Water-air fast scanner and underwater scanning method based on water-air fast scanner

Abstract

The invention discloses a water-air fast scanner and a water bottom scanning method based on the water-air fast scanner, and relates to the field of water body ecological environment investigation. The problems that an existing underwater scanner cannot be suitable for a shallow water area with the water depth smaller than 5 m, GPS position information and a real scanning position are staggered, and the position precision of a bottom map is reduced are solved. The method comprises the following steps: determining a deployment mode of a water-air rapid scanner, and selecting a corresponding baseline length and array spacing; configuring a working mode of the camera array, and synchronizing the camera array and a GPS module clock; the water-air fast scanner records water bottom image information; extracting a video frame; calculating scanning track coordinate data of each camera; interpolation is carried out on the scanning track coordinate data, and a smooth GPS motion track is obtained; acquiring a route track of the camera according to the GPS motion track and the array spacing; and performing reconstruction processing on the route trajectory by adopting SfM to obtain a water bottom image. The method is suitable for the fields of ocean research, geological survey, underwater archaeological and the like.

Description

Technical Field

[0001] The invention relates to the field of water ecological environment investigation, and in particular to an underwater scanning method based on a water-to-air rapid scanner. Background Art

[0002] Rivers, lakes and even shallow seas are habitats for a variety of aquatic organisms, so conducting ecological surveys is crucial to gaining a deeper understanding of the biodiversity within these ecosystems. In addition, by investigating the ecosystems of these waters, we can monitor changes in water quality and detect pollution problems in a timely manner, which helps to ensure the safety of human drinking water and the health of the ecological environment.

[0003] When the existing underwater scanner Speedy Sea Scanner (SSS) and its simplified version Speedy Sea Scanner-portable (SSS-P) are in operation, the SSS is connected to a power boat on the water surface by a cable and towed by the power boat. Due to the limitations of its structure, the camera can only be in a 90-degree posture at all times, so that the lens always faces the bottom of the water to shoot (similar to an orthographic map). When the SSS is towed along the motion trajectory to scan the bottom of the water, the collected image data can be processed to generate an underwater map. However, due to its large size (wingspan of approximately 5m) and its lack of its own power device, the device can only operate in waters where power boats can only navigate (at least wide waters with a depth of more than 5m). The above reasons severely limit the operating range of the device. The device is basically only used in shallow sea waters with a depth greater than 5m, and cannot be used for inland river and lake surveys. In addition, since the main body of the equipment is underwater and is softly connected to a surface vessel by a cable during operation, and the GPS can only be fixed on the vessel, in waters with complex water flow environments, the GPS location information may be misaligned with the actual scanning position of the SSS, and the accuracy of the bottom map position obtained will be reduced. Summary of the invention

[0004] The present invention aims at the problem that the existing underwater scanner Speedy Sea Scanner cannot be applied to shallow sea waters with a depth of less than 5m, and because the main body of the device is underwater, it is softly connected to the surface ship by a cable during operation, and the GPS can only be fixed on the ship. In waters with complex water flow environment, the GPS position information may be misaligned with the real scanning position of the SSS, and the accuracy of the bottom map position obtained is reduced. A water-to-air fast scanner is proposed, and the solution specifically includes:

[0005] A water-air fast scanner, the scanner comprising:

[0006] Mariner 4 drone, GPS module, extension pole and optical camera array;

[0007] The GPS module is set on the cover of the Sailor 4 UAV.

[0008] The extension rod is fixed to the bottom of the Sailor 4 UAV; the optical camera array is arranged along the axial direction of the extension rod.

[0009] Based on the same inventive concept, the present invention also proposes an underwater scanning method based on a water-to-air fast scanner, the method comprising:

[0010] S1: Determine the deployment method of the water-air rapid scanner according to the water environment, and select the corresponding baseline length and array spacing;

[0011] S2: Configure the working mode of the camera array and synchronize the clock of the camera array with the GPS module;

[0012] S3: Fix the GPS module at the center point of the camera array, and calculate the scanning trajectory coordinate data of each camera according to the data collected by the GPS module, the baseline length and the array spacing;

[0013] S4: The water-air rapid scanner is deployed to the scanning water area according to the preset deployment mode, and the water-air rapid scanner records underwater image information;

[0014] S5: extracting video frames according to the acquired underwater image information;

[0015] S6: interpolating the scanning trajectory coordinate data according to the frequency of the extracted video frame to obtain a smooth GPS motion trajectory;

[0016] S7: Obtain the camera's flight path according to the GPS motion trajectory and array spacing;

[0017] S8: Use SfM technology to reconstruct the route trajectory and obtain underwater images.

[0018] Furthermore, a preferred method is proposed, wherein the deployment method in step S1 includes: remote air deployment and ship mode deployment;

[0019] When the deployment method is remote aerial deployment, the baseline is 2m;

[0020] When the deployment mode is ship mode, the baseline length is 2m or 4m;

[0021] When the baseline is 2m, the array spacing is 0.6m;

[0022] When the baseline is 4m, the array spacing is between 0.6m and 1.2m.

[0023] Furthermore, a preferred method is proposed, wherein step S2 comprises:

[0024] Turn on each camera in the camera array in turn;

[0025] Set the camera working mode to 4K video recording and the lens mode to standard shooting mode;

[0026] Place a clock in seconds in front of the camera array, and the clock stays in front of each camera lens for 5 seconds;

[0027] Set GPS to sample spatial position and attitude at a frequency of 1 time / s;

[0028] The clock image captured by the camera is used as synchronization information, combined with the position information sampled by the GPS, to achieve clock synchronization between the camera array and the GPS module.

[0029] Furthermore, a preferred method is proposed, wherein step S4 comprises:

[0030] When performing the task, ensure that each camera is recording video and simultaneously recording GPS data, the GPS data including the latitude and longitude coordinates of each camera's location and the timestamp of the recorded video;

[0031] Every time the clock shows a new time, record the timestamp of the current time and the corresponding GPS data.

[0032] Furthermore, a preferred embodiment is proposed, wherein step S5 comprises:

[0033] Extract video frames from the videos captured by each camera at a frame rate of 4FPS or 5FPS;

[0034] Match each extracted video frame with the clock timestamp recorded during recording;

[0035] Use the GPS data when recording the video to synchronize the timestamp of the video frame with the corresponding GPS data timestamp;

[0036] The synchronized video frames are associated with the GPS data to obtain the geographic location information corresponding to each video frame.

