Interactive digital twinning method for deep sea construction

Abstract

The invention provides an interactive digital twinning method for deep sea construction, and belongs to the technical field of digital twinning. The problem of remote control in seabed geological classification in robot landing and deep sea construction in a deep sea invisible environment is solved. According to the technical scheme, the method comprises the following steps that S1, relevant parameters are given; s2, establishing a deep sea environment digital twinborn model; s3, designing an elevation data external interaction interface, inputting the wavelength of a multi-beam millimeter radar wave, and training a corresponding classifier; and S4, integrating robot parameters, an underwater address segmentation result and embedded sensor feedback data to complete the interactive digital twinborn body. The interactive digital twinborn body has the advantages that the interactive digital twinborn body is completed, and remote control over land and deep sea is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and in particular to an interactive digital twin method for deep-sea construction. Background Art

[0002] At present, there are two schemes for underwater geological segmentation: acoustic systems and radar systems. In the article "Automatic classification and mapping of the seabed using airborne LiDARbathymetry" by Lukasz Janowski et al., it is proposed that airborne measurement systems can conduct surveys in shallow waters faster than acoustic systems. Compared with acoustic systems, airborne measurement systems can conduct surveys faster in shallow waters. The use of aerial survey technology can reduce the cost of mapping large coastal areas. Hydroacoustic surveys are generally considered to fill the gaps in airborne data coverage because there are many characteristics related to the penetration of laser beams, such as siltation, interference levels, water levels, scanning system types, and signal processing methods. Therefore, there is a gap in the literature on the use of ALB. Underwater operations are mainly completed by humans or underwater robots. Underwater robots, also known as unmanned remote-controlled submersibles, are extreme operation robots that work underwater. The underwater environment is harsh and dangerous, and the diving depth of humans is limited, making it difficult to complete deep-sea operation tasks. Existing underwater robots rely too much on real-time scanning of underwater multi-beam radars, and the mission cost is too high.

[0003] How to solve the above technical problems is the subject faced by the present invention. Summary of the invention

[0004] The purpose of the present invention is to provide an interactive digital twin method for deep-sea construction. By designing an interactive digital twin method, workers on land can control underwater robots to carry out oil well exploration, underwater welding and seabed topography changes.

[0005] The purpose of this invention is to propose a highly interactive digital twin method for deep-sea construction environments, aiming to solve several technical problems currently faced by underwater operations from a technical level. These technical problems include but are not limited to:

[0006] High cost of underwater operations: Traditional underwater operations rely on a large number of expensive underwater equipment and professional divers. Not only are the equipment purchase and maintenance costs high, but divers' operations are also strictly limited by time, depth and environment, resulting in high overall operating costs. This invention builds a high-precision digital twin model, allowing workers to remotely control underwater robots on land, reducing the need for expensive divers and improving equipment utilization, thereby significantly reducing operating costs.

[0007] Insufficient real-time performance caused by communication delays in sea-land interaction: Existing underwater operation systems often have problems with large interaction delays and untimely information feedback, which seriously restricts the efficiency and safety of underwater operations. The interactive digital twin method designed by the present invention uses advanced real-time data transmission and processing technology to ensure instant communication between land staff and underwater robots, achieves accurate perception and rapid response to the underwater operation environment, and enhances the real-time interactivity of the operation.

[0008] Uncertainty of underwater environment: The underwater environment is complex and changeable. During the operation, unknown obstacles, water flow interference and other challenges are often encountered, resulting in a low success rate of the operation. This invention integrates high-precision DEM underwater elevation maps and advanced simulation algorithms to simulate various construction scenarios in the digital twin in advance, predict possible problems and optimize response plans, thereby improving the success rate of underwater operations and reducing the risk of failure.

[0009] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is specifically: an interactive digital twin method for deep-sea construction, comprising the following steps:

[0010] S1: Given relevant parameters, including embedded sensor data, excavator related parameters, and multi-beam millimeter radar wave waveforms. Embedded sensor data include underwater pressure, deep sea water velocity, and deep sea water flow direction. Excavator related parameters include excavator working weight, excavator engine power, and excavator bucket capacity. Multi-beam millimeter radar waves illuminate the seabed terrain through the orthogonality of the transmitting and receiving sector pointing, giving the water depth values ​​of hundreds of seabed measured points in the vertical plane perpendicular to the heading, and depicting the three-dimensional characteristics of the seabed terrain through specific waveforms.

