Single-beam unmanned ship cross-section measurement path planning method based on bank slope

Abstract

The present application relates to the field of surveying and mapping navigation technology, in particular to a single-beam unmanned ship cross-section measurement path planning method based on bank slope gradient. In the present application, by real-time solving the specific slope of the river bank and combining the structural parameters of the unmanned ship and the sensor performance, an optimal boundary distance considering the data acquisition quality and navigation safety can be dynamically generated for each cross-section, and then the boundary distance is adaptively compensated and corrected according to the severity of the continuous change of the river bank shape, thereby effectively avoiding the risk of grounding or collision caused by the too close navigation track to the bank in the complex and variable terrain area, while ensuring that the measurement equipment operates in the optimal working water depth range to obtain high-quality data, and after determining the accurate measurement start and end points of all cross-sections, a path planning algorithm is used to find the shortest connection path that traverses all end points, forming a continuous operation route with the shortest total distance, thereby improving the overall operation efficiency of cross-section measurement.

Description

Technical Field

[0001] This invention relates to the field of surveying and navigation technology, and in particular to a single-beam unmanned surface vessel cross-section measurement path planning method based on shore slope. Background Technology

[0002] The field of surveying and navigation technology is a comprehensive technical field that integrates geographic information acquisition and positioning control. It mainly involves using various sensor devices, positioning systems and data acquisition methods to accurately determine the location of natural geographic environments or artificial facilities, perform spatial geometric modeling and navigation path planning.

[0003] Among them, the single-beam unmanned surface vessel (USV) cross-section measurement path planning method refers to using a single-beam echo sounder mounted on an unmanned surface vessel platform to carry out cross-section terrain data collection operations through preset survey lines or fixed tracks.

[0004] The main shortcomings of the pre-set fixed track planning method used in the existing technology are the rigidity of the path and the lack of adaptability to complex riverbank environments. Since the slope morphology of natural river channels is usually not uniform, in actual operation, a fixed near-shore distance cannot simultaneously meet the safety avoidance requirements of steep slopes and the data integrity requirements of gentle slopes. For example, in concave bank areas, the pre-set track may be too close to the underwater steep slope, which may cause the unmanned vessel to run aground. In the gentle slope area of ​​convex bank, the same track may be too far from the effective depth measurement range, resulting in the loss of key topographic data on the bank. In addition, the connection of each cross-section line usually relies on simple sequential connection without considering the spatial layout and distance relationship between cross-sections. This causes the unmanned vessel to generate a lot of redundant and ineffective voyages when changing measurement cross-sections, especially in wide or winding river sections, which greatly reduces the overall efficiency of measurement operations and increases time and energy costs. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a single-beam unmanned surface vessel cross-section measurement path planning method based on bank slope.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a single-beam unmanned surface vessel cross-section measurement path planning method based on shore slope gradient, comprising the following steps:

[0007] S1: Obtain orthophotos and digital elevation models of the river channel locations corresponding to the unmanned vessel cross-section measurement path, and construct the initial layer dataset;

[0008] S2: Refer to the initial layer dataset to extract the centerline and water level line of the river channel corresponding to the unmanned vessel cross-section measurement path, calculate the riverbank slope angle, and generate a cross-section feature dataset;

[0009] S3: Obtain the unmanned ship structure parameter set, and call the section feature data set to calculate the optimal safety ranging of the section, correct the boundary distance of the unmanned ship section measurement path through adaptive error compensation on the river bank slope gradient, and output the ranging correction result;

[0010] S4: Refer to the ranging correction result, load the river center line, water level line and section point position data in AutoCAD in sequence, construct the basic point structure of the section measurement path, and aggregate as a section measurement boundary point set;

[0011] S5: Based on the section measurement boundary point set, connect each group of left and right bank boundary points to form a section line, and generate an unmanned ship navigation path.

[0012] As a further scheme of the application, the initial layer data set includes an orthographic image layer, a digital elevation model layer and spatial registration control information, the section feature data set includes river center line data, water level line data and a river bank slope angle set, the ranging correction result includes left bank correction ranging value, right bank correction ranging value and section stake index value, the section measurement boundary point set includes left bank boundary point coordinate set, right bank boundary point coordinate set and section number information, and the unmanned ship navigation path includes section connection line sequence, path node number set and path layer file code.

[0013] As a further scheme of the application, the obtaining step of the initial layer data set is specifically:

[0014] S111: Collect the orthographic image of the river location corresponding to the unmanned ship section measurement path through the unmanned aerial vehicle camera, obtain the LiDAR point cloud of the river location, and construct the river original image and point cloud set;

[0015] S112: Call the LiDAR point cloud in the river original image and point cloud set, perform ground point and non-ground point classification, construct a grid according to the set spatial resolution, calculate the elevation value of each grid node, and generate a river digital elevation model;

[0016] S113: According to the orthographic image in the river original image and point cloud set and the corresponding river digital elevation model, calibrate the pixel coordinates and geographic coordinates of a plurality of homonymous points and calculate the coordinate transformation parameters, project all pixel points of the orthographic image to the corresponding geographic space position of the river digital elevation model, and construct an initial layer data set.

[0017] As a further scheme of the application, the obtaining step of the section feature data set is specifically:

[0018] S211: Perform contrast enhancement and edge sharpening processing on the orthophoto in the initial layer data set, identify the water area boundary according to the image texture and edge features, extract the center line and water level line of the river channel position corresponding to the cross-section measurement path of the unmanned ship, and establish the river channel characteristic line geometry data;

[0019] S212: Call the digital elevation model in the initial layer data set, combine the river channel characteristic line geometry data calibrated bank slope area, calculate the elevation change rate of each elevation data point in the horizontal and vertical directions, use the trigonometric function conversion to obtain the slope value, and generate the global bank slope gradient grid information;

[0020] S213: According to the center line in the river channel characteristic line geometry data, generate a cross-section vertical line according to the preset stake number interval, and determine the intersection of the vertical line and the left and right bank water level line, call the global bank slope gradient grid information, extract the river bank slope angle of each intersection position, and construct the cross-section characteristic data set.