[0037] Furthermore, a preferred method is proposed, wherein the step S3 of calculating the scanning trajectory coordinate data of each camera includes:

[0038] Get the tangent vector V i :

[0039] V i =P i+1 -P i =(x i+1 -x i ,y i+1 -y i )=(v x ,v y );

[0040] Among them, P i is the i-th point in the GPS trajectory, P i+1 is the i+1th point in the GPS track. In the WGS 84 coordinate system, x i P i The horizontal coordinate of the point, y i P i The vertical coordinate of the point, x i+1 P i+1 The horizontal coordinate of the point, y i+1 P i+1 The vertical coordinate of the point, v x is the X-axis component of the tangent vector, v y is the Y-axis component;

[0041] Get the baseline direction vector

[0042]

[0043] According to the baseline length L, tangent vector and baseline direction vector, calculate when the GPS is at point P i The specific coordinates of each camera at that time, among which the coordinates of the camera closest to GPS C are:

[0044] C=(c x ,c y )=(x i -0.5L·v y ,y i +L·v x );

[0045] Among them, c x is the horizontal coordinate of a camera in the camera array in the WGS 84 coordinate system at the current moment, c y Is the vertical axis.

[0046] Furthermore, a preferred method is proposed, wherein step S6 comprises:

[0047] Aligning the timestamp in the scanning trajectory coordinate data with the timestamp of the video frame;

[0048] Interpolation is performed between adjacent scanning trajectory points according to the piecewise cubic Hermite interpolation method;

[0049] The interpolated scanning trajectory data is integrated with the video frame data to obtain a smooth GPS motion trajectory.

[0050] Furthermore, a preferred embodiment is proposed, wherein step S7 comprises:

[0051] According to the relationship between the normal line of the camera array and the tangent line of the GPS motion trajectory and the array spacing value, the route trajectory of each camera relative to the GPS motion trajectory is calculated;

[0052] The calculated GPS position information of each camera is added to the corresponding video frame.

[0053] Furthermore, a preferred method is proposed, wherein step S8 comprises:

[0054] Use the SfM algorithm to extract feature points in the image and match them between consecutive images;

[0055] Using GPS route trajectory information, estimate the camera position and orientation for each image;

[0056] Reconstruct the 3D point cloud using the matched feature points and the camera’s positioning information;

[0057] An underwater map is generated based on the three-dimensional point cloud information.

[0058] The present invention is beneficial in that:

[0059] The present invention solves the problem that the existing underwater scanner Speedy Sea Scanner cannot be applied to shallow sea waters with a water depth of less than 5m, and because the main body of the equipment is underwater, it is softly connected to the surface vessel by a cable during operation, and the GPS can only be fixed on the vessel. In waters with complex water flow environments, the GPS position information may be misaligned with the actual scanning position of the SSS, resulting in reduced accuracy of the bottom map position.

[0060] The water-to-air rapid scanner described in the present invention is different from existing scanners. The designed water-to-air rapid scanner is specially designed for shallow sea waters with a water depth of less than 5m, which expands the scope of application of underwater scanners. By installing the GPS module on the cover of the Sailor 4 UAV, the shielding of the GPS signal by the underwater equipment body is avoided, and the problem of misalignment between the GPS position information and the actual position of the scanner due to the complex water flow environment is reduced. By arranging the optical camera array along the axial direction of the extension rod, the underwater scene can be captured more accurately and the position accuracy of the scanned map can be improved. The use of a drone as a scanning platform increases the flexibility and maneuverability of the equipment, making scanning tasks in waters more efficient and faster.

[0061] The water-air rapid scanner described in the present invention uses the commercial-grade unmanned aerial vehicle Sailor 4 waterproof unmanned aerial vehicle produced by Swepu Company as the carrier of the optical camera array, thereby realizing rapid deployment. The Sailor 4 waterproof unmanned aerial vehicle is a unmanned aerial vehicle designed for water and air operations, with a relatively rich selection of accessories, such as a ship-shaped float, a camera extension rod, etc., which can meet a variety of surface or underwater operation needs. However, the unmanned aerial vehicle currently only supports mounting a single waterproof gimbal and an optical camera, and there are no extension accessories related to the optical camera array. When a single camera is used for underwater scanning and map reconstruction, the efficiency will be seriously reduced, and even the scanning task cannot be completed within the battery life. The present invention improves the unmanned aerial vehicle Sailor 4 and designs an optical camera array, whose submergence depth is about 0.3m. The optical camera array is connected to the Sailor 4 unmanned aerial vehicle through an extension rod, and the optical camera array is suspended underwater through the extension rod to better capture the underwater scene. In the array, lightweight and stable sports cameras such as GoPro or Osmo Action are mainly selected as optical cameras. The skeleton structure of the camera array is mainly composed of 3D printed parts and carbon fiber tubes. The overall design of the structure focuses on lightness, easy disassembly and adjustment. The GPS module is installed on the cover of the drone. The GPS module is responsible for obtaining the global positioning system information of the current position in order to accurately locate the scanning area and avoid the interference of the underwater subject on the GPS signal. The optical camera array is arranged along the axial direction of the extension rod to collect information such as underwater terrain and objects. These data are transmitted to surface ships or ground stations by drones for subsequent processing and analysis.

[0062] The underwater scanning method based on the water-to-air rapid scanner described in the present invention first determines the deployment mode of the water-to-air rapid scanner according to the water environment, and selects the appropriate baseline length and array spacing, so that it can adapt to different water environments, thereby improving the applicability of the scanning. By configuring the working mode of the camera array, the camera array is synchronized with the clock of the GPS module to ensure that the recording of the underwater image and the acquisition of GPS data are coordinated, which is conducive to the accuracy of subsequent data processing. The GPS module is fixed at the center point of the camera array, and the scanning trajectory coordinate data of each camera is calculated in combination with the baseline length and array spacing, and interpolated to obtain a smooth GPS motion trajectory, thereby improving the accuracy and reliability of the trajectory data. The route trajectory is reconstructed and processed by SfM (Structure from Motion), so as to efficiently obtain underwater images. SfM can restore the three-dimensional structure of the scene by analyzing the images taken by the camera at different positions, thereby realizing the accurate restoration of the underwater landform.