[0011] S2: Establish a digital twin model of the deep-sea environment, combine it with the robot model in Blender, and build a sea-land remote control system to simplify the complexity of deep-sea operation processes and reduce the cost of deep-sea operations. The robot model in Blender consists of the skeleton of the following parts: excavator base, excavator crawler, and excavator bucket. The sea-land remote control system transmits commands to the robot through cables through the control system on the operating ship, and the robot transmits data to the operating ship through cables.

[0012] S3: Design the external interaction interface of elevation data, input the wavelength of multi-beam millimeter radar wave, train the corresponding classifier, realize the geological segmentation of deep-sea seabed, and ensure that the landing operation of the operation robot can be monitored. The external interaction interface of elevation data is completed by C# script, and the wave surface data is obtained by scanning the airborne multi-beam millimeter radar, and the wave surface data is transmitted back to the system in the form of a multi-dimensional matrix. The zoning of the bed shape by deep-sea seabed geological segmentation includes demarcating the bed shape boundary according to depth, slope and width (including the division of edge, slope base, bed shape ridge and trough axis), analyzing the seabed buoyancy, and analyzing the bathymetry based on a scale of 1:5000 or larger. The slope, width and bathymetry profiles are generated using Blue Marble Geographics 21.1 software. The seabed of the working sea area is scanned by the radar carried by the robot. For the feature extraction and selection of radar data, the actual control points on the ground are determined. GEOBIA (geographic object-based image analysis) image segmentation and classification. The obtained seabed map is summarized and the accuracy is evaluated. A digital elevation model (DEM) is constructed for data control. The scanned terrain data was imported into the DEM model and the scanned model was classified. Nine natural types of bedforms and three types of artificially intervened bedforms were mainly distinguished, including: (a) Seabed areas with visible interpolation remnants. (b) Sandy seabed areas with slightly undulating seabed within sandbars and depressions between sandbars. (c) Seabed with slightly undulating sandy seabed areas within sandbars and depressions between sandbars, but with artifact interference. (d) Flat sandy seabed. (e) Seabed with flat sandy surface characteristics but with artificial interference. (f) Mixed seabed with small hills and depressions formed due to bottom irregularities. (g) Rough sandy seabed. (h) Corrugated seabed partially parallel to coastal currents. (i) Large corrugated seabed completely parallel to coastal currents. (j) Submerged granite breakwater with some biological habitats. (k) Trapezoidal concrete modules between submerged breakwaters. (l) Flat seabed composed of strongly compressed peat.

[0013] S4: Integrate robot parameters, underwater address segmentation results, and embedded sensor feedback data to complete an interactive digital twin, realize remote control on land and deep sea, and solve the problem of remote control in deep sea invisible environments. Embedded sensors feed back data in the form of matrices of different types of signals (underwater pressure, underwater flow velocity, underwater flow direction, terrain elevation) through cables. The feedback data is used to edit the corresponding environment in Unity in real time through the Update function.

[0014] Furthermore, in step S1:

[0015] S1.1: Embedded sensor data include underwater pressure, deep sea water velocity, and deep sea water flow direction.

[0016] S1.2: Excavator related parameters include excavator operating weight, excavator engine power, and excavator bucket capacity.

[0017] S1.3: Multi-beam millimeter radar waves illuminate the seabed topography through the orthogonality of the transmitting and receiving sector directions, giving the water depth values ​​of hundreds of measured points on the seabed in the vertical plane perpendicular to the heading, and depicting the three-dimensional characteristics of the seabed topography through specific waveforms.

[0018] Furthermore, in step S2:

[0019] S2.1: The robot model in Blender consists of the following skeleton parts: excavator base, excavator tracks, and excavator bucket.

[0020] S2.2: The sea-land remote control system transmits instructions to the robot through cables via the control system on the work vessel, and the robot transmits data to the work vessel through cables.

[0021] Furthermore, in step S3:

[0022] S3.1: The external interaction interface of elevation data is completed by C# script. The wave surface data is obtained through multi-beam millimeter radar scanning and transmitted back to the system in the form of a multi-dimensional matrix.

