Cofferdam filling method based on unmanned ship underwater measurement and unmanned aerial vehicle oblique photography

Abstract

This invention discloses a cofferdam construction method based on underwater surveying by unmanned surface vessels (USVs) and oblique photography by unmanned aerial vehicles (UAVs), comprising the following steps: 1) selection of USVs and UAVs; 2) unified aerial-water collaborative topographic mapping of the entire area; 3) generation of an integrated aerial-water three-dimensional topographic model of the cofferdam construction area; 4) optimization of the topographic model; 5) layered filling construction with real-time dynamic monitoring; and 6) slope trimming and quality acceptance. This application solves the problems of low measurement efficiency, insufficient data accuracy, and high safety risks by selecting USVs and UAVs, conducting unified aerial-water collaborative topographic mapping of the entire area, generating an integrated aerial-water three-dimensional topographic model of the cofferdam construction area, optimizing the topographic model, carrying out layered filling construction with real-time dynamic monitoring, and performing slope trimming and quality acceptance. It has the advantages of improving surveying efficiency and accuracy, achieving full-process data fusion, thereby improving filling quality and construction efficiency.

Description

Technical Field

[0001] This invention relates to a method for constructing a cofferdam, and more particularly to a method for constructing a cofferdam based on underwater surveying by an unmanned vessel and oblique photography by an unmanned aerial vehicle. Background Technology

[0002] With the increasing number of water conservancy projects in national engineering construction, cofferdams, as temporary water-retaining structures for water conservancy hubs, river management, port terminals, and other projects, directly determine the safety and construction period of the main project through their filling quality and construction efficiency.

[0003] Traditional cofferdam construction often involves manually navigating boats and using plumb bobs and single-beam echo sounders to measure underwater topography point by point, while the shoreline and the surface of the fill body rely on manual RTK station marking. This traditional construction method has three major drawbacks: First, the measurement efficiency is extremely low, requiring a large investment of manpower and equipment for a single large-area measurement, and is greatly affected by the flood season; second, the data accuracy is insufficient, with manual measurement and low coverage density, making it easy to miss shallows and ditches in the underwater topography, and the surface cannot reflect micro-topographic undulations, resulting in inaccurate calculation of fill volume and large deviations in boundary control; third, the safety risks are prominent, with manual wading operations in shallow water and fast-flowing areas prone to accidents such as capsizing and falling into the water.

[0004] With technological advancements, UAV oblique photogrammetry and unmanned surface vessel (USV) underwater surveying are increasingly being applied in engineering surveying. UAV oblique photogrammetry can quickly acquire large-scale 3D models of the water surface, while USVs equipped with multibeam sonar can scan underwater terrain with high precision. However, current technologies mostly involve single devices operating independently, failing to form an integrated "air-water" collaborative surveying system. Furthermore, the survey data is only used for preliminary topographic mapping and is not deeply integrated into the entire process of cofferdam construction, including scheme optimization, process control, dynamic monitoring, and acceptance verification. This results in the data's value not being fully realized, and problems such as blind construction, frequent overfilling / underfilling, and lagging quality control still exist. Summary of the Invention

[0005] The purpose of this invention is to provide a cofferdam construction method based on underwater surveying by unmanned vessel and oblique photography by drone, which improves surveying efficiency and accuracy, reduces safety risks, and achieves full-process data fusion, thereby improving filling quality and construction efficiency.

[0006] The objective of this invention is achieved through the following technical solution: a cofferdam construction method based on underwater measurement by unmanned surface vessels and oblique photography by unmanned aerial vehicles, the method comprising the following steps: 1) Selection of unmanned ships and drones; 2) Unified topographic mapping of the entire region, combining air and water surveying; 3) Generate an integrated three-dimensional terrain model of the entire air-water area of ​​the cofferdam construction zone; 4) Optimize the terrain model; 5) Layered filling construction with real-time dynamic monitoring; 6) Acceptance of slope trimming and shaping quality.

[0007] In step 1), the unmanned vessel is equipped with a single / multi-beam receiver, a global navigation satellite system, an attitude sensor, and a wireless communication module; the unmanned aerial vehicle (UAV) is a multi-rotor oblique photography UAV, equipped with a high-definition camera, a global navigation satellite system, and RTK.

[0008] To further describe, in step 2), the area affected by the filling is the UAV mapping area, and the underwater cofferdam filling area and the near-shore shallow water area are the UAV mapping area, with the cofferdam axis as the boundary. When performing oblique photography with UAVs, select sunny, windless or light-wind periods, plan a five-directional aerial photography route combining orthogonal and oblique views, set the flight altitude to 80-150m, the directional overlap rate to 80%, and the lateral overlap rate to 75%, autonomously fly to collect high-definition images, process them with software to construct a real-world 3D model of the water area, and extract data such as shoreline boundaries, original ground elevation, and terrain slope. When conducting underwater surveying operations, the unmanned vessel plans survey lines parallel to the cofferdam axis with a spacing of 5 to 10 meters. The unmanned vessel autonomously navigates and cruises, continuously scanning the underwater terrain with multibeam sonar, and simultaneously collecting water depth and bottom point cloud data. After attitude correction and coordinate transformation, a high-precision underwater digital elevation model is generated to identify the distribution of dam foundation trenches and shoals.

[0009] In step 3), the above-water 3D model and the underwater high-precision digital elevation model are unified into the same engineering coordinate system. By matching feature points at the junction, they are integrated to form a three-dimensional terrain model of the entire air-water integrated area of ​​the cofferdam construction area. The original earthwork volume is accurately calculated, providing data support for the filling scheme design.

[0010] Further, in step 4), the cofferdam axis is divided into several filling sections, each section is filled in layers, the starting and ending elevations of filling, slope trimming parameters, earthwork sources and transportation routes of each section are clearly defined, and the construction schedule and key points of quality control are determined.

[0011] Step 5) further includes the following steps: (1) Layered filling: dump trucks are used for unloading, excavators are used for filling, bulldozers are used for leveling, and road rollers are used for compaction in layers. After each layer is filled, "air-water" coordinated monitoring is carried out immediately. (2) Rapid inspection by drone: oblique photography of the surface of the filled layer is carried out to generate a three-dimensional model of the surface, compare it with the design elevation, detect the surface flatness and slope gradient, identify the underfilled area, overfilled area and local collapse hazards, and generate a deviation distribution map. (3) Underwater verification by unmanned vessel: The underwater cofferdam filling area and the underwater part of the front edge of the filling body are re-measured to monitor the settlement of the cofferdam filling area and the scouring of the front edge of the filling body. The data are compared with the previous data to analyze the impact of underwater topographic changes on the stability of the cofferdam. (4) Real-time data transmission and analysis: The monitoring data from UAVs and unmanned vessels are processed and compared with the design to generate a monitoring report; when the monitoring data exceeds the preset deviation threshold, an early warning is issued immediately to guide the on-site construction personnel to adjust the filling parameters.