[0021] As a further scheme of the present application, the acquisition step of the distance measurement correction result is specifically:

[0022] S311: Obtain the unmanned ship structure parameter set, including the transducer draft depth, transducer to bow horizontal distance and minimum effective detection depth, and refer to the left and right bank river bank slope angle corresponding to each cross-section in the cross-section characteristic data set, calculate the optimal safe distance of the left and right bank of the cross-section;

[0023] S312: According to the river bank slope angle in the cross-section characteristic data set, calculate the change degree of the river bank slope along the center line direction, and compare the change degree with the preset interval, select the corresponding error compensation amount, and establish the adaptive error compensation parameter;

[0024] S313: Call the optimal safe distance of the left and right bank of the cross-section and the adaptive error compensation parameter, adjust the original distance measurement value of each cross-section and the matching compensation amount, correct the boundary distance of the cross-section measurement path of the unmanned ship, and output the distance measurement correction result.

[0025] As a further scheme of the present application, the acquisition step of the cross-section measurement boundary point set is specifically:

[0026] S411: Load the river center line, water level line coordinate data and corresponding cross-section point position data in AutoCAD in sequence, spatially align the three, determine the intersection of the cross-section point and the water level line, and construct the basic point structure of the cross-section measurement path;

[0027] S412: call the boundary distance of each cross-section survey path in the distance correction result, and perform an offset operation equal to the corrected distance value to the outside of the shore along the normal direction of the intersection of the cross-section point in the base point structure of the cross-section survey path and the water level line, to calculate the coordinates of each cross-section left and right boundary point;

[0028] S413: integrate the coordinates of each cross-section left and right boundary point in sequence according to the cross-section stake number, and generate a cross-section survey boundary point set.

[0029] As a further scheme of the present application, the obtaining step of the unmanned ship navigation path is specifically:

[0030] S511: pair the left and right bank boundary points belonging to the same cross-section stake number in the cross-section survey boundary point set, connect each group of paired boundary points to form an independent cross-section line, and establish a cross-section line path set;

[0031] S512: define the endpoints of each cross-section line in the cross-section line path set as path nodes, calculate the navigation distance between the endpoints of two cross-section lines as the connection weight, and use the Dijkstra algorithm to traverse the minimum navigation distance path of all cross-sections to determine the cross-section survey optimization connection sequence;

[0032] S513: according to the cross-section survey optimization connection sequence, perform sequence reorganization on all cross-section lines in the cross-section line path set, and sequentially connect each cross-section line after reorganization with the optimized connection path therebetween to generate an unmanned ship navigation path.

[0033] Compared with the prior art, the present application has the following advantages and positive effects:

[0034] In the present application, by real-time solving the specific slope of the river bank and combining the structural parameters of the unmanned ship and the sensor performance, an optimal boundary distance considering data acquisition quality and navigation safety can be dynamically generated for each cross-section, and then the boundary distance is adaptively compensated and corrected according to the severity of the continuous change of the river bank shape, thereby effectively avoiding the risk of grounding or collision caused by too close navigation to the shore in complex terrain areas, while ensuring that the measurement equipment operates in the best working water depth range to obtain high-quality data, and after determining the accurate measurement start and end points of all cross-sections, a path planning algorithm is used to find the shortest connection path traversing all endpoints, and finally a continuous operation route with the shortest total distance is formed, which significantly improves the overall operation efficiency of cross-section survey and reduces unnecessary energy consumption. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 is a schematic diagram of the main steps of the present application;

[0036] Figure 2This is a flowchart of step S1 of the present invention;

[0037] Figure 3 This is a flowchart of step S2 of the present invention;

[0038] Figure 4 This is a flowchart of step S3 of the present invention;

[0039] Figure 5 This is a flowchart of step S4 of the present invention;

[0040] Figure 6 This is a flowchart of step S5 of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0042] Please see Figure 1 This invention provides a technical solution: a single-beam unmanned surface vessel cross-section measurement path planning method based on shore slope, comprising the following steps:

[0043] S1: Obtain orthophotos and digital elevation models of the river channel locations corresponding to the unmanned vessel cross-section measurement path, and construct the initial layer dataset;

[0044] S2: Refer to the initial layer dataset to extract the center line and water level line of the river channel corresponding to the unmanned vessel cross-section measurement path, calculate the riverbank slope angle, and generate a cross-section feature dataset;

[0045] S3: Obtain the set of unmanned vessel structural parameters, call the cross-sectional feature dataset to calculate the optimal safe distance measurement of the cross-section, and correct the boundary distance of the unmanned vessel cross-section measurement path by adaptive error compensation of the riverbank slope, and output the distance correction result.

[0046] S4: Based on the distance measurement correction results, load the river centerline, water level line and cross-section point location data in AutoCAD in sequence to construct the basic point structure of the cross-section measurement path and summarize it into a cross-section measurement boundary point set;

[0047] S5: Based on the boundary point set of cross-section measurement, connect each set of left and right bank boundary points to form a cross-section line and generate the unmanned vessel navigation path.

[0048] The initial layer dataset includes orthophoto layers, digital elevation model layers, and spatial registration control information. The cross-section feature dataset includes river centerline data, water level data, and a set of riverbank slope angles. The distance correction results include corrected distance values ​​for the left bank and the right bank, and cross-section station index values. The cross-section measurement boundary point set includes the coordinate set of the left bank boundary points and the coordinate set of the right bank boundary points, as well as cross-section number information. The unmanned vessel navigation path includes a sequence of cross-section connecting lines, a set of path node numbers, and path layer file encoding.

[0049] Please see Figure 2 The specific steps for obtaining the initial layer dataset are as follows:

[0050] S111: Collect orthophotos of the river channel location corresponding to the unmanned vessel's cross-section measurement path using a drone camera, and obtain LiDAR point clouds at that river channel location to construct the original river channel image and point cloud set.

