Mudflat topographic surveying and mapping method and system based on water surveying and mapping unmanned ship

Abstract

The invention relates to the technical field of unmanned ship surveying and mapping, in particular to a mud flat topography surveying and mapping method and system based on a water surveying and mapping unmanned ship, and the method comprises the following steps: S1, setting the area range of mud flat topography surveying and mapping; s2, collecting spectral information of different wavebands to generate water surface feature data of water surface brightness, texture and coverage states; s3, obtaining topographic data of the bottom of the mud flat, and carrying out real-time calibration on the sonar data; s4, performing manifold alignment on the water surface feature data and the sonar water bottom data; s5, constructing a three-dimensional terrain model of the mud flat area; and S6, comparing the generated three-dimensional terrain model with historical surveying and mapping data, and dynamically adjusting the terrain model through an adaptive calibration algorithm. According to the invention, through autonomous navigation, real-time data calibration and multi-dimensional data fusion technologies, high-precision, dynamic and visual surveying and mapping of the mud flat terrain are realized, and the data acquisition stability and the accuracy of the terrain model are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned vessel surveying and mapping, and in particular to a method and system for surveying and mapping tidal flat topography based on an unmanned vessel for water surveying and mapping. Background Art

[0002] Tidal flat topography mapping has important applications in marine engineering, environmental monitoring, resource management and other fields. Traditional tidal flat mapping methods usually rely on manually operated survey ships or remote sensing equipment to obtain data through manual observation or satellite imagery. However, these methods face problems such as limited data collection range, poor real-time performance, and low mapping accuracy. Especially in tidal flat areas where water levels change frequently and the terrain is complex, traditional mapping methods are difficult to meet the requirements of high precision and high efficiency. In recent years, with the rapid development of unmanned ship technology, the use of unmanned ships for tidal flat topography mapping has become a trend. Relying on the autonomous navigation of unmanned ships and a variety of mapping equipment, a certain degree of automated water mapping can be achieved.

[0003] However, existing technologies still face numerous technical challenges in the application of unmanned vessels in mudflat mapping. On the one hand, the surface and bottom characteristics of mudflat areas change frequently, and the complex hydrological conditions make the data collection process susceptible to factors such as water surface fluctuations, resulting in deviations in bottom data. Existing equipment struggles to achieve real-time calibration and high-precision mapping. On the other hand, the effective fusion of surface and bottom feature data and the accurate construction of three-dimensional terrain models remain key challenges. Traditional data processing methods are unable to accurately align and fuse surface and bottom data, resulting in a lack of integrity and three-dimensionality in the terrain model. Therefore, there is a significant need to develop a mudflat topographic mapping method and system based on an unmanned vessel for water mapping. Summary of the Invention

[0004] Based on the above objectives, the present invention provides a method and system for mapping mudflat topography based on an unmanned water mapping vessel.

[0005] A method for mapping tidal flat topography based on an unmanned water mapping vessel comprises the following steps:

[0006] S1: Based on the inertial navigation system and differential GPS module, the initial position of the unmanned vessel is determined and the area for mudflat topography mapping is set; the autonomous navigation control system is activated to navigate the unmanned vessel to the mapping starting point;

[0007] S2: During the navigation process of the unmanned vessel, the multispectral imaging equipment continuously scans the mudflat surface and collects spectral information of different bands to generate surface feature data of water brightness, texture and coverage status, providing a reference for subsequent underwater data calibration;

[0008] S3: Synchronously activate the sonar scanning device on the unmanned vessel during navigation to obtain the terrain data of the mudflat bottom, and use the water surface feature data to calibrate the sonar data in real time to eliminate deviations caused by water surface disturbances;

[0009] S4: Perform manifold alignment on the water surface feature data in S2 and the sonar bottom data in S3. The implicit mapping relationship between the water surface and bottom feature data is constructed through the manifold alignment algorithm to achieve alignment and fusion of the two sets of data in the same latent space.

[0010] S5: Based on the water surface and bottom feature data obtained through manifold alignment and fusion, differential operations are performed to generate contour maps and depth distribution maps of the tidal flat area, thereby constructing a three-dimensional terrain model of the tidal flat area and achieving three-dimensional visualization of the tidal flat area.

[0011] S6: Compare the generated 3D terrain model with historical surveying and mapping data, dynamically adjust the terrain model through an adaptive calibration algorithm, analyze and calibrate the changing trend of the tidal flat terrain to ensure the accuracy and timeliness of the model.

[0012] Optionally, the S1 specifically includes:

[0013] S11: The inertial navigation system on the unmanned vessel collects the speed, acceleration, and heading data of the unmanned vessel's current position, and makes a preliminary position estimate based on the collected data to determine the approximate position range of the unmanned vessel, providing a basis for position calibration;

[0014] S12: Fusing the unmanned ship position data initially estimated by the inertial navigation system with the high-precision geographic coordinates obtained by the differential GPS module, accurately calibrating the unmanned ship position through the real-time correction signal of the differential GPS module to obtain the high-precision initial position of the unmanned ship;

[0015] S13: Based on the position data provided by the differential GPS, the target boundary coordinates of the tidal flat topography mapping area are input, the boundary of the mapping area is established through the navigation control system of the unmanned vessel, and a mapping route plan including the starting point is generated;

[0016] S14: Load the boundary coordinates of the surveying area and the preset navigation route, enable the path tracking module to match the real-time position of the unmanned vessel with the target path, and adjust the navigation direction of the unmanned vessel in real time to ensure that the unmanned vessel navigates to the surveying starting point;

[0017] S15: According to the dynamic adjustment instructions of the path tracking module, the unmanned vessel automatically navigates along the set route to the starting point of the surveying area, and records the arrival position of the unmanned vessel to verify the accuracy of autonomous navigation and provide a starting point benchmark for subsequent surveying.

[0018] Optionally, the S2 specifically includes:

[0019] S21: After the unmanned vessel enters the mudflat mapping area, it starts the multispectral imaging equipment and sets the target spectral band parameters, including visible light of 400-700nm, near-infrared light of 700-900nm, and short-wave infrared light of 900-1700nm, to ensure that the reflection information of the water surface in different spectral ranges is captured;

[0020] S22: When the unmanned vessel is traveling along the preset route, the continuous scanning function of the multispectral imaging device is activated to collect multi-band image data of various areas on the mudflat surface in real time using the set spectral bands, and ensure that the scanning rate matches the speed of the unmanned vessel;

[0021] S23: Separate the multi-band spectral data obtained by scanning into bands one by one, and extract the corresponding brightness, texture and coverage feature information in each band;

[0022] S24: The separated preliminary feature data of each band are fused, and feature association processing is performed according to the band weight distribution to generate comprehensive water surface feature data including water surface brightness, texture and coverage status.