[0012] Further, step 6) also includes the following steps: (1) After the cofferdam is filled to the design elevation, the overall slope is trimmed; based on the slope tilt photography data obtained by UAV tilt photography, the slope is trimmed by excavator and manual labor to ensure that the slope and flatness meet the design requirements. (2) Full-area completion survey: Re-launch UAV and unmanned boat to carry out full-coverage surveying, construct an integrated air-water three-dimensional model after the cofferdam is formed, and extract completion data such as the top elevation of the cofferdam, slope gradient, axis position, and filling volume; (3) Quality acceptance and data archiving: Compare the as-built 3D model and measurement data with the design model, verify various quality indicators, and automatically generate an acceptance report; (4) Organize and archive the measurement data, model files, monitoring records and acceptance data of the entire construction process to form a digital construction archive.

[0013] By adopting the above-mentioned technical solutions, the cofferdam filling method based on unmanned vessel underwater measurement and UAV oblique photography provided in this application solves the problems of low measurement efficiency, insufficient data accuracy and high safety risks in traditional methods. It has the advantages of improving surveying efficiency and accuracy, reducing safety risks, and realizing full-process data fusion, thereby improving filling quality and construction efficiency. Attached Figure Description

[0014] The accompanying drawings of this invention are described below: Figure 1 : A schematic diagram of the aerial photography range of unmanned surface vessels and drones; Figure 2 Schematic diagram of drone flight path planning; Figure 3 Schematic diagram of unmanned vessel route planning; Figure 4 : Raw data map of unmanned surface vessel surveying; Figure 5: An integrated air-water three-dimensional terrain model for unmanned ships and drones. Detailed Implementation

[0015] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings. However, the present invention is not limited to these embodiments. Any improvements or substitutions based on the basic spirit of these embodiments shall still fall within the scope of protection claimed by the claims of the present invention.

[0016] Example 1: As Figure 1 , 2 As shown in Figures 3, 4, and 5, a cofferdam construction method based on underwater measurement by unmanned surface vessels and oblique photography by unmanned aerial vehicles (UAVs) includes the following steps: 1) Selection of unmanned ships and drones; 2) Unified topographic mapping of the entire region, combining air and water surveying; 3) Generate an integrated three-dimensional terrain model of the entire air-water area of ​​the cofferdam construction zone; 4) Optimize the terrain model; 5) Layered filling construction with real-time dynamic monitoring; 6) Acceptance of slope trimming and shaping quality.

[0017] In implementing this method, the first step is to select between unmanned surface vessels (USVs) and unmanned aerial vehicles (UAVs). USV selection can be based on a manual assessment of its payload capacity, endurance, and seaworthiness in specific waters. UAV selection can be based on a manual assessment of its flight altitude, payload capacity, and compatibility with camera interfaces. Before operation, lens distortion correction and GPS static calibration are performed on the UAVs; sonar depth sounding accuracy calibration and attitude error compensation are conducted on the USVs to ensure the equipment's measurement accuracy meets requirements. The UAV oblique photogrammetry 3D model has a planar error ≤ ±3cm and an elevation error ≤ ±5cm, while the USV's underwater topographic measurement accuracy reaches 1cm + 0.1% × water depth.

[0018] In step 1), the unmanned vessel is equipped with a single / multi-beam receiver, a global navigation satellite system, an attitude sensor, and a wireless communication module; the unmanned aerial vehicle (UAV) is a multi-rotor oblique photography UAV, which is equipped with a high-definition camera, a global navigation satellite system, and RTK.

[0019] Specifically, the unmanned surface vessel (USV) is an unmanned surface vehicle capable of autonomous navigation on the water surface and carrying measurement equipment. Its implementation can include a catamaran design for better stability and adaptability to complex aquatic environments; or a monohull design to enhance maneuverability and facilitate operation in narrow or obstacle-filled areas. The single / multi-beammeter is an acoustic device used for underwater topographic surveying. A single-beammeter measures water depth directly below by transmitting and receiving a single sound beam, suitable for linear measurements or preliminary surveys; a multi-beammeter transmits a fan-shaped array of sound beams, covering a wider underwater area at once, enabling the acquisition of high-density underwater 3D point cloud data, suitable for detailed underwater topographic modeling. The global navigation satellite system (GNSS) provides accurate geographic location information. It can employ a GPS receiver or a GNSS receiver supporting multiple constellations (such as GPS, GLONASS, BeiDou, and Galileo) to improve positioning accuracy and reliability. The attitude sensor monitors the USV's attitude parameters, such as roll, pitch, and heading, in real time. It can employ an inertial measurement unit (IMU) to sense hull motion by integrating accelerometers and gyroscopes; or a motion reference unit (MRU), which typically includes more advanced sensors and algorithms to provide more stable attitude data and perform vertical compensation. The wireless communication module is used to enable data transmission and command reception between the unmanned surface vessel (USV) and the shore-based control station. It can employ a long-range radio module, such as a UHF / VHF band data radio, to ensure stable communication over long distances; or a cellular network / satellite communication module to achieve wider-range data transmission and remote control.

[0020] The aforementioned drone is an unmanned aerial vehicle capable of autonomous flight. The multi-rotor oblique photography drone is a drone equipped with multiple rotors (e.g., quadcopter, hexacopter, or octocopter) specifically designed for oblique photography tasks. It can utilize a commercially available industrial-grade multi-rotor platform, incorporating a professional camera and flight control system to achieve oblique photography functionality; or it can employ a custom-designed, high-payload multi-rotor drone to accommodate heavier camera payloads and longer flight times. The high-definition camera is used to acquire high-resolution image data. It can employ a professional-grade digital SLR camera (DSLR) or mirrorless camera, equipped with a large-format sensor to acquire high-quality images; or it can employ an industrial-grade aerial mapping camera with high resolution and global shutter capabilities to meet high-precision mapping requirements. The Global Navigation Satellite System (GNSS) provides the drone with precise location information. It can employ a standard GPS module integrated into the flight controller; or it can employ a dual-frequency GNSS receiver to improve positioning accuracy and anti-interference capabilities. The RTK (Real-Time Kinematic) technology is a real-time differential positioning technology capable of providing centimeter-level or even millimeter-level positioning accuracy. The implementation can include sending differential correction data to the UAV via a ground base station, or obtaining correction data from a virtual reference station via a network RTK service, thereby significantly improving the geolocation accuracy of UAV imagery.

