Seawall measurement method and system based on multi-source data fusion

Abstract

The invention belongs to the technical field of hydraulic engineering measurement, and discloses a seawall measurement method and system based on multi-source data fusion, and the method comprises the steps: collecting geological data and hydrological data through a data collection module; the route planning module formulates a route; the overwater data acquisition equipment and the underwater data acquisition equipment are used for monitoring; the auxiliary equipment provides a high-precision positioning reference; the meteorological data module obtains meteorological data; the data processing module processes the data; the three-dimensional modeling module carries out point cloud registration and DEM model construction; the comprehensive analysis module performs comprehensive evaluation and predicts potential problems and risks; when an abnormal condition or a potential risk is monitored, the early warning module gives an alarm. Multi-source data acquisition equipment such as unmanned aerial vehicle aerial survey, a laser radar, a multi-beam depth finder, a side-scan sonar, a shallow profiler and a micro-motion exploration instrument are integrated, seawall overwater and underwater full-dimensional data acquisition, processing, modeling and volume calculation are realized, and high precision and reliability of measurement results are ensured.

Description

Technical Field

[0001] The present application relates to the field of water conservancy engineering measurement technology, and more specifically, to a seawall measurement method and system based on multi-source data fusion, which is suitable for dynamic monitoring of deepwater dams and refined volume measurement of multi-phase filling projects. Background Art

[0002] Seawalls are subject to the influence of the natural environment and their condition is constantly changing. A single data point cannot fully reflect the true condition of a seawall. Different data collection methods have their own advantages and disadvantages. For example, sonar accurately measures underwater topography but is unable to measure above-water terrain. Drone aerial photography can capture large-scale images but struggles to penetrate deeply underwater.

[0003] The document with the prior art publication number CN112629597B provides a dike overtopping measurement system. The system includes an overtopping collection box installed behind the crest of the target monitoring dike, a water outlet at the bottom of the overtopping collection box connected to the water inlet of the flow meter group, a signal transmission end of the flow meter group connected to the overtopping amount monitoring server, and the overtopping amount monitoring server connected to the remote monitoring terminal. The rainfall measuring device is installed behind the crest of the target monitoring dike, and the signal transmission end of the rainfall measuring device is connected to the overtopping amount monitoring server. The overtopping collection box collects the overtopping water on the dike site, collects flow data based on the flow meter group, combines it with the rainfall collected by the rain gauge, and calculates the real-time overtopping amount through the overtopping amount monitoring server. The remote monitoring terminal displays the overtopping amount in real time via the Internet, thus realizing on-site monitoring and remote access of the dike overtopping amount during typhoons. The system is characterized by full automation, high precision, and good durability.

[0004] Although the above-mentioned existing technical solutions can achieve the relevant beneficial effects through the existing technical structure, they still have the following defects: 1. The efficiency of single-point measurement is low, and a typical project takes 75 days to complete; 2. The above-water and underwater data are discontinuous, with errors often reaching ±20cm; 3. The multi-source data alignment deviation exceeds ±10cm, and there are obvious cracks at the edges of the model.

[0005] In view of this, we propose a seawall measurement method and system based on multi-source data fusion. Summary of the Invention

[0006] 1. Technical problems to be solved

[0007] The purpose of this application is to provide a seawall measurement method and system based on multi-source data fusion, which solves the technical problems raised in the above-mentioned background technology, and realizes the technical effect of full-dimensional data collection, processing, modeling and volume calculation of seawalls above and below the water by integrating multi-source data acquisition equipment such as drone aerial survey, lidar, multi-beam echo sounder, side-scan sonar, shallow layer profiler, micro-motion exploration instrument, etc., with professional data processing software and rigorous quality control process, to ensure the high precision and reliability of measurement results.

[0008] 2. Technical solution

[0009] The technical solution of this application provides a seawall measurement system based on multi-source data fusion, including:

[0010] Data collection module: collects geological data and hydrological data at the seawall, including historical data, and labels the data.

[0011] Route planning module: Based on the geological and hydrological data at the seawall, a scientific and reasonable operation route is formulated for the above-water and underwater data acquisition equipment.

[0012] Water data collection equipment: This includes a water drone equipped with a Zenmuse P1 camera and a Zenmuse L1 LiDAR. This drone acquires orthophotos and point cloud data for the water.

[0013] Auxiliary equipment: Equipped with a CORS station to provide high-precision positioning benchmarks for data collection by drones and survey ships, ensuring that data can be uniformly aligned to the CGCS2000 coordinate system and the 1985 National Height Datum.

[0014] Underwater data acquisition equipment: including survey vessels equipped with multi-beam echo sounders, side-scan sonars, shallow-sediment profilers, underwater drones and micro-motion detectors to monitor the underwater portion of the seawall.

[0015] Data processing module: processes the collected data, including point cloud data processing, posture correction, image processing and data registration;

[0016] 3D Modeling Module: Performs point cloud registration and DEM model construction. An improved ICP algorithm is used to register multi-source point cloud data. A DEM model with a grid spacing of ≤2m is constructed based on the TIN algorithm. The embankment is divided into four blocks: east, north, south, and west. Block boundaries are marked to improve modeling accuracy in complex terrain.

[0017] Volume calculation module: A grid method was used to calculate volume, setting a 2m x 2m grid and calculating the fill volume for each block. ArcGIS cut-and-fill analysis tools were used for cross-validation with Trimble RealWorks point cloud volume extraction results.

[0018] Meteorological data module: includes a weather instrument to obtain real-time meteorological data, including information such as rain, snow, wind and earthquakes.

[0019] Comprehensive analysis module: This module combines surface and underwater data acquisition equipment with meteorological data to promptly detect abnormalities, conduct a comprehensive assessment of the seawall, and predict potential problems and risks.

[0020] Early warning module: including alarm, which will issue an alarm in time when abnormal conditions or potential risks are detected;

[0021] Result output module: Visualize the measurement results and display the DEM model of the seawall, volume calculation results, etc. in a three-dimensional visual way.

[0022] PLC control center: network connection with surface data acquisition equipment, underwater data acquisition equipment, meteorological data module, data collection module, route planning module, auxiliary equipment, data processing module, three-dimensional modeling module, volume calculation module, comprehensive analysis module and early warning module.

[0023] The present invention provides a seawall measurement method based on multi-source data fusion, comprising the following steps:

[0024] S1. The data collection module collects geological data and hydrological data at the seawall, including historical data, and annotates the data;

[0025] S2, the route planning module formulates a scientific and reasonable operation route for the above-water data acquisition equipment and the underwater data acquisition equipment based on the geological data and hydrological data at the seawall;

[0026] S3. The surface data acquisition equipment and the underwater data acquisition equipment monitor according to the planned route; the auxiliary equipment provides a high-precision positioning benchmark for the surface data acquisition equipment and the underwater data acquisition equipment; the meteorological data module obtains meteorological data in real time.

[0027] S4, the data processing module processes the collected data, including point cloud data processing, posture correction, image processing and data registration;

[0028] S5, 3D modeling module performs point cloud registration and DEM model construction;

[0029] S6. The volume calculation module uses the grid method to calculate the volume and the filling volume of each block.

[0030] S7, the comprehensive analysis module combines surface data acquisition equipment, underwater data acquisition equipment and meteorological data to promptly detect abnormal conditions, conduct a comprehensive assessment of the seawall, and predict potential problems and risks;

[0031] S8. When abnormal conditions or potential risks are detected, the early warning module will issue an alarm in time;

[0032] S9. The result output module displays the measurement results visually.

