A reservoir capacity dynamic monitoring method and system based on a unmanned aerial vehicle

Abstract

The embodiment of the application provides a reservoir capacity dynamic monitoring method and system based on a UAV, which comprises the following steps: obtaining a reservoir surface layer feature data set through a UAV, determining a water area boundary and scanning a reservoir bottom according to the data set, obtaining underwater topographic elevation and silt distribution data, and constructing a three-dimensional topographic model; determining topographic and silt change data based on dynamic change data of the three-dimensional topographic model of a continuous time sequence. Through change rate analysis, silt space distribution characteristics are obtained, and the amount of silt transport and the silt deposition space rule are determined in combination with water flow velocity. Key driving factors are screened through correlation analysis, a high-risk silt area is determined, silt change in the area is monitored, and thus the reservoir capacity change is mastered. The application can accurately and dynamically monitor reservoir siltation, clearly determine a high-risk area and a change mode, construct a dynamic database, update the capacity change in real time, and provide important technical support for reservoir management and flood control and disaster reduction.

Description

TECHNICAL FIELD

[0001] The embodiment of the application relates to the technical field of geographic information monitoring, and particularly relates to a reservoir capacity dynamic monitoring method and system based on a unmanned aerial vehicle. BACKGROUND

[0002] As a core water conservancy project for water resource regulation and flood control and disaster reduction, the dynamic change of the reservoir capacity is directly related to the regional water supply safety, agricultural irrigation efficiency and flood control system stability. With the intensification of climate change and the increase of basin development intensity, the problem of reservoir siltation is becoming increasingly prominent. The sediment deposition leads to the reduction of reservoir capacity and the decline of regulation and storage capacity, which not only affects the service life of the project, but also poses a potential threat to the downstream flood control safety. Therefore, accurately and efficiently monitoring the dynamic change of the reservoir topography and siltation material has become a key requirement in the field of water conservancy management.

[0003] The current reservoir capacity monitoring means has significant technical limitations: firstly, the traditional manual measurement relies on equipment such as total station and depth finder, and a large number of monitoring points need to be set in the reservoir area. This not only consumes time and effort and has limited coverage, but also is limited by hydrological conditions (such as high water level during flood season and turbulent water flow), making it difficult to achieve high-frequency and full-water-area monitoring, resulting in insufficient data integrity and inability to reflect the spatial distribution heterogeneity of siltation material; secondly, although fixed-point sensors can achieve continuous data collection, the monitoring range is restricted by the density of hardware arrangement, and for reservoir areas with complex terrain and uneven siltation distribution, monitoring blind spots are easily generated. Moreover, the sensors are immersed in water for a long time, which has high maintenance cost and the data accuracy is easily affected by water quality; thirdly, the existing technology mainly stays at the static description of siltation change at a single time point, lacks continuous comparison of multi-period data, and is difficult to quantify the siltation rate and long-term trend, and cannot reveal the spatial law of siltation combined with driving factors such as rainfall and water flow; fourthly, the traditional method does not establish a siltation risk model, which cannot accurately identify high-risk areas, resulting in insufficient pertinence of silt removal engineering and serious waste of resources.

[0004] In addition, the existing three-dimensional terrain modeling technology mainly relies on a single data source (such as only waterborne remote sensing or underwater sonar), and the fusion degree of waterborne and underwater data is low, which is difficult to build a complete three-dimensional model of the reservoir area, resulting in large calculation error of siltation volume; at the same time, the analysis of siltation driving mechanism lacks systematicness, and historical hydrological data and spatial distribution characteristics are not integrated, which cannot accurately judge the sediment transport path and deposition law, restricting the scientificity of the dynamic prediction of reservoir capacity.

[0005] Therefore, how to break through the technical bottleneck of the traditional monitoring method, realize the high-precision, full-cycle and intelligent monitoring of the reservoir topography and siltation material through multi-source data fusion, multi-period dynamic analysis and driving factor correlation modeling, has become an urgent requirement to improve the management level of reservoirs and ensure the safety of the project. SUMMARY

[0006] The embodiment of the present application provides a reservoir capacity dynamic monitoring method and system based on a UAV, to solve the problem that the analysis of the silt driving mechanism lacks systematicness, historical hydrological data and spatial distribution characteristics are not integrated, the silt transport path and deposition rule cannot be accurately judged, and the scientificity of the reservoir capacity dynamic prediction is restricted.

[0007] In a first aspect, the embodiment of the present application provides a reservoir capacity dynamic monitoring method based on a UAV, comprising:

[0008] S100, obtaining a reservoir surface feature data set by a UAV, wherein the reservoir surface feature data set comprises reservoir water surface image data;

[0009] S200, determining water area boundary information according to the reservoir surface feature data set, performing depth scanning on the reservoir bottom according to the water area boundary information, obtaining underwater topographic elevation data and silt distribution data, and constructing a reservoir three-dimensional topographic model;

[0010] S300, determining reservoir topographic change data and silt change data according to dynamic change data of the reservoir three-dimensional topographic model in a continuous time sequence;

[0011] S400, performing change rate analysis on the reservoir topographic change data and the silt change data to obtain spatial distribution characteristics of the reservoir silt change;

[0012] S500, determining a silt transport amount according to the spatial distribution characteristics and a reservoir flow velocity, and determining a spatial rule of silt deposition according to the silt transport amount;

[0013] S600, determining key driving factors affecting silt deposition according to the spatial rule of silt deposition, wherein the driving factors include rainfall and flow velocity;

[0014] S700, determining a high siltation risk area according to the key driving factors, and monitoring reservoir capacity change according to silt change in the high siltation risk area.

[0015] In a second aspect, the embodiment of the present application provides a reservoir capacity dynamic monitoring system based on a UAV, comprising:

[0016] A data acquisition module is configured to obtain a reservoir surface feature data set by a UAV carrying a multi-source remote sensing device, wherein the reservoir surface feature data set comprises reservoir water surface image data;

[0017] A topographic modeling module is configured to determine water area boundary information according to the reservoir surface feature data set, perform depth scanning on the reservoir bottom according to the water area boundary information, obtain underwater topographic elevation data and silt distribution data, and construct a reservoir three-dimensional topographic model;

[0018] a dynamic analysis module configured to determine reservoir topography change data and sediment change data according to dynamic change data of the three-dimensional topography model of the reservoir on a continuous time sequence;

[0019] a sedimentation velocity analysis module configured to analyze the change rate of the reservoir topography change data and the sediment change data to obtain spatial distribution characteristics of reservoir sediment change;

[0020] a sedimentation and migration law analysis module configured to determine a sediment migration amount according to the spatial distribution characteristics and water flow velocity of the reservoir, and determine a spatial law of sediment deposition according to the sediment migration amount;

[0021] a correlation analysis module configured to determine key driving factors affecting sediment deposition according to the spatial law of sediment deposition; the key driving factors include rainfall and water flow velocity;

[0022] a dynamic monitoring module configured to determine a high-risk sedimentation area according to the key driving factors, and monitor reservoir capacity change according to sediment change in the high-risk sedimentation area.

[0023] Compared with the prior art, the reservoir capacity dynamic monitoring method and system based on a UAV provided by the embodiments of the present application have the following beneficial effects:

[0024] (1) The present application can realize accurate dynamic monitoring of reservoir sedimentation, and by combining multi-source remote sensing of a UAV with underwater sonar detection, fusing water surface images and underwater elevation data to construct a three-dimensional topography model, and combining multi-time period data acquisition and difference analysis, the dynamic change of reservoir topography and sedimentation can be accurately captured, overcoming the problems of incomplete data coverage and limited accuracy of traditional methods, and improving the comprehensiveness and accuracy of monitoring.

[0025] (2) The present application can determine the monthly rate of sedimentation change, significant areas and spatial distribution characteristics through change rate analysis and spatial distribution heterogeneity analysis; by combining water flow velocity simulation and sediment migration amount estimation, the spatial law of sediment deposition can be determined, and by driving factor correlation test, the influence weight of each factor on sedimentation can be determined, making the analysis more scientific and deep, and improving the depth and reliability of sedimentation change analysis.

[0026] (3) The present application can accurately identify the distribution range and change mode of the high-risk sedimentation area by using a UAV carrying a high-precision laser radar to collect key data in the area corresponding to the driving factor with a weight exceeding a threshold, and combining time trend fitting and abnormal point detection technology, thereby providing accurate targets for risk prevention and control, and accurately identifying the high-risk sedimentation area.

[0027] (4) The application constructs a dynamic monitoring database, associates the classification result with the real-time collected data, can judge the change state of the reservoir capacity caused by siltation in real time, generates dynamic monitoring update data and real-time reports, provides timely and reliable basis for reservoir desilting planning, water resource optimal allocation, flood control decision, etc., helps to guarantee the safety of reservoir engineering 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 DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0029] Figure 1 The flow chart of the reservoir capacity dynamic monitoring method based on the unmanned aerial vehicle provided by an embodiment of the present application is shown in the figure.

