Ship panoramic image information acquisition method and system

Abstract

The application discloses a kind of ship panoramic image information acquisition method and system, method includes: according to the size of ship size Determine the installation number and installation position of camera around ship;From the top of ship, the overhead image of ship when ship is parked in wharf is acquired, while the image around ship collected by each camera is acquired;The first projection change matrix of each ship around image to ship overhead image is calculated, and each ship around image is projected to get overhead projection according to the first projection change matrix Fusion weight between adjacent overhead projection is calculated according to the overlapping area between adjacent overhead projection, and all overhead projection is fused into panoramic image figure;The second projection change matrix of each ship around image to panoramic image figure is calculated, in the actual work of ship, based on the second projection change matrix, the image around ship collected by each camera is mapped to get real-time panoramic image figure.The application improves the efficiency and quality of panoramic image information acquisition.

Description

Technical Field

[0001] This invention belongs to the field of machine vision technology, specifically relating to a method and system for acquiring panoramic image information of ships. Background Technology

[0002] The ability of intelligent ships to perceive their surroundings is a core technology of intelligent navigation modules and a prerequisite for ensuring safe navigation. Commonly used intelligent ship perception technologies include radar, Automatic Identification System (AIS), Global Positioning System (GPS), and visual imaging systems. Current research on intelligent ship perception mainly focuses on distant water areas, while research on perception of the waters surrounding the ship in close proximity remains lacking. In open waters, radar, AIS, and visual tracking systems can quickly detect distant targets, supporting the decision-making of early warning systems and enabling timely risk avoidance. However, when navigating narrow waters or during berthing and unberthing, real-time dynamic perception of the ship's surroundings is necessary. Current methods rely on personnel around the ship to observe and broadcast the surrounding situation to the bridge via wireless communication. This method requires multiple people to collaborate, is inefficient, and involves uncertainty in information transmission between personnel, failing to accurately and effectively transmit information to the command system.

[0003] To overcome the shortcomings of manual lookout in transmitting information, technologies for acquiring image information of the waters surrounding ships at close range have emerged. Existing technology, such as Chinese patent document CN2020114281345, discloses an automatic panoramic image stitching system for ships. This system uses retractable cameras mounted around the ship, one at the fore and aft, and two on each side, for a total of six cameras. A dynamic stitching and fusion method is used to generate the panoramic image. The camera height is adjusted based on the stitching seams between adjacent cameras. The images are wirelessly transmitted to a computing platform. By stitching and fusing images in pairs, and then fusing them with the next camera, a complete panoramic image is finally stitched together using this progressive fusion method. However, this solution has the following drawbacks:

[0004] (1) The solution uses wireless transmission, which cannot continuously and efficiently transmit multiple video data; (2) The design uses 6 cameras, and fixed cameras cannot meet the needs of different ship application scenarios; (3) The solution automatically adjusts the camera height by stitching between two cameras, but this adjustment cannot guarantee that all camera heights will meet the ideal height when finally converged; (4) The stepwise fusion scheme cannot guarantee that all camera images will be mapped to a unified imaging plane at the same time, that is, the top-down imaging plane; (5) The entire calculation process is real-time dynamic calculation, and this method cannot guarantee the calculation speed when dealing with multiple cameras. Summary of the Invention

[0005] One of the objectives of this invention is to provide a method for acquiring panoramic image information of ships, thereby improving the efficiency and quality of panoramic image information acquisition.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for acquiring panoramic ship image information, the method comprising:

[0008] The number and location of cameras to be installed around the ship are determined based on the ship's size.

[0009] The system captures an overhead view of the ship docked at the pier from directly above, while also acquiring images of the ship's surroundings from various cameras.

[0010] Calculate the first projection transformation matrix from the images around each ship to the ship's top view image, and project the images around each ship onto the imaging plane where the ship's top view image is located according to the first projection transformation matrix to obtain the top view projection image;

[0011] The fusion weight between adjacent top-view projection images is calculated based on the overlapping area between them, and all top-view projection images are fused into a single panoramic image based on the fusion weight.

[0012] Calculate the second projection transformation matrix from the images around each ship to the panoramic image. In actual ship operation, the images around the ship captured by each camera are mapped based on the second projection transformation matrix to obtain a real-time panoramic image.