[0051] A DJI M300RTK hexacopter drone equipped with a Zenmuse P1 camera was used to collect data at a constant speed along a pre-set flight path at an altitude of 100 meters above the Yellow River channel from K0+500 to K1+200 in Kaifeng. The high overlap rate (85% forward and 70% lateral) was designed to ensure that any ground point could be captured by multiple images when generating a 3D model using photogrammetry, providing sufficient data redundancy to guarantee model accuracy and integrity and avoid data gaps. The data collection yielded 450 raw orthophotos with a resolution of 2.5 cm covering approximately 0.5 square kilometers of the river channel. Simultaneously, the onboard LiDAR system scanned the same area at a transmission frequency of 500,000 points per second, acquiring approximately 50 million points with 3D coordinates. , , The lidar point cloud contains intensity information, among which This represents the east-west (e.g., east longitude) coordinates of a point in a geographic coordinate system. Represents north-south coordinates (such as North latitude). This represents the elevation value of the point. The absolute accuracy of the point cloud is 5 cm in the horizontal direction and 3 cm in the vertical direction. The data acquisition was carried out under clear and windless conditions. All acquired orthophotos and LiDAR point cloud data were stored together to construct the original river channel image and point cloud set.

[0052] S112: Call the original river image and the LiDAR point cloud in the point cloud set, perform ground point and non-ground point classification, construct a grid according to the set spatial resolution, calculate the elevation value of each grid node, and generate a digital elevation model of the river.

[0053] The study utilized original river channel imagery and 50 million LiDAR point clouds from a point cloud dataset. Each point cloud was processed, and classification was performed by analyzing the differences in elevation values ​​between the point cloud and its neighboring points. Specifically, an elevation difference threshold of 1.5 meters and a slope threshold of 15 degrees were set. These thresholds were based on prior knowledge of the terrain features in the study area: the elevation changes of natural surfaces (such as beaches and watersides) in river channels are usually relatively continuous and gentle. Abrupt local elevation changes exceeding 1.5 meters or local slopes exceeding 15 degrees are highly likely caused by non-surface features (such as trees on the riverbank or structures on the slope). When the average elevation difference between a point and its eight neighboring points exceeds 1.5 meters or the calculated local slope exceeds 15 degrees, that point is classified as a non-ground point; the remaining points are classified as ground points. After this process, the number of ground points was approximately 35 million. Next, on the horizontal plane (… - A two-dimensional grid with a spatial resolution of 0.5 meters by 0.5 meters is constructed on the plane (i.e., the horizontal projection plane of the ground), covering the entire river channel area. For each node in the grid, all ground points falling within that grid cell are retrieved, and the elevation values ​​of these ground points are extracted. For example, a grid node located at coordinates (123.5, 456.7) has five ground points within its corresponding 0.5m × 0.5m cell, with elevations of 370.12m, 370.15m, 370.11m, 370.13m, and 370.14m. The arithmetic mean of these five elevation values ​​((370.12+370.15+370.11+370.13+370.14) / 5) is calculated to obtain 370.13m. This value is assigned to the grid node. This calculation is repeated for all grid nodes to ultimately generate a digital elevation model (DEM) of the river channel that represents the continuous elevation changes of the river surface.

[0054] S113: Based on the original river image and the orthophoto in the point cloud set and the corresponding digital elevation model of the river, the pixel coordinates and geographic coordinates of multiple points of the same name are calibrated and the coordinate transformation parameters are calculated. All pixels of the orthophoto are projected to the corresponding geographic spatial location of the digital elevation model of the river to construct the initial layer dataset.

[0055] Based on the original river channel image and the unprocessed orthophoto from the point cloud dataset, and referring to the digital elevation model (DEM) of the river channel generated by S112, 10 evenly distributed and clearly defined corresponding points were selected on the image and DEM. For example, the corner point of a fixed rock measuring 1 meter × 1 meter on the riverbank is marked with pixel coordinates (10520, 8734) on the orthophoto. By querying the geographic coordinates of the same location on the DEM, its geographic coordinates are found to be (458330.25). 3855120.70 371.55 The pixel coordinate set and geographic coordinate set of these 10 pairs of points with the same name are processed, and an affine transformation model is applied to calculate six coordinate transformation parameters. , , , , , ),in, and Controlling images in and Scaling of direction and Controlling the rotation and shearing deformation of images, and Then control the image in and Directional translation. These parameters are solved using the least squares method to minimize the sum of squared projection errors from geographic coordinates to pixel coordinates for all corresponding points. The obtained transformation parameters are used in a projection process that iterates through every pixel in the orthophoto, for example, the pixel with coordinates (1, 1), converting its pixel coordinates to corresponding geographic coordinates using the calculated transformation parameters. These geographic coordinates are then overlaid onto the river channel digital elevation model, and the elevation value at that location is retrieved, thus accurately projecting the pixel from 2D image space to its 3D geographic location. This process is performed on all 200 million pixels of the image, constructing an initial layer dataset where each pixel has accurate 3D geographic coordinates.

[0056] Please see Figure 3 The specific steps for obtaining the cross-sectional feature dataset are as follows:

[0057] S211: Perform contrast enhancement and edge sharpening on the orthophotos in the initial layer dataset, identify the water area boundary based on image texture and edge features, extract the center line and water level line of the river corresponding to the unmanned vessel cross-section measurement path, and establish the geometric data of the river feature line.

[0058] Image processing operations were performed on the orthophotos in the initial layer dataset. First, histogram equalization was applied to expand the dynamic range of the image's grayscale levels, thereby enhancing the contrast between water and land areas. Next, a 3×3 Laplacian operator with kernel functions [[0, -1, 0], [-1, 5, -1], [0, -1, 0]] was used to convolve the enhanced image to sharpen the edges of objects such as riverbanks and sandbars. Based on the low reflectivity and smooth texture of the water areas in the processed image, and the high reflectivity and complex texture of the riverbank areas, the boundaries of the water areas were automatically identified by setting a color segmentation threshold. This threshold was set so that the red, green, and blue (RGB) channel values ​​of all pixels were below 60. The threshold is set based on the following: Pixel value sampling analysis of known water and land areas in the acquired images revealed that water bodies, due to their physical characteristics (absorbing most visible light), generally have reflectance values ​​below 60 in all three color channels, while land features (such as soil and vegetation) have much higher reflectance and thus much higher RGB values. Therefore, this threshold effectively distinguishes the water-land boundary. The extracted boundaries are processed into vector line features, where the centerline representing the main channel flow direction and the left and right bank water boundary lines representing the current water level are derived separately. For example, the centerline is formed by a series of coordinate points (…). , ), ( , ), …( , ) constitutes, of which ( , ) represents the first line that forms the center line The planar coordinates of each vertex. Similarly, the water level lines on the left and right banks are also composed of their respective coordinate point sequences, and these coordinate data together establish the geometric data of the river channel feature lines.