[0033] 3. Beneficial effects

[0034] One or more technical solutions provided in the technical solution of this application have at least the following technical effects or advantages:

[0035] 1. This invention achieves integrated "air-space-ground-water" coverage through the synergistic integration of surface and underwater measurement, eliminating blind spots in traditional single-device measurements (such as complex terrain and shallow areas). This ensures seamless integration of multi-source data and avoids assessment distortions caused by coordinate system deviations. Multi-device collaborative scheduling: Surface drones and survey vessels operate in different time zones and regions (e.g., drones operate in low winds early in the morning, while survey vessels scan shallows during high tide), reducing airspace / water conflicts.

[0036] 2. Dynamically adapt to the environment: Automatically optimize routes based on real-time data such as tides, wind speed, and water depth, improving operational efficiency while avoiding the risk of equipment running aground or collisions.

[0037] 3. Achieve multi-dimensional comprehensive assessment and real-time risk warning; integrate optical imaging (crack identification), lidar (settlement monitoring), multi-beam (scour pit analysis) and meteorological data (typhoon, rainstorm), calculate the seawall health score in real time through a weighted average model (or machine learning algorithm), and automatically identify anomalies such as seepage damage and foundation scour.

[0038] 4. Foresighted risk prediction: Based on historical data and climate change scenarios, predict future erosion trends and settlement rates, identify weak sections in advance, and create a time window for preventive maintenance.

[0039] 5. Through visualization methods such as DEM models, point cloud rendering, and underwater terrain animation, the system intuitively presents seawall structural defects (such as crack distribution maps) and earthwork volume statistics (block-by-block volume reports), allowing even non-professionals to quickly grasp key information. Data backtracking and trend analysis, comparing measurement data from different periods, and automatically generating deformation trend reports assist management departments in formulating differentiated maintenance strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of the seawall measurement method based on multi-source data fusion disclosed in this application.

[0041] Figure 2 This application uses the Qingdao Ruitai Shipbuilding Seawall Project as an example of a dam volume statistics table.

[0042] Figure 3This application discloses a schematic diagram of a multi-beam bathymetric topography rendering using the Qingdao Ruitai Shipbuilding Seawall Project as an example.

[0043] Figure 4 This is a schematic diagram of the dam block division using the Qingdao Ruitai Shipbuilding Seawall Project as an example disclosed in this application. DETAILED DESCRIPTION

[0044] The present application is further described in detail below with reference to the accompanying drawings.

[0045] Reference Figure 1 The present application provides a seawall measurement system based on multi-source data fusion, including:

[0046] Data Collection Module: This module collects geological and hydrological data from the seawall, including historical data, and annotates the data. It also obtains key information such as the seawall's underwater stratigraphic structure and the thickness of the embankment's fill layer. It also obtains detailed underwater landform information to help identify unique landforms near the seawall, such as reefs and scour pits.

[0047] Route Planning Module: Based on the geological and hydrological data at the seawall, a scientific and reasonable operation route is formulated for the above-water and underwater data acquisition equipment to ensure that the measurement work is efficient, comprehensive and safe.

[0048] Water data collection equipment includes a water drone equipped with a Zenmuse P1 camera (45 megapixels, supporting oblique photography) and a Zenmuse L1 LiDAR (240kHz scanning frequency, 240,000 points / second per echo, supporting three echoes). The drone is a DJI M350RTK, featuring high-precision positioning and stable flight performance. Flight parameters were set at an altitude of 80m, a heading overlap of 75%, a sideways overlap of 55%, and a ground sampling distance (GSD) of ≤2cm. This was used to acquire orthophotos and point cloud data for the water portion.

[0049] Auxiliary Equipment: Equipped with a CORS station, it provides a high-precision positioning reference for drone data collection, ensuring that data is uniformly aligned to the CGCS2000 coordinate system and the 1985 National Elevation Datum. This auxiliary CORS station plays a critical positioning support role in the entire seawall measurement system. By receiving satellite signals and fusing them with reference station data, the CORS station provides high-precision, real-time positioning information for surface drones and underwater data collection equipment. During surface drone data collection, the CORS station ensures the drone's positioning accuracy, ensuring that the collected orthophotos and point cloud data accurately reflect the seawall's actual location and shape. The CORS station is equally important for underwater data collection equipment, providing a precise positioning reference for survey vessels and underwater drones. This ensures that data collected by equipment such as multibeam echo sounders and side-scan sonars can be accurately aligned to a unified coordinate system, enabling seamless integration of underwater and surface data from the seawall and building a complete and accurate three-dimensional model of the seawall.

[0050] Underwater data acquisition equipment: including survey vessels equipped with multi-beam echo sounders, side-scan sonars, shallow-sediment profilers, underwater drones and micro-motion detectors to monitor the underwater portion of the seawall.

[0051] Multibeam echo sounder: Reson SeaBat T50P, with a frequency of 200kHz and a resolution of 0.75cm, is used to accurately measure underwater topography.

[0052] Side-scan sonar: EdgeTech 4200FS is selected, with a range of ≥150m and a resolution of ≤5cm, which can obtain detailed information on underwater terrain.

[0053] Shallow layer profiler: It has a penetration depth of ≥40m and a resolution of ≤2cm, and is used to detect underwater stratum structures.

[0054] Underwater drone: Equipped with a multi-beam device with a resolution of 0.75cm, it can perform full-coverage scanning of underwater areas, filling the measurement blind spots of conventional equipment in complex underwater areas.

[0055] Micro-seismic exploration instrument: frequency 10-200Hz, detection depth ≥30m, collected data is used to correct the thickness of the embankment fill layer.

[0056] Data processing module: processes the collected data, including point cloud data processing, posture correction, image processing and data registration;

[0057] Point cloud data processing: Lidar360 software is used to perform noise filtering using the DBSCAN algorithm, and outliers are removed using the 3σ criterion to ensure the accuracy of the point cloud data.

[0058] Attitude Correction: Hypack software is used to perform attitude correction on data collected by underwater measurement equipment, maintaining an attitude angle error of ≤0.02°. Multibeam echosounders, side-scan sonars, subsurface profilers, and multibeam equipment onboard underwater drones can experience attitude deviations in the data collected while operating underwater due to factors such as water currents and equipment movement. Hypack software is used to perform attitude correction on these data, maintaining an attitude angle error of ≤0.02° and ensuring the spatial accuracy of the measured data.

[0059] Image processing: Geometric distortion correction is performed on orthophotos based on the RPC model, with RMSE ≤ 1 pixel after correction; the NDVI index is used to remove vegetation coverage areas in orthophotos to improve image quality.

[0060] Data registration: Multi-source data are uniformly registered to the CGCS2000 coordinate system and the 1985 National Elevation Datum through the CORS station to ensure that the plane coordinate error is ≤±2cm.

[0061] 3D modeling module: perform point cloud registration and DEM model construction;

[0062] Point cloud registration: ICP algorithm is used to register multi-source point cloud data to ensure that the registration error is ≤±3cm.

[0063] DEM model construction: Based on the TIN algorithm, a DEM model with a grid spacing of ≤2m was constructed, with an elevation error of ≤±5cm. The levee was divided into four blocks: east, north, south, and west for block modeling, and block boundaries were marked to improve modeling accuracy in complex terrain.

[0064] Volume calculation module: Use the grid method to calculate the volume, set a 2m×2m grid, and use a specific formula to calculate the filling volume of each block.

[0065] Cross-validation: ArcGIS cut-and-fill analysis tools were used to cross-validate the volume extraction results of Tianbao RealWorks point cloud, with a tolerance of ±0.5m. 3 .

[0066] Error correction: For areas with slopes ≥15°, irregular triangulated networks are used to correct volume errors and improve the accuracy of volume calculations.

[0067] Meteorological data module: includes a weather instrument to obtain real-time meteorological data, including information such as rain, snow, wind and earthquakes.