[0023] Optionally, the S23 specifically includes:

[0024] S231: Separate the multi-band spectral data collected by the multispectral imaging device by band, extract various image data including visible light, near infrared, and short-wave infrared bands; and perform convolution operation on the spectral intensity data of each band to ensure that the data of each band is independently separated;

[0025] S232: In the visible light band image, the brightness value of each position is determined by calculating the red, green, and blue light intensity values of the image to analyze the distribution characteristics of the water surface reflection intensity; and the brightness data is statistically analyzed to generate a brightness histogram feature;

[0026] S233: Analyze the texture characteristics of the water surface in near-infrared band images based on the gray-level co-occurrence matrix method. First, the co-occurrence matrix is calculated based on the predetermined pixel spacing and direction angle. Then, a series of texture features are extracted from the co-occurrence matrix, including energy, contrast, entropy, and uniformity, to describe the ripples and surface structure of the water surface.

[0027] S234: In the short-wave infrared band image, by calculating the ratio of the reflected light to the incident light intensity at each position on the water surface and setting a threshold based on the difference in reflectivity between the water surface and the covering object, the data is binarized to identify the coverage status of sediments and floating objects on the water surface and generate coverage feature data.

[0028] Optionally, the S3 specifically includes:

[0029] S31: After the unmanned vessel enters the mudflat mapping area, it activates the sonar scanning device and sets measurement parameters such as scanning frequency, depth range, and echo sampling rate to ensure continuous scanning of the mudflat bottom within the designated route and collect underwater terrain reflection data;

[0030] S32: During the navigation of the unmanned vessel, the sonar scanning device continuously transmits acoustic wave signals and receives the bottom reflected echo signals, converts the bottom echo intensity and echo delay time into bottom depth data, and generates a real-time bottom terrain information matrix D(x, y), where x and y are position coordinates;

[0031] S33: Using the water surface brightness, texture, and coverage characteristic data obtained in step S2, the sonar echo is calibrated in real time to eliminate the echo deviation caused by water surface disturbances. The calibration factor C is determined by comparing the actual path of the sonar signal with the ideal path. The calculation formula is: Among them, V actual Indicates the echo speed of the sonar signal after being disturbed by the water surface, V ideal Indicates the ideal echo velocity under undisturbed conditions; the calibration factor C is applied to the bottom depth data at each location to update the depth value D calibrated (x, y), the formula is: D calibrated (x, y) = D(x, y) × (1-C);

[0032] S34: The calibrated bottom depth data D calibrated )x, y) is stored as calibrated terrain data and is continuously updated within the surveying area to ensure the accuracy of the underwater terrain data during the navigation of the unmanned vessel.

[0033] Optionally, the S4 specifically includes:

[0034] S41: normalizing the water surface feature data obtained in S2 and the bottom feature data obtained in S3 respectively, converting all feature data into the same numerical range to ensure that the data have the same dimension;

[0035] S42: Mapping the normalized water surface feature data and bottom feature data to a high-dimensional feature space to generate a water surface feature vector and a bottom feature vector, respectively. Each position coordinate corresponds to a feature vector, including brightness, texture, coverage status, and bottom depth information, to ensure that the two sets of data have a corresponding relationship in the high-dimensional space.

[0036] S43: Using the manifold alignment algorithm, the implicit mapping relationship between the two sets of feature data is constructed by minimizing the distance between the feature vectors of the water surface and the bottom;

[0037] S44: Projecting the water surface feature data and the bottom feature data after the mapping relationship is constructed into a common latent space, and performing data fusion in the space. The fused feature vector includes the combined features of the water surface brightness, texture, coverage status and bottom depth information, forming the comprehensive terrain feature data of the mudflat area;

[0038] S45: Output the fused feature data according to the position coordinates to generate a unified terrain feature data matrix of the tidal flat area, providing complete input data for the subsequent three-dimensional terrain model construction.

[0039] Optionally, the S5 specifically includes:

[0040] S51: Based on the water surface and bottom feature data after manifold alignment and fusion, obtain the water surface elevation value and the bottom depth value of each position coordinate; and subtract the water surface elevation value from the bottom depth value to calculate the relative height difference of each coordinate point;

[0041] S52: Based on the calculated relative height difference, a contour generation algorithm is used to classify the Nuotu area and divide the regional boundaries of different height levels; and a linear interpolation method is used to smooth the height difference between adjacent coordinate points to generate continuous contour lines, thereby forming a contour map of the Nuotu area;

[0042] S53: generating a depth distribution map of the mudflat area based on the distribution of the bottom depth values; classifying the depth distribution of each position coordinate, and marking different depth areas by color coding;

[0043] S54: Combine the contour map and depth distribution map to construct the water surface elevation value and bottom depth value of each coordinate point into a three-dimensional coordinate point. Use the grid reconstruction algorithm to fit the three-dimensional coordinate points to form a continuous terrain surface to construct a complete three-dimensional model of the mudflat area.

[0044] 55: Input the 3D terrain model into the visualization system, enhance the three-dimensional sense of the terrain model through coloring and lighting effects, and generate a three-dimensional visualization image containing height, depth and contour information.

[0045] Optionally, the S54 specifically includes:

[0046] S541: Obtain the water surface elevation value and the water bottom depth value of each position coordinate from the contour map and the depth distribution map of the mudflat area; and construct a three-dimensional coordinate point based on the water surface and water bottom height information;

[0047] S542: performing initial grid division on all three-dimensional coordinate points in the tidal flat area at a preset distance, dividing the tidal flat area into a number of adjacent grid cells; each grid cell contains a number of three-dimensional coordinate points;

[0048] S543: Perform plane fitting on the three-dimensional coordinate points within each grid cell using the least squares method;

[0049] S544: After the grid cell fitting is completed, bilinear interpolation is performed on the boundaries of adjacent grid cells to make the boundary height values of adjacent grid cells transition smoothly and eliminate discontinuities;

[0050] S545: Integrate the fitting and interpolation results of all grid cells to construct a complete three-dimensional terrain model containing all three-dimensional coordinate points of the mudflat area.

[0051] Optionally, the S6 specifically includes:

[0052] S61: Compare the generated three-dimensional terrain model of the tidal flat area with the historical surveying and mapping data, extract the current height value and the historical height value at each position coordinate, and identify terrain changes by calculating the difference Δz(x, y) between the current height value and the historical height value;

[0053] S62: Based on the height difference Δz(x, y) of each coordinate point, the terrain change is classified into different modes, including rising, falling, and stable. The change type is determined by setting a threshold T: when Δz(x, y)>T, it is classified as rising; if Δz(x, y)<-T, it is classified as falling; if |Δz(x, y)|≤T, it is classified as stable;

[0054] S63: Use the adaptive calibration algorithm to dynamically adjust the area with significant changes; for each coordinate point, apply the adaptive calibration coefficient and update the current height value to make it close to the historical data or in line with the change trend; the updated height value z adjusted The formula for (x, y) is:

[0055] z adjusted (x, y) = z current (x, y) + α·Δz(x, y), where α is an adaptive calibration coefficient with a value range of 0 < α < 1 and is automatically adjusted according to the change pattern;

[0056] S64: All adjusted height values z adjusted (x, y) integration to generate a calibrated 3D terrain model.