[0021] Through the aforementioned technical solutions, this application addresses the accuracy and efficiency issues in collaborative surveying by specifying the equipment configuration of unmanned surface vessels (USVs) and unmanned aerial vehicles (UAVs), ensuring seamless integration of air-water data. The single / multi-beam scanners mounted on the USVs can accurately scan underwater topography, effectively avoiding missed measurements of shallows or ditches; the Global Navigation Satellite System (GNSS) provides high-precision positioning, helping to unify the coordinate system of air-water measurement data; attitude sensors correct the vessel's attitude in real time, significantly reducing underwater measurement errors; and the wireless communication module enables real-time data transmission and collaborative control, improving operational efficiency. Simultaneously, the use of multi-rotor oblique photography UAVs, with their stability and flexibility, allows them to adapt to flight operations in complex environments; high-resolution cameras acquire high-resolution images, providing a foundation for constructing accurate 3D models of the water; and the GNSS combined with RTK technology provides centimeter-level high-precision positioning for the UAVs, greatly enhancing the accuracy and reliability of water-based surveying. The collaborative work of these devices not only improves the overall efficiency and data quality of the surveying but also reduces the need for manual intervention, thereby effectively reducing safety risks during operations and providing a solid data foundation for subsequent optimization of cofferdam filling schemes and construction control.

[0022] To further describe, in step 2), the area affected by the filling is the UAV mapping area, and the underwater cofferdam filling area and the near-shore shallow water area are the UAV mapping area, with the cofferdam axis as the boundary. When performing oblique photography with UAVs, select sunny, windless or light-wind periods, plan a five-directional aerial photography route combining orthogonal and oblique views, set the flight altitude to 80-150m, the directional overlap rate to 80%, and the lateral overlap rate to 75%, autonomously fly to collect high-definition images, process them with software to construct a real-world 3D model of the water area, and extract data such as shoreline boundaries, original ground elevation, and terrain slope. When conducting underwater surveying operations, the unmanned vessel plans survey lines parallel to the cofferdam axis with a spacing of 5 to 10 meters. The unmanned vessel autonomously navigates and cruises, continuously scanning the underwater terrain with multibeam sonar, and simultaneously collecting water depth and bottom point cloud data. After attitude correction and coordinate transformation, a high-precision underwater digital elevation model is generated to identify the distribution of dam foundation trenches and shoals.

[0023] Specifically, the area affected by the cofferdam construction is designated as the UAV mapping area, while the underwater cofferdam construction area and near-shore shallow water area are designated as the UAV mapping area. This technical feature aims to clarify the regional division strategy for air-water collaborative mapping, ensuring comprehensive coverage and optimized resource allocation. The cofferdam axis, as the core baseline for engineering construction, clearly and logically separates the above-water and underwater areas. The impact area typically refers to land and water-land interface areas that may be affected or require monitoring during cofferdam construction. This area has complex terrain and significant elevation changes, making it suitable for large-scale, high-efficiency oblique photogrammetry by UAVs. However, the underwater cofferdam construction area and near-shore shallow water area are obscured by water, making direct mapping by traditional UAVs impossible. Furthermore, the water depth varies drastically, requiring high measurement accuracy. Therefore, mapping by UAVs equipped with specialized underwater surveying equipment is more suitable. For example, based on the design width of the cofferdam and the construction safety distance, the land area within a certain range on both sides of the cofferdam axis can be designated as the UAV mapping area; at the same time, the area below the cofferdam axis and extending into the water, including possible shoals and underwater channels, can be designated as the UAV mapping area.

[0024] When conducting oblique photography operations using drones, selecting sunny, windless, or light-wind periods is crucial. This technical characteristic specifies the environmental conditions for such operations, aiming to ensure image data quality and flight safety. Sunny days provide ample illumination, avoiding interference from shadows and clouds on image color and texture, thus guaranteeing image clarity. Windless or light-wind periods effectively reduce drone attitude jitter during flight, lowering the risk of image blur and distortion, while also facilitating stable aircraft control and maintaining flight path accuracy. For example, before the operation, weather forecast information for the next few hours or days can be obtained through a meteorological forecasting system, allowing for the selection of a period with wind speeds below level three, no precipitation, and good visibility. Alternatively, the drone's built-in wind speed sensor can monitor wind speed in real time, automatically pausing or adjusting the flight plan when the wind speed exceeds a preset threshold (wind speed > 5 m / s).

[0025] The planning of orthophoto + oblique five-axis aerial photography flight paths describes the method of planning UAV flight paths, aiming to acquire multi-angle, all-round ground feature imagery data. An orthophoto flight path typically refers to the UAV camera lens shooting vertically downwards to obtain a vertical view of the ground features, used to generate a high-precision orthophoto map. An oblique five-axis aerial photography flight path adds four lateral (front, back, left, right) oblique shots to the orthophoto flight path, ensuring that each ground feature is captured from multiple angles. For example, professional flight path planning software can be used to automatically generate a grid-like orthophoto flight path within the survey area. Based on this, instructions can be set for the camera to tilt in four lateral directions at each waypoint or at regular intervals, ensuring complete image coverage.

[0026] Setting the flight altitude to 80–150 m, the forward overlap rate to 80%, and the lateral overlap rate to 75% defines the key parameters for UAV oblique photogrammetry, aiming to balance the efficiency, accuracy, and completeness of data acquisition. The flight altitude of 80–150 m is a relatively moderate range, ensuring the coverage of a single image, improving operational efficiency, while also ensuring that the ground resolution of the image meets the accuracy requirements for 3D model construction. The forward overlap rate of 80% and the lateral overlap rate of 75% ensure sufficient overlap between adjacent images, which is the basis for image matching, feature extraction, and 3D reconstruction. For example, in flight path planning software, by inputting parameters such as the expected ground resolution, camera focal length, and sensor size, the software will automatically calculate the flight altitude that meets the overlap rate requirements.

[0027] Autonomous flight for high-definition image acquisition emphasizes the automation and data quality of UAV operations. Autonomous flight refers to the UAV completing the entire flight and image acquisition process according to preset routes and parameters without human intervention, significantly improving operational efficiency and safety. High-definition imagery refers to images with high resolution and clarity, capable of capturing subtle features of ground objects, providing high-quality raw data for subsequent 3D model construction and data extraction. For example, UAV systems are typically equipped with high-precision GNSS modules and inertial measurement units (IMUs), combined with flight control algorithms, to achieve centimeter-level or sub-meter-level autonomous navigation and positioning, ensuring the accuracy of the flight path.

[0028] The technical feature of constructing a realistic 3D model of a water area through software processing describes the processing flow and output of data collected by drones. Software processing typically refers to using professional photogrammetry or 3D modeling software to perform a series of processing steps on the acquired high-definition images, such as aerial triangulation, dense point cloud generation, mesh construction, and texture mapping, ultimately generating a realistic 3D model of the water area with realistic textures and geometry. For example, commercial software such as ContextCapture, Pix4Dmapper, and DJI Terra can be used for processing.

[0029] Extracting data such as shoreline boundaries, original ground elevation, and terrain slope is a key feature that identifies crucial geographic information extracted from a realistic 3D model. Shoreline boundaries pinpoint the precise location of the land-water interface, crucial for the planning and construction of cofferdam projects. Original ground elevation data provides fundamental information reflecting surface undulations and can be used to calculate earthwork volume and analyze terrain features. Terrain slope data reflects the steepness of the surface, providing a basis for selecting construction machinery and developing construction plans. For example, shoreline boundaries can be manually or semi-automatically delineated using tools in 3D modeling software.

