Underwater long-range scanning system based on active calibration board and calibration method thereof

Abstract

The application discloses an underwater long-distance scanning system based on an active calibration plate and a calibration method thereof, relates to the technical field of underwater optical measurement and three-dimensional reconstruction, and comprises a laser emission module which is arranged on the bow of an underwater vehicle and is composed of a laser, a galvanometer one and a galvanometer two, wherein the laser emits a point laser beam, the point laser beam is scanned into a line laser by the galvanometer one, a line laser scanning effect is formed, and the angle of the line laser is changed by the galvanometer two. The active calibration plate adopts active light-emitting design, each light-emitting point can be independently lighted as a circular calibration feature point and cover the frosted glass to soften the light, the signal-to-noise ratio of the feature point in the image can be enhanced, the problems of weak natural light, serious water absorption and scattering interference in the underwater environment can be overcome, and in combination with the circular light spot form, the center coordinates can be accurately extracted in image processing, high-precision feature positioning is realized, and reliable feature data basis is provided for light plane fitting and system calibration.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of underwater optical measurement and three-dimensional reconstruction, in particular to an underwater long-distance scanning system based on an active calibration board and a calibration method thereof. BACKGROUND

[0002] Underwater topographic survey is an important basic work in the fields of ocean exploration, seabed resource exploration, underwater cultural heritage investigation and underwater engineering construction. In recent years, underwater line laser scanning measurement technology based on laser structured light is an important means of obtaining seabed topography. This technology projects a laser line on the underwater target surface and uses a camera to collect the deformed image of the laser line to obtain the seabed topography and spatial structure information. In actual deep sea exploration tasks, the scanning system is usually installed on a manned submersible. In order to improve the efficiency of large-scale seabed coverage, the submersible needs to maintain a long working distance from the seabed (i.e. long-distance scanning) to obtain a larger field of view without frequent adjustments to the flight path and depth.

[0003] However, in order to ensure the measurement accuracy under long-distance scanning conditions, the system needs to be configured with a large baseline length to significantly improve the geometric sensitivity of triangulation. Due to the limited installation space of the submersible equipment, the laser and the camera are usually arranged at the bow and stern of the boat body to obtain sufficient baseline length. This installation method and long-distance imaging distance make calibration difficult. The main problem is that there is limited space underwater, which makes it impossible to perform long-distance underwater calibration. In addition, due to the absorption and scattering of light by water in the underwater long-distance calibration scenario, the brightness of the traditional passive imaging calibration board is extremely dark underwater, and the calibration image is blurred. Therefore, the present application proposes an underwater long-distance scanning system based on an active calibration board and a calibration method thereof to solve the above problems. The system uses an auxiliary camera to indirectly shorten the baseline and thus shorten the calibration distance. The active calibration board and its laser single-point calibration method are used to enhance the underwater image and reduce the impact of underwater light scattering. SUMMARY

[0004] To solve the above technical problems, the present application is implemented by the following technical solutions: In a first aspect, an underwater long-distance scanning system based on an active calibration board includes the following modules:

[0005] A laser emission module is mounted on the bow of an underwater submersible and consists of a laser, a galvanometer one and a galvanometer two. The laser emits a point laser beam, which is scanned into a line laser by the galvanometer one to form a line laser scanning effect. The galvanometer two changes the angle of the line laser to realize the rotation of the light plane, ensuring that a clear line laser can be generated under long-distance and large baseline conditions.

[0006] The active calibration board is arranged in a target underwater environment, and a plurality of arrayed circular light-emitting holes are formed on the board surface, and a point light source is arranged in each light-emitting hole, and a ground glass is arranged outside the hole to soften the light, so as to actively emit light, overcome the problem of weak underwater light and water light absorption, and each point light source is used as a circular point calibration feature point, and the geometric feature of the circular point calibration feature point is the center position of the light-emitting circle;

[0007] The imaging system is composed of a main camera and an auxiliary camera, wherein the main camera is mounted on the stern of the underwater vehicle, and the auxiliary camera is mounted on the bottom of the underwater vehicle and close to the laser emission module of the bow, the main camera is used to collect laser line images during actual scanning, and the underwater terrain is inverted, and the auxiliary camera is used in the calibration process, and the imaging of the laser line on the active calibration board is captured;

[0008] The data processing system includes a laser control module, a calibration board control module and an image acquisition and processing module, which is used to control the point / line scanning mode of the laser, control the point-by-point light emission of the active calibration board, collect images and perform background removal, multi-frame averaging, circle center fitting and light plane fitting calculation, realize the automatic calibration process, and the least square method is used to fit the line laser plane equation and the rotation axis.

[0009] In the second aspect, the underwater long-distance scanning system calibration method based on the active calibration board is used to realize the underwater long-distance scanning system based on the active calibration board, and includes the following steps:

[0010] Step 1: In the target underwater environment, the internal parameters of the main camera and the auxiliary camera are calibrated by using Zhang Zhengyou method, and the relative position and attitude relationship between the two is calibrated by using a stereo matching method, and a unified camera coordinate system is established;

[0011] Step 2: The background image is shot without laser and light, so as to remove the underwater light scattering background, the point light source in each circular point calibration feature point of the active calibration board is actively lighted, the auxiliary camera shoots multiple frames and removes the background, the accurate position of the center of the circular point calibration feature point is obtained by fitting, the spatial pose parameters of the calibration board relative to the camera coordinate system are estimated by using PnP method, and the plane equation of the calibration board in the world coordinate system is further derived;

[0012] Step 3: The galvanometer drives the point laser to project on the active calibration board point by point, multiple frames are shot at each point, the background is removed and synthesized, the complete image of the whole line laser on the active calibration board is obtained, and the intersection line of the laser plane and the calibration board plane is obtained;

[0013] Step 4, by changing the active calibration board posture and repeating the calibration board plane and line laser image extraction and synthesis process, a plurality of line laser images are obtained, and the least square method is used to fit the spatial equation of the line laser plane;

[0014] Step 5, adjust the galvanometer two to change the laser plane angle, repeat the light plane calibration process, obtain a plurality of line laser plane equations, and calculate the intersection line as the rotation axis of the galvanometer two;

[0015] Step 6, according to the rotation axis of the galvanometer two and the line laser plane equation, the light plane at any angle is derived, and combined with the pose relationship of the double cameras, the spatial position parameters of the main camera and the light plane are determined, and converted to the main camera coordinate system, completing the full parameter calibration of the long-distance scanning system.