[0057] A tidal flat topography mapping system based on an unmanned water surveying vessel is used to implement the above-mentioned tidal flat topography mapping method based on an unmanned water surveying vessel, and includes the following modules:

[0058] Positioning and navigation module: including inertial navigation device and differential GPS unit, used to determine the initial position and navigation path of the unmanned vessel in the mudflat area, set the boundary of the survey area, and ensure that the unmanned vessel navigates to the survey starting point;

[0059] Water surface feature acquisition module: includes multispectral imaging equipment, which is used to collect multi-band spectral information of the mudflat water surface during the navigation of the unmanned vessel, and obtain water surface feature data such as brightness, texture and coverage status of the water surface;

[0060] Underwater terrain scanning module: Contains a sonar scanning device for synchronously collecting bottom terrain depth data in the mudflat area and using water surface feature data for real-time calibration to eliminate the impact of water surface disturbances on bottom data;

[0061] Data processing module: used to receive and process water surface feature data and underwater terrain data, construct an implicit mapping relationship between water surface and underwater feature data through manifold alignment algorithm, and generate aligned comprehensive terrain feature data;

[0062] 3D model construction module: used to generate contour maps and depth distribution maps of the tidal flat area through differential calculation based on the comprehensive terrain feature data output by the data processing module, and to generate an initial 3D terrain model of the tidal flat area based on 3D coordinate points;

[0063] Dynamic calibration module: Based on the initial 3D terrain model generated by the 3D model construction module, combined with historical surveying and mapping data, the model is dynamically adjusted through an adaptive calibration algorithm. By comparing the differences between the current terrain data and the historical data, the terrain model is updated and the calibrated model is transmitted to the display module;

[0064] Display module: used to receive the calibrated three-dimensional terrain model output by the dynamic calibration module, and visualize the three-dimensional terrain information of the mudflat area, including height, depth and contour changes, to provide real-time terrain data support for surveying and mapping personnel.

[0065] Beneficial effects of the present invention:

[0066] The present invention effectively solves the problem of environmental interference in water surface and bottom feature data in mudflat topography mapping by introducing technologies such as unmanned boat autonomous navigation, real-time data calibration, and fusion of multispectral imaging and sonar scanning. It collects water surface brightness, texture, and coverage status and other features through multispectral imaging, combines sonar scanning to obtain bottom topography data, and performs data fusion through manifold alignment algorithm. It can accurately align and eliminate the influence of water surface disturbance on bottom data, greatly improving the stability of data acquisition and mapping accuracy. At the same time, the real-time surface and bottom data calibration mechanism enables the mapping results to quickly respond to dynamic changes in the mudflat environment, ensuring the reliability of mapping in complex environments.

[0067] The present invention generates contour maps and depth distribution maps through high-precision differential operations. Combined with the automatic construction of three-dimensional terrain models and adaptive calibration algorithms, it can accurately reconstruct the three-dimensional terrain of the tidal flat area and realize complete three-dimensional visualization of the terrain of the tidal flat area. This method can not only intuitively display the terrain undulations and depth changes, but also has the ability to dynamically update and analyze change trends. It provides high-precision terrain data support for marine environmental monitoring and engineering planning, and effectively enhances the application value of tidal flat terrain mapping. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0069] Figure 1 Schematic diagram of a method for surveying and mapping tidal flat topography according to an embodiment of the present invention;

[0070] Figure 2 Schematic diagram of a tidal flat topography mapping system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0072] It should be noted that references in the specification to "one embodiment," "an embodiment," "an exemplary embodiment," "some embodiments," etc. indicate that the described embodiments may include specific features, structures, or characteristics, but not every embodiment necessarily includes such specific features, structures, or characteristics. In addition, when specific features, structures, or characteristics are described in conjunction with an embodiment, it is within the knowledge of persons skilled in the relevant art to implement such features, structures, or characteristics in conjunction with other embodiments (whether or not explicitly described).

[0073] In general, terms can be understood, at least in part, from their use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending at least in part on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0074] like Figure 1 As shown, a method for mapping mudflat terrain based on an unmanned water mapping vessel includes the following steps:

[0075] S1: Based on the inertial navigation system and differential GPS module, the initial position of the unmanned vessel is determined and the area for mudflat topography mapping is set; the autonomous navigation control system is activated to navigate the unmanned vessel to the mapping starting point;

[0076] S2: During the navigation process of the unmanned vessel, the multispectral imaging equipment continuously scans the mudflat surface and collects spectral information of different bands to generate surface feature data of water brightness, texture and coverage status, providing a reference for subsequent underwater data calibration;

[0077] S3: Synchronously activate the sonar scanning device on the unmanned vessel during navigation to obtain the terrain data of the mudflat bottom, and use the water surface feature data to calibrate the sonar data in real time to eliminate deviations caused by water surface disturbances;

[0078] S4: Perform manifold alignment on the water surface feature data in S2 and the sonar bottom data in S3. An implicit mapping relationship between the water surface and bottom feature data is constructed through the manifold alignment algorithm, achieving alignment and fusion of the two sets of data in the same latent space, thereby enhancing the recognition accuracy of the terrain features in the mudflat area.

[0079] S5: Based on the water surface and bottom feature data obtained through manifold alignment and fusion, differential operations are performed to generate contour maps and depth distribution maps of the tidal flat area, thereby constructing a three-dimensional terrain model of the tidal flat area and achieving three-dimensional visualization of the tidal flat area.

[0080] S6: Compare the generated 3D terrain model with historical surveying and mapping data, dynamically adjust the terrain model through an adaptive calibration algorithm, analyze and calibrate the changing trend of the tidal flat terrain to ensure the accuracy and timeliness of the model.

[0081] S1 specifically includes:

[0082] S11: The inertial navigation system on the unmanned vessel collects the speed, acceleration, and heading data of the unmanned vessel's current position, and makes a preliminary position estimate based on the collected data to determine the approximate position range of the unmanned vessel, providing a basis for position calibration;

[0083] S12: Fusing the unmanned ship position data initially estimated by the inertial navigation system with the high-precision geographic coordinates obtained by the differential GPS module, accurately calibrating the unmanned ship position through the real-time correction signal of the differential GPS module to obtain the high-precision initial position of the unmanned ship;

[0084] S13: Based on the position data provided by the differential GPS, the target boundary coordinates of the tidal flat topography mapping area are input, the boundary of the mapping area is established through the navigation control system of the unmanned vessel, and a mapping route plan including the starting point is generated to cover the entire tidal flat topography mapping range;

[0085] S14: Load the boundary coordinates of the surveying area and the preset navigation route, enable the path tracking module to match the real-time position of the unmanned vessel with the target path, and adjust the navigation direction of the unmanned vessel in real time to ensure that the unmanned vessel navigates to the surveying starting point;

[0086] S15: According to the dynamic adjustment instructions of the path tracking module, the unmanned vessel automatically navigates to the starting point of the surveying area along the set route, and records the arrival position of the unmanned vessel at the same time to verify the accuracy of autonomous navigation and provide a starting point benchmark for subsequent surveying. Through the above steps, the unmanned vessel can use the inertial navigation system and differential GPS module to accurately determine the initial position, and automatically navigate to the surveying starting point with the support of the autonomous navigation control system, realizing high-precision setting and stable positioning of the mudflat terrain surveying area.

