution model of the reservoir surface, and determining the final distribution characteristic information of the reservoir surface according to the three-dimensional distribution model.

[0021] For areas missing from the preliminary feature distribution information due to cloud cover, equipment failure, and other factors (e.g., blind spots less than 0.5 square kilometers), neighborhood-based interpolation algorithms (e.g., inverse distance weighted method) were used to infer feature values ​​in the missing areas using valid surrounding data, increasing the dataset completeness to over 98%. The infilled data, which includes intact water surface imagery, surrounding terrain elevations, and extracted topographic features, comprehensively reflects the spatial distribution of the reservoir surface. Based on the complete dataset, a 3D reconstruction tool (e.g., Context Capture) was used to construct a 3D distribution model of the reservoir surface. This model visually displays 3D features such as water surface elevation, bank slope, and depth anomalies. Key insights extracted from the 3D model include the precise boundaries of the reservoir surface, the slope classification of the bank (e.g., 0-5° for flat areas and 5-15° for gentle slopes), and the location and extent of potential risk points (e.g., cracks and subsidence areas), providing a foundation for subsequent water boundary determination and underwater scanning.

[0022] During the terrain feature extraction process, automated algorithms identify abnormal features such as cracks and subsidence on the reservoir surface. For example, if the extraction results indicate a crack approximately 0.3 meters wide in a certain area, the area will be marked as a potential risk point in the preliminary feature distribution information. This method helps quickly locate problem areas and provides a basis for reservoir safety assessment.

[0023] If there are missing data in the preliminary feature distribution information, such as incomplete image data in a certain area due to cloud cover, a neighborhood-based interpolation algorithm can be used to fill in the gaps. For example, if the missing area is approximately 0.5 square kilometers, the completeness of the second dataset of surface features after interpolation increases to over 98%, preventing the impact of missing data on subsequent modeling.

[0024] During the 3D reconstruction phase, specialized software was used to construct a three-dimensional distribution model of the reservoir surface based on the second dataset of surface features. If the model reveals unusual water depth variations in a particular area, possibly related to sediment accumulation, that area will be marked as a key monitoring point in the final distribution feature information. This three-dimensional model intuitively displays the spatial characteristics of the reservoir surface, providing a scientific basis for management decisions.

[0025] The above-mentioned technical means of this embodiment ultimately yield highly accurate and comprehensive reservoir surface distribution feature information, significantly improving the efficiency and reliability of reservoir safety monitoring. For example, in one application, this method enabled the early detection of potential leakage risk areas, avoiding potential economic losses and fully demonstrating the practical value of this technology in reservoir management.

[0026] S200, determining water boundary information based on the reservoir surface characteristic dataset, performing a depth scan of the reservoir bottom based on the water boundary information, obtaining underwater terrain elevation data and sediment distribution data, and constructing a three-dimensional terrain model of the reservoir; The S200 specifically includes: S210, determining water boundary information based on the reservoir surface feature dataset, performing a depth scan of the reservoir bottom using underwater sonar detection equipment, and obtaining original elevation point cloud data and bottom reflectance feature data of the underwater terrain; Specifically, an edge detection algorithm extracts the boundary between the water surface and land based on the reservoir surface image data and surrounding terrain elevation data from the reservoir surface feature dataset. This accuracy can be controlled to within 0.2 meters. This is used to clearly define the boundaries of the reservoir's underwater scanning range, ensuring that the scanning area does not deviate from the target waters. This boundary information effectively guides the scanning path planning of the multi-beam sonar equipment, ensuring that the scanning area does not deviate from the target waters.

[0027] For deep scanning of the reservoir bottom, multi-beam sonar equipment uses its transducer array to emit multi-angle sound wave signals distributed in a fan shape underwater. These signals reflect back to the device after contacting the reservoir bottom. By accurately calculating the propagation time and angle of the sound waves and combining them with real-time water sound velocity data, the device can simultaneously obtain the original elevation point cloud data and bottom reflection characteristic data of the underwater terrain. The original elevation point cloud data consists of the three-dimensional coordinates of a large number of discrete points, which can accurately reflect the undulations of the underwater terrain. The bottom reflection characteristic data records the intensity of the reflected sound waves in different areas, providing a basis for subsequent judgment of the bottom material properties.

[0028] S220, obtaining an echo intensity value in the bottom sediment reflection characteristic data; if the echo intensity value exceeds a preset echo intensity threshold, determining that the corresponding bottom material is bedrock terrain; otherwise, determining that the corresponding bottom material is a sediment distribution area, and obtaining a bottom material classification result and sediment layer thickness distribution data; Suppose, in a given scan, a sonar device acquires data at a frequency of 10 times per second, following a pre-set parallel or radial pattern, covering an area of ​​approximately 5 square kilometers at the bottom of a reservoir. The resulting point cloud data has a depth resolution of 0.1 meters. This high-frequency acquisition method can capture subtle changes in underwater topography, such as thin layers of sediment and small depressions, laying a solid data foundation for subsequent analysis. When determining the type of bottom material based on the reflection characteristic data, classification can be performed based on differences in echo intensity. Specifically, a pre-set echo intensity threshold is set. If the echo intensity in a certain area exceeds this threshold, the corresponding bottom material is classified as bedrock. If it is below this threshold, it is determined to be an area of ​​sediment. This generates bottom material classification results and sediment thickness distribution data. Suppose, during a scan of a reservoir, approximately 30% of the bottom area has low echo intensity, initially indicating sediment coverage. Further analysis reveals that the average sediment thickness in these areas is approximately 0.5 meters. This classification method can quickly identify the material distribution at the reservoir bottom.

[0029] S230, generating a continuous reservoir bottom terrain elevation network using a spatial interpolation method based on the bottom material classification result, and spatially registering the reservoir bottom terrain elevation grid with the reservoir water surface image data to obtain an integrated reservoir elevation dataset; In order to generate a continuous reservoir bottom terrain elevation grid using the spatial interpolation method, the Kriging interpolation method can be used to process the point cloud data. This method interpolates the discrete point cloud data by analyzing the spatial correlation between sample points to generate a continuous and uniformly accurate elevation grid. Assuming that in a certain processing, the interpolated grid resolution reaches 1 meter, it can better reflect the subtle undulating characteristics of the underwater terrain and effectively fill the blank areas that may exist in data collection.

[0030] When spatially registering the underwater elevation grid with the surface image data, professional coordinate transformation tools can be used to unify the two into the same geographic coordinate system for precise alignment. Assuming that the spatial error after registration is controlled within 0.3 meters, the surface image data and the underwater elevation data can be seamlessly connected to form a complete integrated reservoir elevation dataset, providing a unified and accurate spatial reference for subsequent comprehensive modeling.

[0031] S240, constructing a triangulated network for the integrated elevation dataset of the reservoir, and layer-labeling the bottom network according to the silt layer thickness distribution data, to obtain a three-dimensional terrain model of the reservoir including the surface morphology of the reservoir and the undulations of the underwater terrain.