[0023] S3.2: The robot uses algorithms to segment and process deep-sea seabed geology.

[0024] S3.3: Segmentation of deep seabed geology The zoning of bedforms includes delineation of bedform boundaries based on depth, slope and width (including the division of edges, slope bases, bedform ridges and trough axes), analysis of seafloor buoyancy, and analysis of bathymetry based on a scale of 1:5000 or larger. Slope, width and bathymetry profiles were generated using Blue Marble Geographics 21.1 software. Nine types of natural bedforms and three types of artificially intervened bedforms were mainly distinguished.

[0025] Furthermore, in step S3.2:

[0026] S3.2.1: Scan the seabed of the working area using airborne radar.

[0027] S3.2.2: Construct a digital elevation model (DEM) for data control. For DEM moving average smoothing, it is generally as shown in formula (1):

[0028]

[0029] In formula (1)

[0030] Zn+1 / 2

[0031] is the smoothed elevation value of the center point of the moving window, n=9, 25, 49...correspond to windows of 3×3, 5×5, 7×7... respectively; z i is the altitude value at the i-th point in the window, W i is the weight of the point, for W i , generally there is formula (2):

[0032]

[0033] In formula (2), d i is the distance between the ith point and the center point, m = 0, 1, 2;

[0034] S3.2.3: Import the scanned terrain data into the DEM model and classify the scanned model.

[0035] Furthermore, in step S3.3:

[0036] S3.3.1: The detailed classification of bed shapes can be divided into the following categories:

[0037] (1) The seafloor area with visible interpolation remnants. As shown in Figure (a)

[0038] (2) There are slightly undulating sandy seabed areas within the sandbars and depressions between the sandbars. See Figure (b)

[0039] (3) The surface features are the same as those in Figure (b), but there is a seabed with artifact interference. As shown in Figure (c)

[0040] (4) Flat sandy seabed. As shown in Figure (d)

[0041] (5) The surface features are the same as those in Figure (d), but there is artificially disturbed seabed. As shown in Figure (e)

[0042] (6) A mixed seabed with small hills and depressions formed due to the irregularity of the bottom surface. As shown in Figure (f)

[0043] (7) Rough sandy seabed. As shown in Figure (g)

[0044] (8) The wavy seabed is partially parallel to the coastal currents. See Figure (h)

[0045] (9) Large wavy seabed parallel to the coastal currents. See Figure (i)

[0046] (10) Submerged granite breakwater with some biological habitats. See Figure (j)

[0047] (11) Trapezoidal concrete modules between submerged breakwaters. See Figure (k)

[0048] (12) A flat seabed composed of strongly compressed peat. See Figure (l)

[0049] Furthermore, in step S3.4:

[0050] S3.4.1: Feature analysis extraction and selection. The DEM processed in S3 forms the basis for extracting secondary features, including simple statistical ALBs such as standard deviation and variance, geomorphic attributes, etc. The software used is ArcGIS10.8 and SAGA 7.9.0. Feature selection refers to the selection of features extracted from DEM and suitable for automatic classification.

[0051] S3.4.2: Determine actual control points on the ground. Use hand-drawn maps of topography to generate topographic control points. Use the "Create Random Points" tool in ArcGIS software to randomly determine terrain samples. These samples are generated by manual classification within the study site, with a minimum distance of 1 meter for each point and no less than 2 to 4 meters from the study site boundary. The number of control points for each study site is approximately 5,000, and the distribution of control points between terrain categories varies proportionally according to the percentage of aerial coverage of all units. The terrain samples of each study site are divided into training subsets and test subsets in a ratio of 70:30.

[0052] This step was performed randomly using the Subset Features tool in ArcGIS.

[0053] S3.4.3: GEOBIA (Geographic Object Based Image Analysis) image segmentation and classification. GEOBIA is a method for processing remote sensing image analysis. This method mainly uses eCognition software. Image segmentation and classification is

[0054] The main component of the GEOBIA method. Image segmentation works by dividing an image into homogeneous segments or image objects with similar spectral characteristics to replace pixels for further analysis. Due to the high resolution of current remote sensing images, the most detailed measurements represented by pixel values ​​may be interfered by various environmental and physical factors, resulting in data errors.