[0016] Preferably, in step 1, the specific process of establishing a unified camera coordinate system is:

[0017] In the target underwater environment, the Zhang Zhengyou calibration method is used to obtain the internal parameters of the main camera and the auxiliary camera, including focal length, principal point coordinates and distortion coefficient, which can accurately correct lens distortion and improve image geometric fidelity;

[0018] The relative position and attitude relationship between the main camera and the auxiliary camera is calibrated by a stereo vision matching method, a rigid body transformation model is established between them, a high-precision double camera space relationship is established, and three-dimensional measurement consistency is ensured;

[0019] Based on the established rigid body transformation model, the main camera and the auxiliary camera are unified to the same world coordinate system, forming a unified camera system space reference, realizing cross-camera data fusion, and providing a stable space reference for subsequent calibration.

[0020] Preferably, the step 2 specifically comprises:

[0021] Under the condition of no laser irradiation and no light emission of the active calibration board, multiple frames of background images are captured by the auxiliary camera as the reference background image for image processing, and the random noise of the environment is effectively suppressed by multiple frame averaging, improving the signal-to-noise ratio of the background image;

[0022] Control the active calibration board to light up each circular light-emitting calibration feature point in turn, and synchronously collect multiple frames of illumination images by the auxiliary camera when each calibration feature point is lit, so that each light-emitting calibration feature point is fully exposed and covered by multiple frames, avoiding missing or blurring;

[0023] Subtract the reference background image from each calibration feature point corresponding to the multiple frames of illumination images, and average the multiple frames of results to obtain a clear calibration feature point image after background removal, effectively removing stray light and scattering interference, and highlighting the calibration feature point outline for subsequent center extraction.

[0024] Preferably, the step 2 further comprises:

[0025] For each background-removed clear calibration feature point image, a center positioning algorithm is used to extract the center pixel coordinates of the light-emitting circle in the image, so as to improve the accuracy and anti-noise ability of the extraction of the calibration point;

[0026] By means of the camera internal parameter and the calibrated camera system space reference, the center pixel coordinates of each circle are back-projected into a unified world coordinate system, so as to establish a high-precision mapping relationship of the feature points in the three-dimensional space coordinate system;

[0027] A three-dimensional space position set of all light-emitting circle centers of the active calibration board in the world coordinate system is obtained, and the PnP method is used to estimate the space pose parameters of the calibration board relative to the camera coordinate system in combination with the known geometric structure of the calibration board, so as to further derive the plane equation of the calibration board plane in the world coordinate system, which provides a strong geometric constraint and a uniform data distribution space reference for subsequent light plane calibration.

[0028] Preferably, in the step 3, a complete image of the line laser on the active calibration board is obtained, so that the intersection line of the laser plane and the calibration board plane is obtained, and the specific process is as follows:

[0029] The galvanometer drives the point laser to perform point-by-point scanning projection on the surface of the active calibration board, and the auxiliary camera synchronously collects multiple images during the projection of each point, so as to effectively avoid the interference of single-frame image noise and ensure the reliability and redundancy of the imaging of the laser point;

[0030] The multiple images corresponding to each point are subjected to background subtraction and multi-frame average processing, so that the clear imaging positions of the laser points on the calibration board are obtained, the signal-to-noise ratio and positioning accuracy of the laser points are significantly improved, and the influence of stray light and scattering under water is suppressed.

[0031] The laser point images at different imaging positions are synthesized to form a complete projection image of the line laser on the surface of the active calibration board, so as to ensure the continuity and geometric consistency of the laser line on the calibration board and provide complete spatial geometric information for subsequent light plane calibration.

[0032] Preferably, the step 4 specifically comprises:

[0033] The attitude of the active calibration board relative to the underwater vehicle is adjusted, and the steps 2 and 3 are repeatedly executed to obtain multiple groups of active calibration board images under different attitudes, so as to improve the spatial diversity of the data and ensure that the light plane calibration has better spatial adaptability;

[0034] The complete projection image of the line laser on the active calibration board and the three-dimensional coordinates of all light-emitting circle centers under each attitude are synchronously obtained, so as to ensure that the laser and the feature points are strictly corresponding in space and time and to enhance the spatial consistency of subsequent fitting;

[0035] By matching and aligning the calibration plate plane equations with the line laser projection images from multiple sets of data, a data sample set for optical plane fitting is constructed, forming a high-precision geometric constraint dataset, which provides a reliable spatial distribution basis for plane fitting.

[0036] Preferably, step 4 further includes:

[0037] Based on the data sample set, the initial spatial plane equation of the line laser plane in the world coordinate system is fitted using the least squares method to obtain the initial plane, which provides a geometric benchmark for subsequent optimization and ensures that the fitting starting point is accurate.

[0038] The fitting results are corrected by iterative optimization method to improve the accuracy and robustness of the line laser plane equation, enhance its adaptability to noise and outliers, and output the optimized line laser plane equation as one of the calibration results of the laser projection geometry in the long-distance scanning system, forming a reliable laser plane model that directly supports subsequent 3D reconstruction and measurement.

[0039] Preferably, step 5 specifically includes:

[0040] Adjust the second galvanometer to change the projection angle of the laser plane in space, and repeat the optical plane calibration process in step 4 to obtain optical plane samples at different angles, providing diverse data support for rotation axis calibration;

[0041] By obtaining multiple line laser plane equations at different angles, a set of line laser plane equations is formed, providing sufficient geometric constraints for subsequent calculations and improving the stability and accuracy of the rotation axis solution;

[0042] Calculate the intersection line of all line laser plane equations in the set of line laser plane equations, and calibrate it as the spatial straight line equation of the rotation axis of the second galvanometer in the world coordinate system. Solving the intersection line enables the accurate extraction of the rotation axis of the galvanometer and establishes a reliable geometric relationship between the optical plane and the rotating mechanism.

[0043] Preferably, step 6 specifically includes:

[0044] Based on the rotation axis equation of the second galvanometer and the calibrated linear laser plane equation, a kinematic model of the optical plane changing with the angle of the second galvanometer is established to ensure that the laser plane can be accurately calculated under any rotation angle, thereby achieving high-precision control of the dynamic scanning posture.

[0045] By combining the pose transformation relationship between the main camera and the auxiliary camera, the linear laser plane equation is transformed into the expression in the main camera coordinate system, unifying the system measurement benchmark, realizing the direct application of laser plane parameters in the main camera coordinate system, and avoiding the accumulation of coordinate transformation errors;

[0046] The final output includes a complete set of system calibration parameters, including main camera parameters, line laser plane equations, and rotation axis parameters, completing the geometric calibration of the long-distance underwater scanning system and forming a complete and reusable calibration database to support long-term stable operation and high-precision 3D reconstruction.

