Underwater immersed tube positioning monitoring methods, docking monitoring methods and related equipment

Abstract

This invention provides an underwater immersed tube positioning and monitoring method, a docking monitoring method, and related equipment. By pre-setting a black and white checkerboard pattern on the immersed tube and using an underwater robot equipped with a binocular camera to photograph and identify the checkerboard pattern, reliable calculation of the immersed tube's position and attitude can be achieved in complex underwater environments. Compared to existing assisted positioning methods using active laser targets, this invention utilizes a black and white checkerboard pattern as a passive visual marker, which has the advantages of simple structure, low manufacturing cost, convenient deployment, and no need for retrieval. It can effectively reduce the investment in underwater construction auxiliary equipment and work processes. Combined with underwater robot operations, it can also reduce the risks of manual underwater operations and improve the safety, practicality, and engineering economy of immersed tube positioning and monitoring.

Description

Technical Field

[0001] This invention relates to the field of underwater positioning technology, and more specifically, to underwater immersed tube positioning and monitoring methods, docking monitoring methods, and related equipment. Background Technology

[0002] In existing technologies, underwater immersed tunnel construction is widely used in cross-river and cross-sea tunnels, oil and gas transportation, and other fields. Its core involves prefabricating pipe sections on land, then floating, sinking, and docking them to finally lay the pipeline underwater. This construction method has advantages such as minimal impact on waterways, strong environmental adaptability, and controllable construction period, making it one of the preferred solutions for large-scale underwater infrastructure construction. However, precise underwater docking is one of the core challenges that needs to be addressed in underwater immersed tunnel construction. Developing high-precision, high-reliability methods for measuring the relative position and attitude of immersed tunnel sections is of great significance for improving construction efficiency and quality and avoiding economic losses.

[0003] With the development of underwater robots and computer vision technology, underwater binocular vision measurement technology has been widely applied in marine engineering. Currently, the most widely used underwater robot in underwater operations is the Remotely Operated Vehicle (ROV), which can operate in the water for extended periods and possesses functions such as hovering and orientation, enabling it to approach targets for binocular vision measurement. However, the water in the area where immersed tunnel sections are docked is usually turbid, resulting in low contrast and high noise in optical images. Existing methods struggle to extract effective features for high-precision 3D reconstruction of the immersed tunnel sections, and consequently, cannot effectively assess the relative pose relationships of the tunnel sections. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide an underwater immersed tube positioning and monitoring method, a docking monitoring method and related equipment to overcome the shortcomings of the existing technology.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution: Firstly, an underwater immersed tube positioning and monitoring method is applied to an underwater robot equipped with a binocular camera; a black and white checkerboard pattern is provided on the immersed tube; the black and white checkerboard pattern is fixedly installed at a predetermined installation reference position on the immersed tube and has a pre-calibrated installation relationship with the immersed tube coordinate system; the method includes the following steps: S11. Use the binocular camera to capture the black and white checkerboard to obtain a left-eye checkerboard image and a right-eye checkerboard image; extract left corner data from the left-eye checkerboard image and right corner data from the right-eye checkerboard image; match the left corner data and right corner data to obtain at least three non-collinear matching corner data. S12. Using the calibration parameters of the binocular camera, calculate the matching corner point data for each of the three points, transform the matching corner point data into the coordinate system of the left camera, and base the calculation on the side length of the black and white checkerboard grid. Constrain the transformed matching corner point data to obtain the optimized coordinates of each matching corner point in the left eye camera coordinate system; S13. Using the optimized coordinates of any three non-collinear matching corner points, calculate the analytical expression of the left eye plane in the left eye camera coordinate system of the black and white checkerboard; based on the current attitude of the underwater robot, transform the analytical expression of the left eye plane from the left eye camera coordinate system to the global fixed coordinate system, and calculate the absolute attitude of the immersed tube in the global fixed coordinate system according to the relative pose relationship between the black and white checkerboard and the immersed tube.

[0006] Secondly, an underwater immersed tube docking monitoring method is applied to an underwater robot equipped with a binocular camera. Two immersed tubes to be docked are each marked with a black and white checkerboard pattern. The checkerboard pattern is fixed at a predetermined installation reference position on the immersed tube and has a pre-calibrated installation relationship with the tube coordinate system. The method includes the following steps: S21. Using the binocular camera, capture images of the two black and white checkerboard squares to obtain corresponding left-eye checkerboard square images and right-eye checkerboard square images; extract left corner data from the left-eye checkerboard square image, extract right corner data from the right-eye checkerboard square image, and match the left corner data and the right corner data to obtain at least three non-collinear matching corner data. S22. Using the calibration parameters of the binocular camera, calculate the matching corner point data for each of the three points, transform the matching corner point data into the coordinate system of the left camera, and base the calculation on the side length of the black and white checkerboard grid. Constrain the transformed matching corner point data to obtain the optimized coordinates of each matching corner point in the left eye camera coordinate system; S23. Using the optimized coordinates of any three non-collinear matching corner points, calculate the planar analytical expressions of the two black and white checkerboard grids in the left eye camera coordinate system, and determine the planar normal vector of the corresponding black and white checkerboard grids based on the planar analytical expressions. S24. Based on the optimized coordinate values ​​of the matching corner points, the plane normal vector, and the angle relationship between the preset direction and the line connecting the corner points in the black and white checkerboard coordinate system, calculate the direction vector of the x-axis of the two black and white checkerboard grids in the left eye camera coordinate system; calculate the angle between the central axes of the two immersed tubes according to the x-axis direction vectors of the two black and white checkerboard grids; and calculate the relative distance between the two immersed tubes according to the predetermined relative pose relationship between the black and white checkerboard grids and the corresponding immersed tubes. S25. Based on the current posture of the underwater robot, the pose information of the two black and white checkerboard grids in the left eye camera coordinate system is transformed into the global fixed coordinate system, and the absolute posture of the two submerged tubes in the global fixed coordinate system is calculated according to the relative pose relationship between the black and white checkerboard grids and the submerged tubes.

[0007] Thirdly, a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the underwater immersed tube positioning and monitoring method described in the first aspect, or the steps of the underwater immersed tube docking and monitoring method described in the second aspect.

[0008] Fourthly, a computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the underwater immersed tube positioning and monitoring method described in the first aspect, or the steps of the underwater immersed tube docking and monitoring method described in the second aspect.

