A UAV-based non-linear image stitching method, device and terminal equipment

Abstract

The present invention discloses a non-linear image stitching method, device and terminal device based on an unmanned aerial vehicle. For a large number of images with complex association relationships, a tree structure is constructed for image stitching, and then the mapping matrix between images is calculated, and the mapping matrix of the first image and the matrix of each image are calculated twice to determine the size of the image outer frame, so as to create a blank bottom with the determined image outer frame, and finally all the pixels of the images are traversed and drawn on the blank bottom according to the tree structure, thereby obtaining the total image. Compared with the prior art of performing one-dimensional continuous stitching along the shooting path, the present invention forms a two-dimensional tree structure from the one-dimensional relationships between multiple images, and can project and stitch together the images with complex stitching relationships onto a plane well, improving the stitching efficiency of a large number of disordered images, and can be widely applied to the field of stitching of unmanned aerial vehicle aerial images.

Description

Technical Field

[0001] The present invention relates to the field of UAV aerial photography, and particularly to a non-linear image stitching method, device and terminal device based on a UAV. Background Art

[0002] The full name of a UAV is "unmanned aerial vehicle" (UAV). By using radio remote control equipment and a designed program to control the UAV for flight operations, it has the advantages of small size, low cost, and flexible movement. With the continuous development of UAV technology, panoramic shooting of scenes by UAVs has become a conventional technical means in professional photography technology teams. By mounting corresponding photographic equipment on the UAV body, the combination of UAV technology and photography technology is realized, enabling the shooting of high-difficulty pictures, such as aerial pictures of cities. The application of UAVs has provided great help for shooting diversified and high-quality photos, and UAV photography technology has been widely applied in natural disaster prevention, natural resource measurement, urban planning, etc., and has been recognized by the obtained results. The technology of UAVs has been gradually improved, and the photos obtained by on-board cameras are of higher quality. For large-scale shooting operations, UAV photography technology is usually combined with post-stitching technology to synthesize large-scale images. The common image stitching technology is to obtain images through a shooting mode of determining coordinates and grid fixed points. After shooting is completed at a predetermined position in a predetermined orientation, image stitching is performed, and a large-scale image can be obtained after simple data correction. However, this requires high requirements for the pre-shooting position and shooting orientation, and cannot handle the situation of stitching messy pictures.

[0003] For the non-linear image stitching method based on a UAV, the prior art: "Video image adaptive stitching method and device based on video frame matching information" proposed by Shen Wei and Li Ruicheng (CN201810876608.9). This technology can perform one-dimensional continuous stitching along the shooting path for any image obtained from an aerial video through a free linear continuous stitching method, and aerial photography can be carried out without pre-setting position information and shooting order, reducing the pre-preparation work.

[0004] The inventor of the present invention found the following technical problems in the process of implementing the present invention: The "Video image adaptive stitching method and device based on video frame matching information" can only perform one-dimensional image stitching along the shooting direction of the video, and cannot handle the image stitching problems with some complex linear / non-linear stitching relationships. The prior art cannot combine multiple complex stitching relationships into the same stitching relationship. Therefore, in this case, the stitched images obtained by the prior art may have problems such as slow stitching and stitching errors. Summary of the Invention

[0005] An embodiment of the present invention provides a non - linear image stitching method, device and terminal device based on an unmanned aerial vehicle (UAV), which realizes the stitching of multiple images with complex association relationships by constructing a tree structure.

[0006] To solve the above problems, an embodiment of the present invention provides a non - linear image stitching method based on an unmanned aerial vehicle, including:

[0007] Obtain the captured images, and according to the image information of the captured images, obtain the stitching order of the captured images;

[0008] Obtain a tree structure according to the stitching order, and obtain child - node images, parent - node images and root - node images;

[0009] Calculate the image mapping matrix from all child - node images to all parent - node images according to the tree structure;

[0010] Set the root - node image as a zero mapping matrix, calculate the matrices of all child - node images according to the image mapping matrix, and obtain the first total image frame; perform a secondary operation on the mapping matrix of the root - node image and the matrices of the child - node images according to the first total image frame, so as to obtain the second total image frame;

[0011] Create a blank bottom according to the second total image frame, traverse the pixels of the captured images respectively according to the image sorting of the tree structure, and draw them on the blank bottom to obtain a total image; wherein, the pixels are obtained from the image information.