[0047] This invention provides an underwater long-range scanning system and calibration method based on an active calibration plate. It has the following beneficial effects:

[0048] (I) The underwater long-range scanning system and its calibration method based on the active calibration plate. The active calibration plate adopts an active light-emitting design. Each light-emitting point can be independently lit as a circular calibration feature point and covered with frosted glass to soften the light. This can enhance the signal-to-noise ratio of feature points in the image and overcome the problems of weak natural light, severe water absorption and scattering interference in the underwater environment. In addition, the design combined with the circular light spot form makes it easy to accurately extract the center coordinates in image processing, realize high-precision feature positioning, and provide a stable and reliable feature data foundation for light plane fitting and system calibration.

[0049] (II) The underwater long-range scanning system and its calibration method based on the active calibration plate can capture the image of the laser line on the calibration plate completely during the close-range calibration process by setting an auxiliary camera and installing it close to the laser emission module. This effectively avoids the problem of excessive calibration distance caused by the excessively long baseline between the main camera and the laser. With the point-by-point scanning of the galvanometer and the attitude adjustment of the galvanometer, the system can simulate long-range scanning conditions in a limited water area and complete the full parameter calibration of the large baseline system, breaking through the dependence of traditional calibration methods on the actual operating distance.

[0050] (III) The underwater long-range scanning system and its calibration method based on the active calibration plate are based on multi-angle line laser plane calibration data. The least squares method is used to fit the line laser plane equation, and the kinematic model of the light plane changing with the two angles of the galvanometer is established by combining the Rodriguez rotation formula. The rotation axis is determined by calculating the intersection of multiple light planes, and the nonlinear optimization method is used to improve the calibration accuracy of the axis system. Finally, the real-time accurate calculation of the light plane under any scanning angle is realized, providing accurate geometric constraints for long-range three-dimensional reconstruction.

[0051] (iv) The underwater long-range scanning system based on the active calibration plate and its calibration method calibrate the rigid body transformation relationship between the main camera and the auxiliary camera through the stereo matching method, unify all observation data to the world coordinate system with the main camera as the reference, ensure that the data of the two cameras are strictly aligned in space, provide a reliable conversion basis for the parameter transfer from the calibration of the auxiliary camera to the actual scanning of the main camera, and ensure the consistency of the whole system at the geometric level. Attached Figure Description

[0052] Figure 1This is a schematic diagram illustrating the workflow of the underwater long-range scanning system and calibration method based on an active calibration plate according to the present invention.

[0053] Figure 2 This is a schematic diagram of the underwater vehicle mounting according to the present invention;

[0054] Figure 3 This is a schematic diagram of the laser, galvanometer one, and galvanometer two of the present invention for laser scanning. Figure 1 ;

[0055] Figure 4 This is a schematic diagram of the laser, galvanometer one, and galvanometer two of the present invention for laser scanning. Figure 2 ;

[0056] Figure 5 This is a schematic diagram of the structure of the active calibration plate of the present invention;

[0057] Figure 6 This is a schematic diagram of the internal structure of the active calibration plate of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1, please refer to Figures 1 to 6 This invention provides a technical solution: an underwater long-range scanning system based on an active calibration plate, comprising the following modules:

[0060] The laser emission module, mounted on the bow of the underwater vehicle, consists of a laser, a first galvanometer, and a second galvanometer. The laser emits a point laser beam, which is scanned into a line laser beam by the first galvanometer, creating a line laser scanning effect. The second galvanometer then changes the angle of the line laser, rotating the optical plane to ensure that a clear line laser beam can still be generated even at long distances and with a large baseline. During calibration, it can scan point by point, facilitating complete capture of the laser line during close-range calibration. By controlling the second galvanometer to change the optical plane, it can adapt to different scanning angle requirements.

[0061] An active calibration board is deployed in the target underwater environment. Its surface has multiple arrayed circular light-emitting holes, and each light-emitting hole contains a point light source. The holes are covered with frosted glass to soften the light for active light emission, overcoming the problems of weak underwater light and water absorption. Each point light source serves as a circular calibration feature point. The geometric feature of the circular calibration feature point is the center position of its emitting circle. By lighting up point by point, the auxiliary camera can sequentially capture and fit the center position. The circular light spot facilitates the accurate extraction of the center in the underwater scattering environment, thereby improving the signal-to-noise ratio of the calibration image. It supports high-precision calibration point extraction under multiple views and attitudes, and is suitable for long-distance system calibration in limited underwater spaces.

[0062] The imaging system consists of a main camera and an auxiliary camera. The main camera is mounted on the stern of the submersible, while the auxiliary camera is installed on the bottom of the submersible near the laser emission module at the bow. The main camera is used to acquire laser line images during actual scanning and to perform underwater terrain inversion. The auxiliary camera is used during calibration to help capture the image of the laser line on the active calibration plate. Since the main camera is far from the laser, it is difficult to see both the laser line and the calibration plate at the same time during close-range calibration. Therefore, the auxiliary camera is installed close to the laser to solve the problem of field of view obstruction during calibration in a limited pool for long-range systems. Furthermore, the main and auxiliary cameras establish their pose relationship through stereo matching calibration, which facilitates the transmission of calibration results to the main camera.

[0063] The data processing system includes a laser control module, a calibration board control module, and an image acquisition and processing module. It controls the laser's point / line scanning mode and controls the active calibration board to emit light point by point to acquire images and perform calculations such as background removal, multi-frame averaging, circle center fitting, and light plane fitting to realize an automated calibration process. It can also fit the line laser plane equation and rotation axis using the least squares method to finally establish the spatial geometric relationship between the main camera and the laser plane, supporting long-distance underwater scanning.

[0064] Furthermore, this invention also provides a calibration method for an underwater long-range scanning system based on an active calibration plate, used to implement the aforementioned underwater long-range scanning system based on an active calibration plate, comprising the following steps:

[0065] Step 1: In the target underwater environment, the intrinsic parameters of the main camera and the auxiliary camera are calibrated using the Zhang Zhengyou method, and the relative position and attitude relationship between them is calibrated using the stereo matching method to establish a unified camera coordinate system. In the target underwater environment, the intrinsic parameters of the main camera and the auxiliary camera, including focal length, principal point coordinates and distortion coefficients, are obtained using the Zhang Zhengyou calibration method. This can accurately correct lens distortion and improve the geometric fidelity of the image. The relative position and attitude relationship between the main camera and the auxiliary camera are calibrated using the stereo vision matching method, and a rigid body transformation model between them is established. A high-precision dual-camera spatial relationship is established to ensure the consistency of three-dimensional measurement. Based on the established rigid body transformation model, the main camera and the auxiliary camera are unified to the same world coordinate system to form a unified camera system spatial reference, realizing cross-camera data fusion and providing a stable spatial reference for subsequent calibration.