[0068] Comprehensive analysis module: This module combines surface and underwater data acquisition equipment with meteorological data to promptly detect abnormalities, conduct a comprehensive assessment of the seawall, and predict potential problems and risks.

[0069] Early warning module: including alarm, which will issue an alarm in time when abnormal conditions or potential risks are detected;

[0070] The Result Output Module visualizes measurement results, presenting the seawall's DEM model and volume calculations in a 3D format, allowing users to intuitively understand the seawall's topography and engineering data. It also provides data analysis capabilities, such as comparing measurement data from different periods to analyze seawall deformation trends and performing statistical analysis on volume data for each block.

[0071] PLC control center: network connection with surface data acquisition equipment, underwater data acquisition equipment, meteorological data module, data collection module, route planning module, auxiliary equipment, data processing module, three-dimensional modeling module, volume calculation module, comprehensive analysis module and early warning module.

[0072] Furthermore, the data processing module processes the collected data, including point cloud data processing, posture correction, image processing and data registration; including the following steps:

[0073] 1. Point cloud data processing (tool: Lidar360):

[0074] 1. Noise filtering (DBSCAN algorithm)

[0075] 1.1: Import the original point cloud data into the Lidar360 software and select the "Point Cloud Processing" module.

[0076] 1.2: Enable DBSCAN filtering function and set parameters:

[0077] Neighborhood radius (Epsilon): Adjust according to the point cloud density (for example, 0.1-0.5 meters is recommended for vehicle-mounted point clouds, and 0.5-2 meters is recommended for airborne point clouds).

[0078] Minimum number of points (MinPts): Usually set to 3-5 points (needs to be greater than the minimum number of samples in the algorithm dimension space).

[0079] 1.3: Run the algorithm to automatically identify and mark noise points (outlier clusters), manually check the filtering results, and retain the main point cloud.

[0080] 2. Outlier removal (3σ criterion):

[0081] 2.1: Perform statistical analysis on the filtered point cloud in Lidar360 and calculate the mean (μ) and standard deviation (σ) of the point cloud in the X, Y, and Z axis directions.

[0082] 2.2: Set the threshold range: μ±3σ, and points outside the range are judged as outliers.

[0083] 2.3: Use the "Delete by Range" tool in the "Point Cloud Editing" function of the software to remove outliers and output the net point cloud data.

[0084] 2. Posture Correction (Tool: Hypack)

[0085] 1. Device data import

[0086] 1.1: Import the raw data (including attitude sensor data) from multibeam echo sounder, side scan sonar, shallow subsurface profiler or underwater drone into Hypack software.

[0087] 1.2. Match device timestamps to ensure that sensor data (such as compass and attitude angle) is synchronized with measurement data.

[0088] 2. Attitude error correction:

[0089] 2.1: In the Hypack "Sensor Calibration" module, select the corresponding device type (such as multibeam) and enter the device installation parameters (such as roll, pitch, and heading offsets).

[0090] 2.2: Enable the automatic correction function. The software calculates the attitude angle error compensation value based on the inertial navigation model (INS) and iteratively optimizes it until the attitude angle error is ≤ 0.02°.

[0091] 2.3: Generate the corrected data file, which contains the georeferenced coordinates and corrected attitude parameters.

[0092] 3. Image processing (based on RPC model and NDVI index):

[0093] 1. Geometric distortion correction (RPC model):

[0094] 1.1. Open the orthophoto in remote sensing processing software (such as ENVI / ArcGIS) and load the RPC parameters (rational polynomial coefficients) of the image.

[0095] 1.2. Select ≥20 ground control points (GCPs) evenly distributed throughout the image coverage area with sub-meter accuracy.

[0096] 1.3. Use the "RPC Orthorectification" tool to establish an error compensation model based on GCPs, calculate the resampled pixel coordinates, and output the corrected image. The root mean square error (RMSE) must be ≤ 1 pixel.

[0097] 2. Vegetation cover removal (NDVI index):

[0098] 2.1. Calculate the NDVI index in the image.

[0099] 2.2. Set the NDVI threshold (usually ≥0.2 for vegetation area) and generate vegetation mask through threshold segmentation.

[0100] 2.3. Use masking tools to remove vegetation-covered areas, or fill non-vegetation pixels through interpolation to improve the recognition of image objects.

[0101] 4. Data Registration

[0102] 1. Preparation for coordinate system conversion:

[0103] 1.1. Obtain real-time or post-processed differential data from the CORS station (Continuously Operating Reference Station), including the conversion parameters (seven parameters or three parameters) from WGS84 coordinates to CGCS2000.

[0104] 1.2. Organize the original coordinate systems (such as WGS84, local coordinate systems) and elevation benchmarks (such as Yellow Sea elevation) of multi-source data (point clouds, images, and measurement data).

[0105] 2. Plane coordinate registration:

[0106] 2.1. In the data processing software (such as GlobalMapper / QGIS), select the "Coordinate Conversion" function and enter the conversion parameters provided by the CORS station.

[0107] 2.2: Perform batch coordinate conversion on multi-source data to unify the plane coordinates to the CGCS2000 coordinate system, and control the plane error ≤±2cm.

[0108] 2.3: Verify the registration accuracy by superimposing checkpoints (ground landmarks with known coordinates). If the error exceeds the limit, re-optimize the transformation parameters.

[0109] 3. Unification of elevation benchmarks:

[0110] 3.1. Use the geoid model (e.g., EGM2008) provided by the CORS station to convert the original elevation data (e.g., GPS geodetic height) to the 1985 National Elevation Datum.

[0111] 3.2. Tide level correction is performed on underwater measurement data (such as multi-beam water depth), and the tidal influence is eliminated by combining the tide station data to ensure that the elevation accuracy meets the requirements.

[0112] 5. Quality control output:

[0113] For point cloud data, check point cloud density and noise residue, and output LAS / PLY format files.

[0114] For attitude correction data, a metadata file containing the corrected attitude angle is generated, with an error report attached.

[0115] For image data, output orthophotos without vegetation cover (TIFF / JPEG format), with RPC parameters and NDVI processing records.

[0116] The registered data are uniformly stored as vector / raster data in the CGCS2000 coordinate system, with a coordinate conversion log attached.

[0117] Furthermore, the 3D modeling module performs point cloud registration and DEM model construction, including the following steps:

[0118] 1. Point cloud registration: Use tools such as CloudCompare and TerraSolid to perform point cloud registration;

[0119] 1. Data preprocessing:

[0120] 1.1. Import multi-source point cloud data (such as drone LiDAR, ground station scanning) into the processing software, with the unified format being LAS / LAZ.

[0121] 1.2. Check the point cloud coordinate system and ensure that all data has been converted to the same coordinate system (such as CGCS2000).

[0122] 1.3: Downsample the preprocessed point cloud (such as voxel grid filtering) to reduce the amount of calculation while retaining feature points.

[0123] 2. Coarse registration (initial alignment):

[0124] 2.1. Select the "Point Cloud Registration" module in the software and load the point cloud data to be registered.

[0125] 2.2. Use feature extraction algorithms (such as ISS key point detection) to identify feature points in the point cloud (such as corner points and intersection points of seawall buildings).

[0126] 2.3 Select at least three pairs of points with the same name (such as ground landmarks, obvious ground features), complete the initial alignment, and roughly overlap the point clouds.