[0012] As an improvement of the above solution, the step of traversing the pixels of the captured images respectively according to the image sorting of the tree structure and drawing them on the blank bottom to obtain a total image is specifically:

[0013] Obtain the pixel and position information of each captured image according to the tree structure, and draw them on the blank bottom to obtain a preliminary total image;

[0014] Identify the pixels in the overlapping area of the preliminary total image, calculate the mixing ratio by using the sin function for the distance from the pixels to each edge of the area, and complete color mixing.

[0015] As an improvement of the above solution, the step of obtaining a tree structure according to the stitching order, and obtaining child - node images, parent - node images and root - node images is specifically:

[0016] In the tree structure obtained from the stitching order, there is a connection relationship between each captured image; wherein, each captured image is a child - node image, the previous image of each captured image is a parent - node image, and the captured image without a previous image is a root - node image.

[0017] As an improvement to the above solution, set the root node image as a zero mapping matrix, calculate the matrices of all child node images according to the image mapping matrix, and obtain the outer frame of the first total image, specifically:

[0018] After setting the root node image as a zero mapping matrix, according to the image mapping matrix, right-multiply the matrix of each child node image by the matrix of the parent node image to calculate the matrix of each child node;

[0019] Calculate the four corner vertex data of the image according to the matrix of each child node, so as to obtain the area and origin position of the outer frame of the first total image.

[0020] As an improvement to the above solution, perform a secondary operation on the mapping matrix of the root node image and the matrices of the child node images according to the outer frame of the first total image, so as to obtain the outer frame of the second total image, specifically:

[0021] According to the origin position of the outer frame of the second total image, recalculate the mapping matrix of the root node image, right-multiply the matrix of each child node image by the matrix of the parent node image again, and recalculate the matrix of each child node;

[0022] Calculate the four corner vertex data of the image according to the matrix of each recalculated child node, so as to obtain the area and origin position of the outer frame of the second total image.

[0023] As an improvement to the above solution, calculate the image mapping matrix from all child node images to all parent node images according to the tree structure, specifically:

[0024] Use the AKAZE feature point extraction method and the RANSAC feature point matching method to calculate the image mapping matrix from all child node images to all parent node images.

[0025] Correspondingly, an embodiment of the present invention provides a non-linear image stitching device based on an unmanned aerial vehicle, including: an extraction module, a construction module, a calculation module, a framing module, and a stitching module;

[0026] Among them, the extraction module is used to obtain the captured image and obtain the stitching order of the captured image according to the image information of the captured image;

[0027] The construction module is used to obtain a tree structure according to the stitching order, and obtain child node images, parent node images, and root node images;

[0028] The calculation module is used to calculate the image mapping matrix from all child node images to all parent node images according to the tree structure;

[0029] The fixed frame module is used to set the root node image as a zero mapping matrix, calculate the matrices of all child node images according to the image mapping matrix, and obtain the first total image outer frame; perform a secondary operation on the mapping matrix of the root node image and the matrices of the child node images according to the first total image outer frame, so as to obtain the second total image outer frame.

[0030] The splicing module is used to create a blank bottom according to the image outer frame, traverse the pixels of the captured images respectively according to the image sorting of the tree structure, and draw them on the blank bottom to obtain the total image; wherein, the pixels are obtained from the image information.

[0031] As an improvement of the above solution, the construction module is used to obtain the tree structure according to the splicing order, and obtain the child node images, parent node images and root node images, specifically:

[0032] In the tree structure obtained from the splicing order, each of the captured images is connected; wherein, each of the captured images is a child node image, the previous image of each of the captured images is a parent node image, and the captured image without a previous image is a root node image.

[0033] Correspondingly, an embodiment of the present invention provides a computer terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the non-linear image splicing method based on an unmanned aerial vehicle as described in the present invention.

[0034] Correspondingly, an embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium includes a stored computer program. Wherein, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the non-linear image splicing method based on an unmanned aerial vehicle as described in the present invention.