[0009] In summary, this invention offers the following advantages: By pre-setting a black and white checkerboard pattern on the immersed tube and utilizing an underwater robot equipped with a binocular camera to photograph and identify the checkerboard pattern, reliable calculation of the immersed tube's position and attitude can be achieved in complex underwater environments. Compared to existing assisted positioning methods using active laser targets, this invention utilizes a black and white checkerboard pattern as a passive visual marker, offering advantages such as simple structure, low manufacturing cost, convenient deployment, and no need for retrieval, effectively reducing investment in underwater construction auxiliary equipment and operational procedures. Furthermore, combining this with underwater robot operations reduces the risks associated with manual underwater operations, improving the safety, practicality, and economic efficiency of immersed tube positioning and monitoring. Attached Figure Description

[0010] Figure 1 This is a flowchart of the underwater immersed tube positioning and monitoring method of the present invention; Figure 2 This is a flowchart of the underwater immersed tube docking monitoring method of the present invention; Figure 3 This is an internal structural diagram of the computer device in an embodiment of the present invention; Figure 4 This is a schematic diagram of the underwater submerged pipe positioning in an embodiment of the present invention; Figure 5 This is a schematic diagram of a black and white checkerboard pattern in an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the conversion of a black and white checkerboard pattern into a binary image in Embodiment 1 of the present invention; Figure 7 This is a schematic diagram of the white region dilation of a binary image in Embodiment 1 of the present invention; Figure 8 This is a schematic diagram of convex hull detection in Embodiment 1 of the present invention; Figure 9 This is a schematic diagram of corner detection performed on a defined circular area in Embodiment 1 of the present invention; Figure 10 This is a schematic diagram of the image coordinate system and the left eye camera coordinate system in Embodiment 1 of the present invention; Figure 11 This is a schematic diagram illustrating the acquisition of the direction vector of the x-axis of the chessboard grid in the camera coordinate system in Embodiment 2 of the present invention. Detailed Implementation

[0011] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein.

[0012] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0013] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0014] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0015] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0016] The above description is merely a specific embodiment of this application. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. The protection scope of this application should be determined by the protection scope of the claims.

[0017] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0018] Example 1 To address the aforementioned problems, this invention provides a method for underwater submerged pipe positioning and monitoring, such as... Figure 4 As shown, the underwater immersed tube positioning and monitoring method provided in this embodiment is applied to an underwater robot. The underwater robot is used to carry image acquisition equipment and approach the immersed tube docking area to perform underwater observation tasks. Preferably, the underwater robot is an unmanned remotely operated vehicle (ROV) with multi-degree-of-freedom propulsion capabilities, enabling it to perform forward, backward, surfacing, diving, lateral movement, and hovering operations underwater, thereby allowing the mounted binocular camera to maintain a relatively stable observation posture near the immersed tube interface. The binocular camera is fixedly mounted on the underwater robot and includes a left eye camera and a right eye camera, which are spaced apart along a preset direction to form a binocular baseline. The binocular camera has undergone underwater calibration before being put into use to obtain intrinsic parameters, extrinsic parameters, and distortion parameters. To ensure stable recognition of the black and white checkerboard pattern, the binocular cameras are preferably installed within the forward field of view of the underwater robot, facing the checkerboard pattern set at the interface of the immersed tube. During the shooting process, the underwater robot is controlled to move to a suitable observation position, and its heading, pitch, and roll angles are adjusted to ensure that the black and white checkerboard pattern falls completely within the overlapping imaging area of ​​the left and right cameras. The left and right cameras simultaneously capture images of the black and white checkerboard pattern, obtaining left and right checkerboard images respectively, for subsequent corner extraction, corner matching, and spatial pose calculation.

[0019] like Figure 4 , Figure 5 As shown, the immersed tube is provided with a black and white checkerboard pattern. The checkerboard pattern is fixed at a predetermined installation reference position on the immersed tube and has a pre-calibrated installation relationship with the immersed tube coordinate system. Specifically, the checkerboard pattern is fixed near the interface area of ​​the immersed tube, preferably on the end face of the immersed tube for docking, the surface of the connecting plate parallel to the end face, or fixed on an installation reference surface with a predetermined geometric relationship to the axis of the immersed tube. To ensure that subsequent visual measurement results can accurately represent the spatial positional relationship of the immersed tube, during installation, the plane containing the checkerboard pattern is made parallel or coincident with the corresponding installation reference surface of the immersed tube, and the preset coordinate axis direction of the checkerboard pattern is kept parallel, coincident, or has a pre-calibrated fixed angle relationship with the central axis direction of the immersed tube. Preferably, the x-axis direction of the checkerboard pattern is parallel to the central axis direction of the immersed tube, the origin of the checkerboard pattern is set at a predetermined corner point, and the positional offset of this origin relative to the reference point of the immersed tube is determined by measurement or pre-calibration, thereby establishing a one-to-one correspondence between the checkerboard coordinate system and the immersed tube coordinate system.

[0020] The black and white checkerboard pattern can be fixed to the surface of the immersed tube or its auxiliary installation components through at least one of the following methods: bolting, welding, gluing, snap-fitting, or bracket installation. It is preferably made of waterproof, pressure-resistant, and deformation-resistant materials to ensure that it does not experience significant displacement, warping, or detachment during transportation, sinking, and underwater docking. Furthermore, the size, grid side length, and installation position of the black and white checkerboard pattern can be preset according to the working distance, field of view, and docking accuracy requirements of the binocular camera to ensure complete imaging of the checkerboard pattern at the predetermined shooting position and to guarantee the stability of corner point extraction and 3D reconstruction.

[0021] like Figure 1 As shown, the underwater immersed tube positioning and monitoring method includes the following steps: S1. Control the underwater robot equipped with binocular cameras to move to the vicinity of the immersed tube interface area, and adjust the robot's attitude and hovering position so that the black and white checkerboard pattern set on the immersed tube is within the effective observation range of the binocular cameras. During shooting, the left and right cameras simultaneously acquire images of the black and white checkerboard pattern, obtaining left-eye and right-eye checkerboard patterns respectively. Preferably, during the shooting process, an LED supplementary lighting device can be used to illuminate the black and white checkerboard area to improve the brightness and contrast of the image in turbid water environments, thereby providing clear and stable image data for subsequent corner point extraction and matching. In this embodiment, the black and white checkerboard pattern does not need to emit light itself and interacts with the robot for positioning; therefore, the black and white checkerboard pattern does not need to be retrieved, effectively reducing positioning costs.

[0022] In this embodiment, since the left and right checkerboard images are captured by the left and right cameras of the binocular camera from different observation positions, the same checkerboard corner will be imaged in both images. To achieve corner matching and 3D coordinate calculation, the checkerboard corners need to be accurately extracted from both images. Although the black and white checkerboard itself has advantages such as clear black and white distinction, distinct boundaries, and regular corners, in an underwater environment, factors such as water turbidity, suspended particle scattering, uneven local lighting, and background noise interference can easily lead to missed detections, false detections, or significant positioning errors when directly performing corner identification on the original images. Therefore, this embodiment employs a step-by-step corner extraction method for the left and right checkerboard images.