[0063] The underwater scanning method based on the water-to-air rapid scanner described in the present invention determines the deployment mode of the water-to-air rapid scanner according to the water environment, selects the appropriate baseline length and array spacing to adapt to the specific water environment, and ensures the effectiveness of the scanning. The working mode of the camera array is configured, and the camera array is synchronized with the clock of the GPS module to ensure that the recording of the underwater image is synchronized with the acquisition time of the GPS data for subsequent data processing. The scanning trajectory coordinate data of each camera is calculated according to the data collected by the GPS module, the baseline length and the array spacing, and the interpolation technology is used to make the motion trajectory smoother and improve the accuracy of the trajectory data. The route trajectory is reconstructed by SfM, and the three-dimensional structure of the underwater landform is obtained by analyzing the images taken by the camera at different positions, and finally a high-precision underwater image is obtained. Through the water-to-air rapid scanner, combined with a reasonable deployment method and data processing steps, efficient and accurate scanning of the bottom of the water area is achieved. The method aims to solve the positioning accuracy problem existing in traditional underwater scanning, is suitable for shallow sea waters with a water depth of less than 5m, and ensures that the acquired underwater image data has high precision and high resolution to meet the fine requirements of underwater topography in the fields of marine research, geological survey, underwater archaeology, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a schematic diagram of the underwater scanner described in Embodiment 1;

[0065] Figure 2 This is a flow chart of the underwater scanning method based on the water-air fast scanner described in the second embodiment;

[0066] Figure 3 This is a water-off effect diagram for testing the remote deployment function when the 2m baseline is mounted as described in Implementation Method 3;

[0067] Figure 4 This is a physical picture of the equipment when the 4m baseline is mounted as described in Implementation Method 3;

[0068] Figure 5 This is a schematic diagram of equipment operation and scanning test when a 4m baseline is mounted as described in Implementation Method 3;

[0069] Figure 6 This is a schematic diagram of identifying and detecting river bottom garbage according to the eleventh embodiment;

[0070] Figure 7 This is the underwater surface map described in Implementation Example 11. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0072] Implementation method 1, see Figure 1 The water-air rapid scanner described in this embodiment comprises:

[0073] Mariner 4 drone, GPS module, extension pole and optical camera array;

[0074] The GPS module is arranged on the cover of the Sailor 4 UAV.

[0075] The extension rod is fixed to the bottom of the Sailor 4 UAV; the optical camera array is arranged along the axial direction of the extension rod.

[0076] The scanner described in this embodiment is different from the existing scanners. The water-to-air fast scanner designed in this embodiment is specially designed for shallow sea waters with a depth of less than 5m, which expands the scope of application of underwater scanners. By installing the GPS module on the cover of the Sailor 4 UAV, the shielding of the GPS signal by the underwater equipment body is avoided, and the problem of misalignment between the GPS location information and the actual position of the scanner due to the complex water flow environment is reduced. By arranging the optical camera array along the axial direction of the extension rod, the underwater scene can be captured more accurately and the position accuracy of the scanned map can be improved. Using a drone as a scanning platform increases the flexibility and maneuverability of the equipment, making scanning tasks in waters more efficient and faster.

[0077] This embodiment uses the commercial-grade unmanned aerial vehicle Sailor 4 waterproof drone produced by Swepu as the carrier of the optical camera array, thereby achieving rapid deployment. The Sailor 4 waterproof drone is a drone designed for water and air operations. It has a relatively rich selection of accessories, such as a boat-shaped float, a camera extension rod, etc., which can meet a variety of surface or underwater operation needs. However, the drone currently only supports mounting a single waterproof gimbal and an optical camera, and there are no extension accessories related to the optical camera array. When a single camera is used for underwater scanning and map reconstruction, it will seriously reduce efficiency, and even fail to complete the scanning task within the battery life. This embodiment improves the unmanned aerial vehicle Sailor 4 and designs an optical camera array, whose submergence depth is about 0.3m. The optical camera array is connected to the Sailor 4 drone through an extension rod, and the optical camera array is suspended underwater through the extension rod to better capture the underwater scene. In the array, lightweight and stable sports cameras such as GoPro or Osmo Action are mainly selected as optical cameras. The skeleton structure of the camera array is mainly composed of 3D printed parts and carbon fiber tubes. The overall design of the structure focuses on the characteristics of lightness, easy disassembly and easy adjustment. The GPS module is installed on the canopy of the drone. The GPS module is responsible for obtaining the global positioning system information of the current position in order to accurately locate the scanning area and avoid the interference of underwater subjects on the GPS signal. The optical camera array is arranged along the axial direction of the extension rod to collect information such as underwater terrain and objects. These data are transmitted to the surface ship or ground station through the drone for subsequent processing and analysis.

[0078] When the camera array is within the load limit of the Sailor 4 (2m baseline), the scanner described in this embodiment can take off from land, fly into the water to collect data, and then leave the water and return. In many specific waters, efficient data collection can be achieved by using the remote deployment function. For example, when collecting river data, the river bank is about 1.5m away from the water surface. Ordinary equipment needs to be placed in the water manually and with great difficulty, but the scanner described in this embodiment can fly directly into the water from the air, collect data, and then fly out of the water and return to land. In order to realize this remote deployment function, the main body of the camera array adopts a lightweight design. In actual applications, the skeleton adopts a hollow carbon fiber tube (the water in the skeleton can be emptied as quickly as possible when the device leaves the water to ensure the balance of the body), and the connectors are made of 3D printed materials. The camera uses a lightweight sports camera such as GoPro 12, which minimizes the weight while ensuring data quality.

[0079] The scanner described in this embodiment provides an underwater device that can scan efficiently and accurately in shallow waters. By solving the limitations of existing scanners in small water depth environments, reducing the impact of GPS misalignment on scanning accuracy, and improving the accuracy of underwater maps, underwater surveys and map making in a wider range can be achieved. The scanner described in this embodiment can be applied to fields such as marine research, underwater archaeology, and environmental monitoring.

[0080] Implementation Method 2: See Figure 2 This embodiment describes an underwater scanning method based on a water-to-air rapid scanner, the method comprising:

[0081] S1: Determine the deployment method of the water-air rapid scanner according to the water environment, and select the corresponding baseline length and array spacing;

[0082] S2: Configure the working mode of the camera array and synchronize the clock of the camera array with the GPS module;

[0083] S3: Fix the GPS module at the center point of the camera array, and calculate the scanning trajectory coordinate data of each camera according to the data collected by the GPS module, the baseline length and the array spacing;

[0084] S4: The water-air rapid scanner is deployed to the scanning water area according to the preset deployment mode, and the water-air rapid scanner records underwater image information;

[0085] S5: extracting video frames according to the acquired underwater image information;

[0086] S6: interpolating the scanning trajectory coordinate data according to the frequency of the extracted video frame to obtain a smooth GPS motion trajectory;

[0087] S7: Obtain the camera's flight path according to the GPS motion trajectory and array spacing;

[0088] S8: Use SfM technology to reconstruct the route trajectory and obtain underwater images.