[0127] 3. Precise registration: Use the improved ICP algorithm for precise registration;

[0128] 3.1. Start the improved ICP (Iterative Closest Point) algorithm and set the parameters: Maximum number of iterations is 100-200 (to ensure convergence). Distance threshold is 0.1-0.5 meters (adjusted according to point cloud density). Convergence tolerance is 0.01 meters (stop when the difference in transformation matrix between two iterations is less than this value). The improved ICP algorithm model is:

[0129] Let the source point cloud be P = {pi} n i＝1 , the target point cloud is Q = {q j}m j＝1 , by iteratively solving the optimal rotation matrix R and translation vector t, the following objective function is minimized:

[0130] min R,t {(1 / n)∑ pi∈P ||pi-(R×q j +t)|| 2};q j ∈NN(p i ); where P is the source point cloud, pi is the i-th point in the source point cloud, and its coordinates are (x i ,y i ,z i )(3D space point). Q is the target point cloud (point cloud data used as reference), consisting of m points. j is the jth point in the target point cloud, with coordinates (x j ′,y j ′,z j ′)(three-dimensional space point). R is the rotation matrix (3×3 orthogonal matrix), which is used to describe the rotation transformation from the source point cloud to the target point cloud. t is the translation vector (3×1 vector), which is used to describe the translation transformation from the source point cloud to the target point cloud. NN(p i ) is the nearest neighbor of the source point pi in the target point cloud Q, that is, the point q closest to pi j ||·|| is the Euclidean distance (L2 norm), which is used to calculate the spatial distance between two points. min R,t Is the optimization goal, solve the rotation matrix R and translation vector t that minimize the objective function. Parameter settings:

[0131] Maximum number of iterations: 100 ≤ Iterations ≤ 200 (to ensure convergence);

[0132] Distance threshold: 0.1m ≤ distance threshold d thres ≤0.5m (adjusted according to point cloud density);

[0133] Convergence tolerance ∈: ∈≤0.01m.

[0134] 3.2. Run the ICP algorithm. The software automatically calculates the optimal rotation matrix (R) and translation vector (t) to minimize the distance between the source point cloud and the target point cloud.

[0135] 3.3. Check the registration results and observe the degree of fit of the overlapping areas using visualization tools (such as point cloud overlay display).

[0136] 4. Accuracy evaluation optimization:

[0137] 4.1. Calculate the registration error: Evenly select ≥10 checkpoints in the overlapping area and measure the difference in their coordinates between the two point clouds. The average error must be ≤±3cm and the maximum error must be ≤±5cm.

[0138] 4.2. If the error exceeds the limit, adjust the ICP parameters (such as reducing the distance threshold) or add coarse registration points and iterate again.

[0139] 2. DEM model construction: Use Global Mapper, ArcGIS and other tools to build DEM model:

[0140] 1. Data preprocessing and block planning:

[0141] 1.1. Divide the registered point cloud data into four blocks: east, north, south, and west according to the direction of the embankment, ensuring that the boundaries of each block are clear (such as turning points and structural changes).

[0142] 1.2. Classify the point cloud of each block and extract ground points (filter non-ground points such as buildings and vegetation).

[0143] 2. TIN Model Construction: Select the "Create TIN" function in the software to generate a triangulated irregular network (TIN) based on the ground points. Set the parameters: the maximum triangle side length is ≤ 2 meters (to control the grid density). Constraints include the embankment boundary and terrain feature lines (such as the crest line and the toe line) to ensure that the model conforms to the actual terrain.

[0144] 3. DEM Grid Generation: Generate a regular grid DEM from the TIN model. Set the grid spacing to ≤ 2 meters (e.g., 1 meter, 0.5 meter, adjust based on accuracy requirements). Use Inverse Distance Weighted (IDW) or Natural Neighbor interpolation to ensure smooth transitions. Repeat the above steps for each block to generate four sub-DEMs.

[0145] 4. Block splicing and boundary processing:

[0146] 4.1: Splice the four sub-DEMs into a complete DEM using the "Mosaic" or "Merge" function in the software.

[0147] 4.2: Processing block boundaries: Smooth the boundary areas (e.g., using a Gaussian filter) to eliminate splicing artifacts. Check the elevation continuity at the boundaries to ensure there are no obvious steps.

[0148] 5. Accuracy verification and correction:

[0149] 5.1. Use measured elevation points (such as RTK measurement data) to verify DEM accuracy. The requirements are: elevation error ≤ ±5cm (95% confidence interval). Maximum error ≤ ±10cm (outliers require investigation, such as incorrect point cloud classification).

[0150] 5.2: If the error exceeds the limit, reclassify the point cloud in the local area or adjust the TIN parameters and iterate the optimization.

[0151] 3. Output: Generate a 3D terrain model (e.g., OBJ, 3DS format), overlay orthophoto textures, and visually inspect terrain details. Output the registered overall point cloud (LAS / LAZ format) with a transformation matrix record file. Generate a registration error report containing checkpoint coordinates, error statistics tables, and visualization charts. Output a block DEM (GeoTIFF format) with metadata (coordinate system, grid spacing, etc.). Output a vector file (SHP format) marking the block boundaries, including polygons for the east, north, south, and west blocks.

[0152] Furthermore, the volume calculation module includes the following steps:

[0153] 1. Grid volume calculation steps

[0154] 1. Data preparation: Import the DEM model (e.g., GeoTIFF data in block format) and the design terrain data (e.g., CAD design drawings, BIM models, or target elevation surfaces). Ensure that the coordinate systems of the two types of data are consistent (e.g., both are CGCS2000) and cover the same area.

[0155] 2. Gridding: In computing software (e.g., Global Mapper, ArcGIS), set up a 2m x 2m square grid. Divide the computational area into a regular grid, with each grid node corresponding to the elevation of the DEM and the designed terrain. Output the grid node coordinates and the corresponding measured and designed elevations.

[0156] 3. Single grid volume calculation:

[0157] For each grid unit, the fill / cut height is calculated, and based on the fill / cut height and the grid area, the fill / cut volume of a single grid is calculated.

[0158] 4. Summarize volume by block: According to the four blocks of east, north, south and west divided in the early stage, count the fill / cut volume of all grids in each block respectively, and output the block volume report, including parameters such as fill / cut volume and total area of ​​each block.

[0159] 2. Cross-validation:

[0160] 1. Use ArcGIS tools for cut and fill analysis: Load the measured DEM and designed terrain surface into ArcGIS. Use the "Surface Difference Analysis" tool, set the calculation range to the entire area, and directly generate cut and fill volume statistics (including total volume and block volume).

[0161] 2. Use Trimble RealWorks tools to extract point cloud volume:

[0162] Import the registered point cloud data and the designed terrain model into RealWorks. Use the "Volume Calculation" module to calculate the cut and fill volumes based on the spatial relationship between the point cloud and the designed surface (ensure that the point cloud covers the entire area).

[0163] 3. Tolerance Comparison Verification: Compare the calculation results of the two tools; check the difference in volume of each block simultaneously. If the error in a block exceeds the limit, it needs to be marked as "area to be corrected." Output a cross-validation report, recording the results of each tool, the difference, and whether it passed the verification.

[0164] 3. Error correction (for slope ≥ 15°):

[0165] 1. Region Identification: Use the "Slope Analysis" tool in the DEM to extract steep slope areas with a slope of ≥15° (such as embankment slopes and sudden changes in terrain). Export the vector boundaries of the steep slope areas (in SHP format) as the target range for triangulated irregular network (TIN) correction.

[0166] 2. TIN model construction: In steep slope areas, a TIN model is generated based on the original point cloud data (not DEM grid data): terrain feature lines (such as embankment top line and slope foot line) and high-density point clouds are retained to ensure that the triangulation network fits the actual terrain.

[0167] The triangle side length is adjusted according to the point cloud density (usually ≤ 2m, consistent with the grid resolution).