[0087] S2 specifically includes:

[0088] S21: After the unmanned vessel enters the mudflat mapping area, it starts the multispectral imaging equipment and sets the target spectral band parameters, including visible light of 400-700nm, near-infrared light of 700-900nm, and short-wave infrared light of 900-1700nm, to ensure that the reflection information of the water surface in different spectral ranges is captured, thereby enhancing the distinction between the brightness, texture and coverage status of the water surface;

[0089] S22: As the unmanned vessel travels along the preset route, the continuous scanning function of the multispectral imaging device is activated to collect multi-band image data of various areas on the mudflat surface in real time using the set spectral bands. The scanning rate is ensured to match the speed of the unmanned vessel to obtain seamless multi-band spectral data coverage.

[0090] S23: Separate the multi-band spectral data obtained by scanning band by band, and extract the corresponding brightness, texture, and coverage feature information in each band. Specifically, the visible light band data is used for brightness analysis to detect the reflection intensity of the water surface; the near-infrared band data is used for texture detection to distinguish water surface ripples; and the short-wave infrared band data is used for coverage state analysis to identify sediments and floating objects on the water surface, thereby forming preliminary feature data for each band.

[0091] S24: The separated preliminary feature data of each band are fused and feature association processing is performed according to the band weight distribution to generate comprehensive water surface feature data including water surface brightness, texture and coverage status, so as to ensure that the data can accurately reflect the overall characteristics of the water surface and provide stable reference data for the subsequent calibration of underwater data; through the above steps, the multispectral imaging equipment can accurately collect multi-band spectral information of the mudflat water surface during the navigation of the unmanned ship, analyze the water surface brightness, texture and coverage status characteristics in detail, and generate complete water surface feature data, thereby providing a solid data foundation for the calibration of underwater data and terrain reconstruction in mudflat topography mapping.

[0092] S23 specifically includes:

[0093] S231: Separate the multi-band spectral data collected by the multispectral imaging device by band, extract various image data including visible light, near infrared, and short-wave infrared bands; and perform convolution operation on the spectral intensity data of each band to ensure that the data of each band is independently separated;

[0094] S232: In the visible light band image, the brightness value of each position is determined by calculating the red, green, and blue light intensity values of the image to analyze the distribution characteristics of the water surface reflection intensity; and the brightness data is statistically analyzed to generate a brightness histogram feature to represent the brightness distribution of the water surface in the visible light band;

[0095] S233: Analyze the texture characteristics of the water surface in near-infrared band images based on the gray-level co-occurrence matrix method. First, the co-occurrence matrix is calculated based on the predetermined pixel spacing and direction angle. Then, a series of texture features are extracted from the co-occurrence matrix, including energy, contrast, entropy, and uniformity, to describe the ripples and surface structure of the water surface.

[0096] S234: In the short-wave infrared band image, by calculating the ratio of the reflected light to the incident light intensity at each position on the water surface and setting a threshold based on the difference in reflectivity between the water surface and the covering object, the data is binarized to identify the coverage status of sediments and floating objects on the water surface and generate coverage feature data.

[0097] The specific steps to extract the corresponding brightness, texture and coverage features in each band are as follows:

[0098] Separation and processing of multi-band spectral data: The multi-band spectral data collected by the multi-spectral imaging device is separated by band, the image data of the visible light band, near infrared band and short-wave infrared band are extracted, and the spectral intensity matrix I is processed by convolution operation. λ (x, y), where I λ(x, y) represents the spectral intensity value at position (x, y), λ is the spectral band, and x and y represent the horizontal and vertical coordinates of the image, respectively, to ensure that each band data is independently separated;

[0099] Extraction of brightness features: In the visible light band image, brightness calculation is used to extract the water surface reflection intensity. The formula is: Where L(x, y) represents the brightness value at the position (x, y), R(x, y), G(x, y), and B(x, y) represent the light intensity values of the red, green, and blue components at the position (x, y), respectively. Statistical analysis is performed on the brightness matrix L(x, y) to obtain the brightness histogram features to characterize the reflection intensity distribution of the water surface in the visible light band.

[0100] Extraction of texture features: In the near-infrared band image, the gray-level co-occurrence matrix is used to analyze the texture features of the water surface. First, the gray-level co-occurrence matrix P(d, θ) is calculated, where P(d, θ) represents the co-occurrence probability of gray values with a spacing of d and a direction angle of θ. Then, the following texture feature values are extracted from P(d, θ):

[0101] Energy E: Where i and j represent grayscale values, and N is the number of grayscale levels;

[0102] Contrast C:

[0103] Entropy H:

[0104] Uniformity U: These texture feature values E, C, H, and U are used to describe the ripples and surface structure states of the water surface;

[0105] Extraction of coverage features: In shortwave infrared band images, the water surface coverage status is identified by reflectivity calculation; the reflectivity calculation formula is: Where R(x, y) represents the reflectivity at position (x, y, I ref (x, y) is the reflected light intensity at position (x, y), I inc (x, y) is the incident light intensity at position (x, y); then, based on the reflectivity difference between the water surface and the covering object, a threshold is set for binarization processing to identify the coverage status of sediments and floating objects on the water surface and generate a coverage feature matrix; through the above steps, after the multi-band spectral data is separated by band, the brightness, texture and coverage feature information of the water surface in each band can be extracted, providing accurate water surface basic data for mudflat topography mapping.

[0106] S3 specifically includes:

[0107] S31: After the unmanned vessel enters the mudflat mapping area, it activates the sonar scanning device and sets measurement parameters such as scanning frequency, depth range, and echo sampling rate to ensure continuous scanning of the mudflat bottom within the designated route and collect underwater terrain reflection data;

[0108] S32: During the navigation of the unmanned vessel, the sonar scanning device continuously transmits acoustic wave signals and receives the bottom reflected echo signals, converts the bottom echo intensity and echo delay time into bottom depth data, and generates a real-time bottom terrain information matrix D(x, y), where x and y are position coordinates;

[0109] S33: Using the water surface brightness, texture, and coverage characteristic data obtained in step S2, the sonar echo is calibrated in real time to eliminate the echo deviation caused by water surface disturbances. By comparing the actual path of the sonar signal with the ideal path, the calibration factor C is determined. This factor represents the degree of influence of water surface disturbances. The calculation formula is: Among them, V actual Indicates the echo speed of the sonar signal after being disturbed by the water surface, V ideal Indicates the ideal echo velocity under undisturbed conditions; the calibration factor C is applied to the bottom depth data at each location to update the depth value D calibrated (x, y), the formula is: D calibrated (x, y) = D(x, y) × (1-C). Through the above calibration process, the bottom depth data D(x, y) is adjusted to remove the deviation caused by water surface disturbance;

[0110] S34: The calibrated bottom depth data D calibrated (x, y) is stored as calibrated terrain data and continuously updated within the surveying area to ensure the accuracy of the underwater terrain data during the navigation of the unmanned vessel. Through the above steps, the sonar scanning device can efficiently collect mudflat underwater terrain data during the navigation of the unmanned vessel, and use water surface feature data for real-time calibration to eliminate the influence of water surface disturbances and generate high-precision underwater terrain data.