[0030] Figure 2 The structural block diagram of the reservoir capacity dynamic monitoring system based on the unmanned aerial vehicle provided by the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0032] Figure 1 The flow chart of the reservoir capacity dynamic monitoring method based on the unmanned aerial vehicle according to the embodiment of the present application is shown in the figure. Figure 1 The method comprises the following steps.

[0033] S100, obtaining a reservoir surface feature data set by an unmanned aerial vehicle, wherein the reservoir surface feature data set comprises reservoir water surface image data; specifically comprising:

[0034] Specifically, in this embodiment, a multi-source remote sensing image device is carried by the unmanned aerial vehicle to comprehensively scan the surface of the reservoir and the terrain within a certain range (such as 1 kilometer) around the reservoir. The multi-source remote sensing image device can include a visible light camera, an infrared camera, and a laser radar, etc. The visible light camera obtains water surface image data of the reservoir, which can clearly present information such as the state of the water surface and the boundary. The laser radar obtains surrounding terrain elevation data, which accurately measures the terrain undulation. When scanning, the unmanned aerial vehicle flies according to a preset flight route to ensure that there is no missing area.

[0035] As a preferred embodiment, the S100 specifically includes:

[0036] S110, by a multi-source remote sensing image device carried by the unmanned aerial vehicle, the surface of the reservoir and the surrounding terrain within a preset range are comprehensively scanned to obtain water surface image data of the reservoir and the surrounding terrain elevation data. The surrounding terrain within the preset range can be selected as a land area of 1-3 kilometers outside the water boundary of the reservoir. The specific range is adjusted according to the size of the reservoir to ensure that the catchment area and the terrain transition zone that may affect the reservoir are covered. The unmanned aerial vehicle flies according to a preset flight route (such as a grid-shaped or a snake-shaped route). The flight speed is controlled at about 50 kilometers / hour. The reservoir area of about 10 square kilometers is covered. The image resolution reaches 0.1 meters per pixel. The elevation data accuracy is controlled within 0.05 meters to ensure that there is no blind area in scanning, and there is an overlap rate of 10%-20% between adjacent scanning bands to ensure the data splicing accuracy.

[0037] S120, the water surface image data of the reservoir and the surrounding terrain elevation data are denoised and corrected to obtain a first image data set of the surface of the reservoir and a first data set of the terrain elevation. The first image data set of the surface of the reservoir and the first data set of the terrain elevation are spatially registered and data fused. The terrain features of the fused data are extracted to determine the preliminary feature distribution information of the surface layer of the reservoir.

[0038] In image processing, for the denoising and correction of image data, a deep learning-based image enhancement algorithm is used to remove noise. At the same time, a geometric correction method is used to eliminate image distortion caused by changes in the flight attitude of the unmanned aerial vehicle. For example, if the original image has a shadow area caused by uneven lighting, the influence of the shadow will be significantly reduced after processing. The clarity of the first image data set of the surface of the reservoir is improved by about 30%, which provides more reliable data support for subsequent feature extraction.

[0039] The spatial registration is realized by a geographic information system (GIS) tool (such as ArcGIS), and the coordinate systems of the water surface image and the terrain elevation data (such as the WGS84 coordinate system) are unified, with a registration error controlled within 0.2 meters; the data fusion is to integrate the two types of data into a unified data set, so that the water surface features are associated with the surrounding terrain (for example, the slope change of the reservoir bank and the connection relationship with the water surface can be clearly presented). Key features are extracted from the fused data by using automatic algorithms (such as edge detection and threshold segmentation), including abnormal terrains such as cracks, settlement zones and abrupt changes of the reservoir bank, and the boundary between the water surface and the land, to form preliminary feature distribution information.

[0040] In S130, data filling is performed on the preliminary feature distribution information to obtain a complete reservoir surface layer feature data set, and based on the reservoir surface layer feature data set, three-dimensional modeling is performed to obtain a three-dimensional distribution model of the reservoir surface, and the final distribution feature information of the reservoir surface is determined according to the three-dimensional distribution model.

[0041] For missing areas (such as blind areas with an area less than 0.5 square kilometers) in the preliminary feature distribution information caused by cloud cover, equipment failure, etc., a neighborhood-based interpolation algorithm (such as inverse distance weighted method) is used to estimate the feature values of the missing areas using the surrounding effective data, so that the data set integrity is improved to more than 98%. The filled data includes non-missing water surface image, surrounding terrain elevation and extracted terrain features, which can fully reflect the spatial distribution of the reservoir surface layer. Based on the complete data set, a three-dimensional reconstruction tool (such as Context Capture) is used to construct a three-dimensional distribution model of the reservoir surface, and the model can intuitively display the three-dimensional features such as water surface elevation, bank slope and water depth anomaly area. The core conclusions extracted from the three-dimensional model include the accurate boundary of the reservoir water surface, the slope classification of the bank terrain (such as 0-5° for gentle area and 5-15° for gentle slope area), the location and range of potential risk points (such as crack area and settlement area), which provide a basis for subsequent water boundary determination and underwater scanning.

[0042] In the terrain feature extraction process, automatic algorithms are used to identify abnormal features such as cracks and settlement of the reservoir surface layer. If the extraction result shows that there is a crack with a width of about 0.3 meters in a certain area, the area will be marked as a potential risk point in the preliminary feature distribution information, which helps to quickly locate the problem area and provides a basis for reservoir safety evaluation.

[0043] If there is data missing 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 for filling. For example, the missing area is about 0.5 square kilometers, and after interpolation, the integrity of the second data set of the surface layer features is improved to more than 98%, avoiding the impact of data missing on subsequent modeling.

[0044] In the three-dimensional reconstruction stage, based on the second data set of surface features, a three-dimensional distribution model of the reservoir surface is constructed using professional software. If the model shows abnormal changes in water depth in a certain area, it may be related to the accumulation of bottom sediments, and the area will be marked as a key monitoring point in the final distribution feature information. This three-dimensional model intuitively displays the spatial features of the reservoir surface, providing a scientific basis for management decisions.

[0045] Through the above technical means of the embodiment, the final obtained reservoir surface distribution feature information has high accuracy and comprehensive coverage, which can significantly improve the efficiency and reliability of reservoir safety monitoring. For example, in a certain application, the potential leakage risk area was found in advance based on this method, avoiding possible economic losses, fully reflecting the practical value of the technology in reservoir management.

[0046] S200, determining water area boundary information according to the reservoir surface feature data set, performing depth scanning on the reservoir bottom according to the water area boundary information, obtaining underwater topographic elevation data and sediment distribution data, and constructing a reservoir three-dimensional terrain model;

[0047] The S200 specifically includes:

[0048] S210, determining water area boundary information according to the reservoir surface feature data set, and performing depth scanning on the reservoir bottom through an underwater sonar detection device to obtain original elevation point cloud data and bottom material reflection feature data of the underwater topography;

[0049] Specifically, based on the reservoir water surface image data and surrounding terrain elevation data in the reservoir surface feature data set, the intersection line of the water surface and the land area is extracted by an edge detection algorithm. Its accuracy can be controlled within 0.2 meters, which is used to clearly define the range boundary of the reservoir underwater scanning, ensuring that the scanning area does not deviate from the target water area. This boundary information can effectively guide the scanning path planning of the multi-beam sonar device, ensuring that the scanning area does not deviate from the target water area.

[0050] For the depth scanning of the multi-beam sonar device on the reservoir bottom, the multi-beam sonar device transmits multi-angle acoustic wave signals in a fan-shaped distribution to the underwater through its transducer array. These signals are reflected back to the device after contacting the reservoir bottom. The device can simultaneously obtain original elevation point cloud data and bottom material reflection feature data of the underwater topography by accurately calculating the sound propagation time and angle, and combining with the real-time measured water sound speed data. The original elevation point cloud data is composed of a large number of discrete three-dimensional coordinates, which can accurately reflect the undulating state of the underwater topography. The bottom material reflection feature data records the intensity of the reflected sound waves in different areas, providing a basis for subsequent judgment of the bottom material properties.

[0051] S220, obtaining the echo intensity value in the bottom material reflection characteristic data, if the echo intensity value exceeds the preset echo intensity threshold, judging the corresponding bottom material as bedrock terrain, otherwise judging the corresponding bottom material as silt distribution area, obtaining the bottom material classification result and silt layer thickness distribution data;

[0052] Suppose in a certain scan, the sonar equipment scans the bottom of the reservoir with a data acquisition frequency of 10 times per second, according to the preset parallel or radial route, covering an area of about 5 square kilometers, the depth resolution of the obtained point cloud data reaches 0.1 meters, this high frequency acquisition method can capture the subtle changes of underwater terrain such as thin layer of silt accumulation, small concave, etc., laying a solid data foundation for subsequent analysis. When judging the type of bottom material through bottom material reflection characteristic data, classification can be carried out according to the difference of echo intensity, that is, a preset echo intensity threshold is set, if the echo intensity value of a certain area exceeds the threshold, it is judged that the corresponding bottom material is bedrock terrain, if it is lower than the threshold, it is determined as silt distribution area, thus obtaining the bottom material classification result and silt layer thickness distribution data, suppose in the scanning of a certain reservoir, it is found that about 30% of the bottom area has low echo intensity, which is preliminarily judged as silt coverage, and further analysis shows that the average thickness of silt in these areas is about 0.5 meters, this classification method can quickly identify the material distribution of the reservoir bottom.