[0168] 3. Volume correction calculation:

[0169] Recalculate cut and fill quantities for steep slope areas using the TIN model:

[0170] Compare the elevation difference between the TIN model and the designed terrain, and calculate the volume of each triangle. The volume is calculated according to the following formula: V = ∑ n i＝1 [(Z i -Z0)A i ]; calculate the filling volume of each block, where V is the total filling and cutting volume of the steep slope area; Z i is the measured elevation of the TIN model of the i-th triangle unit, which comes from the actual terrain elevation value of the triangle unit in the irregular triangulated network model generated from the original point cloud data. Z0 is the design terrain elevation corresponding to the i-th triangle unit, which is the design elevation value of the location determined according to the design drawing or the target elevation surface. i is the horizontal projection area of ​​the i-th triangular unit, the projected area of ​​the triangular unit on the horizontal plane (if the terrain has a slope, the slope area needs to be converted into the horizontal projection area for calculation). The corrected steep slope area volume is replaced with the original grid method result, and the fill and cut volume of the entire area is re-summed.

[0171] 4. Secondary verification: The corrected volume results were cross-validated again using ArcGIS and RealWorks, with a focus on checking whether the difference in the steep slope area was reduced to within the tolerance range (±0.5m 3 ).

[0172] If there is still an error, repeatedly adjust the TIN number (such as encrypting the triangulation nodes) until the accuracy requirements are met.

[0173] 4. Output: Output volume calculation report, visualization results and data files;

[0174] Volume calculation report: Contains cut and fill volumes for each block using the grid method and the TIN correction method, as well as a comparison table of total volume. Cross-validation results (differences between tools, and whether tolerance tests were passed).

[0175] Visualization results: Thematic map of cut and fill distribution (distinguishing between fill and excavation areas). Comparative visualization of the TIN model and the original DEM for the steep slope area.

[0176] Data files: Volume data of partition blocks (Excel format), modified TIN model (DXF or TIN format) and vector boundaries of steep slope areas.

[0177] Furthermore, the comprehensive analysis module combines surface data acquisition equipment (Zenmuse P1 camera and Zenmuse L1 lidar), underwater data acquisition equipment (multibeam echosounder, side-scan sonar, subsurface profiler, and micro-seismic survey instrument) and meteorological data to promptly detect anomalies, conduct a comprehensive assessment of the seawall, and predict potential problems and risks. The module includes the following steps:

[0178] 1. Data integration standardization:

[0179] 1.1. Unification of spatial benchmarks:

[0180] Coordinate system: Convert all data to a unified geographic coordinate system (such as WGS84 or National 2000 coordinate system) to ensure spatial alignment of surface, underwater and meteorological data.

[0181] Time synchronization: Time-calibrate multi-source data according to the acquisition timestamp and establish a time series database.

[0182] 1.2 Data format conversion and cleaning:

[0183] Format Processing: Optical images / LiDAR point clouds are converted to GeoTIFF, LAS / LAZ formats, with metadata (coordinates, time, and sensor parameters). Acoustic data (multibeam and side-scan sonar) are processed into XYZ format or grid data (e.g., ASCIIGrid) using software such as Hypack and QPS. Meteorological data is structured and stored as CSV or database tables (e.g., SQLite), with fields including time, parameter name, and value.

[0184] Perform outlier removal operations, repair missing values ​​through statistical analysis (such as the 3σ rule) or spatial interpolation, and eliminate obvious erroneous data (such as wild points in bathymetric data).

[0185] 2. Comprehensive analysis:

[0186] 2.1. Assessment of the current status of seawall structures:

[0187] 2.1.1. Water data applications:

[0188] Optical image analysis: Image processing algorithms (such as the NDVI vegetation index) are used to identify abnormal vegetation on the embankment slope (withered, sparse), and to determine whether there is a risk of soil exposure or leakage; image feature extraction technology (edge ​​detection, threshold segmentation) is used to identify obvious defects such as cracks and collapses in the embankment.

[0189] LiDAR analysis: Generates digital surface models (DSM) and digital terrain models (DTM), calculates embankment settlement and slope changes, and identifies local deformation areas (such as areas with settlement rates > 5 mm / year).

[0190] 2.1.2 Underwater Data Applications:

[0191] Fusion of multi-beam and side-scan sonar: superimposes water depth data and acoustic images to delineate the scour range of the seawall foundation (e.g., areas with scour pit depth > 2m and distance to the seawall foot < 10m), and analyzes the relationship between scour trends and water flow dynamics (e.g., tidal direction, wave energy).

[0192] Fusion of shallow subsurface and microseismic data: Combines stratigraphic profiles and shear wave velocity to assess foundation soil stability (e.g., liquefaction risk of loose sand layers, compressibility of soft soil layers), and identify potential sliding surfaces or areas with insufficient bearing capacity.

[0193] 2.2 Meteorological-hydrodynamic coupling analysis:

[0194] 2.2.1. Simulating Extreme Weather Impacts: Based on historical typhoon and storm surge data, a wave-tidal current-seawall interaction model (e.g., the SWAN wave model coupled with the Delft 3D hydrodynamic model) was developed to simulate seawall forces (e.g., wave runup, overtopping, and base stress) under different meteorological conditions. Simulation results were compared with measured data (e.g., LiDAR-monitored wave erosion on the seawall crest) to assess the seawall's resilience.

[0195] Set dynamic risk warning indicators: Establish multi-parameter association rules, such as:

[0196] When the wind speed is >20m / s and multi-beam monitoring detects a sudden drop in water depth in front of the dike (intensified scour), the "foundation scour risk" warning is triggered.

[0197] When the rainfall is greater than 50 mm / 24 h and the micro-motion monitoring detects a decrease in the shear wave velocity of the embankment soil (increase in water content), the "seepage damage risk" warning is triggered.

[0198] 3. Abnormal situation identification: abnormal situation identification;

[0199] Embankment cracks and settlement: Multi-temporal difference analysis and image semantic segmentation are used to identify embankment cracks and settlement anomalies. Key data indicators include: LiDAR point cloud ΔZ > warning threshold (e.g., 30 mm) and sudden changes in optical image features.

[0200] Foundation scour: Identify foundation scour anomalies by calculating underwater terrain change rate and estimating scour pit volume. Key data indicators include: multi-beam water depth deepening (ΔH>1m) and the appearance of new scour pits on side-scan sonar.

[0201] Soil permeability and liquefaction: Identify soil permeability and liquefaction anomalies through stratum structure comparison and soil dynamic parameter critical value determination methods. Key data indicators include: abnormal reflections (piping channels) in shallow stratum profiles and a shear wave velocity drop of >10%;

[0202] Vegetation anomalies: Identify vegetation anomalies through vegetation index time series analysis and thermal infrared image temperature anomalies; key data indicators include: NDVI values ​​below the healthy threshold and a sudden drop in vegetation coverage;

[0203] 4. Comprehensive assessment and risk prediction of seawalls: This system combines data from the Zenmuse P1 camera and Zenmuse L1 lidar, multibeam echosounder, side-scan sonar, subsurface profiler, micro-seismic survey instrument, and meteorological data to conduct comprehensive assessment and risk prediction of seawalls.

[0204] 4.1. Health status grading assessment:

[0205] 4.1.1 Construction of indicator system: Establish multi-level evaluation indicators, such as:

[0206] Structural safety: crack density, settlement rate, foundation scour depth.

[0207] Material properties: concrete strength (based on image spectrum inversion), soil density (micro-vibration wave velocity).

[0208] Environmental response: hydrodynamic loads and shoreline change rates under extreme weather conditions.