[0023] Specifically, such as Figure 6 As shown, local average adaptive thresholding is first applied to the left and right checkerboard images respectively. Since the brightness of different areas often varies during underwater photography, directly using a globally uniform threshold for binarization can easily lead to inaccurate segmentation of the black and white checkerboard patterns in some areas. However, by using local average adaptive thresholding, the threshold can be adaptively determined based on the grayscale distribution of the local neighborhood of the image, thus better preserving the brightness differences between the black and white checkerboard regions and obtaining the corresponding binarized image. In this binarized image, the black and white checkerboard regions are distinguished, providing a foundation for subsequent image patch recognition.

[0024] Subsequently, as Figure 7 As shown, the binarized image is dilated. By appropriately dilating the white areas, the boundary range of the black areas can be reduced accordingly, thereby further separating black checkerboard regions that might otherwise be close together due to image noise, edge adhesion, or local blurring, reducing the adhesion between adjacent black blocks. After dilation, each black checkerboard region in the image can be presented as a relatively independent black image block, which is beneficial for subsequent geometric recognition of individual checkerboard units.

[0025] Based on this, such as Figure 8As shown, convex hull calculation and polygon detection are performed on each black image block after dilation. Specifically, the boundary contour of each continuous black image block is first extracted, then its convex hull is calculated based on the boundary contour, and polygon fitting is performed on the convex hull contour to determine whether the corresponding image block has quadrilateral features. Since an individual black square in a checkerboard pattern should ideally appear as a regular quadrilateral, constraints such as aspect ratio, area range, and side length ratio can be further combined to filter the detected quadrilateral regions, eliminating pseudo-quadrilaterals formed by noise, stray shadows, background textures, or other non-target regions, thereby retaining candidate checkerboard units that meet the quadrilateral constraints.

[0026] Furthermore, such as Figure 9 As shown, after obtaining candidate checkerboard cells, each quadrilateral is treated as an independent cell, and its vertex distribution is used to determine whether there are neighboring quadrilateral cells around it. If neighboring quadrilaterals exist, it means that the two quadrilaterals correspond to adjacent color blocks in the checkerboard, and theoretically, they share a common corner point. At this time, the adjacent vertices of the two adjacent quadrilaterals can be connected, and a candidate search circle can be constructed using the connecting line as the diameter of the circle. Since the true common corner point between adjacent checkerboard cells must fall within the area defined by this circle, the corner search range can be narrowed from the entire image to a local high-confidence area. Then, the original image is returned within the circular area, and the Harris corner detection algorithm is used for more refined corner detection. Compared with the aforementioned coarse localization method based on region contours, the Harris algorithm can make full use of local grayscale variation information to achieve sub-pixel level corner localization, thereby obtaining a unique optimal corner point within the circular area.

[0027] After determining the region where the common corner points are located, fine corner detection is then performed on the original image within this local area to obtain the coordinates of the checkerboard corner points. Fine detection is preferably performed on the original image rather than the dilated binary image because the original image retains richer grayscale variation information, which improves the sub-pixel accuracy of corner point localization. Through this method of first coarse localization and then fine detection, the true corner positions on the checkerboard can be identified more accurately in complex underwater environments. Finally, the checkerboard corner coordinates extracted from the left-view checkerboard image are used as the left corner data, and the checkerboard corner coordinates extracted from the right-view checkerboard image are used as the right corner data.

[0028] In this embodiment, after extracting the left and right corner data from the left and right checkerboard images respectively, it is necessary to further determine which points in the two sets of corner points correspond to the same actual checkerboard corner point in order to perform binocular 3D reconstruction. Since the left and right cameras capture the same black and white checkerboard from different perspectives, although the objects captured are the same, the positions of the same corner point in the left and right images are not completely consistent due to factors such as differences in perspective, perspective distortion, changes in underwater lighting, and local occlusion. Furthermore, some corner points may be missed, falsely detected, or have disordered sorting.

[0029] This embodiment further provides a corner point matching step: First, each corner point in the left and right corner point data is numbered. This numbering is preferably based on the corner point's position within its respective chessboard grid, for example, according to the chessboard grid's row and column order, from top to bottom and left to right, or by uniformly sorting the corner points according to a preset origin and coordinate axis direction. Since the left and right images capture the same chessboard grid, corner points in the same arrangement should theoretically correspond to the same physical corner point. Based on this, an initial corner point correspondence between the left and right corner point data can be established. The purpose of this step is to provide a rough but clearly structurally constrained matching result, laying the foundation for subsequent mismatch removal.

[0030] After obtaining the initial corner point correspondences, the RANSAC algorithm is further used to iteratively filter the initial matching results. Specifically, each initial corresponding corner point pair can be used as a candidate sample. Based on the imaging geometry of the stereo camera and known calibration parameters, it is determined whether each corner point pair satisfies epipolar constraints, disparity range constraints, or reprojection error constraints. Based on multiple random samplings, the RANSAC algorithm searches for model parameters that enable the most corner point pairs to satisfy the geometric constraints. Corner point pairs consistent with this model are retained as interior points, while corner point pairs that deviate significantly from the model are considered mismatches and discarded. This step effectively removes erroneous correspondences caused by incorrect corner point extraction, numbering bias, or image noise, thereby improving the accuracy and stability of the matching results.

[0031] In certain situations, such as severe local blurring of the image, occlusion of some checkerboard edges, or uneven underwater lighting leading to unstable corner detection results, the RANSAC algorithm may not be able to automatically complete all corner matching. In such cases, this embodiment further introduces a manual interaction method to confirm the remaining corners. Specifically, the operator can observe the candidate corner positions in the left and right checkerboard images on the host computer interface and manually select the corresponding corners based on the regular arrangement of the checkerboard, or correct the automatic matching results. Manual confirmation compensates for the shortcomings of the automatic algorithm in complex underwater environments, ensuring that a complete and reliable matching relationship is ultimately obtained. After completing the RANSAC screening and manual interaction confirmation, the effective matching corner data between the left and right checkerboard images can be obtained.