[0059] S212: Call the digital elevation model in the initial layer dataset, and combine it with the bank slope area marked by the river feature line geometric data. Calculate the elevation change rate of each elevation data point in the horizontal and vertical directions, use trigonometric functions to convert the slope value, and generate global bank slope raster information.

[0060] The river channel digital elevation model (DEM) from the initial layer dataset is used, and combined with the calibrated bank slope area (the region between the waterline and the riverbank elevation), slope calculation is performed on each DEM data point within this region. This calculation is performed within a 3×3 moving window. Based on the Horn algorithm, the slope is derived by calculating the elevation gradient of the center pixel in the horizontal and vertical directions. Its mathematical model is: Slope ,in, This is the final calculated slope value, in degrees. It is elevation exist The rate of change in the (horizontal) direction is calculated using the following formula: . It is elevation exist The rate of change in the (vertical) direction is calculated using the following formula: In these formulas: to This represents the elevation values ​​of the nine pixels within the 3x3 window; these values ​​are read directly from the DEM. The central pixel is defined as 'Z', and the rest are its eight adjacent pixels in the directions of 'Z'. This is the pixel size of the DEM, i.e., the spatial resolution, which is 0.5 meters in this scenario. The calculation logic is to obtain the horizontal elevation change by subtracting the weighted elevation sum of the three right pixels from the weighted elevation sum of the three left pixels, where the pixels directly horizontally adjacent to the center pixel are included. and It was assigned a weight of 2 to highlight its impact. Similarly, calculate the vertical elevation change. The 8 in the denominator is the total weighted sum, divided by... The changes in pixel space are converted into dimensionless rates of change in actual geographic space. Finally, the total rate of change in elevation is obtained by calculating the square root of the sum of the squares of the rates of change in the two directions, and then the arctangent function is applied. Convert it to a slope value in degrees. For example, a center point is calculated as follows: The value is 0.25. If the value is 0.38, then the slope at that point is... Degree. Repeat this process for all DEM data points within the bank slope area to generate a slope raster covering the entire bank slope.

[0061] S213: Based on the centerline in the geometric data of the river feature line, generate the cross-section vertical line according to the preset station spacing, determine the intersection of the vertical line with the water level lines on the left and right banks, call the slope raster information of the entire riverbank, extract the riverbank slope angle at each intersection point, and summarize to construct the cross-section feature dataset.

[0062] Based on the centerline coordinate sequence from the acquired river channel feature line geometric data, starting from the starting point, a station is set every 100 meters along the centerline, such as K0+500, K0+600, K0+700, etc. At each station, a perpendicular line is generated, perpendicular to the tangent direction of the centerline at that point, and the coordinates of the intersection point of this perpendicular line with the water levels of the left and right banks in the same dataset are calculated. For example, at station K0+500, the perpendicular line intersects the water level of the left bank at point... (coordinate (), intersecting with the right bank water level line at point (coordinate ), where subscript and These represent the left and right banks, respectively. Next, the global bank slope raster information generated by S212 is used to extract and record the intersection points. and intersection The slope angle value at the location of the grid cell. For example, at K0+500, the intersection of the left and right banks. The slope of the location is 25.8 degrees, and the intersection is on the right bank. The slope value at the location is 21.8 degrees. Repeat this extraction operation for all 8 preset cross-sections, summarizing the slopes of each cross-section to construct a cross-section feature dataset. This dataset contains slope data for K0+500 (left bank 25.8 degrees, right bank 21.8 degrees), K0+600 (left bank 22.9 degrees, right bank 21.7 degrees), K0+700 (left bank 23.8 degrees, right bank 21.3 degrees), K0+800 (left bank 23.2 degrees, right bank 22.1 degrees), K0+900 (left bank 22.8 degrees, right bank 29.5 degrees), K1+000 (left bank 22.6 degrees, right bank 26.7 degrees), K1+100 (left bank 19.3 degrees, right bank 21.8 degrees), and K1+200 (left bank 22.9 degrees, right bank 22.0 degrees).

[0063] Please see Figure 4 The specific steps for obtaining the ranging correction results are as follows:

[0064] S311: Obtain the set of unmanned vessel structural parameters, including transducer draft, horizontal distance from transducer to bow and minimum effective detection depth, and calculate the optimal safe distance between the left and right banks of the cross section by referring to the slope angle of the left and right banks of the river corresponding to each cross section in the cross section feature dataset.

[0065] Obtain the set of unmanned surface vessel structural parameters, and refer to the riverbank slope angle corresponding to each cross-section in the cross-sectional feature dataset. Calculate the optimal safe distance measurement between the left and right banks of the cross-section based on the following mathematical model. : In this model, the meanings of each parameter and how to obtain them are as follows: : This is the final calculated optimal safe horizontal distance between the transducer on the unmanned vessel and the waterline, expressed in meters. : Represents a function whose result is calculated. It depends on the multiple parameters within the parentheses. : This is the angle of the riverbank slope, i.e., the angle between the bank slope and the water surface, in degrees, with a range of (0°, 90°). This value is a variable, calculated by substituting the specific slope values ​​of the left and right banks of each cross-section extracted from S213. : is a fixed structural parameter of the unmanned surface vessel, referring to the horizontal distance from the installation position of the transducer on the hull to the bow of the vessel. It is obtained by precise measurement on the hull and its value is 0.525 meters. : This is the draft of the transducer of the unmanned vessel, that is, the vertical distance of the transducer below the water surface in still water. It is obtained through actual measurement and its value is 0.1 meters. : is the shortest effective detection distance of the transducer, referring to the minimum water depth at which the sonar equipment can guarantee the quality of measurement data. This value is determined by the performance of the equipment and is determined through calibration experiments in a calm water surface of a clear water pool. Its value is 0.4 meters.