[0209] 4.1.2. Construct an evaluation model: Use the analytic hierarchy process (AHP) or fuzzy comprehensive evaluation method to perform weighted calculations on each indicator, classify the seawall status into three levels: normal, warning, and dangerous, and output a health score report. The comprehensive evaluation model is:

[0210] S=exp{-∑ n i＝1 [u i (t)X i / ó i ]};∑ n i＝1 [u i (t)] = 1; w i (t) = e -a△ti ;

[0211] u i (t) = [λ i w i (t)] / {∑ n j＝1 [w j (t)λ j ]}; where S is the comprehensive score of seawall health; w i is the spatiotemporal dynamic weight of the i-th indicator; X i is the standardized value of the i-th indicator. The original data is normalized by Z-score and the absolute value is taken to eliminate the dimension effect; i is the coefficient of variation of the i-th indicator; is the ratio of the standard deviation to the mean, reflecting the degree of fluctuation of the indicator data; exp() is the natural exponential function; u i (t) is the dynamic weight of the i-th indicator at time t; λ i is the static weight of the ith indicator, a fixed weight determined by AHP or entropy weight method (does not change over time); w i (t) is the time decay factor of the i-th indicator at time t; e is a natural constant, approximately equal to 2.71828, the base of the exponential function; a is the decay coefficient, which controls the rate of time decay. The larger the coefficient, the faster the influence of historical data disappears; △ti is the data timeliness, which is the difference between the current time and the data collection time of the i-th indicator.

[0212] 4.2 Potential Risk Prediction:

[0213] 4.2.1. Time Series Forecasting: Utilize machine learning algorithms (e.g., LSTM, random forest) to model historical data and predict scour trends, subsidence rates, and other parameters over the next 1-3 years. For example, based on multibeam data from the past 5 years, we can predict the evolution of the terrain in front of the dike and identify areas where deep troughs may form.

[0214] The risk prediction model is: Y^ t =∑ M i＝1 (C i Y t-i );C i =k(1-k) i-1 Where, Y^ t is the predicted value at a certain time t in the future; Y t-i is the data value of the historical period i (the corresponding time is ti, i = 1, 2, ..., m). i is the weight of the historical data of period i, which satisfies the characteristic that recent data has a higher weight. M is the number of historical data periods involved in the weighted calculation (i.e., the length of the time window). k is the smoothing coefficient (value range: 0 < α < 1), which controls the weight decay speed. i is the index;

[0215] 4.2.2 Disaster simulation: Combined with climate change scenarios (e.g., sea level rise of 0.5 m), simulate the response of seawalls under extreme weather conditions and predict high-risk areas (e.g., weak sections with inadequate moisture-proofing standards).

[0216] 5. Output: A three-dimensional geographic information system (3D GIS) platform dynamically displays the health status of seawalls, the distribution of abnormal areas, and risk prediction results. This includes thematic maps of seawall structural defects (with cracks and settlement areas marked), underwater terrain evolution animations (comparing multi-beam data from different periods), and a real-time warning dashboard (displaying meteorological parameters and current risk levels). This provides targeted recommendations to management: Immediately initiate reinforcement measures (e.g., riprap to protect the footing, grouting for anti-seepage) in areas designated as "dangerous." Develop an intensified monitoring plan (e.g., monthly underwater scanning) for areas designated as "warning." Optimize maintenance plans (e.g., pre-deploy anti-scour components) based on prediction results.

[0217] Furthermore, the route planning module develops a scientific and reasonable operation route for the water data collection equipment (surface drone) and underwater data collection equipment (survey ship and underwater drone) based on the geological data and hydrological data at the seawall, including the following steps:

[0218] 1. Data collection and integration: collect geological data, hydrological data and equipment parameters;

[0219] Geological data: Obtain rock and soil stratification (such as sand layer, clay layer), foundation depth, potential crack distribution, and slope stability assessment report along the seawall.

[0220] Hydrological data: Collect tidal patterns (high tide / low tide time, tidal range), water flow speed and direction (tidal current, runoff), wave height and period, and water depth distribution (multi-beam historical data) in the operating area.

[0221] Equipment parameters: Clarify the performance limitations of surface drones (endurance, wind resistance), survey vessels (draft, maneuverability), and underwater drones (diving depth, obstacle avoidance capability).

[0222] 2. Risk Area Identification: High-risk areas are marked based on geological data, such as soft soil foundation sections (prone to settlement), exposed bedrock sections (prone to erosion), and historical landslide sites. Hydrological data is also used to define restricted operating areas: strong current areas are defined as those with current velocities >2m / s for sampling, which may affect underwater drone control; shallow areas are defined as areas with water depths <0.5m, requiring survey vessels to avoid grounding. During typhoon / storm surge warnings, all water operations are suspended.

[0223] 3. Water drone route planning:

[0224] 3.1. Reference route design:

[0225] Longitudinal route: Fly parallel to the axis of the coastal dike, with a spacing of 50-100 meters (adjusted according to image resolution), covering the dike top, dike slope and 100 meters of water near the shore.

[0226] Horizontal route: perpendicular to the direction of the seawall, a cross route is set every 500 meters for three-dimensional modeling and terrain matching.

[0227] Intensification in Special Areas: In risky areas (such as those with dense cracks), flight spacing is reduced to 20 meters, and flights are increased to capture multi-angle imagery. Dynamic flight routes are used in complex terrain (such as jetties and curves), with the drone automatically adjusting its flight altitude based on real-time LiDAR point cloud data (maintaining a height of 10-50 meters above the ground).

[0228] 3.2. Flight parameter setting: Set flight parameters, including speed, altitude, shooting frequency and safety margin;

[0229] Speed ​​and altitude: The flight speed in the conventional section is 8-10m / s and the altitude is 100 meters; the speed in the encrypted section is reduced to 5m / s and the altitude is 50 meters.

[0230] Shooting frequency: The optical camera takes one picture every 2 seconds, and the LiDAR scans 200 lines per second to ensure that the data density meets the needs of later analysis.

[0231] Safety boundary: Set the horizontal distance from the seawall to ≥20 meters (to avoid collision) and the vertical distance from the water surface to ≥30 meters (to prevent wave splash interference).

[0232] 3.3 Dynamic Adjustment: The system uses real-time weather data. If wind speeds exceed level 6 (10.8-13.8 m / s), it automatically switches to low-altitude, low-speed mode (30 meters above sea level, 5 m / s). If wind speeds exceed level 8, the aircraft terminates operations and returns to base. Obstacle avoidance is performed by using onboard visual sensors to identify dynamic obstacles such as seagulls and cables, triggering a detour (±15 meters from the original route).

[0233] 4. Underwater data acquisition equipment route planning:

[0234] 4A. Survey vessel (multi-beam echo sounder, side-scan sonar):

[0235] 1. Full coverage of underwater terrain routes:

[0236] Main survey line: parallel to the direction of the seawall, with a spacing of 10-20 meters (the larger value is taken when the water depth is greater than 10 meters, and the spacing is increased to 5 meters when the water depth is less than 5 meters).

[0237] Check line: perpendicular to the main survey line, set one every five main survey lines for data accuracy verification (water depth error at intersection point must be less than 5cm).

[0238] Special area processing:

[0239] Scour pit / reef area: Use a spiral route (radius 5-10 meters) to scan 360° around the target area.

[0240] Estuary / sea estuary: Adjust the survey line according to the direction of the tide, advancing from the open sea to the embankment during high tide and in the opposite direction during low tide to reduce the impact of water flow on the survey vessel.

[0241] 2. Operation depth control:

[0242] Draft restriction: The minimum safe water depth of the survey vessel is the draft plus 1 meter. The route is dynamically adjusted based on real-time water depth data to avoid shallows (e.g. detour in areas with a water depth of less than 2 meters).

[0243] Safe distance for towing equipment: The side scan sonar towed fish should be kept 5-10 meters (for hard bottom) or 10-20 meters (for soft mud bottom) above the seabed to avoid damage from touching the bottom.

[0244] 3. Tidal collaborative operation:

[0245] During high tide: focus on measuring shallow water areas near the shore (areas with insufficient water depth are only accessible during high tide).