[0032] S2. The calibration parameters of the binocular camera are used to calculate the data of each matching corner point, transforming the matching corner point data into the coordinate system of the left-eye camera. Specifically, in this embodiment, the left-eye and right-eye checkerboard images captured by the binocular camera are essentially two-dimensional planar image information. Therefore, only the position of the corner point in the image plane can be obtained from a single image, and the depth information of the corner point in real space cannot be directly obtained. To recover the three-dimensional position of the checkerboard corner point in space, it is necessary to use the parameters pre-calibrated by the binocular camera to perform three-dimensional calculations on the data of each matching corner point, and represent the calculation results uniformly in the coordinate system of the left-eye camera. In this embodiment, the three-dimensional information is represented based on the three-dimensional coordinate system constructed by the left-eye camera. Those skilled in the art will understand that the right-eye camera can also be used as a reference to construct a corresponding three-dimensional coordinate system. However, for the convenience of subsequent unified calculation, this embodiment preferably uses the coordinate system of the left-eye camera as the expression benchmark for the three-dimensional coordinates, so the construction process of the right-eye camera coordinate system will not be described again.

[0033] Specifically, the calibration parameters of the stereo camera are first obtained, including at least the baseline length of the stereo camera. and focal length Wherein, the baseline length This refers to the distance between the optical centers of the left and right cameras; this parameter reflects the spatial interval between the two cameras in a binocular camera. The focal length... This refers to the focal length parameter in the camera imaging model, used to reflect the projection relationship between image coordinates and spatial coordinates. The baseline length... and focal length All of these parameters can be obtained through pre-calibration using a binocular camera and used as the basis for subsequent 3D reconstruction.

[0034] After obtaining the calibration parameters of the stereo camera, for each matching corner point data The left corner data in the left eye image and the right corner data in the right eye image are determined respectively. Indicates the first The coordinates of the matching corner points in the left eye image coordinate system This represents the coordinates of the same matching corner point in the right eye image coordinate system. Here, and These represent the horizontal and vertical coordinates of the corner point in the left-eye image, respectively. and These represent the horizontal and vertical coordinates of the corner point in the right-eye image, respectively. It should be noted that the left and right corner point data are preferably image coordinates after camera intrinsic parameter correction and distortion correction to improve the accuracy of subsequent 3D calculations.

[0035] Subsequently, based on the principle of binocular imaging, the initial three-dimensional coordinates of the matching corner data in the coordinate system of the left eye camera are calculated. In this embodiment, the formula for calculating the initial three-dimensional coordinates is: ; in, Indicates the first The initial 3D coordinates of the matching corner points in the left eye camera coordinate system; This indicates the lateral coordinates of the corner point in the left-eye camera coordinate system; This indicates the vertical coordinate of the corner point in the left-eye camera coordinate system; This represents the depth coordinate of the corner point in the left eye camera's coordinate system, that is, the distance of the corner point relative to the left eye camera along the optical axis. In the formula, Indicates the first The difference in lateral coordinates of a matching corner point between the left and right images is also called disparity. Disparity is a key quantity in binocular ranging, and its physical meaning is: the closer a point in space is to the binocular camera, the greater the difference in its position between the left and right images; conversely, the farther away it is, the smaller the difference in its position between the left and right images. Therefore, through baseline length... ,focal length and parallax By combining the information from each corner point, the depth information can be obtained. After obtaining the depth information, it is then combined with the two-dimensional coordinates in the left eye image. and This allows for the further calculation of the lateral position of the corner point in the left eye camera's 3D coordinate system. and longitudinal position .

[0036] Through the above calculations, the two-dimensional matching corner data in the left and right eye images can be converted into initial three-dimensional coordinate data in the coordinate system of the left eye camera. This initial three-dimensional coordinate data represents the spatial position of the checkerboard corner points relative to the left eye camera. Furthermore, due to the complexity of the underwater imaging environment, the corner point extraction results in the left and right eye images inevitably contain certain errors, such as pixel-level positioning errors, parallax calculation errors, and reconstruction errors caused by local image blurring. Therefore, the initial three-dimensional coordinates obtained directly often do not strictly satisfy the geometric dimensions of the real checkerboard. Specifically, for two adjacent corner points on the same black and white checkerboard, their actual spatial distance should be equal to the pre-known side length of the checkerboard. The distance between the initial 3D coordinates directly obtained from binocular reconstruction usually deviates from the theoretical value. To improve the accuracy of subsequent planar calculations and immersed tube attitude calculations, this embodiment further uses the side length of a checkerboard grid. Constraint optimization is performed on the transformed matching corner point data to obtain the optimized coordinate values ​​of each matching corner point in the left eye camera coordinate system.

[0037] First, based on the structural characteristics of the black and white checkerboard, distance constraints are established between adjacent corner points on the same black and white checkerboard. This is because the side length of the checkerboard... Known before actual production and installation, the optimized three-dimensional distance between each pair of spatially adjacent chessboard corner points should be equal to the side length of the stated grid. The constraints are as follows. In other words, during the optimization process, the initial three-dimensional coordinates obtained from binocular measurements are no longer relied upon alone. Instead, the geometric prior information of the chessboard itself is introduced, so that the optimized corner point three-dimensional coordinates are more consistent with the spatial structure of the real chessboard.

[0038] In this embodiment, an optimization model with distance constraints is constructed with the objective of minimizing the deviation between the initial 3D coordinates and the optimized 3D coordinates of each matching corner point. Specifically, it is represented as follows: ; in, This represents the total number of matching corner points participating in the optimization; Indicates the first The initial three-dimensional coordinates of the matching corner points in the coordinate system of the left eye camera are calculated from the binocular camera calibration parameters and the matching corner point data in the left and right images. Indicates the first The three-dimensional coordinates of each matching corner point in the optimized left eye camera coordinate system, which is the final optimized value of the matching corner point coordinates required in this embodiment; Indicates the relationship with the first The adjacent corner point of the matching point The 3D coordinates of each matching corner point in the optimized left-eye camera coordinate system; symbol This represents the Euclidean norm, used to express the Euclidean distance between two points in three-dimensional space. The side length of each square in the black and white checkerboard is a fixed value that is known in advance.

[0039] objective function This represents the optimized 3D coordinates of each matching corner point while satisfying the geometric constraints of the chessboard grid. As close as possible to its initial three-dimensional coordinates In other words, the optimization process does not completely change the original reconstruction results, but rather performs small-scale corrections based on the original binocular reconstruction results to reduce the possibility of excessive corner coordinate offsets and ensure that the optimized results remain consistent with the actual measurement data.

[0040] Constraints This means that for any pair of adjacent corner points on the same chessboard grid, the square of their optimized distance in 3D space should be equal to the square of the side length of the chessboard grid. This constraint allows the known structural dimensions of the chessboard grid to be incorporated into the 3D coordinate solution process, thereby correcting the geometric distortion caused by relying solely on image matching and making the optimized corner point distribution more consistent with the regular grid structure of a real chessboard. After establishing the above optimization model, the model can be solved using the least squares method, constraint optimization algorithm, or other commonly used numerical optimization methods in this field to obtain the optimized coordinates of each matching corner point in the left-eye camera coordinate system that satisfy the distance constraint.