[0030] When conducting underwater surveying operations using unmanned surface vessels (USVs), planning survey lines parallel to the cofferdam's axis is a key technical feature that defines the principles for route planning in underwater surveying, aiming to ensure the systematic nature and integrity of underwater topographic data. Planning survey lines parallel to the cofferdam's axis ensures that the survey lines align with the cofferdam's direction, facilitating subsequent alignment and analysis of data with the cofferdam's design axis. For example, by inputting the coordinate information of the cofferdam's axis into the USV control software, the system can automatically generate a series of survey lines parallel to the axis.

[0031] The survey line spacing of 5–10 m specifies the distance between survey lines in underwater measurements using unmanned surface vessels (USVs), aiming to balance data acquisition density and operational efficiency. A 5–10 m spacing ensures sufficient overlap between the coverage areas of the multibeam sonar, resulting in high-density, high-precision underwater topographic data. For example, a 5 m spacing can be chosen for areas with complex underwater topography and high accuracy requirements; for relatively flat areas, it can be appropriately increased to 10 m to improve operational efficiency.

[0032] Autonomous navigation and cruising of unmanned surface vessels (USVs) emphasizes the level of automation and intelligence in USV operations. Autonomous navigation and cruising refers to the ability of USVs to automatically plan paths, avoid obstacles, maintain course and speed, and complete underwater surveying tasks based on pre-set survey lines and mission objectives, utilizing their onboard navigation systems (such as GPS or inertial navigation systems). For example, USVs are typically equipped with high-precision positioning modules and intelligent obstacle avoidance systems, enabling them to perceive their surroundings in real time and make autonomous decisions and path adjustments according to pre-set rules.

[0033] Multibeam sonar continuously scans underwater topography, a technological feature that defines the core equipment and operating method of underwater surveying by unmanned surface vessels (USVs). Multibeam sonar is an advanced underwater depth sounding device capable of simultaneously transmitting and receiving multiple sound beams, forming a fan-shaped coverage area. A single scan can acquire underwater elevation data for a strip-shaped region. For example, a multibeam sonar system typically includes a transducer, a data acquisition unit, and processing software, calculating water depth by measuring the propagation and reflection time of sound waves in the water.

[0034] The simultaneous acquisition of water depth and bottom point cloud data describes the data types acquired by multibeam sonar. Water depth data is a fundamental result of underwater measurement, directly reflecting the elevation information of the seabed. Bottom point cloud data is the raw output of multibeam sonar, composed of a large number of points with three-dimensional coordinates and intensity information, enabling a more detailed description of the geometry and texture features of the seabed. For example, each time a multibeam sonar system transmits and receives sound waves, it records the propagation time, angle, and other information of the sound waves, and calculates the three-dimensional coordinates and echo intensity of the bottom point corresponding to each beam, forming point cloud data.

[0035] This technical feature, encompassing attitude correction and coordinate transformation, describes the key steps in underwater measurement data preprocessing. Attitude correction refers to using attitude information such as roll, pitch, and bow recorded by attitude sensors during the measurement process of the unmanned surface vessel (USV) to correct the water depth data measured by sonar, thereby eliminating the influence of hull attitude changes on the measurement results. Coordinate transformation involves converting the raw measurement data from the sonar or hull coordinate system to a unified engineering coordinate system. For example, attitude correction typically involves algorithmic calculations using specialized post-processing software, combining attitude sensor data and sonar data.

[0036] Generating a high-precision underwater digital elevation model (DEM) is one of the final results of underwater measurement data processing. A high-precision underwater DEM is generated by processing underwater point cloud data (after attitude correction and coordinate transformation) through gridding and interpolation, resulting in a digital model representing underwater topographic elevation information in a regular grid format. For example, algorithms such as Kriging interpolation and inverse distance weighted interpolation can be used to convert discrete point cloud data into a continuous grid DEM.

[0037] Identifying the distribution of weir base trenches and shoals is a key topographic feature extracted from an underwater DEM. Weir base trenches refer to natural or man-made depressions that may exist at the bottom of the cofferdam, and these trenches significantly impact the stability of the cofferdam. Shoals refer to relatively shallow underwater areas that may affect the navigation safety of unmanned vessels. For example, trenches and shoals can be identified through contour analysis, slope analysis, or morphological filtering methods based on the DEM.

[0038] Through the above technical solutions, this application has achieved significant technological progress in unified aerial-water collaborative topographic mapping across the entire area. First, using the cofferdam axis as the boundary, the area affected by the filling is clearly divided into an UAV mapping area, while the underwater cofferdam filling area and near-shore shallow water area are divided into an unmanned surface vessel (USV) mapping area. This regional division strategy fully leverages the mapping advantages of both UAVs and USVs, avoids the limitations of traditional single-equipment operations, ensures seamless coverage of the entire area above and below water, and effectively solves the problems in traditional surveying where underwater topography easily misses shallows and ditches, and the surface of the water is difficult to reflect micro-topographic undulations.

[0039] Secondly, for UAV oblique photography operations, selecting sunny, windless, or lightly windy periods effectively avoided the impact of adverse weather on image quality, ensuring the clarity and accuracy of the raw data. Planning a five-directional aerial photography route combining orthogonal and oblique views, along with optimized settings for flight altitude, heading overlap, and lateral overlap, ensured the acquisition of multi-angle, high-density images, providing a solid foundation for constructing a high-precision 3D model of the waterfront landscape. Autonomous flight acquisition of high-definition images significantly improved operational efficiency and safety, and reduced the need for manual intervention. After software processing, key data such as shoreline boundaries, original ground elevation, and terrain slope could be accurately extracted, providing comprehensive and reliable geographic information support for subsequent cofferdam construction design.

[0040] Furthermore, for unmanned surface vessel (USV) underwater surveying operations, survey lines parallel to the cofferdam axis were planned, with a spacing of 5–10 meters between the lines, ensuring systematic coverage of the underwater area and high precision in data acquisition. The USV's autonomous navigation and cruise, combined with multibeam sonar continuous scanning of the underwater topography, achieved efficient and safe automated underwater surveying, overcoming the safety risks and inefficiencies of manual wading operations. Simultaneously acquired water depth and bottom point cloud data, after professional processing such as attitude correction and coordinate transformation, ultimately generated a high-precision underwater digital elevation model, capable of accurately identifying the distribution of dam foundation trenches and shoals. This detailed underwater topographic data is crucial for the stability and filling quality of the cofferdam, effectively guiding the construction team in precise foundation clearing, backfilling, or reinforcement, avoiding engineering hazards caused by unclear underwater topography.