[0013] Several alternative methods are provided below, but they are not intended as additional limitations on the overall solution above. They are merely further additions or optimizations. Provided there are no technical or logical contradictions, each alternative method can be combined individually with respect to the overall solution above, or multiple alternative methods can be combined with each other.

[0014] Preferably, the camera is a fisheye camera.

[0015] Preferably, determining the number and location of cameras around the ship based on the ship's size includes:

[0016] The installation of cameras is started at the middle of the bow and ended at the middle of the stern. One camera is set at the starting position. Then, the installation positions of cameras are determined sequentially along one side of the ship according to the distribution spacing of the cameras until the ending position is reached. Finally, according to the principle of symmetrical installation of cameras, the installation positions of cameras on the other side of the ship are obtained, and the number and positions of cameras around the ship are finally determined.

[0017] Preferably, the camera distribution spacing d is calculated as follows:

[0018] d = h1tanβ + h2tanβ

[0019] In the formula, h1 and h2 are the installation heights of two adjacent cameras, and the effective field of view of the camera is 2β.

[0020] Preferably, the method of obtaining a top-down view of the ship when it is docked at the pier from directly above the ship includes:

[0021] When a ship is docked at the pier, a drone is used to hover directly above the ship. The drone's camera is vertically downward to obtain a top-down view of the ship. At this time, the effective field of view of the drone's camera includes the effective field of view of all cameras installed on the ship.

[0022] The drone performs two hovering data acquisitions. Based on the ship's top-down view image acquired during the first hovering, the ship's top-down view image acquired during the second hovering is aligned. The two aligned ship's top-down view images are then merged to obtain the final ship's top-down view image.

[0023] Preferably, the calculation of the first projection transformation matrix from the images around each ship to the ship's top-view image includes:

[0024] Based on the camera's calibration parameters, the image around the ship is projected onto a distortion-free projection plane to obtain a distortion-free image;

[0025] Using the ship's top-view image as a reference, select four or more valid matching points from each distortion-free image;

[0026] Calculate the first projection transformation matrix from each distortion-free image to the ship's top view image based on the valid matching points.

[0027] Preferably, the step of calculating the fusion weight between adjacent top-view projection images based on the overlapping area between them includes:

[0028] Take two adjacent top-view projections as I A and I B Among them, p1p2 are close-up top-view projection diagrams I. A The overlapping region boundaries, p3p4 are close to the top-view projection diagram I B Given the overlapping region boundaries, point p5 in the overlapping region is d1 from boundary p1p2 and d2 from boundary p3p4. Then, in the top-view projection diagram I... A and I B The fusion weights are:

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035] In the formula, w A Top-view projection diagram I A The fusion weight, w B Top-view projection diagram I B The fusion weight, w sA Top-view projection diagram I A Image sharpness weight, w sB Top-view projection diagram I B Image sharpness weight, w dA Top-view projection diagram I A Distance weight, w dB Top-view projection diagram I B Distance weights, s A Top-view projection diagram I A Image patch magnification, s B Top-view projection diagram I B The magnification of the image blocks.

[0036] As a preferred embodiment, the formula for calculating the pixel value of the fused image I at point p5 is:

[0037] I(p5)=w A I A (p5)+w B I B (p5)

[0038] In the formula, I(p5) is the pixel value of the fused image I at point p5. A (p5) is a top-view projection diagram, I. A The pixel value at point p5, I B (p5) is a top-view projection diagram I B The pixel value at point p5.

[0039] Preferably, the ship image is cropped from the ship's top-view image and then merged into the blank area of ​​the real-time panoramic image to obtain a panoramic image containing the ship itself.

[0040] This invention provides a method for acquiring panoramic ship image information. Instead of using fixed cameras, it selects appropriate cameras and installation locations based on the ship's size, focusing on the panoramic imaging effect. Furthermore, it eliminates real-time dynamic calculations by employing an image remapping method, thus avoiding complex intermediate calculations, improving computational efficiency, and ensuring real-time performance even with multi-channel video processing.

[0041] The second objective of this invention is to provide a system for acquiring panoramic ship image information, thereby improving the efficiency and quality of panoramic image information acquisition.