[0111] S4 specifically includes:

[0112] S41: normalizing the water surface feature data obtained in S2 and the bottom feature data obtained in S3 respectively, converting all feature data into the same numerical range to ensure that the data have the same dimension so as to facilitate manifold alignment in subsequent steps;

[0113] S42: Mapping the normalized surface feature data and bottom feature data to a high-dimensional feature space to generate a surface feature vector and a bottom feature vector, respectively. Each position coordinate corresponds to a feature vector, including brightness, texture, coverage (from the surface data) and bottom depth information (from the sonar data), ensuring that the two sets of data have a corresponding relationship in the high-dimensional space.

[0114] S43: Using the manifold alignment algorithm, the implicit mapping relationship between the two sets of feature data is constructed by minimizing the distance between the feature vectors of the water surface and the bottom;

[0115] The specific calculation steps of the manifold alignment algorithm are as follows:

[0116] Define feature data set and high-dimensional feature space representation: Assume that the water surface feature data set is M1 = {x1, x2, ..., x n}, the bottom feature dataset is M2={y1,y2,…,y n}, where x i and y i are the water surface and bottom feature vectors, respectively, where i = 1, 2, …, n represents the number of sample points. Each feature vector contains brightness, texture, and coverage information in the visible, near-infrared, and short-wave infrared bands, as well as bottom depth data, ensuring high-dimensional representation of feature data.

[0117] Establish similarity matrix: Construct similarity matrices W and Z for the water surface feature dataset and the bottom feature dataset; for the water surface feature dataset, define the element W in the similarity matrix W ij for: Among them, ||x i -x j || represents the data point x i and x j The Euclidean distance between them, σ is the scale parameter of the similarity matrix; similarly, the element Z in the similarity matrix Z of the bottom feature dataset is ij for: These similarity matrices W and Z are used to measure the local similarity between data points to construct local neighborhood relationships;

[0118] Construction of Laplace matrix: Based on the similarity matrices W and Z, the Laplace matrices of the water surface and bottom feature data sets are calculated; the Laplace matrix L of the water surface feature data set x Calculated as: L x =D x -W, where D x is a diagonal matrix whose diagonal elements D ii is equal to the sum of all elements in the i-th row of W; similarly, the Laplace matrix L of the bottom feature dataset y For: Ly =D y -Z, where D y is the diagonal matrix, D ii is the sum of the elements in the i-th row of Z;

[0119] Minimization problem of mapping function: Define the mapping function f:M1→M2 to minimize the alignment error of surface and bottom feature data in the latent space; the objective function can be expressed as: min f ∑ i,j (f(x i )-f(y j )) 2 (L x +L y ), where f is a mapping function used to map the water surface feature data M1 to the potential space of the bottom feature data M2 to achieve alignment of the two sets of feature data; M1 is a water surface feature data set consisting of multiple water surface feature vectors x i M2 is a water bottom feature dataset, consisting of multiple water bottom feature vectors y j Composition; x i is the i-th feature vector in the water surface feature dataset, which contains brightness, texture and coverage status information; j is the jth feature vector in the bottom feature dataset, which contains terrain information such as bottom depth; i, j are indexes, representing the data points in the water surface and bottom feature datasets respectively; (f(x i )-f(y j )) 2 Represents the mapping function f acting on the water surface feature vector x i and the bottom feature vector y j The square error calculated later is used to measure the distance between the mapped data points; L x is the Laplace matrix of the water surface feature dataset, which is used to represent the local similarity relationship of the water surface data; L y is the Laplacian matrix of the bottom feature dataset, which is used to represent the local similarity relationship of the bottom data. In this formula, the distance between the mapped data points is minimized to ensure the alignment of the surface and bottom feature data in the latent space. The parameters of f are adjusted to minimize the distance error between the surface and bottom data in the latent space to achieve feature data alignment.

[0120] Calculate low-dimensional representation: After solving the above minimization problem, we obtain the mapping function f, which projects the water surface feature data and the bottom feature data into the same latent space to generate aligned low-dimensional feature representations for fusion in subsequent steps.

[0121] S44: Projecting the water surface feature data and the bottom feature data after the mapping relationship is constructed into a common latent space, and performing data fusion in the space. The fused feature vector includes the combined features of the water surface brightness, texture, coverage status and bottom depth information, forming the comprehensive terrain feature data of the mudflat area;

[0122] S45: The fused feature data is output according to the position coordinates to generate a unified terrain feature data matrix for the mudflat area, providing complete input data for the subsequent construction of the three-dimensional terrain model; through the above steps, the manifold alignment algorithm aligns and fuses the water surface and bottom feature data in the same latent space, constructing an implicit mapping relationship between the two sets of data, thereby generating high-precision terrain feature data containing comprehensive features of the water surface and bottom, providing a complete foundation for terrain modeling of the mudflat area.

[0123] S5 specifically includes:

[0124] S51: Based on the water surface and bottom feature data after manifold alignment and fusion, obtain the water surface elevation value H of each position coordinate s (x, y) and the bottom depth value H b (x, y); and subtract the water surface elevation value from the water bottom depth value to calculate the relative height difference ΔH(x, y) of each coordinate point. The formula is: ΔH(x, y) = H s (x,y)-H b (x, y), the relative height difference ΔH(x, y) characterizes the topographic relief of the mudflat area and provides a data basis for generating contour maps and depth distribution maps;

[0125] S52: Based on the calculated relative height difference ΔH(x, y), a contour generation algorithm is used to classify the Nuotu area and delineate the boundaries of regions at different height levels. A linear interpolation method is then used to smooth the height differences between adjacent coordinate points to generate continuous contour lines, thereby forming a contour map of the Nuotu area to display the distribution of the terrain undulations.

[0126] S53: Generate a depth distribution map of the mudflat area based on the distribution of bottom depth values; classify the depth distribution of each location coordinate and mark different depth areas through color coding to ensure intuitive presentation of depth information and provide reference data for subsequent three-dimensional terrain model construction;

[0127] S54: Combine the contour map and depth distribution map to construct the water surface elevation value and bottom depth value of each coordinate point into a three-dimensional coordinate point. Use the grid reconstruction algorithm to fit the three-dimensional coordinate points to form a continuous terrain surface to construct a complete three-dimensional model of the mudflat area.