[0041] S3. After obtaining the optimized coordinate values ​​of each matching corner point in the left-eye camera coordinate system, it is necessary to further utilize these three-dimensional corner point data to calculate the spatial geometric parameters of the plane containing the black and white checkerboard. Since the black and white checkerboard itself is a planar structure, any three non-collinear corner points on it can uniquely determine a spatial plane. Based on this, in this embodiment, any three non-collinear corner point coordinate optimization values ​​are selected from the optimized matching corner point coordinate values, denoted as the first corner point, the second corner point, and the third corner point, respectively. Based on the three-dimensional coordinates of these three corner points in the left-eye camera coordinate system, the plane equation of the plane containing the black and white checkerboard is established, thereby obtaining the analytical expression of the left-eye plane of the black and white checkerboard in the left-eye camera coordinate system.

[0042] Specifically, let the three-dimensional coordinates of the first corner point, the second corner point, and the third corner point in the left eye camera coordinate system be as follows: , , After determining the three non-collinear corner points, a plane equation for the plane containing the black and white checkerboard is established based on these three points. In this embodiment, the plane equation is expressed as: ;in, These represent the three-dimensional coordinates of any point on the plane in the coordinate system of the left eye camera; The parameters to be solved are given. By solving for the parameter values, the equation corresponding to the plane can be determined. Specifically, two direction vectors can be constructed first, pointing from the first corner point to the second corner point and from the first corner point to the third corner point. Then, the normal vector of the plane is obtained from the cross product of these two direction vectors, and thus the plane equation parameters can be obtained. In this embodiment, the parameters to be solved can be calculated using the following formula: ; By using the values ​​of three non-collinear corner points Substituting these values ​​into the formula, we can calculate the analytical expression for the plane.

[0043] Furthermore, the analytical expression for the black and white checkerboard plane obtained in the aforementioned steps is based on the coordinate system of the left-eye camera, reflecting the local spatial position and orientation information of the checkerboard relative to the left-eye camera. However, in actual immersed tunnel construction monitoring scenarios, obtaining only the local pose of the checkerboard relative to the left-eye camera is usually insufficient to directly guide the docking operation. It is necessary to further transform the local measurement results into a pre-set global fixed coordinate system at the construction site to obtain the absolute attitude information of the immersed tunnel in a unified reference system, facilitating comparison with the design position, design direction, or other external measurement results. Therefore, this embodiment further transforms the analytical expression for the left-eye plane from the coordinate system of the left-eye camera to the global fixed coordinate system based on the current attitude of the underwater robot, and combines the pre-known relative pose relationship between the black and white checkerboard and the immersed tunnel to finally calculate the absolute attitude of the immersed tunnel in the global fixed coordinate system.

[0044] Specifically, the current pose information of the underwater robot is first obtained. This current pose information includes the underwater robot's position parameters and attitude parameters in a global fixed coordinate system. The position parameters characterize the underwater robot's spatial position in the global fixed coordinate system, such as three-dimensional coordinates or a Cartesian coordinate system. The attitude parameters characterize the underwater robot's orientation, such as roll angle, pitch angle, and heading angle. The global fixed coordinate system is preferably a pre-established engineering coordinate system, task coordinate system, or other fixed reference coordinate system for the construction area, which does not change with the underwater robot's movement.

[0045] After obtaining the current pose information of the underwater robot, the mounting pose parameters of the left eye camera relative to the underwater robot are further obtained. These mounting pose parameters describe the fixed spatial relationship of the left eye camera relative to the underwater robot's reference coordinate system after it is mounted on the robot's body. Since the left eye camera is mounted on the underwater robot, its coordinate system is usually fixed relative to the underwater robot's coordinate system. These mounting pose parameters can be pre-calibrated after the equipment is installed.

[0046] After obtaining the above two sets of parameters, a coordinate transformation relationship from the left-eye camera coordinate system to the global fixed coordinate system is established based on the current pose information of the underwater robot and the installation pose parameters of the left-eye camera relative to the underwater robot. In simpler terms, this step means: first, knowing "where the left-eye camera is mounted and what its orientation is" relative to the underwater robot, and then combining this with the current "where and what its orientation is" of the underwater robot in the global fixed coordinate system, the position and orientation of the left-eye camera in the global fixed coordinate system can be determined. Based on this coordinate transformation relationship, the spatial points, direction vectors, and planar parameters in the left-eye camera coordinate system can be converted to the global fixed coordinate system. For ease of understanding, the coordinate transformation relationship can be simplified as follows: ;in, This represents the coordinates of a point in the left-eye camera coordinate system. This indicates the coordinates of the point in space after transformation to the global fixed coordinate system; This represents the rotational transformation from the left-eye camera coordinate system to the global fixed coordinate system, used to characterize changes in direction; This represents the translation transformation relationship, used to characterize the positional offset. Through this coordinate transformation relationship, the plane normal vector and plane position parameters corresponding to the left eye plane's analytical expression can be synchronously transformed to the global fixed coordinate system, thereby obtaining the global plane analytical expression of the black and white chessboard in the global fixed coordinate system.

[0047] After obtaining the global planar analytical expression of the black and white checkerboard pattern, and combining it with the pre-determined relative pose relationship between the checkerboard pattern and the immersed tube, the absolute attitude of the immersed tube in the global fixed coordinate system can be further calculated. The relative pose relationship mentioned here refers to the fixed installation relationship between the checkerboard coordinate system and the immersed tube coordinate system. Examples include the parallelism between the plane containing the checkerboard and the end face of the immersed tube, the parallelism between the x-axis of the checkerboard and the central axis of the immersed tube, and the distance offset between the origin of the checkerboard and the pre-defined reference point of the immersed tube. Since this installation relationship is known and remains unchanged after installation, after obtaining the global position and orientation of the checkerboard, the position and orientation of the immersed tube in the global fixed coordinate system can be further determined through coordinate transformation.

[0048] In summary, the underwater immersed tube positioning and monitoring method provided in this embodiment, by pre-setting a black and white checkerboard pattern on the immersed tube and using an underwater robot equipped with a binocular camera to photograph and identify the checkerboard pattern, can reliably calculate the position and attitude of the immersed tube in complex underwater environments. Compared with existing assisted positioning methods using active laser targets, this invention utilizes a black and white checkerboard pattern as a passive visual marker, which has the advantages of simple structure, low manufacturing cost, convenient deployment, and no need for retrieval, effectively reducing the investment in underwater construction auxiliary equipment and the workload. Furthermore, combined with underwater robot operations, it can also reduce the risks of manual underwater operations and improve the safety, practicality, and economic efficiency of immersed tube positioning and monitoring.