[0055] The GEOBIA method generates artifact-free image objects that are similar to human vision. Image objects contain more detailed information than pixels, such as texture and shape features. The GEOBIA method is particularly suitable for high-resolution remote sensing images with heterogeneous pixel information.

[0056] S3.4.4: Map Summarization. After obtaining the topographic map, use the map summarization method to improve the overall readability and quality of the classification results. 2 Small classified artifacts can be merged into adjacent classes, and the map can be generalized using the Eliminate Study Site Part algorithm in ArcGIS software.

[0057] S3.4.5: Accuracy assessment. The overall accuracy is the ratio of correctly classified samples to the total test samples. ArcGIS and R software were used to calculate the error matrix and accuracy assessment results.

[0058] Furthermore, in step S4:

[0059] S4.1: The embedded sensors feed back data in the form of a matrix using cables to transmit different types of signals (underwater pressure, underwater flow velocity, underwater flow direction, terrain elevation).

[0060] S4.2: The feedback data is used to edit the corresponding environment in Unity in real time through the Update function.

[0061] Compared with the prior art, the present invention has the following beneficial effects:

[0062] 1. The present invention introduces an automatic mapping method based on the ALB dataset, and successfully deeply integrates the digital elevation map (DEM) with the ALB dataset. Digital elevation maps have been widely used in land terrain mapping due to their high precision and wide coverage. The present invention seamlessly connects these two data formats through a complex algorithm model, achieving continuous and consistent expression from land to deep-sea terrain. This method changes the current situation where traditional deep-sea seabed terrain analysis relies on manual or semi-automatic processing. As an extension of high-precision land terrain data, the ALB dataset has been creatively applied to the deep-sea environment after screening and adjustment. This method automatically analyzes the complex terrain features in the ALB data, such as submarine mountains, trenches, sediment distribution, etc., through algorithms, to achieve rapid and accurate capture of deep-sea seabed terrain. The fusion of ALB and DEM not only retains the advantages of DEM in vertical accuracy, but also greatly enriches the description dimension of deep-sea seabed terrain with the help of the rich terrain feature information of the ALB dataset. This innovation not only greatly improved the efficiency of data processing, but also significantly reduced human errors. It proved that digital elevation maps and ALB datasets can be used for automatic seabed segmentation before deep-sea construction, opening up a new path for data collection and analysis in the field of marine engineering.

[0063] 2. The present invention introduces an interactive digital twin system, which accurately restores the underwater environment based on embedded sensor technology. The embedded sensor is small in size, low in power consumption, and high in precision. It can be directly deployed on deep-sea equipment or robots to collect and transmit multi-dimensional data such as water temperature, salinity, flow rate, and water quality in real time. These data make the monitoring and modeling of the underwater environment in the digital twin system more refined and comprehensive. Digital twin technology constructs a virtual mirror that is highly consistent with the real underwater environment through high-precision simulation and real-time data feedback. This innovation greatly improves the efficiency and safety of deep-sea exploration and operations. The digital twin system can not only reflect the physical parameters of the underwater environment such as flow rate, flow direction, and pressure conditions in real time, but also simulate complex dynamic changes such as ocean currents and underwater terrain, providing an intuitive and operable virtual experimental platform for robots to carry out deep-sea operations. Through the interactive interface, the operator can adjust the parameter settings of the robot in the virtual environment in real time. In order to further improve the flexibility and safety of deep-sea operations, the present invention constructs a dynamic system of underwater robots through modeling technology. Through this system, operators can precisely control deep-sea robots in the remote control center, realizing automation and intelligence of deep-sea operations. The realization of remote control reduces the workload of manpower preparation before large-scale deep-sea projects and improves operation efficiency and accuracy.

[0064] 3. Integrate underwater terrain data into digital twins. As an accurate mapping of physical entities in virtual space, digital twins can reflect the state changes of the physical world in real time. By integrating DEM underwater elevation maps into digital twin technology, the present invention realizes the comprehensive digitization and visualization of the underwater environment. The high-precision underwater elevation map based on DEM provides a reliable basis for construction design, reducing misjudgment and rework caused by unclear terrain. The underwater terrain editing function in the digital twin further enhances the flexibility and predictability of the construction plan, greatly reduces the risk and cost of underwater construction, and improves construction efficiency and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0066] Figure 1 The multi-beam millimeter radar wave of the present invention illuminates the seabed topography through the orthogonality of the transmitting and receiving sector pointing, provides the water depth values ​​of hundreds of seabed measured points in the vertical plane perpendicular to the heading, and depicts a logical diagram of the three-dimensional characteristics of the seabed topography through specific waveforms.