[0128] 55: Input the three-dimensional terrain model into the visualization system, enhance the three-dimensional sense of the terrain model through coloring and light and shadow effects, generate a three-dimensional visualization image containing height, depth and contour information, realize the intuitive display of the terrain of the tidal flat area, and provide high-precision three-dimensional visualization support for tidal flat terrain mapping; through the above steps, based on the water surface and bottom feature data fused by manifold alignment, generate the contour map and depth distribution map of the tidal flat area through differential operation, and construct a three-dimensional terrain model, realize the complete three-dimensional visualization display of the tidal flat area, and provide accurate terrain information.

[0129] The complete 3D model of the tidal flat area constructed in S54 specifically includes:

[0130] S541: Obtain the water surface elevation value H for each position coordinate (x, y) from the contour map and depth distribution map of the mudflat area s (x, y) and the bottom depth H b (x, y); and according to the height information of the water surface and the bottom, construct a three-dimensional coordinate point (x, y, z), where z represents the height value; for the water surface coordinate point, let z = H s (x, y); for the bottom coordinate point, z = H b (x, y); after constructing the three-dimensional coordinate points, the spatial data of the elevation and depth of the mudflat area are formed, providing input for subsequent grid fitting;

[0131] S542: Initially meshing all three-dimensional coordinate points (x, y, z) in the tidal flat area at a preset distance d, dividing the tidal flat area into a number of adjacent grid cells; each grid cell contains a number of three-dimensional coordinate points, ensuring that coordinate points in the same area are reasonably classified for local fitting;

[0132] S543: For each three-dimensional coordinate point (x i ,y i , z i ) Use the least squares method to fit the plane; assume that the plane equation is: z = ax + by + c, where x and y represent the coordinates on the plane, z is the height, and a, b, and c are the coefficients of the fitting plane equation; the fitting process is performed by minimizing the error function E, which is expressed as: Among them, z i is the height value of the i-th three-dimensional coordinate point, x i and y i is its horizontal coordinate, n represents the total number of coordinate points in the grid unit, and the fitting plane equation is obtained by solving the optimal values of a, b and c to achieve smooth fitting of the coordinate points in the grid;

[0133] S544: After the grid cell fitting is completed, bilinear interpolation is performed on the boundaries of adjacent grid cells to make the boundary height values of adjacent grid cells transition smoothly and eliminate discontinuities. The bilinear interpolation formula is:

[0134] z(x,y)=(1-α)(1-β)z 00 +α(1-β)z 10 +(1-α)βz 01 +αβz 11 , where z(x, y) represents the height value after interpolation, z 00 、z 10 、z 01 and z 11 are the height values of the four vertices of the interpolation area, α and β are the interpolation weights in the x and y directions, respectively. A continuous and smooth terrain surface is generated by interpolation;

[0135] S545: Integrate the fitting and interpolation results of all grid cells to construct a complete three-dimensional terrain model containing all three-dimensional coordinate points of the tidal flat area. This model fully presents the terrain undulations of the tidal flat area, including contour changes and depth distribution, providing a basis for three-dimensional visualization of the tidal flat area. Through the above steps, based on the manifold-aligned and fused water surface and bottom feature data, three-dimensional coordinate points are constructed through differential operations, and a grid reconstruction algorithm is used for fitting and smooth interpolation to construct a complete three-dimensional terrain model of the tidal flat area, achieving the continuity and three-dimensional visualization of the terrain.

[0136] S6 specifically includes:

[0137] S61: Compare the generated 3D terrain model of the tidal flat area with the historical surveying and mapping data, and extract the current height value z at each position coordinate (x, y) current (x, y) and historical height value z history (x, y), and the terrain change is identified by calculating the difference Δz(x, y) between the current altitude value and the historical altitude value. The formula is: Δz(x, y) = z current (x, y)-z history (x, y), where Δz(x, y) represents the change amount and is used to analyze the changing trend of the terrain;

[0138] S62: Based on the height difference value Δz(x, y) of each coordinate point, the terrain change is divided into different modes, including rising, falling, and stable. The change type is determined by setting a threshold T: when Δz(x, y)>T, it is classified as rising; if Δz(x, y)<-T, it is classified as falling; if |Δz(x, y)|≤T, it is classified as stable. The change mode label is attached to the corresponding coordinate point to provide a classification basis for subsequent calibration.

[0139] S63: Use the adaptive calibration algorithm to dynamically adjust the area with significant changes; for each coordinate point, apply the adaptive calibration coefficient and update the current height value to make it close to the historical data or in line with the change trend; the updated height value z adjusted The formula for (x, y) is:

[0140] z adjusted (x, y) = z current (x, y) + α·Δz(x, y), where α is an adaptive calibration coefficient with a value range of 0 < α < 1. It is automatically adjusted according to the change pattern to smoothly follow the terrain change trend and avoid over-adjustment or data fluctuation;

[0141] S64: All adjusted height values z adjusted (x, y) is integrated to generate a calibrated three-dimensional terrain model; this model contains adjusted terrain data of the tidal flat area, reflecting the changing trends and current status of the terrain, and providing accurate model support for dynamic terrain mapping of the tidal flat area; through the above steps, the adaptive calibration algorithm can compare the three-dimensional terrain model with historical mapping data, dynamically adjust the terrain model and accurately reflect the changing trends of the tidal flat terrain, generating a calibrated terrain model, providing high-precision support for terrain monitoring in the tidal flat area.

[0142] like Figure 2 As shown, a tidal flat topography mapping system based on an unmanned water mapping vessel is used to implement the above-mentioned tidal flat topography mapping method based on an unmanned water mapping vessel, and includes the following modules:

[0143] Positioning and navigation module: including inertial navigation device and differential GPS unit, used to determine the initial position and navigation path of the unmanned vessel in the mudflat area, set the boundary of the survey area, and ensure that the unmanned vessel navigates to the survey starting point;

[0144] Water surface feature acquisition module: includes multispectral imaging equipment, which is used to collect multi-band spectral information of the mudflat water surface during the navigation of the unmanned vessel, and obtain water surface feature data such as brightness, texture and coverage status of the water surface;

[0145] Underwater terrain scanning module: Contains a sonar scanning device for synchronously collecting bottom terrain depth data in the mudflat area and using water surface feature data for real-time calibration to eliminate the impact of water surface disturbances on bottom data;

[0146] Data processing module: used to receive and process water surface feature data and underwater terrain data, construct an implicit mapping relationship between water surface and underwater feature data through manifold alignment algorithm, and generate aligned comprehensive terrain feature data;

[0147] 3D model construction module: used to generate contour maps and depth distribution maps of the tidal flat area through differential calculation based on the comprehensive terrain feature data output by the data processing module, and to generate an initial 3D terrain model of the tidal flat area based on 3D coordinate points;

[0148] Dynamic calibration module: Based on the initial 3D terrain model generated by the 3D model construction module, combined with historical surveying and mapping data, the model is dynamically adjusted through an adaptive calibration algorithm. By comparing the differences between the current terrain data and the historical data, the terrain model is updated and the calibrated model is transmitted to the display module;

[0149] Display module: used to receive the calibrated three-dimensional terrain model output by the dynamic calibration module, and visualize the three-dimensional terrain information of the mudflat area, including height, depth and contour changes, to provide real-time terrain data support for surveying and mapping personnel.