[0049] Example 2 Based on Example 1, this application further provides an underwater immersed tube docking monitoring method, applied to an underwater robot equipped with a binocular camera; two immersed tubes to be docked are respectively provided with a black and white checkerboard pattern, the checkerboard pattern is fixedly set at a predetermined installation reference position of the immersed tube, and has a pre-calibrated installation relationship with the immersed tube coordinate system, the method includes the following steps: S1. Use the binocular camera to capture images of the two black and white checkerboard grids respectively, and obtain corresponding left-eye checkerboard grid images and right-eye checkerboard grid images; extract left corner data from the left-eye checkerboard grid image, extract right corner data from the right-eye checkerboard grid image, and match the left corner data and the right corner data to obtain at least three non-collinear matching corner data. like Figure 4 As shown, in this embodiment, the binocular cameras can simultaneously acquire images of two black and white checkerboard patterns on the two immersed tubes to be docked at the same shooting time. That is, both the left and right cameras capture images containing two black and white checkerboard patterns. To address this, the target checkerboard patterns can first be distinguished and selected based on their distribution location, size range, installation area, or preset numbering information in the images. Then, corner point extraction and left / right camera matching processing are performed on the selected checkerboard patterns respectively, thereby obtaining the matching corner point data corresponding to each immersed tube. In this way, measurement information for two immersed tubes can be acquired simultaneously in a single shooting, which is beneficial for improving docking monitoring efficiency.

[0050] Those skilled in the art will understand that, in other embodiments, the underwater robot can be controlled to sequentially photograph the two black and white checkerboard patterns on the two submerged tubes. That is, first, the black and white checkerboard pattern on one submerged tube is photographed, and corner point extraction and matching are completed. Then, the black and white checkerboard pattern on the other submerged tube is photographed and the same processing is performed, ultimately obtaining the matching corner point data corresponding to the two black and white checkerboard patterns respectively. This method helps avoid recognition interference that may be caused when multiple checkerboard patterns appear simultaneously in the same image, and is suitable for application scenarios with limited field of view, significant occlusion, or high image resolution requirements.

[0051] Regardless of whether simultaneous or segmented shooting is used, the core objective is to accurately extract the corner points on the target black and white checkerboard and establish the correspondence between the left and right eyes. The processes of corner point extraction, corner point numbering, matching and filtering, and obtaining matching corner point data can be implemented by referring to the relevant steps in Example 1, and will not be repeated in this example.

[0052] S2. Using the calibration parameters of the binocular camera, calculate the matching corner point data for each point, transform the matching corner point data into the left camera coordinate system, and base the calculation on the side length of the black and white checkerboard grid. Constrain the transformed matching corner point data to obtain the optimized coordinates of each matching corner point in the left eye camera coordinate system; Similarly, in this embodiment, the specific implementation process of the scheme described in step S2 is consistent with that in embodiment one. Specifically, firstly, using the baseline length, focal length, and related calibration parameters obtained from the pre-calibrated binocular camera, three-dimensional calculations are performed on each set of matching corner point data in the left and right eye images, converting the original two-dimensional matching corner point data located in the left and right eye image planes into initial three-dimensional coordinates in the left eye camera coordinate system. Subsequently, combined with the known side lengths of the black and white checkerboard grid... This process establishes distance constraints between adjacent corner points on the same checkerboard grid. The initial 3D coordinates are then constrained and optimized to minimize the deviation between the initial and optimized 3D coordinates of each matched corner point, thereby obtaining optimized corner point coordinate values ​​that satisfy the actual geometric dimensions of the checkerboard grid. This step effectively reduces the 3D reconstruction deviation caused by underwater image noise, corner point extraction errors, and binocular matching errors, improving the accuracy and stability of subsequent planar analytical solutions and immersed tube position and attitude calculations. Since the specific calculation principles, constraint methods, and optimization processes of this step can be directly implemented with reference to the relevant descriptions in Example 1, they will not be repeated in this example. In this embodiment, after completing step S2, the optimized coordinates of each matching corner point on the two black and white checkerboard grids in the coordinate system of the left eye camera are obtained. Based on these optimized three-dimensional corner point coordinates, the spatial parameters of the plane containing the two black and white checkerboard grids, the x-axis direction vector of the checkerboard grids, the relative angle and relative distance between the two immersed tubes, and the absolute attitude of the two immersed tubes in the global fixed coordinate system can be further solved. The specific implementation process is described below in conjunction with steps S3 to S5.

[0053] S3. Using the optimized coordinates of any three non-collinear matching corner points, calculate the planar analytical expressions of the two black and white checkerboard grids in the left eye camera coordinate system, and determine the planar normal vector of the corresponding black and white checkerboard grids based on the planar analytical expressions. The step of solving the planar analytical expressions of the two black and white checkerboard squares in the left-eye camera coordinate system in step S3 can be found in the planar analytical expression solution steps in Example 1, thus obtaining the planar analytical expressions of the two black and white checkerboard squares: ; ; After obtaining the analytical expressions for the two planes, the unit normal vectors of the two planes can be calculated separately: ; ; in, Used to indicate the orientation of the plane containing the first black and white checkerboard square. This is used to represent the orientation of the plane containing the second black and white checkerboard. Step S3 completes the separate solutions for the analytical expressions of the two black and white checkerboard planes and their corresponding normal vectors, and establishes a unified expression between the two planes in the same left-eye camera coordinate system, providing a foundation for the subsequent solution of the docking relationship between the two immersed tubes.

[0054] S4. Based on the optimized coordinates of the matched corner points, the plane normal vector, and the angle relationship between the preset direction and the line connecting the corner points in the black and white checkerboard coordinate system, calculate the direction vectors of the x-axis of the two black and white checkerboard grids in the left eye camera coordinate system; specifically, as follows... Figure 11 As shown, let the coordinates of any two corner points on the first black and white chessboard be respectively in the chessboard coordinate system. , .

[0055] Then vector In the checkerboard coordinate system, the angle between the angle and the checkerboard x-axis is denoted as . Correspondingly, the optimized 3D coordinates of these two corner points in the left eye camera coordinate system are as follows: and Thus, the corresponding space vector is obtained. Since the x-axis of the first black and white checkerboard lies in the first plane, the vector can be... unit normal vector about the first plane Rotation angle The direction vector of the x-axis of the first black and white checkerboard grid in the left-eye camera coordinate system is obtained, denoted as... In terms of implementation, it can be first determined based on the axis of rotation. and rotation angle Constructing quaternions: .

[0056] Then, based on the Rodriguez rotation formula, the quaternion is transformed into a rotation matrix. Finally, using a rotation matrix Acting on vectors This yields a direction vector parallel to the x-axis of the first black and white checkerboard. .