[0089] The method described in this embodiment first determines the deployment mode of the water-to-air rapid scanner according to the water environment, and selects the appropriate baseline length and array spacing, so that it can adapt to different water environments, thereby improving the applicability of the scan. By configuring the working mode of the camera array, the camera array is synchronized with the clock of the GPS module to ensure that the recording of the underwater image and the acquisition of GPS data are coordinated, which helps the accuracy of subsequent data processing. The GPS module is fixed at the center point of the camera array, and the scanning trajectory coordinate data of each camera is calculated in combination with the baseline length and array spacing, and interpolated to obtain a smooth GPS motion trajectory, thereby improving the accuracy and reliability of the trajectory data. The route trajectory is reconstructed and processed by SfM (Structure from Motion), thereby efficiently acquiring underwater images. SfM can restore the three-dimensional structure of the scene by analyzing the images taken by the camera at different positions, thereby achieving accurate restoration of the underwater landforms.

[0090] The method described in this embodiment determines the deployment mode of the water-to-air rapid scanner according to the water environment, selects the appropriate baseline length and array spacing to adapt to the specific water environment, and ensures the effectiveness of the scan. Configure the working mode of the camera array, synchronize the camera array with the clock of the GPS module, and ensure that the recording of the underwater image is synchronized with the acquisition time of the GPS data for subsequent data processing. Calculate the scanning trajectory coordinate data of each camera according to the data collected by the GPS module, the baseline length and the array spacing, and use the interpolation technology to make the motion trajectory smoother and improve the accuracy of the trajectory data. Use SfM to reconstruct the route trajectory, and obtain the three-dimensional structure of the underwater landform by analyzing the images taken by the camera at different positions, and finally obtain a high-precision underwater image. Through the water-to-air rapid scanner, combined with a reasonable deployment method and data processing steps, efficient and accurate scanning of the bottom of the water area is achieved. This method aims to solve the positioning accuracy problem existing in traditional underwater scanning, and is suitable for shallow sea waters with a depth of less than 5m, ensuring that the acquired underwater image data has high precision and high resolution to meet the fine requirements of underwater topography in fields such as marine research, geological survey, and underwater archaeology.

[0091] Implementation method three, see Figure 3 , Figure 4 and Figure 5 This embodiment is a further limitation of the underwater scanning method based on the water-air rapid scanner described in the second embodiment, and the deployment mode in step S1 includes: remote air deployment and ship mode deployment;

[0092] When the deployment method is remote aerial deployment, the baseline is 2m;

[0093] When the deployment mode is ship mode, the baseline length is 2m or 4m;

[0094] When the baseline is 2m, the array spacing is 0.6m;

[0095] When the baseline is 4m, the array spacing is between 0.6m and 1.2m.

[0096] This implementation method can be customized according to the specific water environment and mission requirements through long-distance aerial deployment and the selection of different baselines and array spacings, thereby improving the adaptability and flexibility of scanning. For long-distance aerial deployment, two different baseline lengths (2m and 4m) and a range of array spacing (0.6m-1.2m) were selected when the baseline was 4m, making the method more versatile and applicable to situations of different water depths and complexities. Through the careful setting of the baseline and array spacing, the resolution and scanning efficiency of the underwater image are balanced, ensuring that high-quality underwater image data can be obtained in different scenarios.

[0097] Remote aerial deployment refers to the remote placement of the water-air fast scanner into the water area, and the suspension and movement by aerial means. This deployment method is achieved through drones, which improves the flexibility and efficiency of the operation. Different baseline lengths and array spacing directly affect the spatial layout of the camera array. Choosing appropriate values ​​can balance the performance of the system. Short baselines and small array spacing are suitable for close-range scanning and adapt to smaller waters; long baselines and large array spacing are suitable for long-range scanning and adapt to larger waters. The selection of a baseline of 2m or 4m, and the setting of an array spacing between 0.6m and 1.2m when the baseline is 4m, are the result of careful adjustment according to specific circumstances. This careful adjustment helps to obtain the best underwater scanning effect in different situations. By further limiting the deployment method, the underwater scanning method is optimized to make it more applicable and operational. Specifically, remote aerial deployment improves the flexibility and efficiency of underwater scanning, while the selection of baselines and array spacing is to balance image quality and scanning efficiency in different water environments. The overall goal is to obtain high-quality, high-resolution underwater image data through water-air fast scanners in different water environments to meet the survey needs of rivers, lakes and even shallow sea areas.

[0098] Embodiment 4: This embodiment further limits the underwater scanning method based on the water-air fast scanner described in Embodiment 2, and the step S2 includes:

[0099] Turn on each camera in the camera array in turn;

[0100] Set the camera working mode to 4K video recording and the lens mode to standard shooting mode;

[0101] Place a clock in seconds in front of the camera array, and the clock stays in front of each camera lens for 5 seconds;

[0102] Set GPS to sample spatial position and attitude at a frequency of 1 time / s;

[0103] The clock image captured by the camera is used as synchronization information, combined with the position information sampled by the GPS, to achieve clock synchronization between the camera array and the GPS module.

[0104] By placing a clock in front of the camera array, each camera starts working at the same time, ensuring the time synchronization of each camera in the camera array. This helps to accurately record the images captured by different cameras and improve the data accuracy of underwater scanning. Using the GPS module to sample the spatial position and attitude at a frequency of 1 time / s, the accurate position and orientation of the water-to-air fast scanner in space can be obtained. This is crucial for subsequent image processing and map construction. The clock image captured by the camera is used as synchronization information, combined with the position information sampled by GPS, and the visual information and positioning information are comprehensively utilized to provide diversified synchronization data, which enhances the robustness and reliability of the scanner.

[0105] This embodiment ensures that each camera starts working in sequence by turning on each camera in the camera array in turn, avoiding scanning confusion caused by simultaneous startup. The camera working mode is set to 4K video recording, and the lens mode is set to standard shooting mode to obtain high-resolution underwater images. A clock is placed in front of the camera array, and each camera can start recording images synchronously through the captured clock image. This helps to ensure that the images captured by each camera in the camera array correspond to the same time point, ensuring the time consistency of the data. The GPS module performs spatial position and attitude sampling at a frequency of 1 time / s, providing accurate position and orientation information of the water-air fast scanner in three-dimensional space. Combining the clock image captured by the camera and the position information sampled by GPS, the clock synchronization of the camera array and the GPS module is achieved, ensuring the data consistency of the water-air scanner in time and space. This embodiment ensures the accuracy and consistency of data during underwater scanning. By ensuring that each camera in the camera array starts working at the same time point, and combining the position information sampled by GPS, the water-air fast scanner is highly synchronized in time and space. This helps to generate accurate underwater images and maps.

[0106] Embodiment 5: This embodiment further limits the underwater scanning method based on the water-air fast scanner described in Embodiment 2, and the step S4 includes:

[0107] When performing the task, ensure that each camera is recording video and simultaneously recording GPS data, the GPS data including the latitude and longitude coordinates of each camera's location and the timestamp of the recorded video;

[0108] Every time the clock shows a new time, record the timestamp of the current time and the corresponding GPS data.