[0032] When constructing a triangulated network using a 3D mesh generation tool, specialized geographic information or modeling software can be used to process an integrated elevation dataset, connecting discrete elevation points into a continuous triangulated network structure. Assuming the constructed triangulated network contains approximately 100,000 grid cells, it can meticulously depict the reservoir surface morphology and the undulating characteristics of the underwater terrain, providing intuitive and detailed data support for terrain analysis. When layering the bottom mesh based on the sediment thickness distribution data, different colors or labels can be used to distinguish the bottom mesh according to the different ranges of sediment thickness. For example, areas with a thickness of 0-0.5 meters can be marked as areas of general concern, areas with a thickness of 0.5-1 meter as areas of secondary concern, and areas above 1 meter as areas of key concern. For example, if the sediment thickness in a certain area of ​​the reservoir bottom reaches 1.2 meters, it will be marked as an area of ​​key concern. This layered labeling method allows managers to quickly identify potential problem areas, providing a clear reference for daily reservoir management and dredging planning.

[0033] S300, determining reservoir topography change data and sediment change data based on the dynamic change data of the three-dimensional topography model of the reservoir in a continuous time series; For the comprehensive three-dimensional terrain model, terrain scanning data at different time points are acquired through multi-period data collection to form a continuous time series data set. The terrain elevation difference of each scan is compared, and the elevation change value and sediment volume change calculation results are extracted to obtain preliminary quantitative results of the dynamic changes of reservoir terrain and sediment. Specifically, it includes: S310, acquiring depth scan data of the reservoir bottom from preset time nodes through a multi-period data acquisition method, forming a set of underwater terrain elevation data including multiple time nodes, and obtaining a preliminary time series data set; Specifically, according to the needs of reservoir monitoring, a terrain scanning pattern is repeated at different time points. Its purpose is to capture dynamic changes by comparing data at different times, and at the same time combine it with the hydrological characteristics of the reservoir. For example, before and after the flood season every year (April and October), the water level and siltation status of the reservoir change significantly, which can more accurately reflect the evolution of the terrain; the depth scanning data at the bottom of the reservoir is the original underwater terrain data obtained by equipment such as multi-beam sonar, including elevation point clouds and geological information; the underwater terrain elevation data set is a data set formed by aggregating the scanning data of multiple time nodes. The data of each node is stored in the form of a digital elevation model. The data at each time point covers the entire reservoir to ensure data integrity; the preliminary time series data set is a combination of elevation data arranged in chronological order, which provides a basis for subsequent difference analysis.

[0034] S320, performing layer-by-layer difference calculation on the underwater terrain elevation data of every two adjacent time nodes in the time series data set to determine an elevation difference distribution map; After the reservoir bottom terrain elevation data collected in multiple time periods are sorted in chronological order, they are processed by a special elevation data comparison tool (such as a terrain analysis module developed based on a GIS platform). This tool can calculate the difference between the elevation data at different time points layer by layer and region by region according to the preset elevation hierarchy or geographical division. Suppose that in an actual analysis, the elevation data of two time nodes in April (before the flood season) and October (after the flood season) are selected for comparison. After comparing the two sets of data point by point through the tool, it is found that the elevation of some areas at the bottom of the reservoir has increased by 0.5 meters compared with April. Combined with the hydrological characteristics of the reservoir and the law of sediment deposition, this elevation increase indicates that there may be sediment accumulation in the area. This layer-by-layer comparison method can accurately locate every area where elevation changes occur, whether it is a large-scale siltation belt or a small-scale local deposition, and can be clearly identified. The tool will then generate an elevation difference distribution map with these change data in a visual way. The map uses different colors or color blocks to mark the numerical range and spatial position of the elevation changes, intuitively showing the spatial distribution characteristics of terrain changes in the entire reservoir area; this method can not only quickly identify key areas where siltation problems may exist and reduce the workload of manual investigation, but also provide a reliable basis for subsequent siltation volume calculations, risk assessments, etc. through precise spatial positioning and quantitative data, thereby effectively improving the efficiency and accuracy of the entire monitoring and analysis process.

[0035] S330, if it is determined based on the elevation difference distribution map that the elevation change value of a certain area exceeds a preset elevation change threshold, then calculating the underwater terrain elevation change value of the corresponding area, and calculating the sediment volume change value based on the underwater terrain elevation change value; For example, based on the elevation difference distribution map, if data screening finds that the elevation change values ​​of certain areas exceed the preset judgment threshold (such as 0.3 meters), it means that the terrain changes in these areas have reached a level that requires special attention, and it is necessary to further use volume calculation tools (such as terrain analysis software based on three-dimensional modeling technology) to accurately estimate the sediment volume of these areas; specifically, the volume calculation tool will first lock the boundary range of the changed area according to the elevation difference distribution map, and then build a three-dimensional model based on the elevation data of the area at two time nodes, and obtain the actual volume of the sediment through the spatial volume integral operation of the elevation change part in the model; assuming that the elevation change value of a certain area is found to be 0.4 meters after comparison, and the plane projection area of ​​the area is measured to be 5,000 square meters, through After using the tool's 3D modeling and volume calculations, the silt volume in the area was estimated to be approximately 2,000 cubic meters. A comprehensive analysis was also conducted based on the reservoir's topographical characteristics. For example, the 3D terrain model constructed in the early stages revealed that the terrain in this area has a gentle slope (e.g., a slope between 2 and 5 degrees). According to the laws of sediment movement, the velocity of water in areas with gentle slopes decreases significantly, and the sediment carried is prone to sedimentation. This suggests that siltation in this area is primarily caused by sediment brought in by rainfall or floods upstream, which is then transported and deposited by water. This analytical approach not only quantifies the volume changes of sediment through specific numerical values, but also reveals the causes of siltation in combination with topographic and hydrological characteristics, providing a scientific and specific basis for reservoir management departments to formulate targeted dredging plans and optimize water resource scheduling.

[0036] S340, integrating the elevation difference distribution map and the sediment volume change value, obtaining the dynamic change trends of the reservoir topography and sediment volume in different time periods, and obtaining reservoir topography change data and sediment change data.

[0037] After integrating the elevation difference distribution map and the sediment volume change values, the elevation difference distribution map reflecting the spatial changes in the terrain is correlated and matched with the volume data that quantifies the increase and decrease in sediment. The integrated information is processed through professional data visualization tools (such as Matplotlib, ArcGIS's dynamic mapping module, etc.) to generate a dynamic chart that can intuitively present the changes in the time dimension; assuming that in a certain application, the generated chart uses the time axis (such as month, quarter or year) as the horizontal axis, clearly marks each data collection node, uses the left vertical axis to represent the elevation change value (unit: meter), and the right vertical axis represents the sediment volume change (unit: cubic meter), and uses different colors (such as red for areas with increased elevation and blue for areas with decreased elevation) to mark the changes in different sub-areas of the reservoir At the same time, through broken line or bar graphs, the elevation change trends and sediment volume increases and decreases in different areas are respectively displayed, so that managers can observe at a glance the dynamic trends of the reservoir terrain in different time periods. For example, the elevation of a certain area rises significantly and the sediment volume increases rapidly after the flood season, while the change in the dry season is relatively gentle. This visualization result converts complex spatial data and numerical information into intuitive and easy-to-understand charts, which makes it easier for reservoir managers to quickly grasp the sedimentation status, change speed and potential risks of the entire region and local areas, so that they can formulate and adopt targeted dredging plans (such as giving priority to dredging operations in high-sedimentation areas) or flood control measures (such as predicting the impact of reservoir capacity reduction on flood storage capacity in advance), effectively shortening the time from data acquisition to decision implementation, and significantly improving the efficiency and scientific nature of reservoir management.