[0042] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0043] A ship panoramic image information acquisition system includes a processor and a memory storing a plurality of computer instructions, wherein the computer instructions, when executed by the processor, implement the steps of the ship panoramic image information acquisition method. Attached Figure Description

[0044] Figure 1 This is a flowchart of a method for acquiring panoramic image information of a ship according to the present invention;

[0045] Figure 2 This is a data transformation diagram of a method for acquiring panoramic ship image information according to the present invention;

[0046] Figure 3 (a) is a schematic diagram of the present invention viewed from a perspective parallel to the edge of the ship;

[0047] Figure 3 (b) is a schematic diagram of the present invention from the perspective of the ship's edge;

[0048] Figure 4 This is a schematic diagram of the fused image of the present invention;

[0049] Figure 5 This is a schematic diagram of the image scaling method for the mapping transformation of this invention. Detailed Implementation

[0050] 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.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0052] This invention uses multi-view image fusion technology to solve the perception problem of intelligent ships in close-range scenes. It provides accurate scene information for ships to avoid collisions and berth / unberthen in complex scenes, and can establish a close-range safety early warning system based on the obtained panoramic image information to ensure the navigation safety of intelligent ships.

[0053] like Figures 1-2 As shown, the method for acquiring panoramic ship image information in this embodiment includes the following steps:

[0054] Step 1: Determine the number and location of cameras to be installed around the ship based on the ship's size.

[0055] First, this embodiment determines the choice of camera: a fisheye camera is used as the image data source. Because fisheye cameras have a large field of view, compared to conventional non-fisheye cameras, fewer cameras are needed to obtain images from more perspectives. Furthermore, for subsequent installation and algorithm adjustments, the field of view range of the fisheye camera can be dynamically selected as the stitching area according to requirements.

[0056] Secondly, determine the number of cameras and their installation locations:

[0057] like Figure 3 As shown in (a) and 3(b), the camera installation angles are viewed from a perspective parallel to the ship's edge and from a perspective directly facing the ship's edge. The camera is installed as high as possible, assuming an installation height of h. From the perspective parallel to the ship's edge, the angle between the camera's principal axis and the direction of gravity is α. From the perspective directly facing the ship's edge, the camera's principal axis is parallel to the direction of gravity. Assume the effective field of view of the selected fisheye camera is 2β. Empirically, values ​​of α and β within the range of 45° to 60° are suitable. If the selected angle is too small, the image from each fisheye camera cannot be effectively utilized, resulting in too many cameras being used. If the selected angle is too large, the edge images will be stretched excessively during subsequent projection transformations, leading to unclear images in the edge areas.

[0058] In this embodiment, the middle position of the bow of the ship is taken as the starting position for the installation of the camera, and the middle position of the stern of the ship is taken as the ending position for the installation of the camera. One camera is set at the starting position, and then the installation positions of the cameras are determined sequentially on one side of the ship according to the distribution spacing of the cameras until the ending position is reached. Finally, the installation positions of the cameras on the other side of the ship are obtained according to the principle of symmetrical and uniform distribution of the cameras. Finally, the number and installation positions of the cameras around the ship are determined.

[0059] To accommodate the ship's hull, the installation height of different cameras may vary. Therefore, the camera distribution spacing d in this embodiment is calculated as follows:

[0060] d = h1tanβ + h2tanβ

[0061] In the formula, h1 and h2 are the installation heights of two adjacent cameras, and the effective field of view of the camera is 2β.

[0062] Because of the diverse shapes of ships, the installation approach of this application can be adapted to different ship shapes to develop other installation methods. For example, if the ship is pointed at both the front and rear and transitions to the sides with gentle arcs, the method provided in this embodiment is preferred. If the stern of the ship tends to be flat and has a right angle of nearly 90° with the sides of the ship, the cameras can be determined from the starting position according to the camera distribution spacing until the end of the side of the ship. Then, the installation position of the stern camera is determined according to the width of the stern and the camera distribution spacing.

[0063] It is easy to understand that after determining the position of the camera based on the camera distribution spacing in this embodiment, if there is a monitoring omission area between two adjacent cameras, the camera position needs to be fine-tuned to eliminate the monitoring omission area.