[0109] This embodiment ensures that each image captured by the camera has an associated timestamp and location information by recording GPS data while recording the video, including the latitude and longitude coordinates of each camera position and the timestamp of the recorded video. This helps in the temporal and spatial association of subsequent data processing and analysis. GPS data and video data are recorded at the same time, providing more comprehensive information, including camera positions and timestamps. This is crucial for generating high-quality, information-rich underwater scanning images. Whenever the clock shows a new time, the timestamp of the current time and the corresponding GPS data are recorded, making full use of the clock as a synchronization reference to ensure the accuracy of the timestamp. This helps to maintain consistent time stamps between different cameras.

[0110] When performing underwater scanning operations, ensure that each camera records GPS data while recording video. GPS data includes the latitude and longitude coordinates of each camera's location and the timestamp of the recorded video. This ensures that each video frame has location information and timestamp associated with it. Every time the clock shows a new time, the scanner records the timestamp of the current time and the corresponding GPS data. This recording method associates each timestamp with the corresponding location information, ensuring spatiotemporal consistency. This embodiment achieves comprehensive recording and spatiotemporal consistency of data during underwater scanning. By recording GPS data while recording video, it is ensured that the images captured by each camera have accurate location information and timestamps. This helps with subsequent data processing, map construction and analysis, and provides high-quality, accurate and comprehensive underwater scanning data. At the same time, by recording synchronously with the clock, the consistency of the timestamp is guaranteed.

[0111] Embodiment 6: This embodiment further limits the underwater scanning method based on the water-air fast scanner described in Embodiment 5, and the step S5 includes:

[0112] Extract video frames from the videos captured by each camera at a frame rate of 4FPS or 5FPS;

[0113] Match each extracted video frame with the clock timestamp recorded during recording;

[0114] Use the GPS data when recording the video to synchronize the timestamp of the video frame with the corresponding GPS data timestamp;

[0115] The synchronized video frames are associated with the GPS data to obtain the geographic location information corresponding to each video frame.

[0116] This embodiment summarizes the extraction of video frames from the videos captured by each camera at a frame rate of 4FPS or 5FPS, which helps to reduce the amount of data and improve processing efficiency, while still retaining enough information for underwater image analysis. By matching the extracted video frames with the clock timestamp recorded during recording, combined with the GPS data when the video was recorded, the temporal and spatial consistency of the video frames is achieved. This ensures that each video frame has an accurate timestamp and geographic location information, making subsequent data analysis more reliable and accurate. By associating the synchronized video frames with the GPS data, the geographic location information corresponding to each video frame is obtained.

[0117] Specifically, video frames are extracted from the videos collected by each camera at a frame rate of 4 frames per second (4FPS) or 5 frames per second (5FPS). Such an extraction frequency can not only retain sufficient information, but also reduce the amount of data to a certain extent, which is conducive to subsequent processing and analysis. Each extracted video frame is matched with the clock timestamp recorded during recording to ensure that each video frame has an accurate timestamp. Using the GPS data when recording the video, the timestamp of the video frame is synchronized with the corresponding GPS data timestamp to achieve spatiotemporal synchronization and ensure the spatiotemporal consistency of the video frame. The synchronized video frame is associated with the GPS data to obtain the geographical location information corresponding to each video frame. This makes each video frame associated with a specific geographical coordinate, providing key geographical information for subsequent map construction and terrain analysis. In the underwater scanning process, this embodiment obtains underwater image data with accurate timestamps and geographical coordinates through effective frame extraction, spatiotemporal synchronization and geographical location information acquisition. This provides a high-quality, spatiotemporal consistent data basis for subsequent map construction, underwater terrain analysis, and research in other related fields. This helps to deeply understand the underwater environment and support applications such as marine scientific research and resource exploration.

[0118] Embodiment 7: This embodiment further limits the underwater scanning method based on the water-air rapid scanner described in Embodiment 5. The calculation of the scanning trajectory coordinate data of each camera in step S3 includes:

[0119] Get the tangent vector V i :

[0120] V i =P i+1 -P i =(x i+1 -x i ,y i+1 -y i )=(v x ,v y );

[0121] Among them, P iis the i-th point in the GPS trajectory, P i+1 is the i+1th point in the GPS track. In the WGS 84 coordinate system, x i P i The horizontal coordinate of the point, y i P i The vertical coordinate of the point, x i+1 P i+1 The horizontal coordinate of the point, y i+1 P i+1 The vertical coordinate of the point, v x is the X-axis component of the tangent vector, v y is the Y-axis component;

[0122] Get the baseline direction vector

[0123]

[0124] According to the baseline length L, tangent vector and baseline direction vector, calculate when the GPS is at point P i The specific coordinates of each camera at that time, among which the coordinates of the camera closest to GPS C are:

[0125] C=(c x ,c y )=(x i -0.5L·v y ,y i +L·v x );

[0126] Among them, c x is the horizontal coordinate of a camera in the camera array in the WGS 84 coordinate system at the current moment, c y Is the vertical axis.

[0127] Specifically, the GPS coordinates are transformed from longitude and latitude into WGS 84 / UTMZone X coordinates according to the area selected by the field survey, where X is the number of the area. 1 ,P 2 ,K,P n Define a GPS trajectory where each point P i In the WGS84 / UTM Zone X coordinate system, there are coordinates (x i ,y i ), for the i-th point on the trajectory, its tangent direction can be calculated by comparing it with the next point P i+1 The tangent vector V is obtained by the vector difference between i V i =P i+1 -P i =(x i+1 -xi ,y i+1 -y i )=(v x ,v y ). If the baseline is perpendicular to the tangent vector, then the baseline direction vector According to the baseline length L and the position of each camera in the array, it can be calculated that when the GPS is at point P i The specific coordinates of each camera at that time, for example, the coordinates of the camera closest to the GPS are C = (c x ,c y )=(x i -0.5L·v y ,y+L·v x ).

[0128] Embodiment 8: This embodiment further limits the underwater scanning method based on the water-air fast scanner described in Embodiment 5, and the step S6 includes:

[0129] Aligning the timestamp in the scanning trajectory coordinate data with the timestamp of the video frame;

[0130] Interpolation is performed between adjacent scanning trajectory points according to the piecewise cubic Hermite interpolation method;

[0131] The interpolated scanning trajectory data is integrated with the video frame data to obtain a smooth GPS motion trajectory.