[0057] Similarly, let the coordinates of any two corner points on the second black and white chessboard in the chessboard coordinate system be respectively , After undergoing the same transformation process as the first black and white checkerboard, the direction vector of the x-axis of the second black and white checkerboard can be obtained. .

[0058] In this embodiment, since the x-axis of the first black and white checkerboard pattern is parallel to the central axis of the first immersed tube and the x-axis of the second black and white checkerboard pattern is parallel to the central axis of the second immersed tube during installation, the angle between the x-axis direction vectors of the two black and white checkerboard patterns is the angle between the central axes of the two immersed tubes. This angle is denoted as... The calculation formula is as follows: ;in, This represents the dot product of two direction vectors. Let these represent the magnitudes of the two direction vectors, respectively. This represents the spatial angle between the central axes of the two immersed tubes. This angle directly reflects the degree of directional deviation between the two immersed tubes to be docked.

[0059] Furthermore, based on the predetermined relative pose relationship between the black and white checkerboard pattern and the corresponding immersed tubes, and the immersed tube model, the three-dimensional coordinates of any reference point on the first and second immersed tubes in the left-eye camera coordinate system can be determined. Let the coordinates of the reference point on the first immersed tube be... The coordinates of the reference point on the second immersed tube are: The relative distance between the two immersed tubes It can be represented as: ;in, This represents the Euclidean distance between two reference points, which can be used to characterize the end spacing, offset distance, or other predetermined positional relationship between two immersed tubes. The included angle is obtained simultaneously. and distance This allows for comprehensive monitoring of the docking status of the two immersed tubes.

[0060] S5. Based on the current posture of the underwater robot, the pose information of the black and white checkerboard in the left eye camera coordinate system is transformed into the global fixed coordinate system, and the absolute posture of the two immersed tubes in the global fixed coordinate system is calculated according to the relative pose relationship between the black and white checkerboard and the immersed tube.

[0061] For step S5, after obtaining the pose information of the two black and white checkerboard squares in the left-eye camera coordinate system, the pose information can be further transformed into a global fixed coordinate system based on the current attitude of the underwater robot. Then, according to the relative pose relationship between the black and white checkerboard squares and the immersed tubes, the absolute attitudes of the two immersed tubes in the global fixed coordinate system are calculated respectively. Specifically, firstly, the current pose information of the underwater robot is obtained, including the position parameters and attitude parameters of the underwater robot in the global fixed coordinate system; then, the installation pose parameters of the left-eye camera relative to the underwater robot are obtained. Based on these two parameters, a coordinate transformation relationship between the left-eye camera coordinate system and the global fixed coordinate system can be established. Through this coordinate transformation relationship, the planar parameters, normal vectors, x-axis directions, and feature point coordinates of the first and second black and white checkerboard squares in the left-eye camera coordinate system can be transformed into the global fixed coordinate system. Since the installation relationships between the two black and white checkerboard grids and their respective immersed tubes are known and fixed in advance, once the positions and orientations of the two checkerboard grids in the global fixed coordinate system are obtained, the absolute attitudes of the first and second immersed tubes in the global fixed coordinate system can be calculated respectively. The absolute attitudes mentioned here include the position and orientation parameters of the immersed tubes in the global fixed coordinate system, which can be used to further evaluate the deviation between the immersed tubes and their designed installation positions, guide the docking construction of the immersed tubes, and conduct real-time monitoring during the construction process.

[0062] In summary, this embodiment, based on immersed tube positioning and monitoring, further proposes a method for monitoring immersed tube docking in underwater construction scenarios. This method involves setting up a black and white checkerboard pattern on two immersed tubes to be docked, and using an underwater robot equipped with a binocular camera to acquire images, extract corner points, reconstruct 3D structures, and calculate attitude on the two checkerboard patterns. This allows for the acquisition of the relative angle and distance between the two immersed tubes, enabling real-time monitoring and evaluation of the docking status. Compared to existing technologies that rely on active laser targets for assisted positioning, this application uses a passive black and white checkerboard pattern as a visual marker, offering advantages such as simple structure, convenient deployment, low manufacturing cost, and no need for retrieval, effectively reducing the cost of underwater engineering. Furthermore, this application's calculation of the immersed tube's attitude and position does not depend on a specific orientation of the tube. Even if the two immersed tubes are deflected at different angles in space, their positional relationship can still be accurately calculated through binocular visual measurement results, thus possessing higher adaptability, degree of freedom, and engineering practical value.

[0063] Example 3 A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the underwater submerged pipe positioning and monitoring method as described in Embodiment 1, or the underwater submerged pipe docking and monitoring method as described in Embodiment 2.

[0064] Example 4 In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. When the computer program is executed by the processor, it implements the underwater submerged pipe positioning and monitoring method as described in Example 1, or the underwater submerged pipe docking and monitoring method as described in Example 2.

[0065] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0066] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0067] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0068] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. An underwater immersed tube positioning and monitoring method, applied to an underwater robot, wherein the underwater robot is equipped with a binocular camera; a black and white checkerboard pattern is set on the immersed tube; the black and white checkerboard pattern is fixedly set at a predetermined installation reference position of the immersed tube and has a pre-calibrated installation relationship with the immersed tube coordinate system; characterized in that, The method includes the following steps: S11. Use the binocular camera to capture the black and white checkerboard to obtain a left-eye checkerboard image and a right-eye checkerboard image; extract left corner data from the left-eye checkerboard image and right corner data from the right-eye checkerboard image; match the left corner data and right corner data to obtain at least three non-collinear matching corner data. S12. Using the calibration parameters of the binocular camera, calculate the matching corner point data for each of the three points, transform the matching corner point data into the coordinate system of the left camera, and base the calculation on the side length of the black and white checkerboard grid. Constrain the transformed matching corner point data to obtain the optimized coordinates of each matching corner point in the left eye camera coordinate system; S13. Using the optimized coordinates of any three non-collinear matching corner points, calculate the analytical expression of the left eye plane in the left eye camera coordinate system of the black and white checkerboard; based on the current attitude of the underwater robot, transform the analytical expression of the left eye plane from the left eye camera coordinate system to the global fixed coordinate system, and calculate the absolute attitude of the immersed tube in the global fixed coordinate system according to the relative pose relationship between the black and white checkerboard and the immersed tube.