[0053] S230, according to the classification result of the bottom material, generating a continuous reservoir bottom terrain elevation network by spatial interpolation method, spatially registering the reservoir bottom terrain elevation grid with the reservoir water surface image data to obtain a reservoir integrated elevation data set;

[0054] For generating a continuous reservoir bottom terrain elevation grid by spatial interpolation method, Kriging interpolation method can be used to process point cloud data, which can analyze the spatial correlation between sample points, interpolate the discrete point cloud data, and generate a continuous and uniform elevation grid. Suppose in a certain processing, the resolution of the interpolated grid reaches 1 meter, which can better reflect the subtle features of underwater terrain, effectively filling the possible blank areas in data acquisition.

[0055] When spatially registering the underwater elevation grid with the water surface image data, professional coordinate transformation tools can be used to unify them to the same geographic coordinate system for accurate alignment. Suppose the spatial error after registration is controlled within 0.3 meters, so that the water surface image data and the underwater elevation data can be seamlessly connected to form a complete reservoir integrated elevation data set, providing a unified and accurate spatial reference for subsequent comprehensive modeling.

[0056] S240, triangulation network construction is performed on the integrated elevation data set of the reservoir, and the bottom network is marked in layers according to the thickness distribution data of the silt layer, so as to obtain a three-dimensional terrain model of the reservoir containing the surface morphology of the reservoir and the undulating underwater topography.

[0057] The process of constructing a triangulation network for a three-dimensional grid generation tool can process the integrated elevation data set through professional geographic information or modeling software, connect discrete elevation points into a continuous triangulation network structure, and assume that the constructed triangulation network contains about 100,000 grid cells, which can finely display the morphology of the reservoir surface and the undulating characteristics of the underwater topography, and provide intuitive and detailed data support for terrain analysis. When marking the bottom grid in layers according to the thickness distribution data of the silt layer, different colors or labels can be used to distinguish the bottom grid according to different ranges of silt thickness, for example, the area with a thickness of 0-0.5 meters is marked as a general attention area, the area with a thickness of 0.5-1 meter is marked as a secondary attention area, and the area with a thickness of more than 1 meter is marked as a key attention area. Assuming that the thickness of the silt layer in a certain area of the bottom of a reservoir reaches 1.2 meters, it is marked as a key attention area. This layered marking method can quickly identify potential problem areas for management personnel, and provide clear reference for daily management and desilting planning of the reservoir.

[0058] S300, according to the dynamic change data of the three-dimensional terrain model of the reservoir in the continuous time sequence, determining the reservoir terrain change data and the silt change data;

[0059] For the comprehensive three-dimensional terrain model, the terrain scanning data at different time points is obtained by multi-period data acquisition method to form a continuous time sequence data set, and the elevation difference of the terrain elevation data of each scanning is compared to extract the elevation change value and the silt volume change calculation result, so as to obtain the preliminary quantitative result of the dynamic change of the reservoir terrain and the silt. Specifically, it includes:

[0060] S310, through multi-period data acquisition method, the depth scanning data of the bottom of the reservoir is obtained from the preset time node to form an underwater topographic elevation data set including multiple time nodes, and a preliminary time sequence data set is obtained;

[0061] Specifically, according to the reservoir monitoring requirements, the topographic scanning mode is repeated at different time points, the purpose is to capture the dynamic changes by comparing the data at different times, and at the same time, the reservoir hydrological characteristics are set, such as before and after the flood season (April and October) every year, at this time, the reservoir water level and the siltation state change significantly, which can more accurately reflect the topographic evolution; The depth scanning data of the reservoir bottom is the original underwater topographic data obtained by devices such as multi-beam sonar, including elevation point cloud and geological information; The underwater topographic elevation data set is a data set formed by collecting scanning data at multiple time nodes, the data of each node is stored in the form of digital elevation model, and the data at each time point covers the whole reservoir to ensure data integrity; The preliminary time series data set is an elevation data combination arranged in time sequence, which provides a basis for subsequent difference analysis.

[0062] S320, the underwater topographic elevation data of each adjacent two time nodes in the time series data set is calculated layer by layer, and the elevation difference distribution map is determined;

[0063] After arranging the reservoir bottom topographic elevation data collected at multiple time periods in time sequence, the data is processed by a special elevation data comparison tool (such as a topographic analysis module developed based on a GIS platform), which can perform layer-by-layer and region-by-region difference calculation on the elevation data at different time points according to the preset elevation level or geographical division; Assuming that in a certain actual analysis, the elevation data of two time nodes, April (before the flood season) and October (after the flood season), are selected for comparison, after the tool compares the two groups of data point by point, it is found that the elevation of some areas at the bottom of the reservoir increases by 0.5 meters compared with April, combined with the reservoir hydrological characteristics and sediment deposition law, this elevation increase phenomenon indicates that there may be accumulation of silt in this area; This layer-by-layer comparison method can accurately locate each area where the elevation changes, whether it is a large area of siltation zone or a small range of local deposition, it can be clearly identified, and then the tool will generate an elevation difference distribution map in a visual way, in which the numerical range and spatial position of the elevation change are marked by different colors or color blocks, and the spatial distribution characteristics of the topographic change in the whole reservoir are intuitively displayed; Such a method not only can quickly lock the key areas where siltation problems may exist and reduce the workload of manual investigation, but also can provide reliable basis for subsequent silt volume calculation, risk assessment, etc. through accurate spatial positioning and quantitative data, thereby effectively improving the efficiency and accuracy of the whole monitoring and analysis process.

[0064] S330, if the elevation change value of a certain area exceeds the preset elevation change threshold according to the elevation difference distribution map, the underwater topographic elevation change value of the corresponding area is calculated, and the silt volume change value is calculated according to the underwater topographic elevation change value;

[0065] For example, based on the elevation difference distribution map, if data screening finds that the elevation change value of certain areas exceeds the preset judgment threshold (such as 0.3 meters), it indicates that the topographic change in these areas has reached a level that needs to be focused on, and further precise estimation of the deposition volume of these areas needs to be carried out through the volume calculation tool (such as topographic analysis software based on three-dimensional modeling technology). Specifically, the volume calculation tool will first lock the boundary range of the change area according to the elevation difference distribution map, and then combine the elevation data of the area at two time nodes to construct a three-dimensional model. Through the spatial volume integration operation of the elevation change part in the model, the actual volume of the deposition material is obtained. Assuming that a certain area is found to have an elevation change value of 0.4 meters through comparison, and the planar projection area of the area is estimated to be 5000 square meters, after three-dimensional modeling and volume operation by the tool, the estimated deposition volume of the area is about 2000 cubic meters. At the same time, comprehensive analysis is carried out combined with the characteristics of the reservoir terrain, such as through the three-dimensional terrain model constructed in the early stage, it is known that the terrain slope of the area is relatively gentle (such as the slope is between 2-5 degrees), and according to the law of sediment movement, the flow velocity in the area with a gentle slope will be significantly reduced, and the carried sediment is prone to deposition. Therefore, it can be further judged that the deposition in this area is mainly caused by the deposition of sediment brought by upstream rainfall or flood after being transported by water flow; this analysis method not only quantifies the volume change of the deposition material through specific numerical values, but also reveals the causes of the deposition by combining the terrain and hydrological characteristics, providing scientific and specific basis for the water conservancy management department to formulate targeted dredging plan and optimize water resources scheduling scheme.

[0066] S340, integrating the elevation difference distribution map and the deposition volume change value, obtaining the dynamic change trend of the reservoir terrain and the deposition volume at different time periods, and obtaining the reservoir terrain change data and the deposition change data.

[0067] After integrating the elevation difference distribution map and the silt volume change value, specifically, the elevation difference distribution map reflecting the spatial change of the terrain is associated and matched with the volume data quantifying the increase and decrease of the silt, the integrated information is processed by a professional data visualization tool (such as Matplotlib, the dynamic mapping module of ArcGIS, etc.), and a dynamic chart capable of intuitively presenting the time dimension change is generated; assuming that in a certain application, the generated chart takes the time axis (such as month, quarter or year) as the horizontal coordinate, clearly labels each data collection node, takes the left vertical coordinate to represent the elevation change value (unit: meter) and the right vertical coordinate to represent the silt volume change (unit: cubic meter), and uses different colors (such as red to represent the elevation increase area and blue to represent the elevation decrease area) to label the change of different sub-areas of the reservoir, and through the line chart or the bar chart, the elevation change trend and the silt volume increase and decrease amplitude of different areas are corresponded, so that the manager can intuitively observe the dynamic trend of the reservoir terrain in different time periods, such as the elevation of a certain area significantly increases and the silt volume rapidly increases after the flood season, while the change is relatively flat in the dry season; such visualization result converts the complex spatial data and numerical information into intuitive and easy-to-understand chart forms, which facilitates the reservoir manager to quickly grasp the siltation state, change speed and potential risk of the whole region and local area, so as to timely formulate and take targeted desilting plan (such as preferentially arranging desilting operation for the high siltation area) or flood control measures (such as predicting the influence of the reduction of reservoir capacity on flood regulation capacity in advance), effectively shortening the time from data acquisition to decision implementation, and significantly improving the efficiency and scientificity of reservoir management.