[0132] In this embodiment, the timestamp in the scanning trajectory coordinate data is aligned with the timestamp of the video frame to ensure that the time information of the scanning trajectory and the video frame is consistent. This helps to accurately match the location information of the underwater scan with the corresponding video frame, and improves the consistency and availability of the data. The segmented cubic Hermite interpolation method is used to interpolate between adjacent scanning trajectory points to fill in the missing or sparse parts in the trajectory data. This helps to enrich the trajectory data, making the trajectory more continuous in time, improving the integrity of the trajectory data, and further smoothing the trajectory, eliminating the noise and instability that may be introduced by the interpolation, making the trajectory more stable, more in line with the actual underwater motion conditions, and improving the reliability and authenticity of the trajectory data. The interpolated and smoothed scanning trajectory data is integrated with the video frame data to finally obtain a smooth GPS motion trajectory. The integrated data is more interpretable, making the analysis of the underwater environment more accurate and reliable.

[0133] By aligning the timestamp in the scanning trajectory coordinate data with the timestamp of the video frame, the time information of the two is guaranteed to be consistent. This can be achieved by matching and adjusting the timestamp. Use the linear interpolation method to interpolate between adjacent scanning trajectory points. Linear interpolation is a method of estimating unknown points through the linear relationship between known points, which is used to fill in the missing or sparse parts in the trajectory data. The interpolated trajectory data is processed using the segmented cubic Hermite interpolation algorithm, which can ensure the monotonicity within the interpolation interval and have a trajectory smoothing effect. The interpolated scanning trajectory data is integrated with the video frame data to obtain the final smooth GPS motion trajectory. The integration process involves operations such as matching and aligning the data. This implementation method improves the spatiotemporal consistency and quality of underwater scanning data. Through time alignment, trajectory interpolation and smoothing, the scanning trajectory data is made more continuous, complete and smooth, and better corresponds to the video frame data, thereby providing a more reliable and accurate data basis for subsequent underwater environment analysis, map construction and other applications.

[0134] Embodiment 9: This embodiment further limits the underwater scanning method based on the water-air fast scanner described in Embodiment 2, and the step S7 includes:

[0135] According to the relationship between the normal line of the camera array and the tangent line of the GPS motion trajectory and the array spacing value, the route trajectory of each camera relative to the GPS motion trajectory is calculated;

[0136] The calculated GPS position information of each camera is added to the corresponding video frame.

[0137] In this embodiment, the route trajectory of each camera relative to the GPS motion trajectory is obtained by calculating based on the relationship between the normal of the camera array and the tangent of the GPS motion trajectory and the array spacing value. This helps to accurately determine the position of the camera underwater and improves the accuracy of the scan data. The calculated GPS position information of each camera is added to the corresponding video frame. By associating the GPS position information with the video frame, it is possible to embed geographic coordinate information in the underwater scan data, making the scan data more spatially referenced and facilitating subsequent map construction and analysis. The specific method can be achieved through the seventh embodiment.

[0138] The relationship between the normal of the camera array and the tangent of the GPS motion trajectory is based on the geometric relationship between the underwater camera layout and the direction of the motion trajectory. The array spacing value is the distance between adjacent cameras in the camera array, and this value plays an important role as a geometric parameter in the calculation. By considering the relative position between the cameras, the route trajectory of each camera relative to the GPS motion trajectory can be calculated more accurately. According to the relationship between the normal of the camera array and the tangent of the GPS motion trajectory and the array spacing value, the route trajectory of each camera relative to the GPS motion trajectory is calculated using a geometric calculation method. The calculated GPS position information of each camera is embedded in the corresponding video frame data. This can be achieved by adding the GPS position information of the camera to the metadata of the video frame or by transforming the pixel coordinates. This embodiment introduces accurate geographic coordinate information into the underwater scanning data, so that the scanning data has more geographic spatial reference. By calculating the route trajectory of each camera relative to the GPS motion trajectory and embedding the GPS position information into the video frame, the geographic positioning of the underwater scanning data is achieved. This helps in subsequent map construction, environmental analysis and other applications, and improves the practicality and application scope of underwater scanning data.

[0139] Embodiment 10: This embodiment further defines the underwater scanning method based on the water-air rapid scanner described in Embodiment 2. Step S8 includes:

[0140] Use the SfM algorithm to extract feature points in the image and match them between consecutive images;

[0141] Using GPS route trajectory information, estimate the camera position and orientation for each image;

[0142] Reconstruct the 3D point cloud using the matched feature points and the camera’s positioning information;

[0143] An underwater map is generated based on the three-dimensional point cloud information.

[0144] The SfM (Structure from Motion) algorithm used in this embodiment can extract feature points in the image and match them between consecutive images, thereby achieving high-precision reconstruction of the underwater scene. This can provide a detailed and accurate three-dimensional point cloud, which provides a good foundation for the subsequent generation of underwater maps. The position and direction of the camera are estimated for each image using GPS route trajectory information, so that the generation of the three-dimensional point cloud has global position information. This helps to ensure the accuracy and consistency of the underwater map and improve the reliability of the entire system. The use of a fast water-to-air scanner and the application of the SfM algorithm make the underwater scanning process more real-time and efficient. This is of great significance for application scenarios that require timely acquisition of underwater map information (such as search and rescue, underwater exploration, etc.). Combining the matched feature points with the camera's positioning information helps to better understand the underwater scene. This comprehensive use of information can improve the accuracy and completeness of the underwater map.

[0145] The Structure from Motion described in this embodiment is a computer vision technology that infers the movement of the camera and the three-dimensional structure of the scene by extracting feature points in an image sequence and tracking their movement. Here, SfM is used to extract feature points and match them between consecutive images to restore the three-dimensional structure of the underwater scene. The position and direction of the camera are estimated for each image using GPS route trajectory information. In this way, the image can be positioned in the global coordinate system to ensure that the generated three-dimensional point cloud has accurate geographic location information. Using the matched feature points and the positioning information of the camera, a three-dimensional point cloud of the underwater scene is generated through three-dimensional reconstruction technology. This can be understood as mapping the feature points in images taken from different angles to three-dimensional space, thereby restoring the shape and structure of the underwater scene. Based on the obtained three-dimensional point cloud information, an underwater map can be generated. This process involves converting the information of the three-dimensional point cloud into a visual map form, providing a more intuitive and easy-to-understand underwater scene presentation. By combining the SfM algorithm and GPS information, a high-precision, globally positioned three-dimensional reconstruction of the underwater scene is achieved, and an underwater map with geographic location information is finally generated. This helps in various underwater application fields, such as marine exploration, underwater archaeology, environmental monitoring, etc., provides detailed underwater geographic information, and provides support for scientific research, rescue and planning.

[0146] Implementation method 11: See Figure 6 and Figure 7 This embodiment provides a specific example of the water-air fast scanner described in the first embodiment, specifically:

[0147] The structure of the water-air rapid scanner described in this embodiment includes: an above-water part and an underwater part. The above-water part is a Sailor 4 drone equipped with a float and a precision GPS fixed on the lower cover of the drone. The underwater part is an independently designed optical camera array, and the submerged depth is about 0.3m. The two parts are connected by an extension rod officially provided by the Sailor 4 drone. Both the above-water part and the extension rod adopt an inherent design. When performing related operations, this part will not change parameters. The main parameters that need to be adjusted according to the actual use environment are the relevant parameters of the optical camera array.