[0066] The judgment criterion of this model is Its setting is based on comparing the "critical horizontal distance for hull grounding" with the "bow protrusion length". Expression The calculation is performed when the water depth is exactly equal to the transducer's draft. At that time, the theoretical horizontal distance from the transducer to the waterline. - Case 1 (gentle bank slope): If This means that the critical distance at which the ship can run aground is greater than the length of the bow overhang, at which point the risk of the bow colliding with the shore slope is low. Therefore, the primary objective of path planning is to ensure data quality and safe distance. use Calculations ensured that the water depth directly below the transducer was at least [missing information]. Thus, effective depth sounding data can be obtained. - Scenario 2 (steep bank slope): If This means the shoreline is very steep, the critical distance for the ship to run aground is very small, and the bow faces a high risk of hitting the bottom. In this situation, the ship's navigational safety is the primary consideration, and the model will determine the safe distance. Set directly to This means keeping the transducer at a distance equal to the length of the bow protrusion from the waterline. This is equivalent to positioning the bow directly above the waterline, which is the most direct strategy to avoid collisions.

[0067] S312: Based on the riverbank slope angle in the cross-sectional feature dataset, calculate the degree of change of the riverbank slope along the centerline direction, compare the degree of change with the preset interval, select the corresponding error compensation amount, and establish adaptive error compensation parameters.

[0068] Based on the riverbank slope angle sequence in the cross-sectional feature dataset, the degree of change of the riverbank slope along the centerline is calculated. This degree of change is defined as the absolute value of the difference between the slope values ​​of the same side of the slopes of two adjacent cross-sections. For example, for the left bank, the slope of cross-section K0+600 is 22.9 degrees, and the slope of the previous cross-section K0+500 is 25.8 degrees. Therefore, the degree of change of the slope at cross-section K0+600 is |22.9-25.8|=2.9 degrees. Then, this degree of change is compared with three preset intervals, and the corresponding error compensation amount is selected. The setting of these intervals and compensation values ​​is based on the statistical analysis of historical survey data of this river section and the experience of field reconnaissance, to introduce a safety margin for unforeseen local topographic changes or measurement errors in the measurement path planning. Specifically, the low change interval [0, 2.0) degrees indicates that the slope morphology is continuous and stable, the topographic uncertainty is small, no additional compensation is required, and the compensation amount is 0 meters. The moderate variation range [2.0, 5.0) indicates a certain degree of change in the bank slope morphology. To preventatively increase the safety distance, a compensation of 0.1 meters is introduced. The high variation range [5.0, +∞) indicates drastic changes in the bank slope morphology, with the possibility of collapse or erosion and high uncertainty. Therefore, a larger compensation of 0.2 meters is introduced to ensure absolute safety. Since 2.9 degrees falls within the range [2.0, 5.0), an error compensation of 0.1 meters is selected for the left bank of the K0+600 section. This calculation and comparison are performed on both the left and right banks of all sections to establish a set of adaptive error compensation parameters corresponding to the location of each section.

[0069] S313: Call the optimal safe distance measurement and adaptive error compensation parameters for the left and right banks of the cross section, adjust the original distance measurement value and the matching compensation amount for each cross section, correct the boundary distance of the unmanned vessel cross section measurement path, and output the distance measurement correction result.

[0070] Calling the optimal safe distance measurement of the left and right banks of each cross section The original distance measurement value is adjusted using the adaptive error compensation parameters established in S312. This adjustment is achieved by adding the optimal safe distance measurement value for each cross section to the error compensation amount matched for that cross section. For example, for the left bank of cross section K0+600, the optimal safe distance calculated in S311 is 0.95 meters, and the error compensation amount determined in S312 is 0.1 meters. Therefore, the corrected boundary distance is 0.95 meters + 0.1 meters = 1.05 meters. For the right bank of cross section K0+900, its slope is 29.5 degrees, while the slope of the right bank of the previous cross section K0+800 is 22.1 degrees. The degree of change is |29.5 - 22.1| = 7.4 degrees, which belongs to the high change range. The compensation amount is 0.2 meters, and its original optimal safe distance is 0.70 meters. The corrected boundary distance is 0.70 meters + 0.2 meters = 0.90 meters. The original ranging values ​​of the left and right banks (16 in total) of all cross sections (8 in total) are adjusted with the matching compensation amount to correct the distance between the left and right boundaries of the unmanned vessel cross section measurement path at each cross section, and the corrected ranging results are output.

[0071] Please see Figure 5 The specific steps for obtaining the boundary point set of cross-section measurement are as follows:

[0072] S411: In AutoCAD, load the coordinate data of the river centerline, water level line, and corresponding cross-section point location data in sequence, align the three in space, determine the intersection of the cross-section point and the water level line, and construct the basic point structure of the cross-section measurement path.

[0073] In the AutoCAD software environment, first load the DXF format coordinate data of the river centerline and left and right bank water level lines generated in S211, then load the station location data of each section determined in S213. Use coordinate matching to spatially align these three data points within the same geographic reference system. After alignment, the intersection points of each section's vertical line and the left and right bank water level lines on the drawing are uniquely determined. For example, the vertical line of section K0+500 intersects the left bank water level line at... The point intersects with the right bank water level line. Points, among which Point (subscript) and These represent the left bank and the right bank, respectively, with the number 500 corresponding to station K0+500. This set of intersections... The boundaries between each measurement section and the current water body are precisely defined in space, thereby constructing the basic point structure for the unmanned vessel's cross-section measurement path.

[0074] S412: Call the boundary distance of each cross-section measurement path in the distance measurement correction result, and for the intersection of the cross-section point and the water level line in the basic point structure of the cross-section measurement path, perform an offset operation equal to the correction distance value along the normal direction of the intersection point to the outside of the bank, and calculate the coordinates of the left and right boundary points of each cross-section.