[0064] Step 2: Obtain an overhead view of the ship when it is docked at the pier from directly above, and simultaneously obtain images of the surrounding area of ​​the ship captured by each camera.

[0065] This embodiment uses a drone to acquire an overhead view of the ship. The ship is docked at a pier, which must be a flat, non-convex surface. The drone hovers directly above the ship with its camera pointing vertically downwards to acquire current drone image data (i.e., the ship's overhead view). At this point, the effective field of view of the ship's cameras must be within the drone's field of view. Simultaneously, image data from the ship's cameras (i.e., images of the area surrounding the ship) is also acquired.

[0066] Considering that the ship is docked at the pier, there is only a marked land area on one side and water on the other, two hovering data acquisition experiments are required. To eliminate the deviation between the two drone hovering operations and the ship, the data acquired in the second experiment is processed using the first experiment as the standard to ensure that the ship's position in the image remains unchanged. Specifically, the drone performs two hovering data acquisitions. Based on the ship's top-down view image acquired in the first hovering operation, the ship's top-down view image acquired in the second hovering operation is mapped and aligned. The two aligned ship's top-down view images are then merged to obtain the final ship's top-down view image.

[0067] Step 3: Calculate the first projection transformation matrix from the images around each ship to the ship's top view image, and project the images around each ship onto the imaging plane where the ship's top view image is located according to the first projection transformation matrix to obtain the top view projection image.

[0068] First, the fisheye camera is calibrated using a conventional algorithm (such as the checkerboard calibration method). Then, based on the camera calibration parameters and the selected effective field of view, the image of the ship's surroundings acquired by the fisheye camera is projected onto a distortion-free projection plane to obtain a distortion-free image.

[0069] Using a top-down view of the ship captured by a drone as a reference, at least four valid matching points are selected from each distortion-free image. Based on these valid matching points, a projection transformation matrix is ​​calculated from each camera's distortion-free image to the ship's top-down view. According to the projection transformation matrix, the distortion-free image is projected onto the drone's imaging plane to obtain a top-down projection image. This embodiment uses the drone as a reference to project the camera's distortion-free images onto an imaging plane with a uniform top-down perspective, ensuring the consistency of image transformation.

[0070] In this embodiment, the projection transformation matrix is ​​preferably a homography matrix. The calculation of the homography matrix can be based on existing logic, which will not be elaborated in this embodiment.

[0071] Step 4: Calculate the fusion weight between adjacent top-view projection maps based on the overlapping area between them, and fuse all top-view projection maps into a single panoramic image based on the fusion weight.

[0072] Based on the top-view projection, the overlapping areas between adjacent images are determined, and a fusion algorithm is used to calculate the fusion weight of the overlapping areas. Finally, the weights are used to fuse all the top-view projections into a complete panoramic image.

[0073] In the image stitching process, to eliminate stitching gaps, this embodiment uses a weighted fusion method to fuse overlapping areas. The fusion weight value is determined based on the distance between the pixel position within the overlapping area and the overlapping boundary. The pixel values ​​of the fusion area are added together according to their respective weight values ​​to form the fused image, thereby achieving a smooth transition effect.

[0074] like Figure 4 As shown, in the two fused images I A and I B In the diagram, A and B are non-overlapping regions, and C is the overlapping region between the two images. p1p2 represents the region closest to image I. A The overlapping region boundary, p3p4 is closer to image I B The overlapping region boundary. The distance between point p5 and boundary p1p2 in the overlapping region is d1, and the distance between p5 and boundary p3p4 is d2. Then, the formula for calculating the pixel value of the fused image I at point p5 is:

[0075] I(p5)=w dA I A (p5)+w dB I B (p5)

[0076]

[0077]

[0078] In the formula, I(p5) is the pixel value of the fused image I at point p5. A (p5) is a top-view projection diagram, I. A The pixel value at point p5, I B (p5) is a top-view projection diagram I B The pixel value at point p5, w dA Top-view projection diagram I A Distance weight, w dB Top-view projection diagram I B Distance weights.