[0067] Figure 2 This is the logic diagram for the splicing and construction of the Blender robot model in the present invention.

[0068] Figure 3 This is a logic diagram for the classification of deep seabed geological scans in the present invention.

[0069] Figure 4 The invention provides a detailed diagram of the bedform by which the bedform boundary is demarcated according to the geological segmentation of the deep seabed by the depth, slope and width (including the division of the edge, slope base, bedform ridge and trough axis). DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. Of course, the specific embodiments described here are only used to explain the present invention and are not used to limit the present invention.

[0071] Example 1

[0072] In "Bringing Bathymetry LiDAR to Coastal Zone Assessment: ACase Study in the Southern Baltic" by Pawel Tysiac et al., an innovative approach to identifying natural hazards is presented by combining coastal feature analysis with underwater factors. The approach evaluates the use of bathymetric scans in the Baltic Sea to determine the factors that cause coastal erosion. In addition, the approach performs a geotechnical analysis while taking into account geometric ground changes underwater. This is the first study to use a coastal monitoring approach that combines geotechnical calculations with remote sensing data. However, the technology is costly and generates a large amount of data.

[0073] Example 2

[0074] On the basis of Example 1, based on "Image-based seabed classification: what can we learn from terrestrial remote sensing" by Markus Diesing et al., the existing traditional method is to manually perform identification and classification of seabed landforms from remote sensing data sets based on multiple methods, and manual generation of maps is laborious and time-consuming. The present invention provides a technology for automatic segmentation of the seabed before deep-sea construction using a digital elevation map (DEM) and an ALB data set, which improves the efficiency of seabed map generation.

[0075] Example 3

[0076] Based on Example 2, based on the extraction of seafloor landforms based on attributes and relationships in remote sensing data sets proposed in the article "Tidal inlets in the Anthropocene: Geomorphology and benthic habitats of the Chioggia inlet, Venice Lagoon" by Stefano Fogarin et al., the present invention provides an interactive digital twin system, which uses remote sensing data to scan and edit the seabed through the twin, without the need for subsequent real-time scanning, thereby reducing the cost of seafloor operations.

[0077] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An interactive digital twin method for deep-sea construction, characterized in that: The following steps are involved: S1: Given relevant parameters, including: embedded sensor data, excavator related parameters, multi-beam millimeter radar waveform; S2: Establish a digital twin model of the deep-sea environment, combine it with the robot model in Blender, and build a sea-land remote control system; S3: Design an external interaction interface for elevation data, input the wavelength of multi-beam millimeter radar waves, train the corresponding classifier, realize deep-sea seabed geological segmentation, and ensure that the landing operation of the operation robot can be monitored; S4: Integrate robot parameters, underwater address segmentation results, and embedded sensor feedback data to complete an interactive digital twin and realize remote control on land and deep sea.

2. The interactive digital twin method for deep-sea construction according to claim 1, characterized in that: The step S1 includes the following steps: S1.1: Embedded sensor data include underwater pressure, deep sea water velocity, and deep sea water flow direction; S1.2: Excavator related parameters include excavator working weight, excavator engine power, excavator bucket capacity; S1.3: Multi-beam millimeter radar waves illuminate the seabed topography through the orthogonality of the transmitting and receiving sector directions, giving the water depth values ​​of hundreds of measured points on the seabed in the vertical plane perpendicular to the heading, and depicting the three-dimensional characteristics of the seabed topography through specific waveforms.

3. The interactive digital twin method for deep-sea construction according to claim 1 or 2, characterized in that: The step S2 includes the following steps: S2.1: The robot model in Blender consists of the following skeleton parts: excavator base, excavator tracks, and excavator bucket; S2.2: The sea-land remote control system transmits instructions to the robot through cables via the control system on the work vessel, and the robot transmits data to the work vessel through cables.