[0038] The above steps of this embodiment, through the combination of multi-period data collection and difference analysis, can effectively monitor reservoir topography changes and provide early warning of potential risks, such as siltation leading to reduced storage capacity or flooding hazards. The generation of dynamic change charts transforms complex data into intuitive information, lowers the threshold for understanding, and improves decision-making support capabilities. The overall solution forms a complete closed loop through data collection, difference calculation, volume estimation, and visual presentation, ensuring the accuracy and practicality of reservoir topography monitoring. S400, performing a change rate analysis on the reservoir topography change data and the sediment change data to obtain a spatial distribution characteristic of the reservoir sediment change; In this example, based on the preliminary quantitative results of dynamic changes, a change rate analysis method was used to calculate the monthly rate of change of topography and sediment. At the same time, spatial distribution heterogeneity analysis was used to identify areas of significant change, thereby obtaining the spatial distribution characteristics of reservoir sedimentation changes. Specifically, the following are included: S410, acquiring reservoir topography change data for multiple periods from a preset time series, performing monthly analysis on the reservoir topography change data, calculating the monthly change rate, and obtaining a reservoir topography monthly change rate distribution dataset; In the business scenario of dynamic monitoring of reservoir sedimentation, monthly analysis of terrain change data over multiple time periods is a key link in grasping the short-term evolution of sedimentation. Its core lies in capturing subtle trends in terrain changes through high-frequency data collection and comparison. Specifically, a monthly terrain elevation data collection plan needs to be formulated for the target reservoir. The collection time can be fixed in the middle and late part of each month (to avoid the influence of extreme weather). The entire reservoir area (including the water surface and underwater terrain) is scanned synchronously by drone equipped with a combination of lidar and multi-beam sonar equipment to ensure that the resolution of the elevation data collected each time is consistent (for example, the accuracy of underwater terrain elevation is controlled within 0.05 meters). The data is then organized by month to form a continuous time series data set. Each data entry contains the digital elevation model of the corresponding month, a vector diagram of sediment distribution, etc. During the data processing phase, time interval comparison tools (such as the Python-based time series analysis module, which integrates a sliding window algorithm and a difference calculation function) are used to perform a point-by-point comparison of terrain elevation data from two consecutive months. Spatial registration is first used to ensure the coordinate systems of the two data months are identical. The elevation difference at the same location is then calculated. Combined with the area parameters of the region, the rate of terrain change per unit time (month) is derived. For example, in the scanned data from May to June, the elevation of a reservoir upstream increased from 100.2 meters to 100.4 meters, a difference of 0.2 meters, resulting in a calculated rate of change of 0.2 meters per month. However, from June to July, the elevation of the region increased by only 0.1 meters, resulting in a rate of change of 0.1 meters per month. This monthly rate of change not only provides a visual indication of the speed of terrain change but also allows identification of anomalous signals within short-term fluctuations (for example, a sudden increase in rate in a particular month may be related to sediment input from heavy rainfall). The results will serve as the basis for subsequent spatial distribution analysis and correlation with driving factors, providing a quantitative basis for accurately identifying periods of active sedimentation and providing early warning of high-risk areas.

[0039] S420, dividing the reservoir topography monthly change rate distribution dataset into spatial regions, and if it is determined that the monthly change rate of a certain region after the spatial region division exceeds a preset rate threshold, marking the corresponding region as a significant region, and determining the spatial distribution range of the significant region; Spatial analysis of the monthly change rate distribution dataset is a key step in accurately locating active sedimentation areas. The specific process requires a combination of specialized tools and reservoir geographic characteristics. First, using spatial distribution analysis tools (such as ArcGIS's spatial analysis module or QGIS's zonal statistics function), the reservoir is divided into subregions based on topographic complexity (e.g., grid sizes of 1,000-5,000 square meters). This division takes into account both flow paths and topographic units (e.g., dividing the inlet area, bay area, and main channel area into separate subregions) to ensure relatively uniform topographic characteristics within each subregion. Next, a rate threshold (e.g., 0.15 meters / month) is set based on the reservoir sedimentation risk assessment. This threshold should be informed by historical sedimentation data and management requirements (e.g., the threshold can be lowered to increase sensitivity in key flood control areas). The tool then compares the monthly change rates of each subregion in batches. If a subregion's rate exceeds the threshold (e.g., a rate of 0.2 meters / month in May-June), it is marked as a significant area, and parameters such as its boundary coordinates and area are automatically recorded. Taking a certain reservoir as an example, analysis revealed that significant areas are primarily concentrated near the upstream water inlet, covering an area of ​​approximately 3,000 square meters. Combined with flow simulation data, this area is prone to sediment deposition due to water flow deceleration. Furthermore, during the flood season of May and June, sediment inflow surges, further confirming the high sedimentation activity in this area. This regional division and significance assessment allows monitoring to be focused on these key areas of change. Subsequently, the frequency of data collection in these areas can be increased (for example, from monthly to biweekly), and targeted water flow and sediment transport monitoring can be conducted. This significantly improves the targeting and efficiency of monitoring, providing data support for the precise formulation of dredging plans.

[0040] S430, extracting heterogeneous features from the spatial distribution range of the significant area to obtain heterogeneous distribution data of reservoir topography changes and sediment changes, and generating a heterogeneous feature distribution map based on the heterogeneous distribution data; When analyzing the heterogeneity characteristics of significant regions, this step aims to delve deeper into the spatial differences in topographic changes within the same significant region, breaking through the limitations of focusing solely on overall regional changes, thereby more accurately grasping the detailed patterns of siltation distribution. Specifically, heterogeneity feature extraction tools (such as GIS-based spatial autocorrelation analysis modules and hotspot analysis tools) can be used to first subdivide the significant region into smaller grid cells (e.g., 10 m x 10 m). Statistical analysis of data such as the rate of elevation change and siltation thickness for each subdivided cell can then be performed to identify the spatial differentiation patterns between areas of rapid change and areas of moderate change. Assuming that in the above-mentioned significant area of ​​about 3,000 square meters, calculations using tools found that the 500-square-meter area close to the water inlet had an average monthly elevation change rate of up to 0.3 meters per month due to the direct impact of sediment carried by upstream water, and the sediment was mainly coarse-grained sediment; while the 2,500-square-meter area far from the water inlet was affected by the diffusion of water flow, and the sediment deposition rate gradually decreased, with an average monthly elevation change rate of only 0.05-0.1 meters per month, and the sediment was mainly fine-grained silt. This difference directly indicates that there is obvious unevenness in the silt distribution within this significant area. When generating a heterogeneous distribution map, the tool visually annotates the differences in elevation within each subdivision using a color gradient (e.g., dark red to light red indicates a high to low rate), and overlays auxiliary information such as flow direction arrows. The map clearly shows that, affected by the attenuation of water flow dynamics, the rate of elevation change gradually decreases from the water inlet toward the interior of the region. This means that the upstream area near the water inlet becomes a sedimentation "hotspot" due to strong sediment deposition, while downstream areas experience relatively gentle changes. This refined heterogeneity analysis not only reveals the inherent spatial distribution pattern of sediment deposits within significant regions, with sediment deposits being thick near the source and thinner far from the source, but also provides precise guidance for subsequent targeted measures. For example, prioritized dredging plans for "hotspots" can be combined with flow control to reduce sediment deposition in these areas, thereby improving the scientific nature and efficiency of reservoir management.