[0066] The specific work involves: In actual calibration, the main camera and auxiliary camera are fixed at corresponding positions on the underwater submersible, with their optical axes pointing towards the same target area. Using a traditional passive calibration board with a known pattern deployed in the underwater environment, the main camera and auxiliary camera acquire image sequences from multiple angles. Subsequently, based on Zhang Zhengyou's calibration method, the image sequences captured by each camera are processed independently, and their intrinsic parameters, including focal length, principal point coordinates, and lens distortion coefficients, are extracted and optimized to achieve precise calibration of the internal geometric characteristics of each camera. Based on the completion of the camera intrinsic parameter calibration, the spatial relative relationship between the main camera and auxiliary camera is determined using a stereo matching method. Specifically, feature points with the same set of known three-dimensional coordinates are used in the images from both cameras. Based on the corresponding projection positions in the image and the calibrated camera intrinsic parameters, the rotation matrix and translation vector between the two cameras are calculated to establish a rigid body transformation model describing their pose relationship. This ensures that feature points are clearly imaged within the fields of view of both cameras. The transformation parameters are optimized using corresponding point data under multiple different poses to improve the stability and accuracy of pose calibration. Based on the obtained rigid body transformation model, a unified spatial reference for the camera system is further established. In practice, the coordinate system of the main camera is selected as the reference reference for the world coordinate system. The three-dimensional coordinates of the feature points observed by the auxiliary camera are transformed to this reference through rigid body transformation relationships, achieving spatial unification of the observation data of the two cameras. This results in a common, high-precision spatial reference frame, ensuring that subsequent calibration processes are performed in a consistent three-dimensional coordinate system.

[0067] Step 2: Background images are captured under conditions of no laser and no light to remove underwater light scattering. Point light sources within each calibration feature point of the active calibration plate are used to actively emit light point-by-point. The auxiliary camera captures multiple frames and removes the background. The precise position of the center of each emitting circle within the calibration feature point of the calibration plate is obtained through fitting. Combined with the known geometry of the calibration plate, the PnP method is used to estimate the spatial pose parameters of the calibration plate relative to the camera coordinate system. The plane equation of the calibration plate in the world coordinate system is further derived. Under conditions of no laser illumination and no light emitted by the active calibration plate, multiple background images are captured by the auxiliary camera as... The baseline background image for image processing effectively suppresses random environmental noise and improves the signal-to-noise ratio of the background image through multi-frame averaging. The active calibration board is controlled to sequentially illuminate each circular luminous calibration feature point. When each calibration feature point is illuminated, the auxiliary camera simultaneously acquires multiple frames of illumination images to ensure that each luminous calibration feature point has sufficient exposure and multi-frame coverage, avoiding missed acquisitions or blurring. The baseline background image is subtracted from the multi-frame illumination images corresponding to each calibration feature point, and the multi-frame results are averaged to obtain a clear calibration feature point image after background removal. This effectively removes stray light and scattering interference, highlights the outline of the calibration feature point, and facilitates subsequent circle center extraction.

[0068] The specific tasks are as follows: In the initial stage of system calibration, the auxiliary camera continuously captures multiple frames of background images in an underwater environment without laser illumination and with the active calibration board in a non-illuminating state. It is necessary to ensure stable ambient lighting without significant disturbances. The multiple background images are then time-series averaged to obtain a reference background image with a high signal-to-noise ratio. This reference background image will serve as a background reference for image processing, effectively separating the active luminous calibration feature points of the calibration board from underwater background stray light and scattered noise interference. The active calibration board is then controlled to sequentially illuminate the circular luminous calibration feature points on its surface. During the illumination of each feature point, the auxiliary camera simultaneously acquires multiple frames of illumination images. During image acquisition, the camera and the active calibration board must remain relatively stationary to avoid motion blur. To improve the accuracy of feature localization, after data acquisition, the multiple lighting images corresponding to each calibrated feature point are subjected to pixel-level difference operations with the reference background image to effectively remove background light. Then, the multiple lighting images after background removal are averaged to further suppress random noise, resulting in a series of clear calibrated feature point images after background removal. These clear calibrated feature point images are then enhanced with image enhancement and filtering preprocessing to improve the contrast and edge integrity of the luminous circular regions. Based on this, a circle center fitting algorithm is used to extract the pixel-level center coordinates of each luminous calibrated feature point frame by frame. The stability and reliability of the localization results are ensured through multi-frame coordinate consistency verification. Finally, the precise position data of all luminous calibrated feature points in the image coordinate system are output.

[0069] Furthermore, step 2 also includes: for each clear calibration feature point image after background removal, the center pixel coordinates of the luminous circles in the image are extracted using a circle center localization algorithm to improve the accuracy and noise resistance of calibration point extraction. Through camera intrinsic parameters and the calibrated camera system spatial reference, the center pixel coordinates of each circle are back-projected to a unified world coordinate system to establish a high-precision mapping relationship of feature points in the three-dimensional spatial coordinate system. The three-dimensional spatial position set of the center of all luminous calibration feature points of the active calibration board in the world coordinate system is obtained. Combined with the known geometric structure of the calibration board, the PnP method is used to estimate the spatial pose parameters of the calibration board relative to the camera coordinate system, and the plane equation of the calibration board plane in the world coordinate system is further derived.

[0070] The specific work involves: after background subtraction and image enhancement, extracting the center coordinates of each clear calibration feature point image after background removal using a circle center localization algorithm. Specifically, the clear calibration feature point image is binarized to identify the outline of the circular luminous region. An ellipse fitting method is used to determine the precise pixel center position of each luminous calibration feature point in the image coordinate system. Time-series averaging of multiple frames can be combined to improve positioning accuracy. Outliers are removed through coordinate consistency checks between adjacent frames, resulting in a stable and accurate set of pixel coordinates for all luminous calibration feature points. After obtaining the circle center pixel coordinates in the image coordinate system, they are transformed to a unified world coordinate system. Based on the calibrated camera intrinsic parameters (including focal length, principal point coordinates, and lens distortion coefficients) and the rigid body transformation relationship between the main and auxiliary cameras, back-projection calculations are performed on the circle center pixel coordinates. The intrinsic parameter matrix is ​​used to eliminate lens distortion. The image points are then mapped to the world coordinate system using the camera extrinsic parameters, resulting in the corresponding coordinates of each emitting circle center in three-dimensional space, forming the three-dimensional point cloud data of each feature point on the calibration board. Through the back projection process, the complete three-dimensional spatial position set of all emitting circles center of the active calibration board in the world coordinate system is obtained, which is the feature point set for subsequent optical plane calibration. This feature point set has a clear geometric distribution and high-precision spatial coordinates. Combined with the known geometric structure of the calibration board, the PnP method is used to estimate the spatial pose parameters of the calibration board relative to the camera coordinate system. The plane equation of the calibration board plane in the world coordinate system is further derived. At the same time, it is ensured that the plane equation of the calibration board corresponds to the line laser projection image in time and space. Its consistency is verified by multiple measurements and redundant data, providing a stable and reliable geometric constraint basis for optical plane fitting based on the least squares method, supporting the high-precision calibration of the entire long-distance scanning system.