[0068] The above steps of the embodiment can effectively monitor the change of the reservoir terrain, and early warn the potential risks such as the reduction of reservoir capacity or flood hazards caused by siltation through the combination of multi-period data collection and difference analysis. The generation of the dynamic change chart converts the complex data into intuitive information, reduces the understanding threshold, and improves the decision support capability. The overall scheme forms a complete closed loop from data collection, difference calculation, volume estimation to visualization presentation, ensuring the accuracy and practicality of the reservoir terrain monitoring

[0069] S400, change rate analysis is performed on the reservoir terrain change data and the silt change data to obtain the spatial distribution characteristics of the reservoir silt change;

[0070] In the embodiment, according to the preliminary quantization result of the dynamic change, the monthly rate of the terrain and silt change is calculated by using the change rate analysis method, and the regions with significant change are identified through spatial distribution heterogeneity analysis to obtain the spatial distribution characteristics of the reservoir silt change. Specifically, it includes:

[0071] S410, obtain multi-period reservoir topographic change data from a preset time sequence, perform monthly analysis on the reservoir topographic change data, calculate a monthly change rate, and obtain a reservoir topographic monthly change rate distribution data set;

[0072] In the business scenario of reservoir siltation dynamic monitoring, the monthly analysis of multi-period topographic change data is a key link to grasp the short-term evolution law of siltation, and the core lies in capturing the subtle trend of topographic change through high-frequency data collection and comparison. Specifically, a monthly topographic elevation data collection plan needs to be formulated for the target reservoir, and the collection time can be fixed in the middle and late of each month (avoiding extreme weather influence), and a combination of laser radar and multi-beam sonar devices carried by a drone is used to synchronously scan the whole reservoir (including water surface and underwater topography), so as to ensure that the resolution of the elevation data collected each time is consistent (such as the underwater topographic elevation accuracy being controlled within 0.05 meters), and the continuous time sequence data set is formed monthly, each data entry containing the digital elevation model and siltation distribution vector diagram of the corresponding month. In the data processing stage, with the aid of time interval comparison tools (such as a time sequence analysis module developed based on Python, integrating a sliding window algorithm and a difference calculation function), the topographic elevation data of adjacent two months are compared point by point: first, the coordinate systems of the two months of data are ensured to be completely consistent through spatial registration, and then the elevation difference of the same position is calculated, combined with the area parameter of the region, to obtain the topographic change rate per unit time (month). For example, in the scanning data of a certain reservoir upstream area from May to June, the elevation increases from 100.2 meters to 100.4 meters, with a difference of 0.2 meters, and the change rate is calculated as 0.2 meters / month; while from June to July, the elevation of the region only increases by 0.1 meters, and the rate decreases to 0.1 meters / month. This calculation of monthly change rate can not only intuitively reflect the speed of topographic change, but also identify abnormal signals in short-term fluctuations (such as a sudden increase in rate in a month, which may be related to the input of sediment caused by heavy rainfall), and the result will serve as the basis data for subsequent spatial distribution analysis and driving factor correlation, providing quantitative basis for accurately judging the siltation active period and early warning of high-risk areas.

[0073] S420, performing spatial region division on the reservoir topographic monthly change rate distribution data set, if it is judged that the monthly change rate of a region after 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;

[0074] The spatial distribution analysis of the monthly change rate distribution dataset is a key step in accurately locating the active siltation area. The specific process needs to be combined with professional tools and reservoir geographical characteristics. First, with the help of spatial distribution analysis tools (such as the spatial analysis module of ArcGIS and the partition statistics function of QGIS), the reservoir is divided into several sub-regions according to the complexity of the reservoir terrain (such as a grid size of 1000-5000 square meters). The division should take into account the water flow path and the terrain unit (such as dividing the water inlet area, bay area, and main river area into independent sub-regions), ensuring that the terrain features of each sub-region are relatively uniform. Then, based on the reservoir siltation risk assessment, set the rate threshold (such as 0.15 meters / month), which needs to refer to historical siltation data and management needs (such as for key flood control areas, the threshold can be appropriately lowered to increase sensitivity). Through the tool, the monthly change rate of each sub-region is compared in batches. If the rate of a certain sub-region exceeds the threshold (such as the rate of a certain sub-region in May and June is 0.2 meters / month), it is marked as a significant area, and the boundary coordinates, area, and other parameters of the area are automatically recorded. Taking a reservoir as an example, it is found that the significant areas are mainly concentrated in the upstream water inlet area of the reservoir, with an area of about 3000 square meters. Combined with the water flow simulation data, it can be known that the sediment is easy to deposit due to the reduction of water flow in this area, and the flood season is in May and June, and the amount of incoming sediment increases, further verifying the high siltation activity of this area. Through this regional division and significance judgment, the monitoring focus can be focused on these key change areas. In the future, data collection frequency can be increased in these areas (such as from once a month to once every half month), and water flow and sediment transport monitoring can be carried out accordingly, thereby greatly improving the targeting and efficiency of monitoring and providing data support for accurately developing a dredging plan.

[0075] S430, extracting heterogeneity features of the spatial distribution range of the significant area to obtain heterogeneity distribution data of reservoir terrain changes and siltation changes, and generating a heterogeneity feature distribution map according to the heterogeneity distribution data;

[0076] In the analysis of the heterogeneity of the significant area, this step aims to further explore the spatial differences of the terrain changes within the same significant area, breaking through the limitations of only focusing on the overall changes of the area, so as to more accurately grasp the details of the silt distribution. Specifically, with the help of heterogeneity feature extraction tools (such as GIS-based spatial autocorrelation analysis module, hotspot analysis tool, etc.), the significant area can be subdivided into smaller grid units (such as 10 meters x 10 meters), and then the elevation change rate, silt thickness and other data of each subdivided unit are statistically analyzed to identify the spatial differentiation pattern of the change of the change area and the change area. Assuming that in the above-mentioned significant area of about 3000 square meters, through tool calculation, it is found that the 500 square meter area near the water inlet is directly impacted by the upstream water carrying sediment, with a monthly average elevation change rate of 0.3 meters / month, and the silt is mainly coarse-grained sediment; while the 2500 square meter area away from the water inlet, affected by the diffusion of water flow, the sediment deposition rate gradually decreases, with a monthly average elevation change rate of 0.05-0.1 meters / month, and the silt is mainly fine-grained silt, which directly indicates that the silt distribution in the significant area is obviously uneven. In generating the heterogeneity feature distribution map, the tool will visually mark the change difference of each subdivided unit with a color gradient (such as deep red to light red to represent the rate from high to low), and superimpose auxiliary information such as water flow direction arrow, from which it can be clearly observed that: affected by the attenuation of water flow power, the elevation change rate presents a gradually decreasing trend from the water inlet to the interior of the area, that is, the upstream area near the water inlet becomes the silt "hot spot area" due to the strong sedimentation effect, and the downstream area changes relatively gently. This fine heterogeneity analysis not only reveals the spatial distribution of silt in the significant area "near source thick, far source thin", but also provides accurate guidance for subsequent targeted measures, for example, the "hot spot area" can be given priority to develop a strengthened dredging plan, combined with water flow regulation to reduce sediment deposition in the area, so as to improve the scientificity and efficiency of reservoir management.

[0077] S440, according to the heterogeneity feature distribution map, the monthly change rate is fused with the spatial distribution range of the significant area, and a comprehensive feature distribution map is generated, and the spatial distribution characteristics of the reservoir silt change are determined according to the comprehensive feature distribution map.

[0078] The rate of change in the time dimension is deeply integrated with the significant area distribution in the spatial dimension to achieve the visual integration of multi-dimensional data. Specifically, with the help of professional data integration tools (such as the layer overlay function of ArcGIS, the spatial analysis module of ENVI, etc.), the monthly change rate data and the significant area distribution data need to be preprocessed: through coordinate system unification (such as unified WGS84 coordinate system) to ensure accurate spatial position matching of the two, and then the monthly change rate data is classified according to the numerical range (such as 0-0.1m / month, 0.1-0.2m / month, >0.2m / month) to lay the foundation for subsequent visual labeling.