[0041] S440: Based on the heterogeneous characteristic distribution map, the monthly change rate is integrated with the spatial distribution range of the significant area to generate a comprehensive characteristic distribution map, and the spatial distribution characteristics of the reservoir sediment change are determined based on the comprehensive characteristic distribution map.

[0042] Deeply integrate the rate of change in the temporal dimension with the significant regional distribution in the spatial dimension to achieve visual integration of multi-dimensional data. Specifically, it is necessary to use professional data integration tools (such as ArcGIS's layer overlay function and ENVI's spatial analysis module) to pre-process the monthly rate of change data and the significant regional distribution data: using coordinate system 1 (such as the unified WGS84 coordinate system) to ensure that the spatial positions of the two are accurately matched, and then classify the monthly rate of change data into numerical ranges (such as 0-0.1 meters / month, 0.1-0.2 meters / month, and >0.2 meters / month), laying the foundation for subsequent visual annotation.

[0043] During the fusion processing phase, the tool uses the spatial boundaries of significant areas as a base layer, overlaying monthly rate-of-change data for the corresponding time period. For example, rate-of-change data from May to July (a critical period during the flood season, when sediment input is active) is selected and spatially correlated with previously identified significant areas (such as a 3,000-square-meter area around the upstream water inlet), assigning specific rate values ​​to each subunit of the significant area. A gradient color scale is used to generate the comprehensive feature distribution map: dark colors such as deep red and dark blue are used to mark high-variance areas with rates exceeding a threshold (such as 0.15 meters per month), while light colors such as light yellow and light gray are used to mark low-variance areas with rates below the threshold. A scale bar, legend, and time labels are also added to the map to ensure complete information.

[0044] For example, a comprehensive distribution map of characteristics for a particular reservoir clearly shows that from May to July, the upstream water inlet area maintained a high rate of change of 0.2-0.3 meters per month (dark concentrated area), which closely overlaps with the previously marked significant areas. Meanwhile, the midstream and downstream areas of the reservoir are mostly marked in light colors, with rates generally below 0.1 meters per month. Further analysis reveals that the high-variability areas extend in a strip-like pattern along the direction of water flow, which closely matches the sediment transport pathway, confirming the conclusion that the upstream area continues to experience a high rate of change.

[0045] The core value of this fusion processing lies in integrating dispersed time series data (monthly rates) and spatial distribution data (significant areas) into a single, intuitive chart. This not only preserves the quantitative information of the rate values, but also highlights the spatial distribution patterns, allowing analysts to simultaneously grasp multi-dimensional information such as "where the changes are occurring, how quickly they are occurring, and how long the changes are occurring." For example, by comparing comprehensive charts from different time periods, it is possible to identify long-term stable high-risk areas and short-term fluctuating anomalies. This provides comprehensive data support for the development of differentiated monitoring plans (such as intensified sampling in high-risk areas) and dredging plans, significantly improving the systematicity and accuracy of dynamic sedimentation analysis.

[0046] S500, determining a sediment transport amount based on the spatial distribution characteristics and the reservoir flow velocity, and determining a spatial pattern of sediment deposition based on the sediment transport amount; Specifically, based on the spatial distribution characteristics of reservoir sedimentation changes, historical rainfall data and inflow fluctuation records are integrated, and water velocity distribution simulation technology is used to calculate the water velocity in different areas of the reservoir, estimate the sediment transport volume, draw a sediment thickness distribution map, and determine the spatial pattern of sediment deposition. The S500 specifically includes: S510, determining a temporal distribution characteristic of inflow fluctuations based on historical rainfall data of a preset surrounding area of ​​the reservoir, and determining a dynamic change dataset of the inflow based on the temporal distribution characteristic of the inflow fluctuations; For example, when analyzing rainfall data and inflow fluctuations in areas surrounding a reservoir, key information can be extracted from historical records and further explored in conjunction with specific scenarios. Suppose rainfall data for an area surrounding a reservoir is derived from a time series database from the past five years. Records show that rainfall intensity peaked in July of a certain year, with a maximum daily rainfall of 120 mm. Using a data comparison tool, this rainfall record was correlated with the peak flow data for the same period. It was found that within 24 hours of the peak rainfall, inflows fluctuated significantly, with a peak flow of 500 cubic meters per second. This temporal distribution characteristic demonstrates the direct impact of rainfall on inflows, providing a foundation for the subsequent construction of a dynamically changing dataset.

[0047] S520, performing a water flow path simulation based on the dynamically changing dataset of the inflow flow, simulating the water flow velocity distribution characteristic values ​​of different areas in the reservoir, and obtaining a water flow velocity distribution characteristic map of each area; For example, using a flow path simulation tool combined with terrain elevation data to analyze a dynamically changing dataset of inflows can simulate the distribution of water velocity across different areas within the reservoir. Suppose the terrain upstream of the reservoir is steep, with an elevation difference of 10 meters, while the downstream area is relatively flat. The simulation tool revealed that water velocity is generally faster in the upstream area, averaging 2.5 meters per second, while downstream, it is only 0.8 meters per second. By mapping the water velocity distribution, it is possible to clearly identify areas of velocity differences, laying the foundation for subsequent sediment transport analysis.

[0048] S530, screening out areas where the water velocity distribution characteristic value exceeds a preset water velocity threshold, estimating sediment transport using a sedimentation calculation tool based on key physical parameters of reservoir sediment and the water velocity distribution characteristic map, and obtaining sediment transport distribution data corresponding to each area; the key physical parameters include average particle size, density, and settling velocity; For example, when estimating sediment transport, if the characteristic value of water velocity in a particular area exceeds a preset threshold of 1.5 m / s, further analysis is conducted based on a database of sediment particle characteristics. Assuming the water velocity in the upstream area reaches 2.5 m / s, significantly exceeding the threshold, and considering the average sediment particle size of 0.2 mm in the database, sediment calculation tools (such as the sediment transport module in HEC-RAS software) estimate the sediment transport in that area to be approximately 1,000 tons per month. This distribution data helps identify the primary source areas of sediment.

[0049] S540 , generating a sediment thickness distribution map using a spatial overlay tool based on the sediment transport distribution data and reservoir area division information to obtain a spatial pattern of sediment deposition.

[0050] For example, when generating a sediment thickness distribution map based on sediment transport data, the spatial overlay tool can be used to divide the reservoir into multiple subregions for analysis. Assuming that sediment transport is high in the upstream region, combined with this regional division information, the generated distribution map reveals an average sediment thickness of 0.3 meters in this region, while in the downstream region it is only 0.05 meters. This spatial pattern indicates that sediment is primarily concentrated upstream, providing a basis for targeted management decisions.

[0051] S600, determining key driving factors affecting sediment deposition based on the spatial pattern of sediment deposition; the driving factors include rainfall and water flow velocity; In this embodiment, based on the spatial pattern of sediment deposition, the influence of slope on the sediment deposition location is calculated using the terrain slope impact analysis method, and the correlation test technology of driving factors is used to analyze the correlation between rainfall, water flow velocity and sedimentation, construct a weight distribution matrix, and determine the weight distribution of each driving factor. The above S600 specifically includes: S610, obtaining terrain slope distribution data of different reservoirs, and classifying the slope values ​​of different reservoir areas according to the terrain slope distribution data to obtain slope grade distribution information of each reservoir area; For example, when analyzing terrain slope distribution data within a target reservoir area, relevant information can be first extracted from the database and classified based on specific regional characteristics. Terrain slope databases typically contain elevation change data around and within reservoirs. Slope classification tools can be used to categorize slope values ​​into multiple levels, such as 0-5 degrees for a gentle zone, 5-15 degrees for a moderate slope zone, and over 15 degrees for a steep slope zone. For example, suppose the slope in the upstream area of ​​a reservoir is mostly between 10-20 degrees, meaning it is a steep slope zone, while the slope in the downstream area is mostly between 2-5 degrees, meaning it is a gentle zone. This distribution information lays the foundation for subsequent analysis.