[0041] In summary, this application significantly improves the efficiency, accuracy, and safety of topographic mapping in the early stage of cofferdam construction through a refined air-water collaborative mapping strategy and parameter optimization. It provides high-quality and comprehensive raw data support for subsequent scheme optimization, process control, and quality acceptance, fundamentally solving the core pain points of traditional measurement methods, such as low efficiency, insufficient accuracy, and prominent safety risks.

[0042] In this invention, in step 3), the above-water three-dimensional model and the underwater high-precision digital elevation model are unified into the same engineering coordinate system. By matching feature points at the junction, they are fused to form an integrated air-water three-dimensional terrain model of the cofferdam construction area, accurately calculating the original earthwork volume and providing data support for the filling scheme design.

[0043] Specifically, unifying the surface 3D model and the underwater high-precision digital elevation model into the same engineering coordinate system aims to ensure that all spatial data (whether from UAV oblique photogrammetry or unmanned surface vessel underwater surveying) references a common and consistent coordinate system. This is crucial for eliminating systematic positional errors and inconsistencies between different data sources, thereby avoiding misalignment or mismatch at the land-water interface. Implementation methods can include: using specialized Geographic Information System (GIS) software or surveying data processing software to transform one or more datasets from their original coordinate system to the target engineering coordinate system based on known coordinate system parameters; or, before surveying operations, deploying high-precision Global Navigation Satellite System (GNSS) reference stations within the construction area and utilizing their differential correction data to ensure that UAVs and unmanned surface vessels use a unified engineering coordinate system for positioning during data acquisition, and directly importing this coordinate system during subsequent data processing to reduce transformation errors.

[0044] Building upon this foundation, feature point matching at the interface aims to identify and associate corresponding points or features in overlapping or adjacent areas of different datasets (above-water 3D model and underwater high-precision digital elevation model) to ensure seamless alignment between data. This step resolves potential local geometric inconsistencies or deformations even after coordinate system unification, particularly at the water-land interface, ensuring the geometric continuity of the model. Implementation methods can include: manually selecting prominent feature points at the water-land interface (such as shoreline inflections, protruding rocks, edges of artificial structures, etc.) in data processing software, selecting corresponding points in both the above-water 3D model and the underwater high-precision digital elevation model, and then using these matching points for local translation, rotation, or scaling to achieve precise alignment; or employing image processing and point cloud registration algorithms to automatically identify edge features, texture features, or elevation abrupt changes in the water-land interface area. For example, using the Iterative Closest Point (ICP) algorithm or its variants, the best matching point pairs can be automatically found in the overlapping area, and transformation parameters can be calculated to achieve high-precision automatic registration.

[0045] Subsequently, an integrated air-water 3D terrain model of the cofferdam construction area was created. This model aimed to merge the processed and aligned surface 3D model with the underwater high-precision digital elevation model into a single, comprehensive 3D representation covering the entire cofferdam construction area. This provided a complete and continuous understanding of the terrain across the entire project site, eliminating data gaps and inconsistencies between the water and land environments, and laying the foundation for all subsequent spatial analyses. The implementation methods may include: stitching together the surface 3D model (usually a triangular mesh or point cloud) and the underwater high-precision digital elevation model (usually a regular grid digital elevation model or an irregular triangular mesh) in data processing software after coordinate unification and feature point matching. For overlapping areas, weighted averaging, interpolation, or priority strategies can be used for fusion to finally generate a unified irregular triangular mesh or digital elevation model, which can be further rendered into a 3D visualization model; or, converting both surface and underwater data into a unified point cloud data format, then performing filtering, denoising, and thinning on the entire point cloud, and constructing a unified irregular triangular mesh or regular grid digital elevation model based on the processed point cloud, thereby forming a seamless air-water integrated 3D terrain model.

[0046] Based on this integrated model, the original earthwork volume is accurately calculated. The aim is to accurately determine the volume of earth and rock materials existing within the construction area before any filling operation begins, utilizing a unified, high-precision 3D terrain model. This provides a reliable benchmark for the planning, budgeting, and management of filling materials, directly impacting the accuracy of cost estimation, material procurement, and construction schedule. Implementation methods can include: defining the base area of ​​the cofferdam based on the integrated global 3D terrain model, and using specialized earthwork calculation software to calculate the earthwork volume between the original terrain and a reference plane (such as the design base elevation) within this area using methods such as meshing, cross-sectioning, or triangular meshing; or, directly using the integrated 3D model (such as an irregular triangular mesh model), constructing a closed 3D volume, and calculating the volume difference between it and the design base surface to obtain a more accurate earthwork volume. This method can better handle complex terrain and irregular boundaries.

[0047] Ultimately, the goal is to provide data support for cofferdam design, making comprehensive and accurate topographic data the fundamental input for developing, optimizing, and validating cofferdam filling schemes. This enables engineers to make informed decisions regarding the quantity of filling materials, layering strategies, equipment deployment, and potential challenges, resulting in a more efficient, economical, and safer construction process. Implementation methods can include: importing an integrated 3D topographic model into design software for 3D visualization, allowing engineers to intuitively analyze topographic undulations, slopes, water depths, and other information, and based on this, to develop preliminary concepts and simulations of cofferdam design schemes, such as simulating different filling paths and accumulation effects; or, utilizing the high-precision data provided by the integrated 3D topographic model, combined with engineering design specifications and requirements, to perform parametric design, such as optimizing parameters like the cofferdam's cross-sectional shape, filling layer thickness, and slope gradient based on topographic slope and water depth data, and using software to conduct multi-scheme comparative analysis to select the optimal filling scheme.

[0048] The above technical solution first unifies the above-water 3D model and the underwater high-precision digital elevation model into the same engineering coordinate system, effectively eliminating systematic deviations caused by different coordinate systems and ensuring alignment of above-water and underwater data within a unified reference framework. Based on this, feature point matching at the junction further resolves potential local geometric inconsistencies in the land-water interface area, achieving precise and seamless integration of above-water and underwater data. This results in the construction of a continuous, complete, and high-precision integrated air-water 3D terrain model of the cofferdam construction area. This integrated model comprehensively reflects the actual topography of the construction area, providing a reliable foundation for subsequent engineering analysis. Given this high-precision integrated model, the original earthwork volume can be accurately calculated, significantly improving the accuracy of earthwork volume estimation and avoiding overfilling or underfilling due to inaccurate data, effectively controlling engineering costs and material waste. Simultaneously, based on this accurate earthwork volume and integrated terrain model, it provides solid data support for filling scheme design, enabling the design scheme to fully consider actual terrain conditions, optimize filling strategies, and improve construction efficiency and project quality. Overall, the solution effectively solved the problem of air-water data fusion, improved the accuracy and reliability of data preparation in the early stage of cofferdam construction, and laid a solid foundation for subsequent construction management and quality control.

[0049] Further, in step 4), the cofferdam axis is divided into several filling sections, each section is filled in layers, the starting and ending elevations of filling, slope trimming parameters, earthwork sources and transportation routes of each section are clearly defined, and the construction schedule and key points of quality control are determined.