[0148] In the array, lightweight and stable sports cameras such as GoPro or Osmo Action are mainly selected as optical cameras. The skeleton structure of the camera array is mainly composed of 3D printed parts and carbon fiber tubes. The overall design of the structure focuses on lightness, easy disassembly and adjustment. The number of elements in the camera array is 4, and the baseline length can be adjusted according to the application scenario. It can be selected as 2m or 4m. The array spacing can be adjusted between 0.6-1.2m, corresponding to different equipment deployment requirements, water depth conditions and map coverage areas.

[0149] The underwater scanning method based on the water-air fast scanner includes:

[0150] Step 1: Determine the equipment deployment method and select the appropriate baseline length and array spacing based on the environmental conditions of the surveyed waters, such as openness, approximate water depth, turbidity, etc.

[0151] When the baseline length is 2m, the array spacing is fixed at 0.6m. At this time, the equipment can be deployed remotely by air (take off from land, fly to the survey waters and land, and then fly back to land after the operation). It is mainly used in narrow waters, such as narrow rivers; or in waters where traditional deployment methods are difficult to implement, such as open lakes with a large number of aquatic plants growing on the bottom and no suitable docks or boats. According to tests, when the 2m baseline is used in waters with a depth of 0.7-5m, it ensures both the operational flexibility of the equipment and the efficiency of underwater scanning.

[0152] Due to the load and balance control restrictions of the Mariner 4 UAV, when the baseline length is 4m, the entire device cannot be deployed remotely in the air. The 4m baseline is mainly used in open, shallow waters with relatively large water depths (2-10m), and can perform large-scale scanning and mapping work with the highest efficiency, but it requires a suitable deployment dock or scientific research ship, and is suitable for open lakes or near-shore shallow sea areas. When a 4m baseline is selected, the array spacing can be adjusted from 0.6 to 1.2m, and the image coverage of adjacent array elements should be guaranteed to be 60%-70%. The larger the array spacing, the smaller the coverage, but a larger array spacing can obtain a larger monitoring range.

[0153] Step 2: Configure the camera working mode and synchronize the clocks of each camera in the camera array with the GPS.

[0154] In this step, first turn on the GPS fixed to the lower cover of the drone, and then turn on each camera in turn. The GPS samples the spatial position and attitude at a frequency of 1 time / s. The camera working mode in the camera array is set to 4K video recording, and the lens mode is set to standard shooting mode (do not use wide-angle mode to ensure that the reconstructed map is not distorted). Then, place the same clock accurate to seconds in front of each camera lens in turn, and let the clock stay in front of each camera lens for about 5 seconds. In the subsequent data processing, the clock in the picture can be used as synchronization information to synchronize with the position information sampled by GPS, so as to obtain accurate geographic coordinates for reconstructing the underwater map.

[0155] Step 3: Deploy the device to the scanning water area according to the predetermined deployment method. Through remote control, heat and drive the camera array to move along the preset trajectory, scan the bottom surface, and save the collected data in the form of 4K video.

[0156] Step 4: Recover the device and export the video data from each camera and the scanning track coordinate data of the GPS.

[0157] Step 5: Synchronize the video and GPS data time through the clock synchronization screen in step 2, and extract video frames from the videos captured by each camera at a frame rate of 4FPS or 5FPS (determined by the water conditions and the speed during the operation, for example, when the water conditions are relatively complex and the speed is fast, select 5FPS frame rate).

[0158] Step 6: The GPS is fixed at the center point of the camera array. The scanning trajectory coordinate data of each camera can be calculated based on the GPS data, baseline length and array spacing. After obtaining the GPS motion trajectory, the trajectory is interpolated and smoothed according to the frame rate of the extracted video frame to generate a smoother GPS motion trajectory; then, since the normal of the camera array is always parallel to the tangent of the GPS motion trajectory and the array spacing is known, the route trajectories of the four cameras can be obtained, that is, the GPS position information corresponding to each frame of the video frame image, and the position information is added to each extracted frame image for subsequent underwater map reconstruction.

[0159] Step 7: The underwater map reconstruction algorithm is based on the existing SfM technology. SfM technology is widely used in UAV aerial survey and map drawing. The image with geographic coordinates obtained in step 6 is similar to the UAV aerial survey image. Existing mature aerial survey software (such as Metashape) can be used to stitch the images to generate a complete underwater map.

[0160] Although SfM technology has been widely used in fields such as aerial mapping, it is rarely used in underwater mapping. SSS is the first device to use this technology to reconstruct underwater maps. The water-to-air rapid scanner described in this embodiment is further improved compared to SSS. SSS is towed by cables, with the main body immersed in water, and the GPS is located on the towing vessel. Due to the catenary effect, the GPS position of SSS has a certain deviation (about 3m) from the actual camera array position, which will cause the generated map's geographical location to be offset, and since it is impossible to set ground control points (GCPs), the offset cannot be compensated by subsequent processing methods. Compared to SSS, the underwater camera array of the water-to-air rapid scanner described in this embodiment is rigidly connected to the power part above water, and the GPS position is at the absolute center point of the camera array. Therefore, the reconstructed map strictly matches the geographical location and has higher accuracy.

[0161] In terms of remote deployment, compared with SSS, the water-air fast scanner described in this embodiment is an improved design based on the Sailor 4 UAV platform. Therefore, when the camera array is within the load limit of the Sailor 4 (2m baseline), the present invention can take off from land, fly into the water for data collection, and then leave the water and return. In many specific waters, this remote deployment function can achieve efficient data collection. For example, when collecting river data in the Hongqi River, the river bank is about 1.5m away from the water surface. Ordinary equipment needs to be placed in the water manually and with great difficulty, but the water-air fast scanner described in this embodiment can fly directly into the water from the air, collect data, and then fly out of the water and return to land. In order to realize the remote deployment function, the main body of the camera array adopts a lightweight design, the skeleton adopts a hollow carbon fiber tube (the water in the skeleton can be emptied as quickly as possible when the equipment leaves the water to ensure the balance of the body), and the connector adopts 3D printing materials. The camera adopts a lightweight sports camera such as GoPro 12, which minimizes the weight while ensuring the data quality.