4. The interactive digital twin method for deep-sea construction according to claim 1, characterized in that: The step S3 includes the following steps: S3.1: The external interaction interface of elevation data is completed by C# script, which obtains wave surface data through multi-beam millimeter radar scanning and transmits the wave surface data back to the system in the form of multi-dimensional matrix; S3.2: The robot uses algorithms to segment and process deep-sea seabed geology; S3.3: The zoning of the deep seabed geological segmentation includes defining the bedform boundaries based on depth, slope and width, and analyzing the seafloor buoyancy.

5. The interactive digital twin method for deep-sea construction according to claim 4, characterized in that: In step S3.2, the depth, slope and width include the following steps: S3.2.1: Scan the seabed of the working area by airborne radar; S3.2.2: Construct a digital elevation model (DEM) for data control and smooth the DME moving average, as shown in formula (1): As shown: In formula (1), z′ n+1 / 2 is the smoothed elevation value of the moving window center point, n is the number of all values ​​in the window matrix, and its size is 9, 25, 49... corresponding to 3×3, 5×5, 7×7... windows respectively; z i is the altitude value at the i-th point in the window, W i is the weight of the point, for W i , formula (2): In formula (2), d i is the distance between the ith point and the center point, m = 0, 1, 2, representing different weights at the ith point respectively; S3.2.3: Import the scanned terrain data into the DEM model and classify the scanned model.

6. The interactive digital twin method for deep-sea construction according to claim 4, characterized in that: The step S3.3 includes the following steps: S3.3.1: The detailed classification of bed shapes is as follows: (1) Seafloor areas with visible interpolation remains; (2) There are slightly undulating sandy seabed areas within the sandbars and depressions between the sandbars; (3) The surface features are the same as those in Figure (b), but there is a seabed with artifact interference; (4) Flat sandy seabed; (5) The surface features are the same as those in Figure (d), with the presence of artificially disturbed seabed; (6) A mixed seabed with small hills and depressions formed by irregularities in the bottom surface; (7) Rough sandy seabed; (8) Corrugated seabed partially parallel to coastal currents; (9) Large rippled seabed parallel to the coastal currents; (10) Submerged granite breakwater with some biological habitat; (11) Trapezoidal concrete modules between submerged breakwaters; (12) A flat seabed consisting of strongly compressed peat; S3.3.2: Feature analysis extraction and selection The DEM processed in step S3 formed the basis for extracting secondary features, including statistical ALBs, using ArcGIS 10.8 and SAGA 7.9.

0. Feature selection refers to the selection of features extracted from the DEM and suitable for automatic classification; S3.3.3: Determine the actual control points on the ground. Use the hand-drawn topographic map to generate terrain control points. Use the "Create Random Points" tool in ArcGIS software to randomly determine terrain samples. These samples are generated by manual classification within the study site. The minimum distance of each point is 1 meter and the distance from the study site boundary is not less than 2 to 4 meters. The number of control points for each study site is about 5,000. The distribution of control points between terrain categories varies proportionally according to the percentage of aerial coverage of all units. The terrain samples of each study site are divided into training subsets and test subsets in a ratio of 70:

30. Use the subset feature tool in ArcGIS to randomly perform this step; S3.3.4: GEOBIA image segmentation and classification. The GEOBIA method generates human visual artifact-free image objects. Image objects contain more detailed information than pixels. The GEOBIA method is applicable to high-resolution remote sensing images with heterogeneous pixel information. S3.3.5: Map summarization: After obtaining the topographic map, use the map summarization method to improve the overall readability and quality of the classification results. For areas less than 400m 2 The small classified artifacts were merged into adjacent classes, and the map was generalized using the Eliminate Study Site Part algorithm in ArcGIS software; S3.3.6: Accuracy assessment. The overall accuracy is the ratio of correctly classified samples to the total test samples. ArcGIS and R software are used to calculate the error matrix and accuracy assessment results.

7. The interactive digital twin method for deep-sea construction according to claim 1, characterized in that: The step S4 comprises the following steps: S4.1: The embedded sensors transmit different types of signals such as underwater pressure, underwater flow velocity, underwater flow direction, and terrain elevation in the form of matrix feedback through cables; S4.2: The feedback data is used to edit the corresponding environment in Unity in real time through the Update function.

Images