[0071] Step 3: Control the galvanometer-driven point laser to project point by point onto the active calibration plate. After capturing multiple frames at each point, remove the background and synthesize the images to obtain a complete image of the entire line laser on the active calibration plate. This allows the intersection of the laser plane and the calibration plate plane to be obtained. Control the galvanometer-driven point laser to scan and project point by point onto the surface of the active calibration plate. During the projection of each point, the auxiliary camera simultaneously acquires multiple frames of images, effectively avoiding noise interference from single-frame images and ensuring the reliability and redundancy of laser point imaging. Perform background subtraction and multi-frame averaging on the multiple frames of images corresponding to each point to obtain a clear imaging position of each laser point on the calibration plate, significantly improving the signal-to-noise ratio and positioning accuracy of the laser points and suppressing the effects of stray light and scattering underwater. Connect and synthesize the imaging positions of each laser point in sequence to form a complete projection image sequence of the line laser on the surface of the active calibration plate, ensuring the continuity and geometric consistency of the laser line on the calibration plate and providing complete spatial geometric information for subsequent optical plane calibration.

[0072] The specific work involves: after the active calibration plate is fixed and correctly positioned in the underwater target area, the galvanometer-driven laser is activated to perform point-by-point scanning. The scanning rate is controlled at 5 to 10 points per second to ensure sufficient exposure time for the auxiliary camera at each point. The auxiliary camera uses a global shutter mode, with the frame rate synchronized with the laser point switching. During the continuous projection at each point, 10 to 15 frames of images are acquired. The exposure time is adjusted according to the underwater ambient light intensity. The camera and the active calibration plate remain relatively stationary to avoid image blurring caused by vibration or water flow disturbance. Simultaneously, the image signal-to-noise ratio is monitored in real time to ensure acquisition quality. During the acquisition process, the camera automatically stores the original image sequence corresponding to each point and numbers it according to the scanning order. Pixel-level difference operations are performed between the multiple frames of images corresponding to each laser point and the acquired reference background image to eliminate stray light and scattering noise from the underwater environment. The differencing images are then subjected to median filtering and Gaussian smoothing with a window size of 3×3 pixels and a standard deviation of 0.5 to 1.0 to suppress random noise and enhance the laser point area. Contrast was assessed, and then time-series averaging was performed on the multi-frame difference images corresponding to each point to further reduce noise and obtain a clear grayscale distribution of each laser point in the image. The image center coordinates of the laser points were extracted using a subpixel-level centroid method, with a fitting window size of 7×7 pixels. The coordinates of multiple frames were repeatedly extracted and consistency was checked. Outliers with offsets exceeding 0.5 pixels were removed to determine the stable position of each laser point in the image coordinate system. The laser point image coordinates extracted point by point were spatially synthesized according to the scanning order to form a continuous line laser projection image. The gaps between points were filled using cubic spline interpolation to ensure smooth and continuous line shape. The synthesized line laser image was spatially aligned with the plane equation of the calibration board, and the offset error was controlled within 1 pixel. If the error of a certain section exceeded the limit, the section was re-scanned and extracted locally until the overall projection image met the geometric consistency requirements. The final generated line laser projection image was stored and managed with timestamps and attitude numbers to ensure the repeatability and traceability of the calibration process.

[0073] Step 4: By changing the orientation of the active calibration plate and repeating the process of extracting and synthesizing the plane and line laser images of the calibration plate, multiple sets of line laser images are obtained, and the spatial equation of the line laser plane is fitted using the least squares method.

[0074] Step 5: Adjust the second galvanometer to change the angle of the laser plane, repeat the optical plane calibration process, obtain multiple line laser plane equations, and calculate their intersection line as the rotation axis of the second galvanometer.

[0075] Step 6: Based on the rotation axis of the second galvanometer and the equation of the line laser plane, derive the light plane at any angle, and combine it with the pose relationship of the two cameras to determine the spatial position parameters of the main camera and the light plane. Transform to the main camera coordinate system to complete the full parameter calibration of the long-distance scanning system.

[0076] Example 2, as Figures 1 to 6 As shown, based on Embodiment 1, the present invention provides a technical solution: Step 4 specifically includes: adjusting the attitude of the active calibration plate relative to the underwater vehicle, repeating steps 2 and 3, acquiring multiple sets of active calibration plate images under different attitudes, improving the spatial diversity of the data, ensuring that the optical plane calibration has better spatial adaptability, and simultaneously obtaining the complete projection image of the corresponding line laser on the active calibration plate and the three-dimensional coordinates of all the emitting circles under each attitude, ensuring that the laser and the feature points strictly correspond in time and space, enhancing the spatial consistency of subsequent fitting, matching and aligning the calibration plate plane equation with the line laser projection image in multiple sets of data, constructing a data sample set for optical plane fitting, forming a high-precision geometric constraint dataset, and providing a reliable spatial distribution basis for plane fitting;

[0077] The specific work involves: after completing the acquisition of images and line laser projection data from the single-attitude active calibration board, adjusting the attitude of the active calibration board relative to the underwater vehicle, and performing multiple sets of data acquisition. During operation, the active calibration board is rotated 15° to 30° around its normal axis each time, while keeping the distance between the active calibration board and the camera constant. A total of 6 to 10 sets of different attitudes are acquired. Under each attitude, the acquisition of luminous calibration feature points in step 2 and the laser line scanning in step 3 are repeated, with an interval of no less than 5 seconds between each scan to ensure that the camera and the calibration board are completely stationary. The attitude angle, timestamp, and camera exposure parameters are recorded synchronously to ensure the spatiotemporal correspondence between data sets. After each attitude acquisition is completed, a complete image of the active calibration board (including the coordinates of all luminous center points) and the corresponding line laser projection image on the surface of the active calibration board under that attitude are acquired synchronously. The coordinates of the luminous center points are obtained through camera calibration back projection, and their three-dimensional coordinate accuracy is controlled within ±0.5 mm. The line laser projection image is a continuously synthesized image, and the laser points... With an interval density of no less than 2 points per millimeter, during the data processing stage, the coordinates of the active calibration board and the laser projection image under the same posture are spatially registered. The registration error must be less than 1 pixel. After registration, the three-dimensional point sequence corresponding to each laser line on the active calibration board is extracted to form a data subset of laser line-calibration board feature point pairs under a single posture. The data subsets obtained under multiple postures are integrated to construct a complete sample set for light plane fitting. During integration, the data is uniformly converted to the world coordinate system based on the main camera, and each posture data is indexed using timestamps and posture numbers. The sample set contains no less than 200 effective laser points and their corresponding three-dimensional spatial coordinates. Each effective laser point is distributed in different areas and different depth planes of the active calibration board. Each data point is accompanied by acquisition posture, illumination conditions, and signal-to-noise ratio metadata, thereby providing sufficient, uniformly distributed, and geometrically diverse spatial constraints for the subsequent least-squares fitting of the line laser plane equation, ensuring a stable fitting process and reliable results.