[0079] In the fusion processing stage, the tool will take the spatial boundary of the significant area as the basic layer, and superimpose the monthly change rate data of the corresponding period. For example, select the change rate data from May to July (the key period of flood season, active sediment input), and spatially correlate it with the previously identified significant area (such as the 3000 square meter area around the upstream water inlet) to make each significant area subunit correspond to a specific rate value. When generating the comprehensive feature distribution map, a gradient color scale system is used: dark red, dark blue, and other dark colors are used to mark high change areas with rates exceeding the threshold (such as 0.15m / month), and light yellow, light gray, and other light colors are used to mark low change areas with rates below the threshold. At the same time, scale, legend, and time label are added to the map to ensure complete information.

[0080] Taking the comprehensive feature distribution map of a certain reservoir as an example, it is clearly shown that during the period from May to July, the upstream water inlet area has been continuously changing at a high rate of 0.2-0.3m / month (dark concentration area), and this area is highly coincident with the previously marked significant area. The areas in the middle and downstream of the reservoir are mostly marked with light colors, and the rates are generally below 0.1m / month. Further analysis shows that the shape of the high change area extends in a strip along the flow direction, which is highly consistent with the sediment transport path, confirming the conclusion that the upstream area is a high change rate area.

[0081] The core value of this fusion processing is to integrate scattered time series data (monthly rate) and spatial distribution data (significant area) into a single intuitive chart, which not only retains the quantitative information of the rate value, but also highlights the pattern characteristics of the spatial distribution, allowing analysts to grasp multi-dimensional information such as "where to change, how fast to change, and how long to change". For example, by comparing the comprehensive maps of different periods, long-term stable high-risk areas and short-term fluctuating abnormal areas can be identified, providing comprehensive data support for developing differentiated monitoring schemes (such as intensifying sampling in high-risk areas) and dredging plans, significantly improving the systematicness and accuracy of sedimentation dynamic analysis.

[0082] S500, according to the spatial distribution characteristics and the water flow velocity of the reservoir, determining the sediment transport amount, and according to the sediment transport amount, determining the spatial rule of sediment deposition;

[0083] Specifically, for the spatial distribution characteristics of reservoir sedimentation changes, historical rainfall data and reservoir inflow fluctuation records are integrated, combined with flow velocity distribution simulation technology to calculate the flow velocity in different areas of the reservoir, estimate the sediment transport amount, draw the sediment thickness distribution map, and judge the spatial law of sediment deposition. S500 specifically includes:

[0084] S510, determine the time distribution characteristics of the reservoir inflow fluctuation according to the historical rainfall data of the preset surrounding area of the reservoir, and determine the dynamic change data set of the reservoir inflow according to the time distribution characteristics of the reservoir inflow fluctuation;

[0085] For example, when analyzing the rainfall data of the surrounding area of the reservoir and the fluctuation of the inflow, key information can be extracted from historical records and in-depth discussion can be carried out combined with specific scenarios. Assuming that the rainfall data of the surrounding area of a reservoir comes from a time series database of the past five years, the record shows that the rainfall intensity reached a peak in July of a certain year, and the maximum daily rainfall was 120 millimeters. Through data comparison tools, this rainfall record is associated with the peak flow data in the same period for correlation analysis, and it is found that within 24 hours after the rainfall peak, the inflow fluctuates significantly, and the peak flow reaches 500 cubic meters per second. This time distribution characteristic shows the direct influence of rainfall on inflow, providing a basis for the construction of the subsequent dynamic change data set.

[0086] S520, simulate the flow path according to the dynamic change data set of the reservoir inflow, simulate the flow velocity distribution characteristic values of different areas in the reservoir, and obtain the flow velocity distribution characteristic map of each area;

[0087] For example, for the dynamic change data set of the reservoir inflow, when analyzing with a flow path simulation tool combined with terrain elevation data, the flow velocity distribution in different areas of the reservoir can be simulated. Assuming that the terrain slope of the upstream area of the reservoir is large, with an elevation difference of 10 meters, and the downstream area is relatively flat. Through the simulation tool, it is found that the flow velocity in the upstream area is generally fast, with an average speed of 2.5 meters per second, while the speed in the downstream area is only 0.8 meters per second. After drawing the flow velocity distribution characteristic map, the speed difference area can be clearly identified, laying a foundation for subsequent sediment transport analysis.

[0088] S530, screen out the areas where the flow velocity distribution characteristic values exceed the preset flow velocity threshold, estimate the sediment transport amount by using a deposition calculation tool combined with the flow velocity distribution characteristic map and the key physical parameters of the reservoir sediment, including the average particle size, density, and settling velocity of the particles, to obtain the sediment transport amount distribution data corresponding to each area;

[0089] For example, when estimating the sediment transport amount, if the characteristic value of the flow velocity of a certain area exceeds the preset threshold value of 1.5 m / s, further analysis is performed in combination with the sediment particle characteristic database. Assuming that the flow velocity of the upstream area reaches 2.5 m / s, which is obviously higher than the threshold value, and in combination with the characteristics of the average particle size of 0.2 mm in the database, the sediment transport amount of the area is estimated to be about 1000 tons per month by using a deposition calculation tool (such as the sediment transport module of HEC-RAS software). The acquisition of such distribution data helps to determine the main source area of the silt.

[0090] S540, according to the sediment transport amount distribution data, in combination with the reservoir area division information, a deposition thickness distribution map is generated by using a spatial superposition tool to obtain the spatial law of the silt deposition.

[0091] For example, when generating the deposition thickness distribution map for the sediment transport amount distribution data, the reservoir can be divided into multiple sub-areas for analysis by using the spatial superposition tool. Assuming that the sediment transport amount of the upstream area is high, in combination with the area division information, it is found that the average deposition thickness of the area is 0.3 m after generating the distribution map, while that of the downstream area is only 0.05 m. Such spatial law characteristics show that the silt is mainly concentrated in the upstream area, which provides a targeted basis for management decision-making.

[0092] S600, determining the key driving factors affecting the silt deposition according to the spatial law of the silt deposition; the driving factors include rainfall and flow velocity;

[0093] In this embodiment, according to the spatial law of the silt deposition, the influence of slope on the deposition position of the silt is calculated by using a terrain slope influence analysis method, and the correlation between rainfall, flow velocity and silt is analyzed by using a driving factor correlation test technology, a weight distribution matrix is constructed, and the weight distribution of each driving factor is determined. The S600 specifically includes:

[0094] 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;

[0095] For example, when analyzing the terrain slope distribution data in the target reservoir area, the related information can be extracted from the database, and the classification processing is performed in combination with the specific area characteristics. The terrain slope database usually contains elevation change data around and inside the reservoir, and the slope values can be divided into multiple grades by using a slope classification tool, such as 0-5 degrees for flat area, 5-15 degrees for medium slope area, and more than 15 degrees for steep slope area. Assuming that the slope of the upstream area of a certain reservoir is mostly between 10-20 degrees, which belongs to the steep slope area, and the slope of the downstream area is mostly between 2-5 degrees, which belongs to the flat area. Such distribution information lays a foundation for subsequent analysis.

[0096] It should be noted that the acquisition of the slope grade distribution information can be realized by a geographic information system software, and the classification can be completed by importing the elevation data and setting the classification standard.

[0097] In S620, according to the slope grade distribution information and the historical record of the silt distribution, the corresponding relationship between the slope grade and the deposition position is analyzed by a spatial superposition method, and the influence distribution characteristics of the slope grade on the deposition position of the silt are determined.

[0098] For example, for the analysis of the corresponding relationship between the slope grade and the deposition position of the silt, the slope distribution map can be superimposed and compared with the historical silt distribution record by means of a spatial superposition tool. It is assumed that the historical record shows that the deposition amount of the silt in the steep upstream area is large, and the deposition amount in the gentle downstream area is small, and it is found by the superposition analysis that the greater the slope is, the more the deposition position is concentrated in the turning area of the flow slowing down. This feature shows that the slope has a significant influence on the deposition position of the silt. In implementation, two layers of data can be loaded by a spatial analysis software, and a distribution characteristic map can be generated after setting the superposition rules.

[0099] In S630, rainfall data and flow velocity data are acquired, the correlation coefficient between the driving factors and the silt distribution is calculated by using a correlation test tool with the rainfall data and the flow velocity data as the driving factors, and if the correlation coefficient of a certain driving factor exceeds a preset threshold value, the corresponding driving factor is marked as a key driving factor.

[0100] For example, in the screening of the key driving factors, the rainfall and the flow velocity can be analyzed by a correlation test tool as main data sources. It is assumed that the historical record of a reservoir area shows that the rainfall peak value appears in a certain month, reaching 150 mm / day, and the flow velocity peak value in the same period is 2 m / s, and it is found by the test that the correlation coefficient between the rainfall and the silt distribution is 0.85, which exceeds the preset threshold value 0.7, and therefore is marked as a key driving factor. The correlation test can be realized by a statistical analysis software, and the coefficient result can be automatically generated after inputting two groups of data.