[0050] Specifically, the cofferdam is divided into several filling sections along its axis, aiming to divide the entire cofferdam construction area into multiple smaller, independently manageable construction units along its central axis. This division can be based on the overall length of the cofferdam or the expected filling volume, using equal distances or equal volumes, for example, dividing the area into sections at certain intervals or containing a certain volume of earth and rock. Alternatively, the cofferdam axis can be divided into sections of different lengths or shapes based on factors such as geological conditions, hydrological environment, construction difficulty, or sensitivity to the surrounding environment, to adapt to the construction needs of different areas. Within each filling section, a layered filling method is adopted, meaning the material is laid in continuous horizontal layers rather than being dumped in large quantities all at once. This can be achieved by setting a fixed filling layer thickness, such as 0.5 meters or 1 meter per layer; or by dynamically adjusting the filling layer thickness according to the properties of the filling material, the type of compaction machinery, and the required compaction degree.

[0051] To ensure construction precision, this application specifies the start and end elevations of each filling section. These elevations are the precise vertical boundaries of each filling section, which can be directly extracted and digitized from the cofferdam design drawings, or dynamically calculated using real-time topographic survey data and design models to adapt to minor variations in the actual terrain. Simultaneously, this application also specifies slope trimming parameters, which are specific technical specifications regarding the shape and finish of the cofferdam's side slopes, including slope ratios, surface smoothness requirements, and allowable tolerances. These parameters can be set according to engineering design specifications and the cofferdam's intended use, such as standard slope ratios (e.g., 1:2, 1:3) and slope smoothness requirements; more refined parameters can also be determined based on the physical and mechanical properties of the filling materials and environmental factors such as water erosion. Furthermore, this application specifies the source and transportation route of the earth and rock, i.e., determining the source of the filling materials and planning the most efficient and safest transportation route. This can be achieved by pre-identifying multiple earthwork material yards and assessing their quality, reserves, and transport distances, while simultaneously planning and optimizing transport routes based on on-site road conditions, traffic flow, and vehicle types. Finally, this application identifies key points for construction schedule and quality control, establishing a timeline for the entire construction process and defining critical checkpoints and quality standards. This can be achieved using project management tools such as Gantt charts or the critical path method to plan the start and end times for each section and each layer of filling, and to develop detailed quality control checklists, inspection frequencies, and acceptance criteria for key aspects such as filling material quality, compaction, elevation control, and slope formation.

[0052] The aforementioned technical solution decomposes the cofferdam filling task into controllable sections and layered construction, and pre-defines and plans key parameters and processes, effectively solving the problems of blind spots, low efficiency, and difficulty in quality control that may occur during construction. Specifically, by dividing the cofferdam into several filling sections along its axis, refined construction management can be carried out for different areas, avoiding overall construction chaos and improving the organization and focus of construction. The implementation of layered filling in each section ensures that the quality of each layer is controllable, effectively reducing the accumulation of errors caused by one-time construction and improving the overall stability and uniformity of the filling body. Defining the starting and ending elevations of each section provides a precise height control benchmark for construction, effectively preventing overfilling or underfilling and ensuring that the geometric dimensions of the cofferdam meet design requirements. Defining slope trimming parameters ensures that the slope, flatness, and stability of the slope meet design standards, improving the structural safety and erosion resistance of the cofferdam. Defining the source and transportation route of earth and rock optimizes resource allocation and transportation efficiency, reduces construction costs, and minimizes environmental impact. By defining the construction schedule and key quality control points, a systematic construction process and quality standards were established, making the construction process traceable and controllable, and significantly improving overall construction efficiency and management level. This refined scheme optimization fully utilized the data support of the pre-construction integrated air-water three-dimensional terrain model, making the filling scheme design more precise and laying a solid foundation for subsequent construction process control and dynamic monitoring, ultimately ensuring the quality and schedule of the cofferdam filling.

[0053] Step 5) further includes the following steps: (1) Layered filling: dump trucks are used for unloading, excavators are used for filling, bulldozers are used for leveling, and road rollers are used for compaction in layers. After each layer is filled, "air-water" coordinated monitoring is carried out immediately. (2) Rapid inspection by drone: oblique photography of the surface of the filled layer is carried out to generate a three-dimensional model of the surface, compare it with the design elevation, detect the surface flatness and slope gradient, identify the underfilled area, overfilled area and local collapse hazards, and generate a deviation distribution map. (3) Underwater verification by unmanned vessel: The underwater cofferdam filling area and the underwater part of the front edge of the filling body are re-measured to monitor the settlement of the cofferdam filling area and the scouring of the front edge of the filling body. The data are compared with the previous data to analyze the impact of underwater topographic changes on the stability of the cofferdam. (4) Real-time data transmission and analysis: The monitoring data from UAVs and unmanned vessels are processed and compared with the design to generate a monitoring report; when the monitoring data exceeds the preset deviation threshold, an early warning is issued immediately to guide the on-site construction personnel to adjust the filling parameters.

[0054] Layered filling aims to ensure the quality and stability of cofferdam construction. This is achieved by dividing the total filling height into several layers, compacting and monitoring each layer to prevent cumulative deviations. Specifically, various types of construction machinery can be used in tandem. For example, dump trucks, either rear-dump or side-dump types, can be selected based on site conditions and transportation efficiency to transport earth and rock to designated areas for unloading. Excavators, either tracked or wheeled, can be selected based on the volume of earth and operating environment to perform initial spreading and filling of the earth and rock. Bulldozers, either tracked or wheeled, can be selected based on leveling accuracy and operational efficiency to finely level the surface of the filling layer. Road rollers, including vibratory rollers, static rollers, and pneumatic tire rollers, can be selected based on the properties of the filling material and compaction requirements to compact each layer of the filling material in layers, achieving the designed density. After each layer of filling is completed, "air-water" coordinated monitoring should be carried out immediately to ensure that problems are detected and corrected in time before deviations accumulate.

[0055] The rapid drone inspection aims to quickly and accurately acquire three-dimensional information of the filled layer surface, enabling timely detection and quantification of filling deviations. In practice, drones can perform oblique photography of the filled layer surface. Besides using orthophoto + oblique five-directional aerial photography routes, multi-angle oblique photography or surround oblique photography can also be employed to obtain more comprehensive surface data. The acquired images are processed using photogrammetry software to generate a high-precision three-dimensional surface model, which can be a point cloud, digital surface model, or realistic 3D model. Subsequently, the generated surface 3D model is overlaid and compared with the design elevation data to detect surface flatness and slope gradient. 3D model analysis tools can automatically extract surface slope information and calculate local undulations. Based on the elevation deviation map, preset thresholds are set to automatically identify underfilled areas, overfilled areas, and potential localized collapses exceeding allowable deviations, generating an intuitive color-coded deviation distribution map to help construction personnel quickly locate problem areas.