[0079] As can be seen from the formula, the closer the fusion point p5 is to image I... A On one side, the fused image I A The larger the weight of a given value, the smaller its weight; conversely, the smaller its weight. During the process of mapping images of the area around a ship captured by a camera into a top-down projection image after multiple transformations, the scaling of the images around different ships varies, resulting in inconsistent clarity in different areas of the fused image. For example... Figure 5 As shown, after the image patch is reduced in size, it still retains good clarity. However, when the image patch is enlarged, it becomes blurry.

[0080] Therefore, in the overlapping area of ​​two fused images, one image may have lower sharpness due to magnification, while the other image may have higher sharpness. This embodiment aims to ensure that the fusion process uses the image with the highest possible sharpness, reducing the impact of blurred images on the generated image and thus improving the imaging quality of the fused image. Therefore, an image sharpness weight w is introduced. s This weight is determined based on the magnification factor *s* of the image patch. Image sharpness weight *w* s The calculation formula is:

[0081]

[0082] Final fused image I A and I B In this context, two image sharpness weights w are introduced. sA and w sB Therefore, the calculation of the fusion weight in the improved algorithm of this embodiment is as follows:

[0083]

[0084]

[0085]

[0086]

[0087] In the formula, w A Top-view projection diagram I A The fusion weight, w B Top-view projection diagram I B The fusion weight, w sA Top-view projection diagram I A Image sharpness weight, w sB Top-view projection diagram I B Image sharpness weight, s A Top-view projection diagram I A Image patch magnification, s B Top-view projection diagram I B The magnification of the image blocks.

[0088] In this embodiment, the magnification factor of the image block can be calculated according to existing technology. For example, a number can be assigned to each pixel in the original image, and then statistics can be performed in the transformed new image to calculate the number of different values ​​in the image block, or the magnification factor can be calculated based on the number of pixels in the image before and after the transformation.

[0089] Therefore, the formula for calculating the pixel value of the fused image I at point p5 is:

[0090] I(p5)=w A I A (p5)+w B I B (p5)

[0091] The image fusion algorithm proposed in this embodiment considers both fusion boundary information and image sharpness information, effectively improving the quality of the fused image.

[0092] Step 5: Calculate the second projection transformation matrix from the images around each ship to the panoramic image. In the actual operation of the ship, the images around the ship collected by each camera are mapped based on the second projection transformation matrix to obtain a real-time panoramic image.

[0093] Based on steps 1-4, a projection map of the x and y coordinates of the image pixels from each fisheye camera's original image (i.e., the image around the ship) to the final complete panoramic image is constructed. For ease of processing, the homography matrix can still be selected for the projection map of the x and y coordinates here.

[0094] A blank area will appear in the middle of the merged panoramic image, and the shape of this area will resemble a ship. For aesthetic purposes, to create an effect similar to a real-time aerial view taken by a drone, the ship image is cropped from the aerial view of the ship and then merged into the blank area of ​​the real-time panoramic image to obtain a panoramic image that includes the ship itself.

[0095] In this embodiment, the second projection transformation matrix from the images around each ship to the panoramic image was obtained in the preliminary calibration work. Therefore, in actual work, the second projection transformation matrix is ​​directly used to map the images around the ship acquired by the fisheye camera onto the fused imaging plane for fusion. Then, the cropped ship images are fused into the middle of the generated panoramic image, and finally the required panoramic view image of the ship is obtained in real time, forming a panoramic image similar to the overhead view of a drone. This helps the ship's driver to have a comprehensive understanding of the situation around the ship and improves the navigation safety of the ship in complex environments.

[0096] This embodiment first calibrates the pixel mapping relationship between each fisheye camera and the panoramic view. In deployment, the panoramic view is directly generated through this mapping relationship, eliminating intermediate calculations, significantly accelerating computation, and improving the real-time performance of the algorithm.

[0097] In another embodiment, this application also provides a ship panoramic image information acquisition system, including a processor and a memory storing a plurality of computer instructions, wherein the computer instructions, when executed by the processor, implement the steps of the ship panoramic image information acquisition method.

[0098] For specific limitations regarding the ship panoramic image information acquisition system, please refer to the limitations on the ship panoramic image information acquisition method mentioned above, which will not be repeated here.

[0099] The memory and processor are electrically connected directly or indirectly to enable data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines. The memory stores a computer program that can run on the processor, which implements the method in the embodiments of the present invention by running the computer program stored in the memory.