[0075] The boundary distance values ​​of each cross-section measurement path in the distance measurement correction results are retrieved, and an offset operation is performed on the intersections of each cross-section point and the water level line in the basic point structure constructed in S411. The left bank intersection of cross-section K0+600 is used as an example. For example, its corrected boundary distance is 1.05 meters. First, calculate the distance to this point. The normal direction of the water level line at that point is perpendicular to the tangent of the water level line at that point and points outwards. Then, along this normal direction, the point... The coordinates of the new left boundary point are obtained by shifting it 1.05 meters outward from the shore and performing a weighted coordinate calculation. Similarly, if the corrected distance on the right bank of section K0+600 is 1.12 meters, then the intersection point on its right bank... Offset 1.12 meters outward from the shore along its normal direction to obtain the coordinates of the right boundary point. This offset operation is performed on all 16 points, which are the intersection points of the left and right banks of all 8 cross sections, to calculate the new coordinates of the left and right boundary points for each cross section.

[0076] S413: Summarize the coordinates of the left and right boundary points of each section, integrate them in the order of the section station number, and generate a set of boundary points for section measurement;

[0077] The coordinate data of all 16 cross-section left and right boundary points were collected and then integrated and sorted according to their respective cross-section station numbers. This integration process began with the cross-section with the smallest station number, K0+500, recording the coordinates of its left and right boundary points. This was followed by the coordinates of the left and right boundary points of the cross-section K0+600, and so on, until the cross-section with the largest station number, K1+200. The result was an ordered list of coordinate points, which constitutes the set of cross-section measurement boundary points. The point set is integrated in order of station number, including the coordinates of the left bank boundary point (458310.1, 3855125.4) and the right bank boundary point (458682.5, 3855140.2) of section K0+500, the coordinates of the left and right boundary points (458255.8, 3855210.9) and (458655.1, 3855228.3) of section K0+600, and so on, until the coordinates of the left and right boundary points (457930.4, 3855720.1) and (458296.8, 3855740.5) of section K1+200 are all integrated.

[0078] Please see Figure 6 The specific steps for obtaining the unmanned vessel's navigation path are as follows:

[0079] S511: Pair the boundary points of the cross-section measurement that belong to the same cross-section station number on the left and right banks, connect each pair of boundary points to form an independent cross-section line, and establish a set of cross-section line paths.

[0080] The boundary points of the cross-section measurement are grouped together, and the left and right bank boundary points belonging to the same cross-section station are paired. For example, the left bank boundary point (458310.1, 3855125.4) of cross-section K0+500 is paired with the right bank boundary point (458682.5, 3855140.2). Then, by connecting these two paired boundary points, a straight line segment is formed in the two-dimensional coordinate system. This line segment is the measurement path of cross-section K0+500. This operation is repeated for all 8 station numbers in the cross-section measurement boundary point group, connecting the left and right bank boundary points of cross-section K0+600, K0+700, and so on, until the left and right bank boundary points of cross-section K1+200. This establishes a set of cross-section line paths containing 8 independent cross-section lines.

[0081] S512: Define the endpoints of each cross-section line in the cross-section line path set as path nodes, calculate the sailing distance between the endpoints of two specified cross-section lines as the connection weight, and use Dijkstra's algorithm to traverse the minimum sailing distance path of all cross-sections to determine the optimal connection order of cross-section measurement.

[0082] Each endpoint (left bank boundary point and right bank boundary point) of the established cross-section path set is defined as a path node, forming a set of 16 nodes. To plan a measurement route with the shortest total distance for the unmanned surface vessel (USV), a graph model for algorithm analysis needs to be constructed first. Specifically, the Euclidean distance between any two endpoints that do not belong to the same cross-section line is calculated, and this distance is used as the weight of the edge connecting these two nodes. Physically, this weight represents the distance the USV travels in a straight line from one cross-section endpoint to another. To find the shortest continuous path that traverses all cross-sections, Dijkstra's algorithm is used in this step.

[0083] Dijkstra's algorithm is a classic graph theory algorithm used to find the shortest path from a single source vertex to all other nodes in a graph. In this scenario, the core of the algorithm is to iteratively update the algorithm to gradually determine the shortest travel distance from the starting measurement point (e.g., the right bank boundary point of section K0+500) to each endpoint of the section. The core formula for distance update is as follows:

[0084] ;

[0085] in, : Represents the node that the algorithm is currently processing. In the physical scene, this is the known location of the unmanned vessel (the endpoint of a certain cross section). : Represents a node An adjacent node whose shortest path has not yet been definitively determined represents, in a physical scenario, the unmanned vessel's path from its current position. Depart, to the next possible target section endpoint reached by navigation. : Here Indicates distance, This refers to the journey from the starting point of this measurement mission to the current node. The total length of the shortest path, which has been determined by the algorithm. This is a known value that will not change. This refers to the journey from the starting point of this measurement mission to the target node. The total length of the currently known shortest paths. This value is temporary and may be updated to a smaller value during algorithm iterations. In the formula, it appears on both the left-hand side of the equals sign (representing the updated value) and the... Inside the function (representing the value before the update). : Here Indicates weight. Refers to the connection node and nodes The weight of the edge, in the physical scenario, is the weight of the unmanned boat from the endpoint. sail directly to the endpoint The straight-line distance, which was obtained before the algorithm started by calculating the Euclidean distance between the two points. This is a minimum value function. It compares two values: one is the known reachable node. Path length The other is a newly discovered one, via nodes. Then to the node Total path length The algorithm always chooses the smaller one as the arrival. The new shortest path.

[0086] The algorithm's execution flow is as follows: Starting from a selected starting point, it continuously selects the currently known unvisited node closest to the starting point and performs a "relaxation" operation (i.e., updates the distance to its neighboring nodes using the formula mentioned above) until all 16 path nodes have been visited. The final output of the algorithm is a visiting order containing all 16 nodes, which determines the optimal connection order for cross-sectional measurements.