[0162] The water-air fast scanner described in this embodiment combines the flexibility of waterproof drones operating on the water surface and in the air, and the functionality of optical camera arrays in scanning terrain to achieve map reconstruction. It can be quickly deployed in a variety of complex waters to scan the underwater surface and reconstruct maps. Compared with existing underwater scanning equipment (mainly SSS), the water-air fast scanner described in this embodiment can be used in inland waters, and can also be used in more complex waters, such as narrow rivers, and shallows covered with aquatic plants, and compared with existing equipment, the present invention can achieve an underwater map reconstruction effect with almost the same performance.

[0163] The water-air rapid scanner described in this embodiment was tested in the Hongqi River in Tonghai County, Yuxi City, Yunnan Province. The equipment used a 2m baseline mode to scan the bottom of a river section in front of a sluice in the Hongqi River where garbage is likely to accumulate, and generated Figure 6 The river bottom map with a length of 10m and a width of 4.5m shown in the figure can be used to visually identify and detect river bottom garbage. In addition, the water-air rapid scanner described in this embodiment was also tested in Fuxian Lake, scanning a section of water with lush bottom aquatic plants and an average water depth of 3m, and generating a water bottom surface map as shown in the figure. Figure 7 shown.

[0164] The above further describes in detail the technical solution provided by the present invention in conjunction with the accompanying drawings in order to highlight the advantages and benefits, and is not intended to be a limitation of the present invention. Any modification, combination of implementation modes, improvement and equivalent substitution of the present invention within the spirit and scope of the present invention shall be included in the protection scope of the present invention.

Claims

1. Water-air fast scanner, characterized in that, The scanner comprises: Mariner 4 drone, GPS module, extension pole and optical camera array; The GPS module is set on the canopy of the Sailor 4 UAV. The extension rod is fixed to the bottom of the Sailor 4 UAV; the optical camera array is arranged along the axial direction of the extension rod.

2. An underwater scanning method based on a water-air fast scanner, characterized in that: The method comprises: S1: Determine the deployment method of the water-air rapid scanner according to the water environment, and select the corresponding baseline length and array spacing; S2: Configure the working mode of the camera array and synchronize the clock of the camera array with the GPS module; S3: Fix the GPS module at the center point of the camera array, and calculate the scanning trajectory coordinate data of each camera according to the data collected by the GPS module, the baseline length and the array spacing; S4: The water-air rapid scanner is deployed to the scanning water area according to the preset deployment mode, and the water-air rapid scanner records underwater image information; S5: extracting video frames according to the acquired underwater image information; S6: interpolating the scanning trajectory coordinate data according to the frequency of the extracted video frame to obtain a smooth GPS motion trajectory; S7: Obtain the camera's flight path according to the GPS motion trajectory and array spacing; S8: Use SfM technology to reconstruct the route trajectory and obtain underwater images.

3. The underwater scanning method based on the water-air fast scanner according to claim 2 is characterized in that: The deployment methods in step S1 include: remote aerial deployment and ship mode deployment; When the deployment method is remote aerial deployment, the baseline is 2m; When the deployment mode is ship mode, the baseline length is 2m or 4m; When the baseline is 2m, the array spacing is 0.6m; When the baseline is 4m, the array spacing is between 0.6m and 1.2m.

4. The underwater scanning method based on the water-air fast scanner according to claim 2 is characterized in that: The step S2 comprises: Turn on each camera in the camera array in turn; Set the camera working mode to 4K video recording and the lens mode to standard shooting mode; Place a clock in seconds in front of the camera array, and the clock stays in front of each camera lens for 5 seconds; Set GPS to sample spatial position and attitude at a frequency of 1 time / s; The clock image captured by the camera is used as synchronization information, combined with the position information sampled by the GPS, to achieve clock synchronization between the camera array and the GPS module.

5. The underwater scanning method based on the water-air fast scanner according to claim 2 is characterized in that: The step S4 comprises: When performing the task, ensure that each camera is recording video and simultaneously recording GPS data, the GPS data including the latitude and longitude coordinates of each camera's location and the timestamp of the recorded video; Every time the clock shows a new time, record the timestamp of the current time and the corresponding GPS data.

6. The underwater scanning method based on the water-air fast scanner according to claim 5 is characterized in that: The step S5 comprises: Extract video frames from the videos captured by each camera at a frame rate of 4FPS or 5FPS; Match each extracted video frame with the clock timestamp recorded during recording; Use the GPS data when recording the video to synchronize the timestamp of the video frame with the corresponding GPS data timestamp; The synchronized video frames are associated with the GPS data to obtain the geographic location information corresponding to each video frame.

7. The underwater scanning method based on the water-air fast scanner according to claim 5 is characterized in that: Calculating the scanning trajectory coordinate data of each camera in step S3 includes: Get the tangent vector V i : V i =P i+1 -P i =(x i+1 -x i ,y i+1 -y i )=(v x ,v y ); Among them, P i is the i-th point in the GPS trajectory, P i+1 is the i+1th point in the GPS track. In the WGS 84 coordinate system, x i P i The horizontal coordinate of the point, y i P i The vertical coordinate of the point, x i+1 P i+1 The horizontal coordinate of the point, y i+1 P i+1 The vertical coordinate of the point, v x is the X-axis component of the tangent vector, v y is the Y-axis component; Get the baseline direction vector According to the baseline length L, tangent vector and baseline direction vector, calculate when the GPS is at point P i The specific coordinates of each camera at that time, among which the coordinates of the camera closest to GPS C are: C=(c x ,c y )=(x i -0.5L v y ,y i +L·v x ); Among them, c x is the horizontal coordinate of a camera in the camera array in the WGS 84 coordinate system at the current moment, c y Is the vertical axis.

8. The underwater scanning method based on the water-air fast scanner according to claim 5 is characterized in that: The step S6 comprises: Aligning the timestamp in the scanning trajectory coordinate data with the timestamp of the video frame; Interpolation is performed between adjacent scanning trajectory points according to the piecewise cubic Hermite interpolation method; The interpolated scanning trajectory data is integrated with the video frame data to obtain a smooth GPS motion trajectory. The interpolated and smoothed scanning trajectory data is integrated with the video frame data to obtain a smooth GPS motion trajectory.

9. The underwater scanning method based on the water-air fast scanner according to claim 2 is characterized in that: The step S7 comprises: According to the relationship between the normal line of the camera array and the tangent line of the GPS motion trajectory and the array spacing value, the route trajectory of each camera relative to the GPS motion trajectory is calculated; The calculated GPS position information of each camera is added to the corresponding video frame.

10. The underwater scanning method based on the water-air fast scanner according to claim 2, characterized in that: The step S8 comprises: Use the SfM algorithm to extract feature points in the image and match them between consecutive images; Using GPS route trajectory information, estimate the camera position and orientation for each image; Reconstruct the 3D point cloud using the matched feature points and the camera’s positioning information; An underwater map is generated based on the three-dimensional point cloud information.

Images