2. The underwater submerged pipe positioning and monitoring method according to claim 1, characterized in that, The step of extracting left corner data from the left-eye chessboard image and right corner data from the right-eye chessboard image specifically includes: Local average adaptive thresholding is performed on the left and right chessboard images respectively to obtain the corresponding binarized images. The binarized image is dilated to reduce the black checkerboard area and separate adjacent black image blocks; Convex hull calculation and polygon detection are performed on the dilated black image blocks to screen out candidate checkerboard units that meet the quadrilateral constraint conditions. Based on the adjacency relationship of the candidate chessboard grid units, the region where the common corner point between adjacent chessboard grid units is located is determined; Within the area where the common corner point is located, fine corner point detection is performed on the original image to obtain the coordinates of the checkerboard corner point; The coordinates of the corner points of the chessboard in the left-eye chessboard image are used as the left corner point data, and the coordinates of the corner points of the chessboard in the right-eye chessboard image are used as the right corner point data.

3. The underwater submerged pipe positioning and monitoring method according to claim 2, characterized in that, The process of matching the left corner data and the right corner data to obtain at least three non-collinear matching corner data specifically includes: The left and right corner data are numbered, and an initial corner correspondence is established based on the arrangement of the corners in their respective chessboard squares. The RANSAC algorithm is used to iteratively filter the initial corner point correspondence to remove mismatched corner point pairs that do not meet the geometric constraints of binocular imaging; Based on the corner point correspondence after filtering and confirmation, the effective matching corner point data between the left and right chessboard images are obtained; Select at least three non-collinear corner point pairs from the valid matching corner point data as the matching corner point data used for subsequent 3D parameter solving.

4. The underwater submerged pipe positioning and monitoring method according to claim 1, characterized in that, The step of calculating each matching corner point data using the calibration parameters of the binocular camera and transforming the matching corner point data into the coordinate system of the left eye camera specifically includes: Obtain the baseline length of the stereo camera and the focal length of the binocular camera ; Based on the matching corner data Determine the corresponding left corner point data. and the right corner point data Calculate the initial three-dimensional coordinates of the matching corner data in the left eye camera coordinate system. .

5. The underwater submerged pipe positioning and monitoring method according to claim 4, characterized in that, The side length of the grid based on the black and white checkerboard. Constraints are applied to the transformed matching corner data to obtain optimized values ​​of the matching corner coordinates in the left eye camera coordinate system, including: Based on the side length of the grid of the black and white checkerboard. A distance constraint relationship is established between adjacent corner points on the same black and white chessboard. The initial three-dimensional coordinates of the matching corner point data are optimized so that the three-dimensional distance between adjacent corner points is equal to the side length of the chessboard. ; An optimization model with distance constraints is constructed with the goal of minimizing the deviation between the initial 3D coordinates and the optimized 3D coordinates of each matching corner point. The optimization model is solved to obtain the optimized coordinate values ​​of each matching corner point in the left eye camera coordinate system that satisfy the distance constraint.

6. The underwater submerged pipe positioning and monitoring method according to claim 5, characterized in that, The step of using the optimized coordinates of any three non-collinear matching corner points to solve the analytical expression of the left-eye plane in the left-eye camera coordinate system specifically includes: Select any three non-collinear corner point coordinate optimization values ​​from the matched corner point coordinate optimization values, and denot them as the first corner point, the second corner point, and the third corner point, respectively; Based on the three-dimensional coordinates of the first corner point, the second corner point, and the third corner point in the left eye camera coordinate system, the plane equation of the plane containing the black and white checkerboard is established, including: ； Solve the parameters in the plane equation based on the three-dimensional coordinates of the first corner point, the second corner point, and the third corner point. , , and ; The parameters obtained from the solution , , and The left-eye plane analytical expression of the black and white checkerboard grid in the left-eye camera coordinate system.

7. The underwater submerged pipe positioning and monitoring method according to claim 6, characterized in that, The process of transforming the left-eye plane analytical expression from the left-eye camera coordinate system to the global fixed coordinate system based on the current attitude of the underwater robot specifically includes: Obtain the current pose information of the underwater robot, which includes the position parameters and attitude parameters of the underwater robot in the global fixed coordinate system; Obtain the installation pose parameters of the left eye camera relative to the underwater robot; Based on the current pose information and the installation pose parameters, establish the coordinate transformation relationship from the left eye camera coordinate system to the global fixed coordinate system; Based on the coordinate transformation relationship, the planar parameters in the left eye plane analytical expression are transformed into the global fixed coordinate system to obtain the global planar analytical expression of the black and white chessboard in the global fixed coordinate system.

8. An underwater immersed tube docking monitoring method, applied to an underwater robot equipped with a binocular camera; two immersed tubes to be docked are respectively provided with a black and white checkerboard pattern, the checkerboard pattern being fixedly set at a predetermined installation reference position of the immersed tubes and having a pre-calibrated installation relationship with the immersed tube coordinate system, characterized in that, The method includes the following steps: S21. Using the binocular camera, capture images of the two black and white checkerboard squares to obtain corresponding left-eye checkerboard square images and right-eye checkerboard square images; extract left corner data from the left-eye checkerboard square image, extract right corner data from the right-eye checkerboard square image, and match the left corner data and the right corner data to obtain at least three non-collinear matching corner data. S22. Using the calibration parameters of the binocular camera, calculate the matching corner point data for each of the three points, transform the matching corner point data into the coordinate system of the left camera, and base the calculation on the side length of the black and white checkerboard grid. Constrain the transformed matching corner point data to obtain the optimized coordinates of each matching corner point in the left eye camera coordinate system; S23. Using the optimized coordinates of any three non-collinear matching corner points, calculate the planar analytical expressions of the two black and white checkerboard grids in the left eye camera coordinate system, and determine the planar normal vector of the corresponding black and white checkerboard grids based on the planar analytical expressions. S24. Based on the optimized coordinate values ​​of the matching corner points, the plane normal vector, and the angle relationship between the preset direction and the line connecting the corner points in the black and white checkerboard coordinate system, calculate the direction vector of the x-axis of the two black and white checkerboard grids in the left eye camera coordinate system; calculate the angle between the central axes of the two immersed tubes according to the x-axis direction vectors of the two black and white checkerboard grids; and calculate the relative distance between the two immersed tubes according to the predetermined relative pose relationship between the black and white checkerboard grids and the corresponding immersed tubes. S25. Based on the current posture of the underwater robot, the pose information of the two black and white checkerboard grids in the left eye camera coordinate system is transformed into the global fixed coordinate system, and the absolute posture of the two submerged tubes in the global fixed coordinate system is calculated according to the relative pose relationship between the black and white checkerboard grids and the submerged tubes.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the underwater immersed tube positioning and monitoring method as described in any one of claims 1-7, or the underwater immersed tube docking and monitoring method as described in claim 8.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the underwater immersed tube positioning and monitoring method as described in any one of claims 1-7, or the underwater immersed tube docking and monitoring method as described in claim 8.

Images