[0078] Step 4 also includes: based on the data sample set, using the least squares method to fit the initial spatial plane equation of the line laser plane in the world coordinate system to obtain the initial plane, providing a geometric benchmark for subsequent optimization, ensuring the accuracy of the fitting starting point, correcting the fitting results through iterative optimization methods, improving the accuracy and robustness of the line laser plane equation, enhancing the adaptability to noise and outliers, and outputting the optimized line laser plane equation as one of the calibration results of the laser projection geometry in the long-distance scanning system, forming a reliable laser plane model, which directly supports subsequent 3D reconstruction and measurement;

[0079] The specific work involves: based on the integrated data sample set, standardizing the samples, removing spatial points exceeding the preset confidence interval, and fitting the initial spatial plane equation of the laser plane using the least squares method, specifically in the form of... During the fitting process, the eigenvector corresponding to the smallest eigenvalue of the covariance matrix is ​​calculated through eigenvalue decomposition to obtain the optimal plane normal vector. And determine the constant term based on the sample center point. , , Let these be the coordinates of the geometric center of the sample point set. The sample points are defined as 3D coordinates in the world coordinate system (main camera coordinate system). Before fitting, the sample coordinates need to be transformed to the main camera coordinate system to check the spatial distribution uniformity of the sample points and ensure that the laser points in each region are balanced. The root mean square error of the initial plane fitting is controlled within 0.8 mm. If the error exceeds the threshold, the sample quality needs to be rechecked or additional data needs to be collected. After obtaining the initial line laser plane equation, the plane parameters are optimized using an iterative reweighted least squares method. In each iteration, the weight is calculated based on the distance of each sample point to the current fitting plane. The larger the distance, the smaller the weight. The threshold is set so that the weight of sample points with a plane normal distance of more than 1.5 mm is reduced to below 0.3. The optimization process continues until the change in plane parameters is less than 1 × 10⁻⁶. -5The maximum number of iterations (set to 20) is reached. Simultaneously, a random sampling consistency method is introduced to improve the robustness of the fit. Multiple subsets are extracted from the samples for plane fitting, and the plane with the highest proportion of interior points and the smallest mean square error is selected as the candidate result. The final optimized line laser plane equation satisfies that the average distance from all interior points to the plane does not exceed 0.5 mm, and the maximum distance does not exceed 1.2 mm. The direction cosine of the plane normal vector and the intercept parameter of the plane in the camera coordinate system are recorded. After optimizing the line laser plane equation, the projection error is calculated using reserved sample points (approximately 15%~20% of the total samples) that were not involved in the fitting. This is then independently verified. The verification points cover different areas and depths of the active calibration board. The verification error requirement is that the average projection error does not exceed 0.6 mm, and the maximum error does not exceed 1.5 mm. If the verification passes, the optimized line laser plane equation and its related parameters (including normal vector, plane offset, fitting error distribution, and confidence interval) are recorded as one of the system calibration output results.

[0080] Step 5 specifically includes: adjusting galvanometer two to change the projection angle of the laser plane in space, repeating the optical plane calibration process of step 4 to obtain optical plane samples at different angles, providing diverse data support for rotation axis calibration, obtaining multiple line laser plane equations at different angles, forming a set of line laser plane equations, providing sufficient geometric constraints for subsequent calculations, improving the stability and accuracy of rotation axis solution, calculating the intersection line of all line laser plane equations in the set of line laser plane equations, calibrating it as the spatial straight line equation of the rotation axis of galvanometer two in the world coordinate system, solving the intersection line to achieve accurate extraction of the galvanometer rotation axis, and establishing a reliable geometric relationship between the optical plane and the rotating mechanism;

[0081] The specific work involves adjusting the second galvanometer to change the projection angle of the laser plane in space. Each angle adjustment step is 1° to 3°. After adjustment, the galvanometer must remain stable for at least 5 seconds before data acquisition begins to ensure there is no residual vibration. The entire optical plane calibration process in step 4 is repeated for each angle, including multi-pose acquisition with the active calibration board, laser line scanning and synthesis, and plane fitting and verification. During acquisition, the actual rotation angle reading of the second galvanometer is recorded, with an angle feedback resolution of no less than 0.01°. Simultaneously, ambient light stability and water temperature fluctuations not exceeding ±0.5℃ are maintained to ensure optimal acquisition conditions at different angles. Consistently, at least eight calibration plate data points with different postures are acquired at each angle. The fitting error of the final fitted line laser plane equation must meet the requirements that the average projection error does not exceed 0.6 mm and the maximum error does not exceed 1.2 mm. After completing all angle scans, a series of line laser plane equations are obtained, forming a set of line laser plane equations that can be used for rotation axis calculations. Based on the obtained set of line laser plane equations, the intersection lines between each plane are calculated. The specific process is as follows: two adjacent line laser plane equations are selected in sequence, and their intersection direction vector and the coordinates of the points passing through the line are solved simultaneously. Let the two plane equations be... : and : , The coefficients of the first plane, It is its normal vector. The coefficients of the second plane, Its normal vector, Indicates transpose. Let be the coordinates of a point in space in the world coordinate system, and let the direction vector of their intersection line be the cross product of the normal vectors of the two planes, i.e. Write it in coordinate form. , , , Let the direction vector of the intersection line be . Direction vector The component in the x-axis direction, Direction vector The component in the y-axis direction, Direction vector The component in the z-axis direction, if If the two planes are not parallel, then there exists a unique line of intersection. The coordinates of a point on this line of intersection are obtained by solving a system of linear equations, leading to the parametric equation of this line. This calculation is repeated for all adjacent plane combinations to obtain an initial estimate of the line of intersection. Subsequently, cluster analysis is used to check the spatial consistency of multiple lines of intersection, eliminating outliers with directional deviations exceeding 0.5° or positional deviations exceeding 1 mm. Finally, the spatial linear equation of the second rotation axis of the galvanometer in the world coordinate system is obtained by averaging the direction vectors and point coordinates of all valid lines of intersection, expressed in parametric form. ,in, Let these be the coordinates of the points on the intersection line. Let be the unit direction vector of the two rotation axes of the galvanometer. The real parameter represents the direction. displacement, The coordinates of known points on the intersection line, i.e., known points on the rotation axis, are used. After obtaining the initial rotation axis equation, it is further optimized to improve the calibration accuracy. Based on all acquired line laser plane equations, an optimization objective function with rotation axis parameters as variables is constructed. This function is expressed as the weighted sum of squares of the perpendicularity between the normal vector of each light plane and the direction vector of the rotation axis, and the consistency of the distance from the plane to the axis. The Levenberg-Marquardt nonlinear optimization method is used for iterative solution. During the optimization process, the rotation axis direction vector is constrained to be a unit vector, and the plane fitting residual is introduced as a weight. After optimization, the angle between the direction vector of the rotation axis equation and the normal vector of each light plane is kept in the range of 89.5° to 90.5°, and the standard deviation of the distance from each light plane to the rotation axis does not exceed 0.3 mm. After optimization, the rotation axis equation parameters, including the axis point coordinates, direction cosines, and error distribution, along with the line laser plane equation and camera intrinsic parameters, are recorded as the system calibration output to form a complete scanning system geometric model.