[0101] For example, for the weight distribution of the key driving factors, a weight matrix construction tool can be used for calculation. It is assumed that the rainfall and the flow velocity are respectively given weights of 0.6 and 0.4, and it is found by combining the historical data analysis that when the rainfall weight is high, the silt distribution tends to be in the upstream area. The generation of this weight distribution data helps to clarify the influence of each factor on the spatial distribution law. In implementation, the factor data and the weight rules can be input into a decision support system software to quickly generate the distribution result.

[0102] It should be noted that the weight calculation needs to be adjusted in combination with long-term observation data to ensure that the result is in line with the actual situation.

[0103] S700, determining a high-risk siltation area according to the key driving factor, and monitoring the reservoir capacity change according to the siltation change of the high-risk siltation area.

[0104] In this embodiment, the terrain elevation data in the target area is obtained from the pre-established reservoir terrain database. The elevation difference in the area is classified by using a hierarchical processing tool for the terrain elevation data, and the elevation distribution characteristics of the area are obtained. By using the elevation distribution characteristics and combining the weight value information of the driving factor, if the weight value exceeds the preset threshold value, the corresponding reservoir terrain area is divided by using a region marking tool to determine the boundary of the key monitoring area. The high-precision scanning equipment is used for detailed data collection in the area for the boundary of the key monitoring area, the local terrain detail information of the area is obtained, and the distribution of the high-risk siltation points in the area is judged. According to the local terrain detail information, the boundary division tool is used to define the range of the high-risk siltation points in the key monitoring area, and the specific distribution range of the high-risk siltation area is obtained. The S700 specifically includes:

[0105] S710, for the reservoir terrain area corresponding to the key driving factor, the key data is collected by using the unmanned aerial vehicle carrying the laser radar, the local terrain elevation change information is obtained, and the high-risk siltation area is determined according to the local terrain elevation change information;

[0106] For example, when analyzing elevation data in a reservoir terrain database, the relevant information of the target area can be extracted from the database first, and the elevation difference can be classified and processed by the layering tool. Assuming that the elevation range of a certain reservoir area is between 100 and 500 meters, after layering processing, the elevation can be divided into low, medium and high levels, which are 100-200 meters, 200-350 meters and 350-500 meters respectively. This classification method helps to clarify the elevation distribution characteristics in the region and provides basic data support for subsequent analysis. The layering tool usually relies on a geographic information system platform, and the operation can be completed by importing elevation data and setting classification standards. Among them, the layering tool is an elevation data classification module developed based on the geographic information system (GIS) platform (such as ArcGIS, QGIS), and the core function is to divide continuous elevation values into discrete levels according to the preset standards. When operating, the elevation data in the reservoir terrain database (such as 100-500 meters) needs to be imported first, and then the classification method is set through the tool: equal interval classification (such as fixed interval of 100 meters), natural breakpoint classification (automatic division according to data distribution characteristics) or manual definition of threshold (such as 100-200 meters, 200-350 meters, 350-500 meters in the text). After processing, the elevation layering vector map or raster map is output, which directly shows the spatial distribution of different elevation areas and provides a structured data basis for subsequent analysis of siltation risk combined with driving factors.

[0107] For example, when analyzing elevation data in a reservoir terrain database, the relevant information of the target area can be extracted from the database first, and the elevation difference can be classified and processed by the layering tool. Assuming that the elevation range of a certain reservoir area is between 100 and 500 meters, after layering processing, the elevation can be divided into low, medium and high levels, which are 100-200 meters, 200-350 meters and 350-500 meters respectively. This classification method helps to clarify the elevation distribution characteristics in the region and provides basic data support for subsequent analysis. The layering tool usually relies on a geographic information system platform, and the operation can be completed by importing elevation data and setting classification standards.

[0108] For the combined analysis of elevation distribution characteristics and driving factor weight values, if the weight value of a certain driving factor is 0.75, which exceeds the preset threshold value 0.6, the relevant area needs to be marked as a key area. Assuming that the driving factor is related to the area with high upstream elevation, a higher weight value indicates that the area may have a higher risk of siltation. The area marking tool can be realized through spatial analysis software, and the elevation distribution map and weight data are superimposed to quickly divide the boundary of the key monitoring area. This method can effectively lock the high-risk area and improve the pertinence of monitoring. Among them, the area marking tool is a functional component in spatial analysis software (such as the spatial overlay module of ArcGIS and the thematic mapping tool of ENVI), which is used to spatially correlate the elevation distribution characteristics and driving factor weight data and mark the key area. The specific operation includes: first, unify the coordinate systems of the two (such as WGS84 coordinate system), and then superimpose the elevation distribution map and the driving factor weight layer through the tool to automatically assign special identification (such as red highlight, boundary thickening) to the area with weight value exceeding the preset threshold value (such as 0.6). For example, when the weight value of a certain driving factor (such as upstream sand amount) is 0.75, the tool can quickly locate the elevation area (such as the upstream 350-500 meter area) where the factor has a significant impact, generate a vector boundary map with markings, and realize the visualization locking of high-risk areas.

[0109] After determining the boundary of the key monitoring area, it is particularly important to use high-precision scanning equipment to collect detailed data in the area. Assuming that a certain area upstream of a reservoir is marked as a key monitoring area, using high-precision scanning equipment can obtain local terrain details, such as slope changes or micro-topographic features in a small range. Scanning equipment usually includes a laser range finder or a high-resolution image acquisition device, which can generate accurate topographic data maps through field operations, and then determine the distribution position of high-risk points of siltation. This refined data acquisition helps to more accurately identify potential risk points.

[0110] For local topographic detail information, the boundary division tool is used to define the scope of high-risk points, which can further clarify the specific distribution of risk areas. Assuming that the scanning results show that there are multiple slope mutation points in a certain area, and historical data shows that there are often siltation phenomena near these points, the specific range can be circled through the boundary division tool. In terms of tool operation, usually load detailed data in geographic information platform, set risk judgment standard to automatically generate distribution range map. This way can provide accurate spatial reference for subsequent management measures. Among them, the boundary division tool is a component in the geographic information platform (such as the vector editing tool of QGIS, AutoCAD Civil 3D) used to define the range of risk areas. When operating, first load the local topographic detail data obtained by high-precision scanning (such as slope mutation point, micro concave point coordinates), and then set the risk judgment standard (such as slope > 15°, point with historical siltation record), the tool automatically connects adjacent high-risk points through spatial clustering algorithm to generate closed risk area boundary vector map (such as SHP format). For example, when scanning finds that there are multiple slope mutation points in a certain area and there are historical siltation, the tool can quickly circulate the specific range of the area (such as polygon boundary), providing accurate spatial reference for dredging operation, monitoring equipment layout and other management measures.

[0111] From multiple aspects, the determination of high-risk siltation areas can be analyzed in combination with topographic features and historical siltation records. Assuming that the slope of a certain high-elevation area upstream changes dramatically, and there have been siltation accumulations several times in the past three years, high-precision scanning confirms that there are multiple micro concave points in this area, which are prone to sediment accumulation. At the same time, the weight analysis shows that the driving factor of this area has a significant impact, and after comprehensive judgment it can be listed as the primary monitoring object. This multi-dimensional analysis can support each other, ensuring the rationality of the division result, and providing a reliable basis for reservoir management and reducing resource waste.

[0112] S720, by processing the time series elevation data of the high-risk area, extracting the long-term trend feature, and locating the short-term mutation feature deviating from the long-term change feature through the anomaly point detection method, obtaining the classification result of the siltation change mode of the high-risk area according to the long-term trend feature and the short-term mutation feature;

[0113] In this embodiment, time series elevation data of the target area is obtained from a pre-established reservoir topography database, and a sliding window method is used to segment and smooth the time series elevation data to obtain a denoised elevation change curve. For the denoised elevation change curve, a least squares method is used to fit a long-term trend line, and the elevation residual at each time point is calculated. If the elevation residual exceeds the preset fluctuation threshold, it is marked as an abnormal point. According to the distribution of the abnormal points, combined with the pre-divided high-risk area boundary, a clustering algorithm is used to spatially aggregate the abnormal points to determine the concentration area of the mutation event. Through the superposition of the mutation event concentration area and the local topographic details, a region growing algorithm is used to expand the boundary to determine the sub-area range with significantly increased deposition rate.

[0114] When obtaining time series elevation data of the target area from the reservoir topography database, the elevation records of the upstream area of a certain reservoir in the past five years can be filtered out, with a data time interval of once a month. Assuming that the range of the area elevation data is between 120 meters and 450 meters, there may be noise in the data due to measurement error or short-term weather influence. By using the sliding window method, the time series data is smoothed by taking every three months as a window to obtain a relatively smooth elevation change curve. This method can effectively reduce the interference of short-term fluctuations on subsequent analysis and lay the foundation for extracting long-term trends.