[0246] Low tide period: Move to deep water for operations, or perform equipment maintenance and data preprocessing.

[0247] 4B. Underwater UAVs (shallow layer profilers, micro-motion exploration instruments):

[0248] 1. Target-oriented route:

[0249] Vertical embankment route: Starting from the outside of the seawall foundation, radiating towards the deep sea at intervals of 5 meters, detecting the foundation rock and soil structure (such as loose layer thickness, bedrock interface).

[0250] Grid route: Lay out a 50m×50m grid in suspicious areas (such as abnormal areas in shallow subsurface profiles) and scan each grid to obtain high-resolution data.

[0251] Path optimization: A zigzag round-trip route is adopted to reduce turning time and improve operational efficiency (the endurance time for a single dive is approximately 2 hours).

[0252] 2. Diving and obstacle avoidance strategies:

[0253] Diving depth: adjusted according to water depth, normal operating depth is 10-50 meters, maximum diving depth ≤100 meters (equipment limit).

[0254] Obstacle avoidance mechanism: The forward-looking sonar detects obstacles ahead (such as fishing nets and sunken ships) in real time. When the distance is less than 10 meters, the aircraft automatically hovers and replans the route (the detour angle is ≥30°).

[0255] 3. Data synchronization and relay:

[0256] Every 30 minutes, it surfaces to 5 meters above the water surface, transmits data back to the mother ship via radio, and receives new route instructions.

[0257] Complex terrain areas (such as reefs): A "surface drone + underwater drone" collaborative mode is adopted, with the surface drone locating in real time and guiding the underwater drone to avoid obstacles.

[0258] 5. Multi-device collaboration and conflict avoidance:

[0259] 5.1. Space-time resource allocation:

[0260] Time staggering: Water drones are prioritized for operation during the early morning / evening low wind speed periods (6:00-10:00, 16:00-20:00); survey ships and underwater drones are concentrated on operations during the high tide period during the day, and data processing is carried out at night.

[0261] Spatial zoning: Delineate surface operating areas (50-200 meters above sea level) and underwater operating areas (0-100 meters below sea level) to avoid crossover of equipment routes. When multiple survey vessels are operating, maintain a horizontal spacing of ≥200 meters and a vertical spacing of ≥500 meters to prevent sonar signal interference.

[0262] 5.2 Emergency Response Mechanism: In the event of equipment loss, surface drones will automatically return home (signal loss > 30 seconds), and underwater drones will surface for recovery. In the event of a sudden hydrological change, if a sudden increase in water velocity is detected (e.g., exceeding the equipment safety threshold), all equipment will be immediately evacuated to a sheltered anchorage (≥ 2 km from the seawall).

[0263] 6. Route verification optimization:

[0264] 6.1. Preview Simulation: Use the GIS platform to simulate the route, import the 3D seawall model and water depth data, and visualize the equipment's flight / navigation trajectory to check for coverage blind spots (such as shadows at the dike foot) or safety hazards (such as routes crossing aquaculture areas). Simulation parameters are set at a wind speed of 5 m / s and a wave height of 1 m to verify the equipment's operational feasibility under typical operating conditions.

[0265] 6.2 Field Test Adjustments: After the maiden voyage, analyze data integrity. If insufficient swath overlap (<15%) is observed in the multibeam sounding, increase the route spacing (reduce spacing by 5 meters). Collect operator feedback and optimize route turning points based on actual operating experience (e.g., reduce sharp turns, curvature radius ≥ 50 meters) to reduce equipment energy consumption.

[0266] 6.3. Periodic iteration: The route database will be updated monthly based on the latest geological and hydrological data (such as changes in scour after typhoons), and key areas (such as embankment sections with scour rates greater than 1m / year) will be included in the dynamic monitoring list.

[0267] The present invention provides a seawall measurement method based on multi-source data fusion, comprising the following steps:

[0268] S1. The data collection module collects geological data and hydrological data at the seawall, including historical data, and annotates the data;

[0269] S2, the route planning module formulates a scientific and reasonable operation route for the above-water data acquisition equipment and the underwater data acquisition equipment based on the geological data and hydrological data at the seawall;

[0270] S3. The surface data acquisition equipment and the underwater data acquisition equipment monitor according to the planned route; the auxiliary equipment provides a high-precision positioning benchmark for the surface data acquisition equipment and the underwater data acquisition equipment; the meteorological data module obtains meteorological data in real time.

[0271] S4, the data processing module processes the collected data, including point cloud data processing, posture correction, image processing and data registration;

[0272] S5, 3D modeling module performs point cloud registration and DEM model construction;

[0273] S6. The volume calculation module uses the grid method to calculate the volume and the filling volume of each block.

[0274] S7, the comprehensive analysis module combines surface data acquisition equipment, underwater data acquisition equipment and meteorological data to promptly detect abnormal conditions, conduct a comprehensive assessment of the seawall, and predict potential problems and risks;

[0275] S8. When abnormal conditions or potential risks are detected, the early warning module will issue an alarm in time;

[0276] S9. The result output module displays the measurement results visually.

[0277] Example 1 (Qingdao Ruitai Shipbuilding Seawall Project):

[0278] Equipment deployment: 6 image control points (planar accuracy ±2cm, elevation accuracy ±3cm), measured by Hi-Target iRTK2 Pro.

[0279] Aerial flight: The drone performs terrestrial flight (80m altitude, 70% overlap), simultaneously collecting images and point clouds.

[0280] Data processing: DJI Aerial Triangulation (computation time ≤ 2 hours); Lidar360 point cloud classification (accuracy ≥ 95%);

[0281] Volume calculation: The volume calculation results are shown in Table 1 dam volume statistics (unit: m 3 ).

[0282] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A seawall measurement method based on multi-source data fusion, characterized in that: The following steps are involved: S1, the data collection module collects geological data and hydrological data at the seawall and labels the data; S2, the route planning module sets routes for the above-water data acquisition equipment and the underwater data acquisition equipment; S3, the surface data acquisition equipment and underwater data acquisition equipment monitor according to the planned route; the auxiliary equipment provides high-precision positioning reference; the meteorological data module obtains meteorological data in real time; S4, the data processing module processes the collected data, including point cloud data processing, posture correction, image processing and data registration; S5, 3D modeling module performs point cloud registration and DEM model construction; S6, the volume calculation module uses the grid method to calculate the volume and the filling volume of each block; S7, the comprehensive analysis module combines surface data acquisition equipment, underwater data acquisition equipment, and meteorological data to detect abnormal conditions, conduct a comprehensive assessment of the seawall, and predict potential problems and risks; S8. When abnormal conditions or potential risks are detected, the early warning module will issue an alarm in time; S9. The result output module displays the measurement results visually.

2. The seawall measurement method based on multi-source data fusion according to claim 1 is characterized in that: Step S4 includes the following steps: S41. Point cloud data processing: including noise filtering and outlier removal; S42, attitude correction: import device data and perform attitude error correction; generate a corrected data file containing geographic reference coordinates and corrected attitude parameters; S43, Image processing: including geometric distortion correction and vegetation cover removal: S44, data registration: including coordinate system conversion, plane coordinate registration and elevation benchmark unification; S45. Quality control output: For point cloud data, check the point cloud density and noise residue, and output LAS / PLY format files; for attitude correction data, generate metadata files containing the corrected attitude angles; for image data, output orthophotos without vegetation cover; and store the registered data in a unified manner.

3. The seawall measurement method based on multi-source data fusion according to claim 1 is characterized in that: Step S5 includes the following steps: S51. Point cloud registration: Use CloudCompare tool to perform point cloud registration; S52, DEM model construction: Use Global Mapper tool to construct DEM model; S53. Output: Generate a 3D terrain model, overlay orthophoto textures, visually inspect terrain details, output the overall point cloud after registration, and generate a registration error report.