[0082] Step 6 specifically includes: based on the rotation axis equation of galvanometer 2 and the calibrated line laser plane equation, establishing a kinematic model of the light plane changing with the angle of galvanometer 2, ensuring that the laser plane can be accurately calculated at any rotation angle, realizing high-precision control of dynamic scanning posture; combining the pose transformation relationship between the main camera and the auxiliary camera, transforming the line laser plane equation into the main camera coordinate system, unifying the system measurement benchmark, realizing the direct application of laser plane parameters in the main camera coordinate system, avoiding the accumulation of coordinate transformation errors, and finally outputting a complete system calibration parameter set including main camera parameters, line laser plane equation, rotation axis parameters, completing the geometric calibration of the long-distance underwater scanning system, forming a complete and reusable calibration database, supporting long-term stable operation and high-precision three-dimensional reconstruction of the system;

[0083] The specific work involves: establishing a kinematic model of the optical plane as a function of the angle of the second galvanometer, based on the rotation axis equation parameters of the second galvanometer and the calibrated laser plane equation of a certain reference line. Specifically, the kinematic model of the optical plane as a function of the angle of the second galvanometer is set to 0 when the initial rotation angle of the second galvanometer is 0. Rotation At an angle, the normal vector of the new light plane is calculated using the Rodrigues rotation formula:

[0084] ;

[0085] in, The normal vector of the new light plane after rotation. The unit normal vector of the initial reference light plane. Let be the unit direction vector of the two rotation axes of the galvanometer. For the galvanometer two-axis rotation rotation angle, This is the vector cross product operator. As a vector dot product operator, the light plane remains through the axis of rotation during rotation; therefore, the constant term of the new linear laser plane equation depends on the point on the axis. Sure:

[0086] ;

[0087] In the formula; For the constant term in the new line laser plane equation; Let be the coordinates of a known point on the axis of rotation, satisfying Rotation angle The angle encoder, provided by galvanometer 2, has a resolution of 0.01. o The angle repeatability positioning accuracy is better than 0.05. o After calibration, the model parameters need to be verified through actual scanning to ensure that the error in the relationship between the angle and the change in the light plane is controlled within a deviation of less than 0.1 in the plane normal vector direction. o The planar offset error is less than 0.3 mm; after establishing the kinematic relationship between the optical plane and the rotation axis, the equation of the line laser plane is transformed and expressed in the main camera coordinate system. The rigid body transformation relationship from the auxiliary camera coordinate system to the main camera coordinate system is known to be a rotation matrix. Translation vector This parameter has been calibrated in step 1, and its transformation accuracy meets the requirement that the rotation error is less than 0.05. o The translation error is less than 0.5 mm, for the line laser plane equation expressed in the auxiliary camera coordinate system. Its normal vector The formula for transforming planar parameters to the main camera coordinate system is:

[0088] ;

[0089] ;

[0090] In the formula: This is the normal vector in the main camera coordinate system; This is the rotation matrix from the auxiliary camera coordinate system to the main camera coordinate system; This is the normal vector in the auxiliary camera coordinate system; This refers to the plane constant term in the main camera coordinate system; This refers to the plane constant term in the auxiliary camera coordinate system; The transpose (row vector) of the normal vector in the main camera coordinate system; The translation vector is taken from the auxiliary camera coordinate system to the main camera coordinate system. During the transformation, the normal vector is kept normalized. The transformed linear laser plane equation is directly combined with the main camera intrinsic parameters to form a complete projection and back-projection model from a 2D image to 3D world coordinates. The transformation error should be controlled so that the reconstruction point error of the plane equation in the main camera coordinate system does not exceed 0.5 mm. After completing the optical plane kinematic modeling and coordinate system transformation, all calibration parameters are integrated and output to form a complete geometric calibration result for the long-range underwater scanning system. The output parameter set includes: the main camera intrinsic parameter matrix and distortion coefficient vector, and the rigid body transformation parameters from the auxiliary camera to the main camera. and Parameters of the two rotation axes of the galvanometer and The coefficients of the laser plane equation for the reference line, and the optical plane as a function of angle. The kinematic formulas vary, and all parameters are stored in numerical form in the system calibration file, along with calibration error statistics. The calibration results are verified by independent point cloud projection tests to confirm the accuracy of 3D reconstruction, thus meeting the geometric calibration requirements of long-distance underwater scanning systems.

[0091] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0092] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An underwater long-range scanning system based on an active calibration plate, characterized in that, Includes the following modules: The laser emission module, mounted on the bow of the underwater vehicle, consists of a laser, a galvanometer, and a second galvanometer. The laser emits a point laser beam, which is scanned into a line laser by the first galvanometer to create a line laser scanning effect. The second galvanometer then changes the angle of the line laser to achieve rotation of the optical plane. An active calibration plate is deployed in a target underwater environment. The plate surface has multiple arrayed circular light-emitting holes, and each light-emitting hole contains a point light source. The holes are covered with frosted glass. Each point light source serves as a circular calibration feature point, and the geometric feature of the circular calibration feature point is the center position of its light-emitting circle. The imaging system consists of a main camera and an auxiliary camera. The main camera is mounted on the stern of the underwater vehicle, and the auxiliary camera is installed on the bottom of the underwater vehicle near the laser emission module at the bow. The main camera is used to acquire laser line images during actual scanning and to perform underwater terrain inversion. The auxiliary camera is used during the calibration process to help capture the image of the laser line on the active calibration plate. The data processing system includes a laser control module, a calibration board control module, and an image acquisition and processing module. It is used to control the point / line scanning mode of the laser and control the active calibration board to emit light point by point in order to acquire images and perform calculations such as background removal, multi-frame averaging, circle center fitting, and light plane fitting.