[0115] For the denoised elevation change curve, a least squares method is used to fit a long-term trend line, which can be fitted as a gradually declining trend line for five years of data, reflecting the possible decrease in topography due to sedimentation. Assuming that the elevation value at a certain time point is 380 meters, and the predicted value of the trend line is 390 meters, the residual is 10 meters, and if the preset fluctuation threshold is 8 meters, the point is marked as an abnormal point. This marking method helps to find irregular fluctuations in elevation changes and provides clues for subsequent risk area judgment.

[0116] When spatially aggregating according to the distribution of abnormal points combined with the high-risk area boundary, a clustering algorithm can be used to group abnormal points by spatial position. Assuming that there are 10 abnormal points in the upstream area of a certain reservoir, 8 of which are concentrated in a small range, these points can be aggregated into a mutation event concentration area through clustering analysis. This aggregation method can help focus on the local range that may have significant topographic changes and avoid resource waste caused by scattered analysis.

[0117] By superimposing the concentrated area of mutation events and local terrain details, and using the region growing algorithm to expand the boundary, we can start from the center point of the concentrated area and gradually expand to the surrounding area, including the adjacent area with larger slope or elevation mutation. Assuming that the slope near the center point of a concentrated area reaches 15 degrees, and there are multiple micro concave points around it, then after the algorithm expansion, a sub-area range of about 2 square kilometers can be determined, which is judged as a region with significantly increased deposition rate. This expansion method can more comprehensively cover potential risk points and provide spatial basis for precise management.

[0118] From multiple aspects, the determination of the sub-area with significantly increased deposition rate can be combined with historical deposition records and terrain features for analysis. Assuming that the sub-area has appeared several times in the past two years, and the terrain scanning shows that there are multiple low points inside, which are easy to accumulate material, and the upstream water quantity is large, carrying more sediment, the comprehensive judgment of this area is a high-risk area. This multi-dimensional analysis can verify each other to ensure the rationality of the division and provide reliable support for reservoir deposition prevention and control.

[0119] In the implementation process, for the processing of time series data and the marking of abnormal points, the window size or threshold parameter can be adjusted according to the specific environment of the reservoir. Assuming that a certain reservoir area is greatly affected by seasonal floods, the sliding window can be adjusted to half a year to capture longer-term fluctuation characteristics. This flexibility can improve the adaptability of data analysis and ensure that the results are more in line with actual needs, providing more effective technical support for subsequent risk prevention and control.

[0120] S730, according to the classification result of the deposition change mode, a dynamic monitoring database is constructed, the deposition change mode is matched with the real-time collected underwater terrain elevation data, the real-time state of the reservoir capacity change caused by the reservoir deposition is determined, and dynamic monitoring update data is obtained.

[0121] In this embodiment, the deposition pattern and classification result of the high-risk area are obtained from the pre-established monitoring database. The classification result is preliminarily compared with the real-time terrain data, the change range of the real-time elevation is extracted by the data collection tool, and the initial matching data set corresponding to the area boundary is obtained. According to the initial matching data set, the data correlation tool is used to deeply compare the deposition pattern and the real-time elevation. If the elevation change exceeds the preset threshold, it is marked as an abnormal area, and the potential impact range of the capacity change is determined. Through the corresponding relationship between the abnormal area and the change state, the data fusion tool is used to integrate the dynamic monitoring data stream and the classification result, obtain the key indicators related to the capacity change in the update data, and judge the current state fluctuation of the reservoir. According to the state fluctuation, the data storage tool is used to synchronize the update data and the historical records in the monitoring database, generate a real-time report of dynamic monitoring for the synchronized data, and obtain the final monitoring result related to the deposition change of the high-risk area.

[0122] When obtaining the deposition pattern and classification result of the high-risk area from the pre-established monitoring database, the deposition pattern data of a specific area upstream of a reservoir can be extracted first. Assuming that this area is classified as a high-risk deposition area, historical data shows that its deposition rate increases by about 0.5 meters per year. For the real-time collected terrain data, the current elevation value can be obtained by high-precision surveying and mapping equipment. Assuming that the latest measurement value is 125.3 meters, compared with the last record of 126.1 meters in the database, the change range is 0.8 meters. After preliminary comparison, the real-time elevation change is matched with the area boundary by the data collection tool to form an initial data set, laying a foundation for subsequent analysis.

[0123] When deeply comparing the initial matching data set, assume that the data correlation tool (spatial correlation module of geographic information system (GIS) platform) is used to correlate and analyze the real-time elevation change and the historical deposition pattern. If the preset elevation change threshold is 0.6 meters and the actual change is 0.8 meters, the area is marked as an abnormal area. Further analysis of the potential impact range can be combined with the reservoir capacity change to speculate that this area may cause the upstream water storage capacity to decrease by about 2%, which needs to be paid attention to.

[0124] In integrating dynamic monitoring data streams and classification results, data fusion tools can be used to combine real-time monitoring of flow, sediment content, and other indicators with historical classification results to extract key indicators such as sedimentation rate changes. Assuming that the latest data stream shows that the sediment content has increased by 15% compared to last month, combined with the classification result, it is determined that the current reservoir state fluctuates greatly and may enter a high-risk stage. This approach helps to timely grasp the dynamic changes of the reservoir. In this embodiment, the GIS tool itself has strong spatial data fusion capabilities, and can associate and integrate spatial data of different sources (such as elevation map, regional boundary, and monitoring point distribution) with non-spatial data (such as sediment content and flow time series data).

[0125] In the case of synchronous updating of data and historical records for state fluctuations, data storage tools can be used to upload the latest monitoring data to the database. Assuming that after synchronization, it is found that the elevation has been continuously decreasing for the past three months, with an average decrease of 0.3 meters per month. Based on this, real-time reports are generated, including the sedimentation trend of high-risk areas, such as the sedimentation thickness in a certain area reaching 1.2 meters, which requires priority control measures. This synchronization and report generation method ensures data consistency and traceability.

[0126] From another side, the corresponding relationship between abnormal areas and change states can be analyzed by focusing on the correlation between topographic features and sedimentation patterns. Assuming that the slope in an abnormal area is small, only 5 degrees, which is easy to form deposition, and real-time data shows that the rainfall has been large in recent period, the sediment input increases, and the comprehensive judgment is that the capacity change in this area may further intensify. Such multi-dimensional analysis can provide more comprehensive reference for reservoir management.

[0127] Figure 2 The structure diagram of the reservoir capacity dynamic monitoring system based on the unmanned aerial vehicle provided by the present application is shown in Figure 2 The reservoir capacity dynamic monitoring system based on the unmanned aerial vehicle comprises:

[0128] The data acquisition module 210 is configured to acquire a reservoir surface feature data set by using a multi-source remote sensing device carried by the unmanned aerial vehicle, wherein the reservoir surface feature data set comprises water surface image data of the reservoir.

[0129] The terrain modeling module 220 is configured to determine water area boundary information according to the reservoir surface feature data set, perform depth scanning on the bottom of the reservoir according to the water area boundary information, acquire underwater topographic elevation data and sediment distribution data, and construct a three-dimensional terrain model of the reservoir.

[0130] The dynamic analysis module 230 is configured to determine reservoir topographic change data and sediment change data according to dynamic change data of the three-dimensional terrain model of the reservoir in a continuous time sequence.

[0131] The siltation velocity analysis module 240 is configured to analyze the change rate of the reservoir topography change data and the siltation change data, and obtain the spatial distribution characteristics of the reservoir siltation change;

[0132] The siltation transport law analysis module 250 is configured to determine the silt transport amount according to the spatial distribution characteristics and the reservoir flow velocity, and determine the spatial law of the silt deposition according to the silt transport amount;

[0133] The correlation analysis module 260 is configured to determine the key driving factor affecting the silt deposition according to the spatial law of the silt deposition; the key driving factor includes rainfall and flow velocity;

[0134] The dynamic monitoring module 270 is configured to determine the siltation high-risk area according to the key driving factor, and monitor the reservoir capacity change according to the siltation change of the siltation high-risk area.