[0052] It should be noted that the acquisition of slope grade distribution information can be achieved through geographic information system software. During operation, it is only necessary to import elevation data and set classification standards to complete the classification.

[0053] S620, analyzing the correspondence between slope grade and sedimentation location using a spatial superposition method based on the slope grade distribution information and the historical records of sedimentation distribution, and determining the distribution characteristics of the influence of the slope grade on the sedimentation location; For example, to analyze the correlation between slope grade and sediment deposition locations, spatial overlay tools can be used to overlay a slope distribution map with historical sediment distribution records. Suppose historical records show that sediment deposition is higher in the upstream steep slopes and lower in the downstream flat areas. Overlay analysis reveals that the steeper the slope, the more concentrated the sediment deposition is in the turning point where the water flow slows. This pattern indicates that slope significantly influences sediment deposition locations. To implement this, spatial analysis software can be used to load the two layers of data, set overlay rules, and generate a distribution map.

[0054] S630, obtaining rainfall data and water flow velocity data, using the rainfall data and water flow velocity data as driving factors, and using a correlation test tool to calculate the correlation coefficient between the driving factors and the sediment distribution. If the correlation coefficient of a certain driving factor exceeds a preset threshold, the corresponding driving factor is marked as a key driving factor.

[0055] For example, when screening for key drivers, rainfall and water velocity, as primary data sources, can be analyzed using correlation testing tools. Suppose historical records for a reservoir region show peak rainfall of 150 mm / day in a certain month, while peak water velocity during the same period was 2 m / s. Testing reveals a correlation coefficient of 0.85 between rainfall and sediment distribution, exceeding the preset threshold of 0.7 and thus flagged as a key driver. Correlation testing can be performed using statistical analysis software, which automatically generates coefficient results by inputting two sets of data.

[0056] For example, weight matrix construction tools can be used to calculate the weights assigned to key driving factors. For example, if rainfall and water velocity are assigned weights of 0.6 and 0.4, respectively, historical data analysis reveals that when rainfall is weighted higher, sediment distribution tends to be more upstream. Generating this weighted distribution data helps clarify the impact of each factor on spatial distribution patterns. In practice, factor data and weighting rules can be input into decision support system software to quickly generate distribution results.

[0057] It should be noted that the weight calculation needs to be adjusted in combination with long-term observation data to ensure that the results are in line with reality.

[0058] S700: Determine a high-risk area for sedimentation based on the key driving factors, and monitor changes in reservoir capacity based on changes in sediment in the high-risk area.

[0059] In this embodiment, the terrain elevation data within the target area is obtained based on a pre-established reservoir terrain database, and a hierarchical processing tool is used to classify the elevation differences within the area for the terrain elevation data, so as to obtain the elevation distribution characteristics of the area. Through the elevation distribution characteristics, combined with the weight value information of the driving factor, if the weight value exceeds the preset threshold, the area marking tool is used to divide the corresponding reservoir terrain area to determine the boundary of the key monitoring area. For the boundary of the key monitoring area, a high-precision scanning device is used to collect detailed data within the area, obtain the local terrain detail information of the area, and determine the distribution of high-risk points for siltation within the area. Based on the local terrain detail information, a boundary demarcation tool is used to define the scope of high-risk points for siltation within the key monitoring area to obtain the specific distribution range of the high-risk area for siltation. The S700 specifically includes: S710: For the reservoir terrain area corresponding to the key driving factors, using a drone equipped with a laser radar to collect key data, obtain local terrain elevation change information, and determine high-risk sedimentation areas based on the local terrain elevation change information; For example, when analyzing elevation data from a reservoir terrain database, one can first extract relevant information about the target area from the database and then use a stratification tool to classify elevation differences. For example, if the elevation range of a reservoir area is between 100 and 500 meters, stratification can be used to categorize the elevations into three levels: low, medium, and high, representing 100-200 meters, 200-350 meters, and 350-500 meters, respectively. This classification helps clarify the elevation distribution characteristics within the area and provides basic data support for subsequent analysis. Stratification tools typically rely on a geographic information system (GIS) platform, completing the operation by importing elevation data and setting classification criteria. Developed based on GIS platforms (such as ArcGIS and QGIS), stratification tools are elevation data classification modules. Their core function is to classify continuous elevation values ​​into discrete levels based on preset criteria. To use this tool, you first need to import elevation data from the reservoir terrain database (e.g., values ​​in the 100-500 meter range). Then, you can use the tool to set the classification method: you can choose equidistant classification (e.g., fixed intervals of 100 meters), natural breakpoint classification (automatic classification based on data distribution characteristics), or manually define thresholds (e.g., 100-200 meters, 200-350 meters, and 350-500 meters in the text). After processing, the output is a layered elevation vector or raster map, which visually displays the spatial distribution of different elevation areas, providing a structured data foundation for subsequent analysis of sedimentation risk in conjunction with driving factors.

[0060] For example, when analyzing elevation data in a reservoir terrain database, you can first extract relevant information about the target area from the database and classify the elevation differences using a layered processing tool. Assuming that the elevation range of a reservoir area is between 100 and 500 meters, after layered processing, the elevation can be divided into three levels: low, medium, and high, which are 100-200 meters, 200-350 meters, and 350-500 meters, respectively. This classification method helps to clarify the elevation distribution characteristics within the region and provide basic data support for subsequent analysis. Layered processing tools usually rely on a geographic information system platform, and operations can be completed by importing elevation data and setting classification standards.

[0061] Based on the combined analysis of elevation distribution characteristics and driving factor weights, if a driving factor weight of 0.75 exceeds the preset threshold of 0.6, the relevant area should be marked as a priority. Assuming that this driving factor is associated with areas with higher elevations upstream, a higher weight indicates a higher risk of sedimentation in that area. The regional marking tool, implemented through spatial analysis software, overlays the elevation distribution map with the weight data to quickly delineate the boundaries of key monitoring areas. This method effectively identifies high-risk areas and improves the targeting of monitoring. The regional marking tool, a functional component of spatial analysis software (such as ArcGIS's Spatial Overlay Module and ENVI's Thematic Mapping Tools), spatially links elevation distribution characteristics with driving factor weight data and marks key areas. The specific operation involves first unifying the coordinate systems of the two (e.g., WGS84), then overlaying the elevation distribution map with the driving factor weight layer. Areas with weights exceeding a preset threshold (e.g., 0.6) are automatically assigned a special designation (e.g., red highlighting, bold borders). For example, when the weight value of a driving factor (such as upstream sediment load) is 0.75, the tool will quickly locate the elevation area where the factor has a significant impact (such as the area 350-500 meters upstream), generate a marked vector boundary map, and visually identify high-risk areas.