[0100] The memory may be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory stores the program, and the processor executes the program upon receiving an execution instruction.

[0101] The processor may be an integrated circuit chip with data processing capabilities. The aforementioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0102] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0103] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0104] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A method for acquiring panoramic image information of a ship, characterized in that, The method for acquiring panoramic ship image information includes: The number and location of cameras to be installed around the ship are determined based on the ship's size. The system captures an overhead view of the ship docked at the pier from directly above, while also acquiring images of the ship's surroundings from various cameras. Calculate the first projection transformation matrix from the images around each ship to the ship's top view image, and project the images around each ship onto the imaging plane where the ship's top view image is located according to the first projection transformation matrix to obtain the top view projection image; The fusion weight between adjacent top-view projection images is calculated based on the overlapping area between them, and all top-view projection images are fused into a single panoramic image based on the fusion weight. Calculate the second projection transformation matrix from the images around each ship to the panoramic image. In the actual operation of the ship, the images around the ship collected by each camera are mapped based on the second projection transformation matrix to obtain the real-time panoramic image. The step of determining the number and location of cameras around the ship based on the ship's size includes: The installation starts at the center of the bow of the ship and ends at the center of the stern. One camera is installed at the starting position. Then, the installation positions of cameras are sequentially determined along one side of the ship, following the camera distribution spacing, until the ending position is reached. Finally, based on the principle of symmetrical camera installation, the installation positions of cameras on the other side of the ship are determined, thus finalizing the number and locations of cameras around the ship. The camera distribution spacing... The calculation is as follows: In the formula, and These represent the installation heights of two adjacent cameras, and the effective field of view of the cameras is... .

2. The method for acquiring panoramic ship image information as described in claim 1, characterized in that, The camera is a fisheye camera.

3. The method for acquiring panoramic ship image information as described in claim 1, characterized in that, The method of obtaining a top-down view of the ship when it is docked at the pier from directly above includes: When a ship is docked at the pier, a drone is used to hover directly above the ship. The drone's camera is vertically downward to obtain a top-down view of the ship. At this time, the effective field of view of the drone's camera includes the effective field of view of all cameras installed on the ship. The drone performs two hovering data acquisitions. Based on the ship's top-down view image acquired during the first hovering, the ship's top-down view image acquired during the second hovering is aligned. The two aligned ship's top-down view images are then merged to obtain the final ship's top-down view image.

4. The method for acquiring panoramic ship image information as described in claim 1, characterized in that, The calculation of the first projection transformation matrix from the images around each ship to the ship's top-view image includes: Based on the camera's calibration parameters, the image around the ship is projected onto a distortion-free projection plane to obtain a distortion-free image; Using the ship's top-view image as a reference, select four or more valid matching points from each distortion-free image; Calculate the first projection transformation matrix from each distortion-free image to the ship's top view image based on the valid matching points.

5. The method for acquiring panoramic ship image information as described in claim 1, characterized in that, The step of calculating the fusion weight between adjacent top-view projection maps based on the overlapping area between them includes: Take two adjacent top-view projections as and ,in, For close-up top-view projection The boundary of the overlapping region, For close-up top-view projection The boundary of the overlapping region, and points within the overlapping region. With boundary The distance is , and the boundary The distance is Then, the top-view projection diagram and The fusion weights are: In the formula, Top-down projection The fusion weight, Top-down projection The fusion weight, Top-down projection Image sharpness weights, Top-down projection Image sharpness weights, Top-down projection Distance weights, Top-down projection Distance weights, Top-down projection Image patch magnification, Top-down projection The magnification of the image blocks.

6. The method for acquiring panoramic ship image information as described in claim 5, characterized in that, Fusion Image At point The formula for calculating the pixel value at a given location is: In the formula, For the merged image At point Pixel value at that location, The top-down projection diagram is as follows At point Pixel value at that location, Top-down projection At point The pixel value at that location.

7. The method for acquiring panoramic ship image information as described in claim 1, characterized in that, The ship image is cropped from the ship's top-view image and then merged into the blank area of ​​the real-time panoramic image to obtain a panoramic image containing the ship itself.

8. A system for acquiring panoramic ship image information, comprising a processor and a memory storing a plurality of computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 7.

Images