[0087] Suppose an unmanned surface vessel needs to measure two adjacent cross-sections, K0+500 and K0+600. This involves four path nodes: node The right bank boundary point of section K0+500, node The left bank boundary point of section K0+500, node The right bank boundary point of section K0+600, node The left bank boundary point of section K0+600.

[0088] Objective: To plan a route from node Start, complete traversal and Two measurement paths, with the shortest total voyage.

[0089] First, we need the travel distance between each node, i.e., the edge weight. These distances are calculated from coordinates using the Euclidean distance formula. A reasonable set of simulated data (unit: meters) is used here: From arrive Measuring distance It's 300 meters. From arrive Connection distance It's 110 meters. From arrive Connection distance Set it to a relatively long value, 500 meters. From arrive Connection distance It's 320 meters. From arrive Connection distance It's 100 meters. From arrive Measuring distance It is 280 meters.

[0090] Dijkstra's algorithm will be used to calculate from the starting point. The shortest distance to all other nodes.

[0091] Step 1: Initialize the starting node: Select a node This serves as the starting point for the unmanned vessel. Visited set: Initially empty. Unvisited set: Contains all four nodes. Distance Record Record from the starting point The shortest distance to each node. Initially, Let 0 be the value, and let the distance to all other nodes be unknown, set to infinity. ).so, , , , Predecessor node record Used to trace the final path. Initially, the predecessor of all nodes is empty.

[0092] Step 2: First iteration (processing the current node) The unmanned boat is located at the starting point. The investigation started from All neighbors that can be reached directly from the starting point .

[0093] For the neighbors :

[0094] yes , apply formula .

[0095] Substitute the numerical values ​​into the calculation: The result was 300.

[0096] Therefore, Updated to 300, and recorded. Forerunner for .

[0097] For the neighbors :

[0098] yes , apply formula .

[0099] Substitute the numerical values ​​into the calculation: The result was 110.

[0100] Therefore, Updated to 110 and recorded. Forerunner for .

[0101] For the neighbors :

[0102] yes , apply formula .

[0103] Substitute the numerical values ​​into the calculation: The result was 500.

[0104] Therefore, Updated to 500, and recorded. Forerunner for .

[0105] Iteration ended, node The node that has been processed is moved to the visited set. Now, the node with the smallest distance among the unvisited nodes is selected as the next node to be processed. The current distances between the nodes are... , , The smallest is .

[0106] Second iteration (processing the current node) The next node to be processed is .

[0107] For the neighbors :

[0108] yes , apply formula .

[0109] Substitute the numerical values ​​into the calculation: ,Right now The result was still 300.

[0110] This means from Directly to The path (300 meters) is longer than from... go through Then The path (430 meters) is shorter. Therefore, and No updates will be made.

[0111] For the neighbors (This is a key step in this example):

[0112] yes , apply formula .

[0113] Substitute the numerical values ​​into the calculation: ,Right now The result was 390.

[0114] The calculations here show that, from go through Then The total path length (390 meters) is longer than previously known from Directly to The path (500 meters) needs to be shorter. Therefore, It was updated to 390, and Forerunner It has also been updated to .

[0115] Iteration ended, node This has been processed. Now, the node with the smallest distance among the unvisited nodes is selected. The distance of the currently unvisited node is... , The smallest is .

[0116] Step 4: Third iteration (processing the current node) The next node to be processed is .

[0117] For the neighbors :

[0118] yes , apply formula .

[0119] Substitute the numerical values ​​into the calculation: ,Right now The result was still 390.

[0120] This means from go through arrive The path (390 meters) is longer than from go through Then The path (400 meters) is shorter. Therefore, and No updates will be made.

[0121] Iteration ended, node This has been processed. The last unvisited node is... .

[0122] Step 5: Fourth iteration (processing the current node) Process the last node The algorithm ends.

[0123] After the algorithm finishes execution, the result is obtained from the starting point. Shortest path length to all nodes and the path itself: to Shortest path: Tracing the path: The path is .arrive Shortest path: Tracing the path: The path is .arrive Shortest path: Tracing the path: The path is .

[0124] This result tells us that in order to reach... The shortest route is to start from... arrive , and then from arrive .

[0125] Based on the measurement tasks of the unmanned surface vessel (USV), a measurement sequence with the shortest total voyage can be planned. There are two possible complete measurement schemes: 1. Scheme 1: First measure the K0+600 section, then measure the K0+500 section. The path sequence is as follows: . 2. Option Two: First measure the K0+500 section, then measure the K0+600 section. The path sequence is as follows: . .

[0126] By comparison, the total distance of Option 1 (490 meters) is much shorter than that of Option 2 (680 meters), therefore it is selected as the optimal path. The calculation results of Dijkstra's algorithm (especially determining the arrival time) The optimal path is This is the key basis for making this final decision.

[0127] S513: Based on the optimized connection sequence of cross-section measurement, perform sequence recombination on all cross-section lines in the cross-section line path set, and connect each recombined cross-section line with the optimized connection path between them in sequence to generate the unmanned vessel navigation path;

[0128] Based on the optimized connection sequence for cross-section measurements, a sequence recombination is performed on all eight cross-section lines in the cross-section line path set. If the optimized connection sequence is K0+500, K0+600, ..., K1+200, and it is a zigzag connection, then the recombined sequence is: cross-section line K0+500 (from right endpoint to left endpoint), connecting path (from left endpoint of K0+500 to left endpoint of K0+600), cross-section line K0+600 (from left endpoint to right endpoint), connecting path (from right endpoint of K0+600 to right endpoint of K0+700), and so on. Each cross-section line in the recombined sequence is sequentially connected end-to-end with its optimized connecting path (i.e., the straight line segment between two cross-section endpoints), ultimately generating a single, continuous, and non-intersecting unmanned surface vessel (USV) overall navigation path. This path completely covers the measurement tasks of all cross-sections with the shortest total voyage.