[0062] After defining the boundaries of key monitoring areas, it's crucial to use high-precision scanning equipment to collect detailed data within them. For example, for example, a section upstream of a reservoir has been designated as a key monitoring area. High-precision scanning equipment can capture local topographic details, such as slope variations or micro-topographic features within a small area. Scanning equipment typically includes a laser rangefinder or high-resolution image acquisition device. Through on-site operations, precise topographic data maps can be generated, allowing the location of high-risk siltation points to be determined. This detailed data acquisition helps more accurately identify potential risk points.

[0063] Using detailed local terrain information, the delineation tool can be used to define high-risk areas for siltation, further clarifying the specific distribution of risk areas. For example, if scan results reveal multiple slope abrupt changes within an area, and historical data indicates frequent siltation near these points, the delineation tool can be used to define the specific area. To operate the tool, detailed data is typically loaded into a geographic information platform, and risk assessment criteria are set to automatically generate a distribution map. This approach provides a precise spatial reference for subsequent management measures. The delineation tool is a component of geographic information platforms (such as QGIS's vector editing tools and AutoCAD Civil 3D) used to define the risk area. To operate, detailed local terrain data (such as the coordinates of slope abrupt changes and micro-sags) acquired through high-precision scanning is loaded. Risk assessment criteria are then set (e.g., slopes > 15° and points with historical siltation records). The tool then automatically connects adjacent high-risk points using a spatial clustering algorithm to generate a closed vector map of the risk area boundary (in a format such as SHP). For example, when a scan reveals multiple slope mutation points in an area and a history of siltation, the tool can quickly delineate the specific scope of the area (such as a polygonal boundary), providing an accurate spatial reference for management measures such as dredging operations and the deployment of monitoring equipment.

[0064] The definition of high-risk areas for siltation can be analyzed from multiple perspectives, combining topographical features with historical siltation records. For example, consider an upstream area at a higher elevation with dramatic slope variations and repeated siltation accumulation over the past three years. High-precision scanning confirms the presence of multiple micro-depressions in this area, which are prone to sediment accumulation. Furthermore, weighted analysis reveals significant influence from driving factors in this area, making it a priority target for monitoring. This multi-dimensional analysis supports each other, ensuring the rationality of the delineation results, while also providing a reliable basis for reservoir management and reducing resource waste.

[0065] S720: Processing the time series elevation data of the high-risk area to extract long-term trend characteristics, and using an outlier detection method to locate short-term mutation characteristics that deviate from the long-term change characteristics, and obtaining a classification result of the sedimentation change pattern of the high-risk area based on the long-term trend characteristics and the short-term mutation characteristics; In this embodiment, time series elevation data of the target area are obtained from a pre-established reservoir terrain database, and the time series elevation data are piecewise smoothed using a sliding window method to obtain a denoised elevation change curve. The denoised elevation change curve is fitted with a long-term trend line using the least squares method, and the elevation residual of each time point is calculated. If the elevation residual exceeds a preset fluctuation threshold, it is marked as an outlier. Based on the distribution of the outliers and in combination with the pre-determined boundaries of the high-risk areas, a clustering algorithm is used to spatially aggregate the outliers to determine the concentrated areas of mutation events. By superimposing the concentrated areas of mutation events with local terrain details, a region growing algorithm is used to expand the boundaries to determine the scope of the sub-area where the sedimentation rate is significantly increased.

[0066] When obtaining time-series elevation data for a target area from a reservoir topography database, one can first select elevation records for the past five years upstream of a specific reservoir, with monthly data intervals. Assuming the elevation data for this region ranges from 120 to 450 meters, the data may contain noise due to measurement errors or short-term weather effects. Using a sliding window method, the time series data is smoothed over three-month windows to produce a relatively smooth elevation curve. This method effectively reduces the interference of short-term fluctuations on subsequent analysis and lays the foundation for extracting long-term trends.

[0067] Using the least squares method to fit a long-term trend line to the denoised elevation curve, five years of data can be fitted into a gradually descending trend line, reflecting the possibility of terrain lowering due to sedimentation. For example, if the elevation at a certain point in time is 380 meters, while the trend line predicts 390 meters, with a residual error of 10 meters, and the preset fluctuation threshold is 8 meters, this point will be marked as an outlier. This marking method helps detect irregular fluctuations in elevation changes, providing clues for subsequent identification of risk areas.

[0068] When performing spatial aggregation based on the distribution of outliers and the boundaries of high-risk areas, clustering algorithms can be used to group outliers by spatial location. For example, suppose there are 10 outliers in the upstream area of ​​a reservoir, 8 of which are concentrated in a small area. Cluster analysis can aggregate these points into a single area of ​​concentrated mutation events. This aggregation method can help focus on local areas where significant topographic changes may occur, avoiding the waste of resources associated with dispersed analysis.

[0069] By overlaying the concentrated area of ​​sudden events with local terrain details, the region growing algorithm can be used to expand the boundary, starting from the center of the concentrated area and gradually expanding outward to include adjacent areas with steeper slopes or sudden elevation changes. For example, if the slope near the center of a concentrated area reaches 15 degrees and there are multiple micro-depressions around it, the algorithm can expand to identify a sub-area of ​​approximately 2 square kilometers, indicating a significantly increased sedimentation rate. This expansion method can more comprehensively cover potential risk points and provide a spatial basis for precise management.

[0070] To identify sub-areas with significantly elevated sedimentation rates, we can analyze them from multiple perspectives, combining historical sedimentation records with topographical features. Assuming a sub-area has experienced repeated sediment accumulation over the past two years, and topographic scans reveal numerous low-lying areas prone to sediment accumulation, coupled with high upstream water flow and the resulting high sediment load, this sub-area is considered a high-risk zone. This multi-dimensional analysis can be mutually verified, ensuring the rationality of the delineation and providing reliable support for reservoir sedimentation prevention and control.

[0071] During implementation, the window size and threshold parameters for processing time series data and flagging outliers can be adjusted based on the specific reservoir environment. For example, if a reservoir area is significantly affected by seasonal flooding, the sliding window could be adjusted to six months to capture longer-term fluctuations. This flexibility improves the adaptability of data analysis, ensuring that results are more aligned with actual needs and providing more effective technical support for subsequent risk prevention and control.

[0072] S730, based on the classification results of the siltation change pattern, a dynamic monitoring database is constructed, the siltation change pattern is correlated and matched with the underwater terrain elevation data collected in real time, the real-time status of the reservoir capacity change caused by the reservoir siltation is determined, and dynamic monitoring update data is obtained.

[0073] In this embodiment, the sedimentation pattern and classification results of the high-risk area are obtained from a pre-established monitoring database, and a preliminary comparison is made between the classification results and the terrain data collected in real time. The change range of the real-time elevation is extracted through a data acquisition tool to obtain an initial matching data set corresponding to the boundary of the area. Based on the initial matching data set, a data association tool is used to perform a deep comparison between the sedimentation pattern and the real-time elevation. If the elevation change exceeds a preset threshold, it is marked as an abnormal area, and the potential impact range of the capacity change is determined. Through the correspondence between the abnormal area and the change state, a data fusion tool is used to integrate the dynamic monitoring data stream and the classification results, and the key indicators related to the capacity change in the updated data are obtained to determine the current state fluctuation of the reservoir. Based on the state fluctuation, a data storage tool is used to synchronize the updated data with the historical records in the monitoring database, and a real-time report of dynamic monitoring is generated for the synchronized data to obtain the final monitoring results related to the sedimentation changes in the high-risk area.