[0056] Underwater verification by unmanned surface vessels (USVs) aims to monitor the morphological changes and stability of underwater embankments, particularly settlement and scour, to ensure the construction quality of the underwater portion. In practice, USVs can re-measure the underwater cofferdam filling area and the underwater section at the front of the embankment. In addition to high-density scanning using multibeam echo sounders, side-scan sonar can be combined to acquire underwater topographic images, providing more comprehensive underwater information. By comparing underwater elevation data from different time periods, the settlement of the cofferdam filling area can be accurately calculated. Simultaneously, by comparing underwater topographic changes in the area at the front of the embankment, scour areas and scour volumes can be identified. This preliminary data can include design model data, monitoring data after the completion of the previous layer of filling, or historical monitoring data. Based on these comparative analysis results, the impact of underwater topographic changes on cofferdam stability can be further analyzed, for example, through assessment using hydraulic models or soil mechanics stability analysis software, thereby promptly identifying potential structural risks.

[0057] Real-time data transmission and analysis aims to ensure the timely processing, analysis, and feedback of monitoring data, enabling dynamic adjustment and optimization of the construction process. Specifically, monitoring data collected by drones and unmanned surface vessels can be automatically processed through a cloud platform, including data stitching, noise reduction, coordinate transformation, and model building. The processed data will be compared in real time with the design model, for example, through overlay and difference analysis using a BIM platform or GIS system. Based on the comparison results, the system will automatically generate a monitoring report, which may include deviation maps, statistical data, and a list of problem areas, providing a basis for construction management. When monitoring data exceeds preset deviation thresholds, such as elevation deviation ±10cm or slope deviation ±5°, the system will immediately issue an early warning. Warning information can be promptly communicated to on-site construction personnel through various means, including SMS, app push notifications, and audible and visual alarms. Based on the warning information, construction personnel can dynamically adjust filling parameters, such as adjusting unloading positions, excavator operating ranges, bulldozer leveling paths, and roller compaction passes, thereby achieving precise control and quality optimization of the construction process.

[0058] Through the above technical solutions, this application effectively solves the problems of traditional monitoring methods being insufficiently specific and lacking collaborative real-time monitoring, leading to the inability to identify deviations in a timely manner. During the layered filling process, standardized construction is carried out using mechanized means such as dump trucks, excavators, bulldozers, and road rollers. Crucially, "air-water" collaborative monitoring is conducted immediately after each layer is filled, significantly shortening the monitoring feedback cycle and preventing the accumulation of deviations. Rapid drone inspections can efficiently acquire three-dimensional models of the fill layer surface. By accurately comparing these models with the design elevation, surface flatness and slope gradient can be detected in real time, and underfilled areas, overfilled areas, and potential localized collapse hazards can be intuitively identified, generating deviation distribution maps, making surface quality problems readily apparent. Simultaneously, underwater verification by unmanned surface vessels focuses on the dynamic monitoring of the underwater cofferdam filling area and its leading edge. By re-measuring settlement and scour and comparing it with previous data, the potential impact of underwater topographic changes on cofferdam stability can be detected in a timely manner, effectively compensating for the blind spots in underwater monitoring. Ultimately, through real-time data transmission and analysis, the monitoring data from drones and unmanned vessels are fused and compared with the design. If a preset deviation threshold is exceeded, the system immediately issues an early warning, directly guiding on-site construction personnel to adjust filling parameters. This achieves real-time and intelligent monitoring throughout the entire process, from data acquisition and analysis to decision feedback. This integrated air-water collaborative monitoring mechanism significantly improves the control precision and efficiency of the cofferdam filling process, effectively avoiding overfilling, underfilling, and quality hazards, ensuring the overall quality and safety of the cofferdam project.

[0059] Furthermore, step 6) also includes the following steps: (1) After the cofferdam is filled to the design elevation, the overall slope is trimmed; based on the slope tilt photography data obtained by UAV tilt photography, the slope is trimmed by excavator and manual labor to ensure that the slope and flatness meet the design requirements. (2) Full-area completion survey: Re-launch UAV and unmanned boat to carry out full-coverage surveying, construct an integrated air-water three-dimensional model after the cofferdam is formed, and extract completion data such as the top elevation of the cofferdam, slope gradient, axis position, and filling volume; (3) Quality acceptance and data archiving: Compare the as-built 3D model and measurement data with the design model, verify various quality indicators, and automatically generate an acceptance report; (4) Organize and archive the measurement data, model files, monitoring records and acceptance data of the entire construction process to form a digital construction archive.

[0060] Specifically, after the cofferdam is filled to the design elevation, overall slope trimming aims to precisely adjust the physical shape of the cofferdam to meet the requirements of the design drawings, ensuring its structural stability, hydraulic characteristics, and appearance quality. Construction personnel can use point cloud models or digital surface models generated from UAV oblique photography data to compare with the design model in real time on-site using high-precision positioning equipment (e.g., RTK global navigation satellite system). This guides excavators to perform precise operations, supplemented by manual adjustments to local details to achieve the required slope and flatness. Alternatively, excavators can be equipped with automated control systems that directly receive the design slope lines and elevation information from the processed UAV oblique photography data, enabling semi-automated or fully automated trimming, reducing manual intervention and improving trimming efficiency and accuracy.

[0061] The purpose of the full-area as-built survey is to conduct a comprehensive and high-precision "as-built" mapping of the completed cofferdam to obtain its final geometric shape and location information. A drone equipped with a high-definition camera, a global navigation satellite system (GNSS), and RTK is reactivated to perform oblique photography of the water area along a preset route. Simultaneously, an unmanned surface vessel (USV) equipped with a single / multi-beam scanner, GNSS, attitude sensor, and wireless communication module is activated to perform multi-beam scanning of the underwater area. The two sets of data are then fused using specialized software to generate an integrated air-water 3D model in a unified coordinate system. Alternatively, an unmanned platform with multi-sensor fusion capabilities, such as a hybrid USV capable of carrying underwater sonar and aerial photography equipment, can be used to complete a full-coverage air-water mapping in one go and automatically extract the required as-built data.

[0062] In the quality acceptance and document archiving phase, the core lies in analyzing and evaluating the as-built measurement data to verify whether the cofferdam meets design requirements and quality standards, and to generate standardized acceptance documents. Using professional BIM (Building Information Modeling) or GIS (Geographic Information System) software, the as-built 3D model and measurement data are overlaid with the design 3D model for analysis. The deviation between the two is automatically calculated, and judgments are made based on preset quality indicators (such as elevation deviation, slope deviation, axis deviation, etc.), generating an acceptance report containing information such as deviation analysis and pass rate. Alternatively, a rule-based expert system can be established. As-built data and design requirements are input, and the system automatically performs verification of various quality indicators and automatically fills in data and conclusions according to preset report templates, generating a structured acceptance report.