[0129] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A single-beam unmanned surface vessel (USV) cross-section measurement path planning method based on bank slope gradient, characterized in that, Includes the following steps: S1: Obtain orthophotos and digital elevation models of the river channel locations corresponding to the unmanned vessel cross-section measurement path, and construct the initial layer dataset; S2: Refer to the initial layer dataset to extract the centerline and water level line of the river channel corresponding to the unmanned vessel cross-section measurement path, calculate the riverbank slope angle, and generate a cross-section feature dataset; S3: Obtain the set of unmanned vessel structural parameters, call the cross-sectional feature dataset to calculate the optimal safe distance measurement of the cross-section, and correct the boundary distance of the unmanned vessel cross-section measurement path by adaptive error compensation of the riverbank slope, and output the distance correction result. S4: Referring to the distance correction results, load the river centerline, water level line and cross-section point location data in AutoCAD in sequence to construct the basic point structure of the cross-section measurement path and summarize it into a cross-section measurement boundary point set; S5: Based on the set of boundary points of the cross-section measurement, connect each set of left and right bank boundary points to form a cross-section line and generate the unmanned vessel navigation path. The specific steps for obtaining the ranging correction result are as follows: S311: Obtain the set of structural parameters of the unmanned vessel, including the transducer draft, the horizontal distance from the transducer to the bow and the minimum effective detection depth, and calculate the optimal safe distance between the left and right banks of the cross section by referring to the slope angle of the left and right banks of the river corresponding to each cross section in the cross section feature dataset. S312: Based on the riverbank slope angle in the cross-sectional feature dataset, calculate the degree of change of the riverbank slope along the centerline direction, compare the degree of change with the preset interval, select the corresponding error compensation amount, and establish adaptive error compensation parameters. S313: Call the optimal safe distance measurement of the left and right banks of the cross section and the adaptive error compensation parameters, adjust the original distance measurement value of each cross section and the matching compensation amount, correct the boundary distance of the unmanned vessel cross section measurement path, and output the distance correction result.

2. The single-beam unmanned surface vessel cross-section measurement path planning method based on bank slope according to claim 1, characterized in that, The initial layer dataset includes an orthophoto layer, a digital elevation model layer, and spatial registration control information. The cross-section feature dataset includes river centerline data, water level data, and a set of riverbank slope angles. The distance correction results include corrected distance values ​​for the left bank and the right bank, and cross-section station index values. The cross-section measurement boundary point set includes a set of coordinates for the left bank boundary points, a set of coordinates for the right bank boundary points, and cross-section number information. The unmanned vessel navigation path includes a sequence of cross-section connecting lines, a set of path node numbers, and path layer file encoding.

3. The single-beam unmanned surface vessel cross-section measurement path planning method based on bank slope according to claim 1, characterized in that, The specific steps for obtaining the initial layer dataset are as follows: S111: Collect orthophotos of the river channel location corresponding to the unmanned vessel's cross-section measurement path using a drone camera, and obtain LiDAR point clouds at that river channel location to construct the original river channel image and point cloud set. S112: Call the original river image and the LiDAR point cloud in the point cloud set, perform ground point and non-ground point classification, construct a grid according to the set spatial resolution, calculate the elevation value of each grid node, and generate a digital elevation model of the river. S113: Based on the original river image and the orthophoto in the point cloud set and the corresponding digital elevation model of the river, calibrate the pixel coordinates and geographic coordinates of multiple points with the same name and calculate the coordinate transformation parameters. Project all the pixels of the orthophoto onto the geographic spatial location corresponding to the digital elevation model of the river to construct the initial layer dataset.

4. The single-beam unmanned surface vessel cross-section measurement path planning method based on bank slope according to claim 1, characterized in that, The specific steps for obtaining the cross-sectional feature dataset are as follows: S211: Perform contrast enhancement and edge sharpening processing on the orthophotos in the initial layer dataset, identify the water area boundary based on image texture and edge features, extract the center line and water level line of the river corresponding to the unmanned vessel cross-section measurement path, and establish the geometric data of the river feature line. S212: Call the digital elevation model in the initial layer dataset, and combine it with the bank slope area marked by the river feature line geometric data to calculate the elevation change rate of each elevation data point in the horizontal and vertical directions, use trigonometric functions to convert the slope value, and generate global bank slope raster information. S213: Based on the centerline in the geometric data of the river feature line, generate the cross-section vertical line according to the preset station spacing, determine the intersection of the vertical line with the water level lines on the left and right banks, call the global bank slope raster information, extract the river bank slope angle at each intersection point, and summarize to construct the cross-section feature dataset.

5. The single-beam unmanned surface vessel cross-section measurement path planning method based on bank slope according to claim 1, characterized in that, The specific steps for obtaining the boundary point set of the cross-section measurement are as follows: S411: In AutoCAD, load the coordinate data of the river centerline, water level line, and corresponding cross-section point location data in sequence, align the three in space, determine the intersection of the cross-section point and the water level line, and construct the basic point structure of the cross-section measurement path. S412: Call the boundary distance of each cross-section measurement path in the distance measurement correction result, and for the intersection of the cross-section point and the water level line in the basic point structure of the cross-section measurement path, perform an offset operation equal to the correction distance value along the normal direction of the intersection point to the outside of the bank, and calculate the coordinates of the left and right boundary points of each cross-section. S413: Summarize the coordinates of the left and right boundary points of each section, integrate them in the order of the section station number, and generate a set of section measurement boundary points.

6. The single-beam unmanned surface vessel cross-section measurement path planning method based on bank slope according to claim 1, characterized in that, The steps for obtaining the unmanned vessel's navigation path are as follows: S511: Pair the boundary points of the cross-section measurement that belong to the same cross-section station number on the left and right banks, connect each pair of boundary points to form an independent cross-section line, and establish a set of cross-section line paths. S512: Define the endpoint of each cross-section line in the set of cross-section line paths as a path node, calculate the sailing distance between the endpoints of two specified cross-section lines as the connection weight, and use Dijkstra's algorithm to traverse the minimum sailing distance path of all cross-sections to determine the optimal connection order of cross-section measurement. S513: Based on the optimized connection sequence of the cross-section measurement, perform sequence recombination on all cross-section lines in the cross-section line path set, and connect each recombined cross-section line with the optimized connection path between them in sequence to generate the unmanned vessel navigation path.

Images