2. A calibration method for an underwater long-range scanning system based on an active calibration plate, used to implement the underwater long-range scanning system based on an active calibration plate as described in claim 1, characterized in that, Includes the following steps: Step 1: In the target underwater environment, use Zhang Zhengyou's method to calibrate the intrinsic parameters of the main camera and the auxiliary camera respectively, and use the stereo matching method to calibrate the relative position and attitude relationship between the two to establish a unified camera coordinate system; Step 2: Take background images in the absence of laser and light to remove underwater light scattering background. Actively emit light point by point within each dot of the calibration plate to assist the camera in taking multiple frames and removing the background. Fit the precise position of the center of the dot within the calibration plate to obtain the precise position of the dot within the calibration plate. Combined with the known geometry of the calibration plate, use the PnP method to estimate the spatial pose parameters of the calibration plate relative to the camera coordinate system. Further derive the plane equation of the calibration plate plane in the world coordinate system. Step 3: Control the galvanometer to drive the point laser to project point by point on the active calibration plate. After taking multiple frames at each point, remove the background and synthesize them to obtain a complete image of the entire line laser on the active calibration plate, thereby obtaining the intersection line between the laser plane and the calibration plate plane. Step 4: By changing the orientation of the active calibration plate and repeating the process of extracting and synthesizing the plane and line laser images of the calibration plate, multiple sets of line laser images are obtained, and the spatial equation of the line laser plane is fitted using the least squares method. Step 5: Adjust the second galvanometer to change the angle of the laser plane, repeat the optical plane calibration process, obtain multiple line laser plane equations, and calculate their intersection line as the rotation axis of the second galvanometer. Step 6: Based on the rotation axis of the second galvanometer and the equation of the line laser plane, derive the light plane at any angle, and combine it with the pose relationship of the two cameras to determine the spatial position parameters of the main camera and the light plane. Transform to the main camera coordinate system to complete the full parameter calibration of the long-distance scanning system.

3. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 2, characterized in that: In step 1, the specific process of establishing a unified camera coordinate system is as follows: In the target underwater environment, the Zhang Zhengyou calibration method was used to obtain the intrinsic parameters of the main camera and the auxiliary camera, including focal length, principal point coordinates and distortion coefficients. The relative position and attitude relationship between the main camera and the auxiliary camera is determined by stereo vision matching method, and a rigid body transformation model between the two is established. Based on the established rigid body transformation model, the main camera and auxiliary camera are unified to the same world coordinate system, forming a unified spatial reference for the camera system.

4. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 2, characterized in that: Step 2 specifically includes: Under conditions of no laser illumination and no light emitted by the active calibration board, multiple background images are captured by the auxiliary camera and used as the reference background image for image processing. The active calibration board is controlled to sequentially light up each circular luminous calibration feature point. When each calibration feature point is lit up, the auxiliary camera simultaneously captures multiple frames of illumination images. The reference background image is subtracted from the multiple frames of illumination images corresponding to each calibration feature point, and the results of multiple frames are averaged to obtain a clear calibration feature point image after background removal.

5. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 4, characterized in that: Step 2 also includes: For each clear, labeled feature point image after background removal, a circle center localization algorithm is used to extract the center pixel coordinates of the luminous circle in the image; Using the camera intrinsic parameters and the calibrated camera system spatial reference, the coordinates of each circle's center pixel are back-projected to a unified world coordinate system; The set of three-dimensional spatial positions of all the emitting centers of the active calibration plate in the world coordinate system is obtained. Combined with the known geometry of the calibration plate, the PnP method is used to estimate the spatial pose parameters of the calibration plate relative to the camera coordinate system. The plane equation of the calibration plate plane in the world coordinate system is further derived.

6. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 2, characterized in that: In step 3, a complete image of the entire laser line on the active calibration plate is obtained, thereby acquiring the intersection line between the laser plane and the calibration plate plane. The specific process is as follows: The control galvanometer drives a point laser to scan and project onto the surface of an active calibration plate point by point, while an auxiliary camera simultaneously acquires multiple frames of images during each point projection. Background subtraction and multi-frame averaging are performed on the multi-frame images corresponding to each point to obtain the clear imaging position of each laser point on the calibration plate. The laser point images from different imaging positions are combined to form a complete projection image of the line laser on the surface of the active calibration plate.

7. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 2, characterized in that: Step 4 specifically includes: Adjust the attitude of the active calibration plate relative to the underwater vehicle, repeat steps 2 and 3, and obtain multiple sets of active calibration plate images under different attitudes; For each posture, the complete projection image of the corresponding line laser on the active calibration plate and the three-dimensional coordinates of all the light-emitting centers are obtained synchronously. The calibration plate plane equations from multiple sets of data are matched and aligned with the line laser projection images to construct a data sample set for optical plane fitting.

8. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 7, characterized in that: Step 4 also includes: Based on the data sample set, the initial spatial plane equation of the line laser plane in the world coordinate system is fitted using the least squares method. The fitting results are corrected by iterative optimization, and the optimized line laser plane equation is output as one of the calibration results of the laser projection geometry in the long-distance scanning system.

9. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 8, characterized in that: Step 5 specifically includes: Adjust the second galvanometer to change the projection angle of the laser plane in space, and repeat the optical plane calibration process in step 4. The plane equations of the line laser at multiple different angles are obtained, forming a set of plane equations of the line laser. Calculate the intersection line of all line laser plane equations in the set of line laser plane equations, and calibrate it as the spatial straight line equation of the rotation axis of galvanometer II in the world coordinate system.

10. The calibration method for an underwater long-range scanning system based on an active calibration plate according to claim 9, characterized in that: Step 6 specifically includes: Based on the rotation axis equation of galvanometer 2 and the calibrated linear laser plane equation, a kinematic model of the optical plane as the angle of galvanometer 2 changes is established. By combining the pose transformation relationship between the main camera and the auxiliary camera, the linear laser plane equation is transformed and expressed in the main camera coordinate system; The final output includes a complete system calibration parameter set, including main camera parameters, line laser plane equations, and rotation axis parameters, completing the geometric calibration of the long-range underwater scanning system.

Images