[0135] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A reservoir capacity dynamic monitoring method based on a UAV, characterized in that, The method comprises the following steps: S100, acquiring a reservoir surface feature data set by a UAV, wherein the reservoir surface feature data set comprises reservoir water surface image data; S200, determining water area boundary information according to the reservoir surface feature data set, performing depth scanning on the bottom of the reservoir according to the water area boundary information, acquiring underwater topographic elevation data and sediment distribution data, and constructing a reservoir three-dimensional topographic model; S300, determining reservoir topographic change data and sediment change data according to dynamic change data of the reservoir three-dimensional topographic model in a continuous time sequence; S400, performing change rate analysis on the reservoir topographic change data and the sediment change data to obtain spatial distribution characteristics of reservoir sediment change; S500, determining a sediment transport amount according to the spatial distribution characteristics and reservoir flow velocity, and determining a spatial rule of sediment deposition according to the sediment transport amount; S600, determining key driving factors affecting sediment deposition according to the spatial rule of sediment deposition; The driving factors include rainfall and flow velocity; S700, determining a high sedimentation risk area according to the key driving factors, and monitoring reservoir capacity change according to sediment change in the high sedimentation risk area. 2.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 1, wherein, The S700 specifically comprises: S710, performing key data acquisition on a reservoir topographic area corresponding to the key driving factors by a UAV carrying a laser radar to acquire local topographic elevation change information, and determining a high sedimentation risk area according to the local topographic elevation change information; S720, processing time sequence elevation data of the high risk area to extract long-term trend characteristics, positioning short-term mutation characteristics deviating from the long-term change characteristics by an abnormal point detection method, and obtaining a classification result of a sedimentation change mode of the high risk area according to the long-term trend characteristics and the short-term mutation characteristics; S730, constructing a dynamic monitoring database according to the classification result of the sedimentation change mode, associating and matching the sedimentation change mode with real-time acquired underwater topographic elevation data, determining a real-time state of reservoir capacity change caused by reservoir sedimentation, and obtaining dynamic monitoring update data. 3.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 2, wherein, The S730 specifically comprises: Preliminarily comparing the sedimentation change mode of the high risk area with the acquired underwater topographic real-time elevation data, extracting a change range of the underwater topographic real-time elevation data, and obtaining an initial matching data set corresponding to a boundary of the high risk area; According to the initial matching data set, deeply comparing the sedimentation change mode with the underwater topographic real-time elevation data, if a change of the underwater topographic real-time elevation data exceeds a preset change threshold, marking a corresponding high risk area as an abnormal area, and determining a potential influence range of reservoir capacity change caused by the abnormal area. The abnormal area and the change state are matched, and a classification result of a change mode of the high-risk area is obtained; the data stream of the dynamic monitoring and the classification result of the change mode of the high-risk area are integrated, a key index related to the capacity change is obtained from updated data after the integration, and a state fluctuation of the reservoir is determined; the change state includes an elevation change value, a deposition rate and a matching degree with the change mode of the deposition; the key index includes a deposition thickness, a deposition volume change and a capacity reduction ratio; According to the state fluctuation, the updated data and historical records in a monitoring database are synchronized, a real-time report of the dynamic monitoring is generated for the synchronized data, and a final monitoring result related to the deposition change of the high-risk area is obtained. 4.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 1, wherein, The S100 specifically includes: S110, a multi-source remote sensing image device carried by the unmanned aerial vehicle is used to perform full-coverage scanning on a reservoir surface and a surrounding terrain in a preset range, and reservoir water surface image data and surrounding terrain elevation data are obtained; S120, the reservoir water surface image data and the surrounding terrain elevation data are denoised and corrected to obtain a first image data set of the reservoir surface and a first data set of the terrain elevation; spatial registration and data fusion are performed on the first image data set of the reservoir surface and the first data set of the terrain elevation, terrain feature extraction is performed on the fused data, and preliminary feature distribution information of a reservoir surface layer is determined; S130, data filling is performed on the preliminary feature distribution information to obtain a complete reservoir surface layer feature data set, three-dimensional modeling is performed based on the reservoir surface layer feature data set to obtain a three-dimensional distribution model of the reservoir surface, and final distribution feature information of the reservoir surface is determined according to the three-dimensional distribution model. 5.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method according to claim 1, wherein, The S200 specifically includes: S210, water area boundary information is determined according to the reservoir surface layer feature data set, and a depth scanning of a reservoir bottom is performed by using an underwater sonar detection device to obtain original elevation point cloud data and bottom material reflection feature data of an underwater terrain; S220, an echo intensity value in the bottom material reflection feature data is obtained, if the echo intensity value exceeds a preset echo intensity threshold value, it is determined that the corresponding bottom material is a bedrock terrain, otherwise, it is determined that the corresponding bottom material is a deposition material distribution area, and a bottom material classification result and a deposition layer thickness distribution data are obtained; S230, a continuous reservoir bottom terrain elevation network is generated by using a spatial interpolation method according to the classification result of the bottom material, the reservoir bottom terrain elevation network is spatially registered with the reservoir water surface image data to obtain a reservoir integrated elevation data set; S240, a triangular net is constructed according to the reservoir integrated elevation data set, and the bottom network is layered and marked according to the deposition layer thickness distribution data to obtain a reservoir three-dimensional terrain model including a reservoir surface shape and an underwater terrain fluctuation. 6.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 1, wherein, The S300 specifically includes: S310, depth scanning data of a reservoir bottom is obtained from preset time nodes by using a multi-time period data acquisition method to form an underwater terrain elevation data set including a plurality of time nodes, and a preliminary time series data set is obtained; S320, performing layer-by-layer difference calculation on the underwater topographic elevation data of each adjacent two time nodes in the time series data set to determine an elevation difference distribution map; S330, if it is judged according to the elevation difference distribution map that the elevation change value of a certain region exceeds a preset elevation change threshold, calculating the underwater topographic elevation change value of the corresponding region, and calculating the silt volume change value according to the underwater topographic elevation change value; S340, integrating the elevation difference distribution map and the silt volume change value to obtain the dynamic change trend of the reservoir topography and silt volume at different time periods, and obtaining reservoir topography change data and silt change data. 7.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 1, wherein, The S400 specifically includes: S410, obtaining reservoir topography change data of multiple time periods from a preset time series, performing monthly analysis on the reservoir topography change data, calculating a monthly change rate, and obtaining reservoir topography monthly change rate distribution data set; S420, performing spatial region division on the reservoir topography monthly change rate distribution data set, if it is judged that the monthly change rate of a certain region after 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 heterogeneity features of the spatial distribution range of the significant region to obtain heterogeneity distribution data of reservoir topography change and silt change, and generating a heterogeneity feature distribution map according to the heterogeneity distribution data; S440, according to the heterogeneity feature distribution map, fusing the monthly change rate and the spatial distribution range of the significant region to generate a comprehensive feature distribution map, and determining the spatial distribution characteristics of the reservoir silt change according to the comprehensive feature distribution map. 8.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 1, wherein, The S500 specifically includes: S510, determining the time distribution characteristics of the inflow fluctuation according to the historical rainfall data of the reservoir preset surrounding area, and determining the dynamic change data set of the inflow according to the time distribution characteristics of the inflow fluctuation; S520, simulating the water flow path according to the dynamic change data set of the inflow, simulating the water flow velocity distribution characteristic value of different regions in the reservoir to obtain the water flow velocity distribution characteristic map of each region; S530, screening out the regions whose water flow velocity distribution characteristic values exceed a preset water flow velocity threshold, estimating the sediment transport amount of each region by using a deposition calculation tool through the key physical parameters of the reservoir sediment combined with the water flow velocity distribution characteristic map; the key physical parameters include average particle size, density and settling velocity; S540, generating a deposition thickness distribution map by a spatial superposition tool according to the sediment transport amount distribution data combined with the reservoir region division information to obtain the spatial law of silt deposition. 9.The unmanned aerial vehicle based dynamic reservoir capacity monitoring method of claim 1, wherein, The S600 specifically includes: S610, obtaining topographic slope distribution data of different reservoirs, performing hierarchical processing on the slope values of different reservoir regions according to the topographic slope distribution data to obtain slope grade distribution information of each reservoir region; S620, according to the slope grade distribution information and the historical record of the silt distribution, a corresponding relationship between the slope grade and the deposition position is analyzed by a spatial superposition method, and an influence distribution characteristic of the slope grade on the silt deposition position is determined. S630, rainfall data and water flow speed data are obtained, the rainfall data and the water flow speed data are taken as driving factors, a correlation test tool is used to calculate a correlation coefficient of the driving factors and the silt distribution, and if the correlation coefficient of a certain driving factor exceeds a preset threshold value, the corresponding driving factor is marked as a key driving factor.

10. A dynamic reservoir capacity monitoring system based on a drone, characterized in that, Comprise: A data acquisition module is configured to acquire a reservoir surface feature dataset by a multi-source remote sensing device carried by a UAV, wherein the reservoir surface feature dataset comprises reservoir water surface image data. A terrain modeling module is configured to determine water area boundary information according to the reservoir surface feature dataset, perform depth scanning on a reservoir bottom according to the water area boundary information, acquire underwater terrain elevation data and silt distribution data, and construct a reservoir three-dimensional terrain model. A dynamic analysis module is configured to determine reservoir terrain change data and silt change data according to dynamic change data of the reservoir three-dimensional terrain model in a continuous time sequence. A silt deposition velocity analysis module is configured to analyze change rate of the reservoir terrain change data and the silt change data, and obtain spatial distribution characteristics of reservoir silt change. A silt deposition and migration rule analysis module is configured to determine a sediment migration amount according to the spatial distribution characteristics and reservoir water flow speed, and determine a spatial rule of silt deposition according to the sediment migration amount. A correlation analysis module is configured to determine a key driving factor affecting silt deposition according to the spatial rule of silt deposition, wherein the key driving factor comprises rainfall and water flow speed. A dynamic monitoring module is configured to determine a silt deposition high-risk area according to the key driving factor, and monitor reservoir capacity change according to silt change of the silt deposition high-risk area.

Images