4. The seawall measurement method based on multi-source data fusion according to claim 3 is characterized in that: Step S51 includes the following steps: S511, data preprocessing: import multi-source point cloud data, check the point cloud coordinate system; downsample the preprocessed point cloud; S512, coarse registration: load the point cloud data to be registered, use the ISS key point detection feature extraction algorithm to identify the feature points in the point cloud; select multiple pairs of points with the same name to complete the initial alignment; S513, fine registration: use the improved ICP algorithm for fine registration; S514, precision evaluation and optimization: calculate the registration error. If the error exceeds the limit, adjust the ICP parameters or add coarse registration points and iterate again.

5. The seawall measurement method based on multi-source data fusion according to claim 3 is characterized in that: Step S52 includes the following steps: S521. Data preprocessing and block planning: Divide the registered point cloud data into four blocks: east, north, south, and west, according to the direction of the embankment; classify the point cloud of each block and extract ground points; S522, TIN model construction: Generate irregular triangulated network TIN model based on ground points; S523, DEM grid generation: Generate regular grid DEM from TIN model interpolation; S524, block stitching and boundary processing: stitch the four sub-DEMs into a complete DEM and smooth the boundary areas; S525, Accuracy Verification and Correction: Use measured elevation points to verify DEM accuracy; if the error exceeds the limit, reclassify the point cloud or adjust TIN parameters for the local area, and iterate and optimize.

6. The seawall measurement method based on multi-source data fusion according to claim 1, characterized in that: Step S6 includes the following steps: S61. Grid volume calculation: Divide the calculation area into a regular grid, with each grid node corresponding to the elevation value of the DEM and the designed terrain. Calculate the fill and cut height for each grid cell. Based on the fill and cut height and the grid area, calculate the fill and cut volume for each grid cell. Count the fill / cut volume of all grid cells within the four divided blocks (east, north, south, and west). S62, Cross-validation: Calculate cut and fill volumes using ArcGIS cut and fill analysis and Trimble RealWorks point cloud volume extraction, compare total volume and block volume, and mark any excess areas for correction. S63, Error Correction: Extract the vector boundaries of steep slope areas with a slope of ≥15° from the DEM, generate a TIN model in this area based on the original point cloud data, and calculate the cut and fill volume; replace the original grid method results with the TIN model results, and then cross-validate using ArcGIS and RealWorks; S64. Output of results: Output volume calculation report, visualization results and data files.

7. The seawall measurement method based on multi-source data fusion according to claim 1, characterized in that: Step S7 includes the following steps: S71. Data integration standardization: including data format conversion and cleaning, spatial reference unification, and time synchronization; S72. Conduct comprehensive analysis of seawalls; S73, abnormal situation identification: abnormal situation identification; S74, Comprehensive Seawall Assessment and Risk Prediction: Combines Zenmuse P1 and Zenmuse L1 LiDAR data, multibeam echosounder, side-scan sonar, subsurface profiler, micro-seismic survey instrument data, and meteorological data to conduct comprehensive seawall assessment and risk prediction. S75. Output: Dynamically display the health status of the seawall, distribution of abnormal areas and risk prediction results through the three-dimensional geographic information system platform.

8. The seawall measurement method based on multi-source data fusion according to claim 6, characterized in that: Step S72 includes the following steps: S72.

1. Assessment of the current status of seawall structures: Assess the current status of seawall structures by combining surface and underwater data; S72.1.

1. Maritime data applications: Optical image analysis: Use image processing algorithms to identify abnormal vegetation on embankment slopes and determine whether there is soil exposure or leakage risk; use image feature extraction technology to identify embankment cracks and collapse hazards; LiDAR analysis: Generate digital surface models (DSM) and digital terrain models (DTM), calculate embankment settlement and slope changes, and identify local deformation areas; S72.1.2 Underwater Data Applications: Multi-beam and side-scan sonar fusion: superimpose water depth data and acoustic images to delineate the scour range of seawall foundations and analyze the relationship between scour trends and water flow dynamics; Fusion of shallow subsurface and microseismic data: Combines stratigraphic profiles and shear wave velocity to assess foundation soil stability and identify potential sliding surfaces or areas with insufficient bearing capacity. S72.

2. Meteorological-hydrodynamic coupling analysis: Simulate the impact of extreme weather. Based on historical typhoon and storm surge data, establish a wave-tide-seawall interaction model to simulate the forces acting on the seawall under different meteorological conditions. Compare the simulation results with the measured data to evaluate the seawall's disaster resistance.

9. The seawall measurement method based on multi-source data fusion according to claim 7, characterized in that: Step S74 includes the following steps: S74.

1. Health status grading assessment: Establish multi-level assessment indicators and construct an assessment model. Use the analytic hierarchy process (AHP) to perform weighted calculations on each indicator and classify the seawall status into three levels: normal, warning, and dangerous. Output a health score report. The comprehensive assessment model is: S=exp{-∑ n i＝1 [u i (t)X i / ó i ]};∑ n i＝1 [u i (t)]=1;w i (t)=e -a△ti ; u i (t) = [λ i w i (t)] / {∑ n j＝1 [w j (t)λ j ]}; where S is the comprehensive score of seawall health; w i is the spatiotemporal dynamic weight of the i-th indicator; X i is the standardized value of the i-th indicator; i is the coefficient of variation of the i-th indicator; exp() is the natural exponential function; u i (t) is the dynamic weight of the i-th indicator at time t; λ i is the static weight of the i-th indicator; w i (t) is the time decay factor of the ith indicator at time t; e is a natural constant; a is the decay coefficient; △ti is the difference between the current time and the data collection time of the ith indicator; S74.

2. Potential risk prediction: Use machine learning algorithms to model historical data and predict scour trends and settlement rate parameters for the next 1-3 years; conduct disaster simulations and, combined with climate change scenarios, simulate the response of seawalls under extreme weather conditions and predict high-risk areas.

10. A seawall measurement system based on multi-source data fusion, comprising a PLC control center, surface data acquisition equipment, underwater data acquisition equipment, a meteorological data module, a data collection module, a route planning module, auxiliary equipment, a data processing module, a 3D modeling module, a volume calculation module, a comprehensive analysis module, and an early warning module; characterized by: Data collection module: collects geological and hydrological data at the seawall and labels the data; Route planning module: formulates routes for surface data acquisition equipment and underwater data acquisition equipment; Water data collection equipment: including a water drone equipped with a Zenmuse P1 camera and a Zenmuse L1 LiDAR to obtain orthophotos and point cloud data of the water surface. Auxiliary equipment: equipped with a CORS station to provide high-precision positioning reference for drones and survey ships; Underwater data collection equipment: including survey vessels equipped with multi-beam echo sounders, side-scan sonars, subsurface profilers, underwater drones, and micro-motion survey instruments to monitor seawalls; Data processing module: processes the collected data, including point cloud data processing, posture correction, image processing and data registration; 3D modeling module: perform point cloud registration and DEM model construction; Volume calculation module: Use grid method to calculate volume and fill volume of each block; Meteorological data module: including meteorological instrument, real-time acquisition of meteorological data; Comprehensive analysis module: This module combines surface and underwater data acquisition equipment with meteorological data to promptly detect abnormalities, conduct a comprehensive assessment of the seawall, and predict potential problems and risks. Early warning module: When abnormal conditions or potential risks are detected, timely alarms are issued; Result output module: visually display the measurement results; PLC control center: network connection with surface data acquisition equipment, underwater data acquisition equipment, meteorological data module, data collection module, route planning module, auxiliary equipment, data processing module, three-dimensional modeling module, volume calculation module, comprehensive analysis module and early warning module.

Images