[0150] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0151] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for mapping mudflat terrain based on an unmanned water mapping vessel, characterized in that: The following steps are involved: S1: Based on the inertial navigation system and differential GPS module, the initial position of the unmanned vessel is determined and the area for mudflat topography mapping is set; the autonomous navigation control system is activated to navigate the unmanned vessel to the mapping starting point; S2: During the navigation process of the unmanned vessel, the multispectral imaging equipment continuously scans the mudflat surface and collects spectral information of different bands to generate surface feature data of water brightness, texture and coverage status, providing a reference for subsequent underwater data calibration; S3: Synchronously activate the sonar scanning device on the unmanned vessel during navigation to obtain the terrain data of the mudflat bottom, and use the water surface feature data to calibrate the sonar data in real time to eliminate deviations caused by water surface disturbances; S4: Perform manifold alignment on the water surface feature data in S2 and the sonar bottom data in S3. The implicit mapping relationship between the water surface and bottom feature data is constructed through the manifold alignment algorithm to achieve alignment and fusion of the two sets of data in the same latent space. S5: Based on the water surface and bottom feature data obtained through manifold alignment and fusion, differential operations are performed to generate contour maps and depth distribution maps of the tidal flat area, thereby constructing a three-dimensional terrain model of the tidal flat area and achieving three-dimensional visualization of the tidal flat area. S6: Compare the generated 3D terrain model with historical surveying and mapping data, dynamically adjust the terrain model through an adaptive calibration algorithm, analyze and calibrate the changing trend of the tidal flat terrain to ensure the accuracy and timeliness of the model.

2. The method for mapping tidal flat topography based on an unmanned watercraft according to claim 1, characterized in that: Said S1 specifically includes: S11: The inertial navigation system on the unmanned vessel collects the speed, acceleration, and heading data of the unmanned vessel's current position, and makes a preliminary position estimate based on the collected data to determine the approximate position range of the unmanned vessel, providing a basis for position calibration; S12: Fusing the unmanned ship position data initially estimated by the inertial navigation system with the high-precision geographic coordinates obtained by the differential GPS module, accurately calibrating the unmanned ship position through the real-time correction signal of the differential GPS module to obtain the high-precision initial position of the unmanned ship; S13: Based on the position data provided by the differential GPS, the target boundary coordinates of the tidal flat topography mapping area are input, the boundary of the mapping area is established through the navigation control system of the unmanned vessel, and a mapping route plan including the starting point is generated; S14: Load the boundary coordinates of the surveying area and the preset navigation route, enable the path tracking module to match the real-time position of the unmanned vessel with the target path, and adjust the navigation direction of the unmanned vessel in real time to ensure that the unmanned vessel navigates to the surveying starting point; S15: According to the dynamic adjustment instructions of the path tracking module, the unmanned vessel automatically navigates along the set route to the starting point of the surveying area, and records the arrival position of the unmanned vessel to verify the accuracy of autonomous navigation and provide a starting point benchmark for subsequent surveying.

3. The method for mapping tidal flat topography based on an unmanned watercraft according to claim 1, characterized in that: The S2 specifically includes: S21: After the unmanned vessel enters the mudflat mapping area, it starts the multispectral imaging equipment and sets the target spectral band parameters, including visible light of 400-700nm, near-infrared light of 700-900nm, and short-wave infrared light of 900-1700nm, to ensure that the reflection information of the water surface in different spectral ranges is captured; S22: When the unmanned vessel is traveling along the preset route, the continuous scanning function of the multispectral imaging device is activated to collect multi-band image data of various areas on the mudflat surface in real time using the set spectral bands, and ensure that the scanning rate matches the speed of the unmanned vessel; S23: Separate the multi-band spectral data obtained by scanning into bands one by one, and extract the corresponding brightness, texture and coverage feature information in each band; S24: The separated preliminary feature data of each band are fused, and feature association processing is performed according to the band weight distribution to generate comprehensive water surface feature data including water surface brightness, texture and coverage status.

4. The method for mapping tidal flat topography based on an unmanned watercraft according to claim 1, characterized in that: The S23 specifically includes: S231: Separate the multi-band spectral data collected by the multispectral imaging device by band, extract various image data including visible light, near infrared, and short-wave infrared bands; and perform convolution operation on the spectral intensity data of each band to ensure that the data of each band is independently separated; S232: In the visible light band image, the brightness value of each position is determined by calculating the red, green, and blue light intensity values of the image to analyze the distribution characteristics of the water surface reflection intensity; and the brightness data is statistically analyzed to generate a brightness histogram feature; S233: Analyze the texture characteristics of the water surface in near-infrared band images based on the gray-level co-occurrence matrix method. First, the co-occurrence matrix is calculated based on the predetermined pixel spacing and direction angle. Then, a series of texture features are extracted from the co-occurrence matrix, including energy, contrast, entropy, and uniformity, to describe the ripples and surface structure of the water surface. S234: In the short-wave infrared band image, by calculating the ratio of the reflected light to the incident light intensity at each position on the water surface and setting a threshold based on the difference in reflectivity between the water surface and the covering object, the data is binarized to identify the coverage status of sediments and floating objects on the water surface and generate coverage feature data.

5. The method for mapping tidal flat topography based on an unmanned water surveying vessel according to claim 1, characterized in that: The S3 specifically includes: S31: After the unmanned vessel enters the mudflat mapping area, it activates the sonar scanning device and sets measurement parameters such as scanning frequency, depth range, and echo sampling rate to ensure continuous scanning of the mudflat bottom within the designated route and collect underwater terrain reflection data; S32: During the navigation of the unmanned vessel, the sonar scanning device continuously transmits acoustic wave signals and receives the bottom reflected echo signals, converts the bottom echo intensity and echo delay time into bottom depth data, and generates a real-time bottom terrain information matrix D(x, y), where x and y are position coordinates; S33: Using the water surface brightness, texture, and coverage characteristic data obtained in step S2, the sonar echo is calibrated in real time to eliminate the echo deviation caused by water surface disturbances. The calibration factor C is determined by comparing the actual path of the sonar signal with the ideal path. The calculation formula is: Among them, V actual Indicates the echo speed of the sonar signal after being disturbed by the water surface, V ideal Indicates the ideal echo velocity under undisturbed conditions; the calibration factor C is applied to the bottom depth data at each location to update the depth value D calibrated (x, y), the formula is: D calibrated (x, y) = D(x, y) × (1-C); S34: The calibrated bottom depth data D calibrated (x, y) is stored as calibrated terrain data and continuously updated within the survey area to ensure the accuracy of the underwater terrain data during the navigation of the unmanned vessel.