[0074] When obtaining sedimentation patterns and classification results for high-risk areas from a pre-established monitoring database, one can first extract sedimentation pattern data for a specific area upstream of a reservoir. Assuming this area is classified as a high-risk sedimentation zone, historical data shows that its sedimentation rate is increasing by approximately 0.5 meters per year. For real-time terrain data, the current elevation value can be obtained using high-precision surveying equipment. Assuming the latest measurement is 125.3 meters, compared to the most recent record of 126.1 meters in the database, the change is 0.8 meters. After a preliminary comparison, data collection tools are used to match the real-time elevation changes with the regional boundaries to form an initial dataset, laying the foundation for subsequent analysis.

[0075] When conducting a deep comparison of the initial matching dataset, we hypothetically used a data association tool (the spatial association module of a geographic information system (GIS) platform) to correlate real-time elevation changes with historical sedimentation patterns. If the preset elevation change threshold is 0.6 meters and the actual change is 0.8 meters, the area would be marked as an anomaly. Further analysis of the potential impact, combined with changes in reservoir capacity, suggests that this area could result in a decrease in upstream water storage capacity of approximately 2%, necessitating significant attention.

[0076] When integrating dynamic monitoring data streams with classification results, data fusion tools can be used to combine real-time monitoring indicators such as flow and sediment content with historical classification results to extract key indicators such as changes in sedimentation rates. Assuming the latest data stream shows a 15% increase in sediment content compared to the previous month, combined with the classification results, it can be determined that the current reservoir status is fluctuating significantly and may have entered a high-risk phase. This approach helps to keep abreast of the dynamic changes in the reservoir. In this embodiment, the GIS tool itself has powerful spatial data fusion capabilities, which can associate and integrate spatial data from different sources (such as elevation maps, regional boundaries, and monitoring point distribution) with non-spatial data (such as sediment content and flow time series data).

[0077] To synchronize data and historical records to address fluctuations, a data storage tool can be used to upload the latest monitoring data to the database. For example, suppose synchronization reveals a continuous decrease in elevation over the past three months, with an average monthly drop of 0.3 meters. Based on this, a real-time report is generated, including siltation trends in high-risk areas. For example, if siltation in a particular area reaches 1.2 meters, it indicates that prevention and control measures should be prioritized. This synchronization and reporting method ensures data consistency and traceability.

[0078] Another way to analyze the relationship between abnormal areas and changing conditions is to focus on the relationship between topographic features and sedimentation patterns. For example, suppose a certain abnormal area has a shallow slope of only 5 degrees, making it prone to sedimentation. Furthermore, real-time data indicates recent heavy rainfall and increased sediment input. Comprehensively assessing this area's capacity changes could further exacerbate. This multi-dimensional analysis can provide a more comprehensive reference for reservoir management.

[0079] Figure 2 The structural block diagram of the reservoir capacity dynamic monitoring system based on drone provided by the present invention is shown in FIG. Figure 2 , the reservoir capacity dynamic monitoring system based on drones includes: The data acquisition module 210 is used to obtain a reservoir surface characteristic dataset using a multi-source remote sensing device carried by an unmanned aerial vehicle, wherein the reservoir surface characteristic dataset includes reservoir water surface image data; A terrain modeling module 220 is configured to determine water boundary information based on the reservoir surface characteristic dataset, perform a depth scan of the reservoir bottom based on the water boundary information, obtain underwater terrain elevation data and sediment distribution data, and construct a three-dimensional terrain model of the reservoir; A dynamic analysis module 230 is used to determine reservoir topography change data and sediment change data based on the dynamic change data of the three-dimensional topography model of the reservoir in a continuous time series; Sedimentation rate analysis module 240, used to analyze the change rate of the reservoir topography change data and sediment change data to obtain the spatial distribution characteristics of the reservoir sediment change; Sedimentation transport law analysis module 250, for determining sediment transport according to the spatial distribution characteristics and reservoir water velocity, and determining the spatial law of sediment deposition according to the sediment transport; Correlation analysis module 260, for determining key driving factors affecting sediment deposition based on the spatial pattern of sediment deposition; the key driving factors include rainfall and water flow velocity; The dynamic monitoring module 270 is used to determine high-risk areas for sedimentation based on the key driving factors, and monitor changes in reservoir capacity based on changes in sediment in the high-risk areas for sedimentation.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for dynamic monitoring of reservoir capacity based on drones, characterized in that: include: S100, obtaining a reservoir surface feature dataset by using a drone, wherein the reservoir surface feature dataset includes reservoir water surface image data; S200, determining water boundary information based on the reservoir surface characteristic dataset, performing a depth scan of the reservoir bottom based on the water boundary information, obtaining underwater terrain elevation data and sediment distribution data, and constructing a three-dimensional terrain model of the reservoir; S300, determining reservoir topography change data and sediment change data based on the dynamic change data of the three-dimensional topography model of the reservoir in a continuous time series; S400, performing a change rate analysis on the reservoir topography change data and the sediment change data to obtain a spatial distribution characteristic of the reservoir sediment change; S500, determining a sediment transport amount based on the spatial distribution characteristics and the reservoir flow velocity, and determining a spatial pattern of sediment deposition based on the sediment transport amount; S600, determine the key driving factors affecting sediment deposition based on the spatial pattern of sediment deposition; Said driving factors include rainfall and water velocity; S700: Determine a high-risk area for sedimentation based on the key driving factors, and monitor changes in reservoir capacity based on changes in sediment in the high-risk area.

2. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1 is characterized in that: The S700 specifically includes: S710: For the reservoir terrain area corresponding to the key driving factors, using a drone equipped with a laser radar to collect key data, obtain local terrain elevation change information, and determine high-risk sedimentation areas based on the local terrain elevation change information; S720: Processing the time series elevation data of the high-risk area to extract long-term trend characteristics, and using an outlier detection method to locate short-term mutation characteristics that deviate from the long-term change characteristics, and obtaining a classification result of the sedimentation change pattern of the high-risk area based on the long-term trend characteristics and the short-term mutation characteristics; S730, based on the classification results of the siltation change pattern, a dynamic monitoring database is constructed, the siltation change pattern is correlated and matched with the underwater terrain elevation data collected in real time, the real-time status of the reservoir capacity change caused by the reservoir siltation is determined, and dynamic monitoring update data is obtained.

3. The method for dynamic monitoring of reservoir capacity based on drone according to claim 2 is characterized in that: The S730 specifically includes: Preliminarily comparing the sedimentation change pattern of the high-risk area with the collected real-time underwater terrain elevation data, extracting the change range of the real-time underwater terrain elevation data, and obtaining an initial matching data set corresponding to the boundary of the high-risk area; Based on the initial matching data set, a depth comparison is performed between the sedimentation change pattern and the real-time underwater terrain elevation data. If the change in the real-time underwater terrain elevation data exceeds a preset change threshold, the corresponding high-risk area is marked as an abnormal area, and the potential impact range of the reservoir capacity change caused by the abnormal area is determined; By integrating the correspondence between the abnormal areas and the change states, the dynamic monitoring data stream and the classification results of the sedimentation change pattern in the high-risk areas are integrated, and key indicators related to capacity changes are obtained from the integrated updated data to determine the current state fluctuation of the reservoir; wherein the change state includes the elevation change value, sedimentation rate and the matching degree with the sedimentation change pattern; the key indicators include sedimentation thickness, sedimentation volume change and capacity reduction ratio; According to the state fluctuation, the updated data is synchronized with the historical records in the monitoring database, and a real-time report of dynamic monitoring is generated for the synchronized data to obtain the final monitoring results related to the sedimentation changes in the high-risk areas.

4. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1 is characterized in that: The S100 specifically includes: S110 uses the multi-source remote sensing imaging equipment carried by the drone to conduct a full-coverage scan of the reservoir surface and surrounding terrain within a preset range, obtaining reservoir surface image data and surrounding terrain elevation data; S120, performing denoising and correction on the reservoir surface image data and the surrounding terrain elevation data to obtain a first reservoir surface image dataset and a first terrain elevation dataset; performing spatial registration and data fusion on the first reservoir surface image dataset and the first terrain elevation dataset, extracting terrain features from the fused data, and determining preliminary characteristic distribution information of the reservoir surface; S130, performing data filling on the preliminary characteristic distribution information to obtain a complete reservoir surface characteristic data set, performing three-dimensional modeling based on the reservoir surface characteristic data set to obtain a three-dimensional distribution model of the reservoir surface, and determining the final distribution characteristic information of the reservoir surface according to the three-dimensional distribution model.

5. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1 is characterized in that: The S200 specifically includes: S210, determining water boundary information based on the reservoir surface feature dataset, performing a depth scan of the reservoir bottom using underwater sonar detection equipment, and obtaining original elevation point cloud data and bottom reflectance feature data of the underwater terrain; S220, obtaining an echo intensity value in the bottom sediment reflection characteristic data; if the echo intensity value exceeds a preset echo intensity threshold, determining that the corresponding bottom material is bedrock terrain; otherwise, determining that the corresponding bottom material is a sediment distribution area, and obtaining a bottom material classification result and sediment layer thickness distribution data; S230, generating a continuous reservoir bottom terrain elevation network using a spatial interpolation method based on the bottom material classification result, and spatially registering the reservoir bottom terrain elevation grid with the reservoir water surface image data to obtain an integrated reservoir elevation dataset; S240, constructing a triangulated network for the integrated elevation dataset of the reservoir, and layer-labeling the bottom network according to the silt layer thickness distribution data, to obtain a three-dimensional terrain model of the reservoir including the surface morphology of the reservoir and the undulations of the underwater terrain.

6. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1, characterized in that: The S300 specifically includes: S310, acquiring depth scan data of the reservoir bottom from preset time nodes through a multi-period data acquisition method, forming a set of underwater terrain elevation data including multiple time nodes, and obtaining a preliminary time series data set; S320, performing layer-by-layer difference calculation on the underwater terrain elevation data of every two adjacent time nodes in the time series data set to determine an elevation difference distribution map; S330, if it is determined based on the elevation difference distribution map that the elevation change value of a certain area exceeds a preset elevation change threshold, then calculating the underwater terrain elevation change value of the corresponding area, and calculating the sediment volume change value based on the underwater terrain elevation change value; S340, integrating the elevation difference distribution map and the sediment volume change value, obtaining the dynamic change trends of the reservoir topography and sediment volume in different time periods, and obtaining reservoir topography change data and sediment change data.

7. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1, characterized in that: The S400 specifically includes: S410, acquiring reservoir topography change data for multiple periods from a preset time series, performing monthly analysis on the reservoir topography change data, calculating the monthly change rate, and obtaining a reservoir topography monthly change rate distribution dataset; S420, dividing the reservoir topography monthly change rate distribution dataset into spatial regions, and if it is determined that the monthly change rate of a certain region after the spatial region division exceeds a preset rate threshold, marking the corresponding region as a significant region, and determining the spatial distribution range of the significant region; S430, extracting heterogeneous features from the spatial distribution range of the significant area to obtain heterogeneous distribution data of reservoir topography changes and sediment changes, and generating a heterogeneous feature distribution map based on the heterogeneous distribution data; S440: Based on the heterogeneous characteristic distribution map, the monthly change rate is integrated with the spatial distribution range of the significant area to generate a comprehensive characteristic distribution map, and the spatial distribution characteristics of the reservoir sediment change are determined based on the comprehensive characteristic distribution map.

8. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1, characterized in that: The S500 specifically includes: S510, determining a temporal distribution characteristic of inflow fluctuations based on historical rainfall data of a preset surrounding area of ​​the reservoir, and determining a dynamic change dataset of the inflow based on the temporal distribution characteristic of the inflow fluctuations; S520, performing a water flow path simulation based on the dynamically changing dataset of the inflow flow, simulating the water flow velocity distribution characteristic values ​​of different areas in the reservoir, and obtaining a water flow velocity distribution characteristic map of each area; S530, screening out areas where the water velocity distribution characteristic value exceeds a preset water velocity threshold, estimating sediment transport using a sedimentation calculation tool based on key physical parameters of reservoir sediment and the water velocity distribution characteristic map, and obtaining sediment transport distribution data corresponding to each area; the key physical parameters include average particle size, density, and settling velocity; S540 , generating a sediment thickness distribution map using a spatial overlay tool based on the sediment transport distribution data and reservoir area division information to obtain a spatial pattern of sediment deposition.

9. The method for dynamic monitoring of reservoir capacity based on drone according to claim 1, characterized in that: The S600 specifically includes: S610, obtaining terrain slope distribution data of different reservoirs, and classifying the slope values ​​of different reservoir areas according to the terrain slope distribution data to obtain slope grade distribution information of each reservoir area; S620, analyzing the correspondence between slope grade and sedimentation location using a spatial superposition method based on the slope grade distribution information and the historical records of sedimentation distribution, and determining the distribution characteristics of the influence of the slope grade on the sedimentation location; S630, obtaining rainfall data and water flow velocity data, using the rainfall data and water flow velocity data as driving factors, and using a correlation test tool to calculate the correlation coefficient between the driving factors and the sediment distribution. If the correlation coefficient of a certain driving factor exceeds a preset threshold, the corresponding driving factor is marked as a key driving factor.

10. A dynamic reservoir capacity monitoring system based on drones, characterized in that: include: A data acquisition module is used to obtain a reservoir surface feature dataset using a multi-source remote sensing device carried by an unmanned aerial vehicle, wherein the reservoir surface feature dataset includes reservoir water surface image data; A terrain modeling module is used to determine water boundary information based on the reservoir surface feature dataset, perform a depth scan of the reservoir bottom based on the water boundary information, obtain underwater terrain elevation data and sediment distribution data, and construct a three-dimensional terrain model of the reservoir; A dynamic analysis module is used to determine reservoir topography change data and sediment change data based on the dynamic change data of the three-dimensional topography model of the reservoir in a continuous time series; A sedimentation rate analysis module is used to analyze the change rate of the reservoir topography change data and the sediment change data to obtain the spatial distribution characteristics of the reservoir sediment change; A sediment transport law analysis module is used to determine the sediment transport amount based on the spatial distribution characteristics and the reservoir water flow velocity, and to determine the spatial law of sediment deposition based on the sediment transport amount; A correlation analysis module is used to determine key driving factors affecting sediment deposition based on the spatial pattern of sediment deposition; the key driving factors include rainfall and water flow velocity; The dynamic monitoring module is used to determine the high-risk area for sedimentation based on the key driving factors, and monitor the change of reservoir capacity based on the change of sediment in the high-risk area for sedimentation.

Images