[0063] The purpose of compiling and archiving measurement data, model files, monitoring records, and acceptance documents throughout the entire construction process into a digital construction archive is to systematically manage and permanently preserve all key data generated throughout the project lifecycle, ensuring data integrity, traceability, and availability. This involves establishing a centralized project database or cloud platform to uniformly upload, categorize, and store all digital materials, including raw data collected by drones and unmanned surface vessels, processed 3D models, monitoring reports during construction, repair records, and final acceptance reports, with access control and version control implemented. Alternatively, blockchain technology can be used to encrypt and distribute the data throughout the construction process, ensuring data immutability and traceability, providing a reliable basis for project quality traceability and liability determination.

[0064] Through the aforementioned technical solutions, after the cofferdam is filled to the design elevation, oblique photography data from drones is used to guide slope trimming, making the trimming work more precise, reducing rework, and improving efficiency. In the overall as-built survey phase, drones and unmanned vessels are again used to conduct full-coverage mapping, constructing an integrated air-water 3D model of the cofferdam after its formation. This not only ensures comprehensive coverage and high-precision data acquisition of the above-water and underwater parts of the cofferdam, eliminating blind spots and errors inherent in traditional manual surveying, but also connects with the previous integrated air-water topographic mapping and the establishment of the integrated air-water 3D topographic model, ensuring data consistency and comparability. Comparing the as-built 3D model and measurement data with the design model and automatically generating an acceptance report significantly improves the objectivity, efficiency, and accuracy of the acceptance process, reducing the risk of human error. At the same time, the measurement data, model files, monitoring records and acceptance data of the entire construction process were organized and archived to form a digital construction archive, providing a complete digital history for the project, which is convenient for later query, traceability and management, and improves the overall management level of the project. This effectively solves the problems of low efficiency, insufficient accuracy and chaotic data management in the process of slope trimming and forming quality acceptance of traditional cofferdam filling methods.

Claims

1. A cofferdam construction method based on underwater surveying by unmanned surface vessels and oblique photography by unmanned aerial vehicles, characterized by: The method includes the following steps: 1) Selection of unmanned ships and drones; 2) Unified topographic mapping of the entire region, combining air and water surveying; 3) Generate an integrated three-dimensional terrain model of the entire air-water area of ​​the cofferdam construction zone; 4) Optimize the terrain model; 5) Layered filling construction with real-time dynamic monitoring; 6) Acceptance of slope trimming and shaping quality.

2. The cofferdam construction method based on underwater surveying by unmanned surface vessel and oblique photography by unmanned aerial vehicle as described in claim 1, characterized in that: in In step 1), the unmanned vessel is equipped with a single / multi-beam receiver, a global navigation satellite system, an attitude sensor, and a wireless communication module; the unmanned aerial vehicle is a multi-rotor oblique photography unmanned aerial vehicle, equipped with a high-definition camera, a global navigation satellite system, and RTK.

3. The cofferdam construction method based on underwater surveying by unmanned surface vessel and oblique photography by unmanned aerial vehicle as described in claim 2, characterized in that: in In step 2), the area affected by the filling is the UAV mapping area, and the underwater cofferdam filling area and the near-shore shallow water area are the UAV mapping area. When performing oblique photography with UAVs, select sunny, windless or light-wind periods, plan a five-directional aerial photography route combining orthogonal and oblique views, set the flight altitude to 80-150m, the directional overlap rate to 80%, and the lateral overlap rate to 75%, autonomously fly to collect high-definition images, process them with software to construct a real-world 3D model of the water area, and extract data such as shoreline boundaries, original ground elevation, and terrain slope. When conducting underwater surveying operations, the unmanned vessel plans survey lines parallel to the cofferdam axis with a spacing of 5 to 10 meters. The unmanned vessel autonomously navigates and cruises, continuously scanning the underwater terrain with multibeam sonar, and simultaneously collecting water depth and bottom point cloud data. After attitude correction and coordinate transformation, a high-precision underwater digital elevation model is generated to identify the distribution of dam foundation trenches and shoals.

4. The cofferdam construction method based on underwater surveying by unmanned surface vessel and oblique photography by unmanned aerial vehicle as described in claim 3, characterized in that: in In step 3), the above-water 3D model and the underwater high-precision digital elevation model are unified into the same engineering coordinate system. By matching feature points at the junction, they are integrated to form a three-dimensional terrain model of the entire air-water integrated area of ​​the cofferdam construction area, accurately calculating the original earthwork volume and providing data support for the filling scheme design.

5. The cofferdam construction method based on underwater surveying by unmanned surface vessel and oblique photography by unmanned aerial vehicle as described in claim 4, characterized in that: in In step 4), the cofferdam axis is divided into several filling sections, and each section is filled in layers. The starting and ending elevations of filling, slope trimming parameters, earthwork sources and transportation routes of each section are clearly defined, and the construction schedule and key points of quality control are determined.

6. The cofferdam construction method based on underwater surveying by unmanned surface vessel and oblique photography by unmanned aerial vehicle as described in claim 5, characterized in that, Step 5) further includes the following steps: (1) Layered filling: dump trucks are used for unloading, excavators are used for filling, bulldozers are used for leveling, and road rollers are used for compaction in layers. After each layer is filled, "air-water" coordinated monitoring is carried out immediately. (2) Rapid inspection by drone: oblique photography of the surface of the filled layer is carried out to generate a three-dimensional model of the surface, compare it with the design elevation, detect the surface flatness and slope gradient, identify the underfilled area, overfilled area and local collapse hazards, and generate a deviation distribution map. (3) Underwater verification by unmanned vessel: The underwater cofferdam filling area and the underwater part of the front edge of the filling body are re-measured to monitor the settlement of the cofferdam filling area and the scouring of the front edge of the filling body. The data are compared with the previous data to analyze the impact of underwater topographic changes on the stability of the cofferdam. (4) Real-time data transmission and analysis: The monitoring data from UAVs and unmanned vessels are processed and compared with the design to generate a monitoring report; when the monitoring data exceeds the preset deviation threshold, an early warning is issued immediately to guide the on-site construction personnel to adjust the filling parameters.

7. The cofferdam construction method based on underwater surveying by unmanned surface vessel and oblique photography by unmanned aerial vehicle as described in claim 6, characterized in that, Step 6) further includes the following steps: (1) After the cofferdam is filled to the design elevation, the overall slope is trimmed; based on the slope tilt photography data obtained by UAV tilt photography, the slope is trimmed by excavator and manual labor to ensure that the slope and flatness meet the design requirements. (2) Full-area completion survey: Re-launch UAV and unmanned boat to carry out full-coverage surveying, construct an integrated air-water three-dimensional model after the cofferdam is formed, and extract completion data such as the top elevation of the cofferdam, slope gradient, axis position, and filling volume; (3) Quality acceptance and data archiving: Compare the as-built 3D model and measurement data with the design model, verify various quality indicators, and automatically generate an acceptance report; (4) Organize and archive the measurement data, model files, monitoring records and acceptance data of the entire construction process to form a digital construction archive.

Images