6. The method for mapping tidal flat topography based on an unmanned water surveying vessel according to claim 1, characterized in that: The S4 specifically includes: S41: normalizing the water surface feature data obtained in S2 and the bottom feature data obtained in S3 respectively, converting all feature data into the same numerical range to ensure that the data have the same dimension; S42: Mapping the normalized water surface feature data and bottom feature data to a high-dimensional feature space to generate a water surface feature vector and a bottom feature vector, respectively. Each position coordinate corresponds to a feature vector, including brightness, texture, coverage status, and bottom depth information, to ensure that the two sets of data have a corresponding relationship in the high-dimensional space. S43: Using the manifold alignment algorithm, the implicit mapping relationship between the two sets of feature data is constructed by minimizing the distance between the feature vectors of the water surface and the bottom; S44: Projecting the water surface feature data and the bottom feature data after the mapping relationship is constructed into a common latent space, and performing data fusion in the space. The fused feature vector includes the combined features of the water surface brightness, texture, coverage status and bottom depth information, forming the comprehensive terrain feature data of the mudflat area; S45: Output the fused feature data according to the position coordinates to generate a unified terrain feature data matrix of the tidal flat area, providing complete input data for the subsequent three-dimensional terrain model construction.

7. The method for mapping tidal flat topography based on an unmanned water surveying vessel according to claim 1, characterized in that: The S5 specifically includes: S51: Based on the water surface and bottom feature data after manifold alignment and fusion, obtain the water surface elevation value and the bottom depth value of each position coordinate; and subtract the water surface elevation value from the bottom depth value to calculate the relative height difference of each coordinate point; S52: Based on the calculated relative height difference, a contour generation algorithm is used to classify the Nuotu area and divide the regional boundaries of different height levels; and a linear interpolation method is used to smooth the height difference between adjacent coordinate points to generate continuous contour lines, thereby forming a contour map of the Nuotu area; S53: generating a depth distribution map of the mudflat area based on the distribution of the bottom depth values; classifying the depth distribution of each position coordinate, and marking different depth areas by color coding; S54: Combine the contour map and depth distribution map to construct the water surface elevation value and bottom depth value of each coordinate point into a three-dimensional coordinate point. Use the grid reconstruction algorithm to fit the three-dimensional coordinate points to form a continuous terrain surface to construct a complete three-dimensional model of the mudflat area. 55: Input the 3D terrain model into the visualization system, enhance the three-dimensional sense of the terrain model through coloring and lighting effects, and generate a three-dimensional visualization image containing height, depth and contour information.

8. The method for mapping tidal flat topography based on an unmanned watercraft according to claim 7, characterized in that: The S54 specifically includes: S541: Obtain the water surface elevation value and the water bottom depth value of each position coordinate from the contour map and the depth distribution map of the mudflat area; and construct a three-dimensional coordinate point based on the water surface and water bottom height information; S542: performing initial grid division on all three-dimensional coordinate points in the tidal flat area at a preset distance, dividing the tidal flat area into a number of adjacent grid cells; each grid cell contains a number of three-dimensional coordinate points; S543: performing plane fitting on the three-dimensional coordinate points within each grid cell using the least squares method; S544: After the grid cell fitting is completed, bilinear interpolation is performed on the boundaries of adjacent grid cells to make the boundary height values of adjacent grid cells transition smoothly and eliminate discontinuities; S545: Integrate the fitting and interpolation results of all grid cells to construct a complete three-dimensional terrain model containing all three-dimensional coordinate points of the mudflat area.

9. The method for mapping tidal flat topography based on an unmanned watercraft according to claim 1, characterized in that: The S6 specifically includes: S61: Compare the generated three-dimensional terrain model of the tidal flat area with the historical surveying and mapping data, extract the current height value and the historical height value at each position coordinate, and identify terrain changes by calculating the difference Δz(x, y) between the current height value and the historical height value; S62: Based on the height difference Δz(x, y) of each coordinate point, the terrain change is classified into different modes, including rising, falling, and stable. The change type is determined by setting a threshold T: when Δz(x, y)>T, it is classified as rising; if Δz(x, y)<-T, it is classified as falling; if |Δz(x, y)|≤T, it is classified as stable; S63: Use the adaptive calibration algorithm to dynamically adjust the area with significant changes; for each coordinate point, apply the adaptive calibration coefficient and update the current height value to make it close to the historical data or in line with the change trend; the updated height value z adjusted The formula for (x, y) is: z adjusted (x, y) = z current (x, y) + α·Δz(x, y), where α is an adaptive calibration coefficient with a value range of 0 < α < 1 and is automatically adjusted according to the change pattern; S64: All adjusted height values z adjusted (x, y) integration to generate a calibrated 3D terrain model.

10. A tidal flat topography mapping system based on an unmanned water surveying vessel, used to implement a tidal flat topography mapping method based on an unmanned water surveying vessel as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Positioning and navigation module: including inertial navigation device and differential GPS unit, used to determine the initial position and navigation path of the unmanned vessel in the mudflat area, set the boundary of the survey area, and ensure that the unmanned vessel navigates to the survey starting point; Water surface feature acquisition module: includes multispectral imaging equipment, which is used to collect multi-band spectral information of the mudflat water surface during the navigation of the unmanned vessel, and obtain water surface feature data such as brightness, texture and coverage status of the water surface; Underwater terrain scanning module: Contains a sonar scanning device for synchronously collecting bottom terrain depth data in the mudflat area and using water surface feature data for real-time calibration to eliminate the impact of water surface disturbances on bottom data; Data processing module: used to receive and process water surface feature data and underwater terrain data, construct an implicit mapping relationship between water surface and underwater feature data through manifold alignment algorithm, and generate aligned comprehensive terrain feature data; 3D model construction module: used to generate contour maps and depth distribution maps of the tidal flat area through differential calculation based on the comprehensive terrain feature data output by the data processing module, and to generate an initial 3D terrain model of the tidal flat area based on 3D coordinate points; Dynamic calibration module: Based on the initial 3D terrain model generated by the 3D model construction module, combined with historical surveying and mapping data, the model is dynamically adjusted through an adaptive calibration algorithm. By comparing the differences between the current terrain data and the historical data, the terrain model is updated and the calibrated model is transmitted to the display module; Display module: used to receive the calibrated three-dimensional terrain model output by the dynamic calibration module, and visualize the three-dimensional terrain information of the mudflat area, including height, depth and contour changes, to provide real-time terrain data support for surveying and mapping personnel.

Images