[0035] The embodiments of the present invention have the following beneficial effects:

[0036] The present invention discloses a non-linear image stitching method, device and terminal device based on an unmanned aerial vehicle (UAV). For a large number of images with complex correlation relationships, a tree structure is constructed for image stitching, and then the mapping matrix between images is calculated, and the mapping matrix of the first image and the matrix of each image are calculated again to determine the size of the image outer frame, so as to create a blank bottom with the determined image outer frame, and finally all the pixels of the images are traversed and drawn on the blank bottom according to the tree structure to obtain the total image. Compared with the prior art that performs one-dimensional continuous stitching along the shooting path, the present invention forms a two-dimensional tree structure from the one-dimensional relationships between multiple images, which can project and stitch images with complex stitching relationships onto a plane well, improve the stitching efficiency of a large number of disordered images, and can be widely applied to the field of stitching of UAV aerial images.

[0037] Further, the present invention constructs a tree structure through the connection relationships of each image, and projects each child node image into the coordinate system of the parent node image by right-multiplying the matrix of the child node image by the matrix of the parent node image, so as to project all the images with complex connection relationships onto the same projection plane. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a schematic flowchart of an embodiment of the non-linear image stitching method based on a UAV provided by the present invention;

[0039] Figure 2 is a schematic structural diagram of an embodiment of the non-linear image stitching device based on a UAV provided by the present invention;

[0040] Figure 3 is a schematic flowchart of another embodiment of the non-linear image stitching method based on a UAV provided by the present invention;

[0041] Figure 4 is a schematic diagram of the result of an embodiment of the non-linear image stitching method based on a UAV provided by the present invention;

[0042] Figure 5 is a schematic structural diagram of an embodiment of the terminal device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0044] See Figure 1 ,Figure 1 It is a schematic flowchart of an embodiment of the non - linear image stitching method based on an unmanned aerial vehicle provided by the present invention. As Figure 1 shown, the method includes steps 101 to 105, and the specific steps are as follows:

[0045] Step 101: Obtain the captured images, and according to the image information of the captured images, obtain the stitching order of the captured images.

[0046] In this embodiment, step 101 is specifically: Extract the EXIF information of the obtained captured images, and store them in the planar regions divided according to the coordinate information. At the same time, determine the target regions of each captured image; for each captured image, within the target region of the selected captured image, iteratively calculate the similarity and position information between the selected captured image and other captured images, and obtain the best - stitched images of each image in eight directions. Finally, remove the duplicates of the repeatedly stitched images, so that only one copy of each stitched image is retained, and generate the stitching order of the captured images.

[0047] Step 102: Obtain a tree - like structure according to the stitching order, and obtain the child - node images, parent - node images, and root - node images.

[0048] In this embodiment, step 102 is specifically: In the tree - like structure obtained from the stitching order, there is a connection relationship between each captured image; among them, each captured image is a child - node image, the previous image of each captured image is the parent - node image, and the captured image without a previous image is the root - node image.

[0049] Step 103: Calculate the image mapping matrix from all child - node images to all parent - node images according to the tree - like structure.

[0050] In this embodiment, step 103 is specifically: Use the AKAZE feature - point extraction method and the RANSAC feature - point matching method to calculate the image mapping matrix from all child - node images to all parent - node images.

[0051] Step 104: Set the root - node image as a zero mapping matrix, calculate the matrices of all child - node images according to the image mapping matrix, and obtain the first total image frame; perform a secondary operation on the mapping matrix of the root - node image and the matrices of the child - node images according to the first total image frame, so as to obtain the second total image frame.

[0052] In this embodiment, step 104 is specifically: After setting the root - node image as a zero mapping matrix, according to the image mapping matrix, right - multiply the matrix of each child - node image by the matrix of the parent - node image to calculate the matrix of each child - node.

[0053] Calculate the image four-corner vertex data based on the matrix of each child node, so as to obtain the area and origin position of the first total image outer frame.

[0054] According to the origin position of the second total image outer frame, recalculate the mapping matrix of the root node image, multiply the matrix of each child node image on the right by the matrix of the parent node image again, and recalculate the matrix of each child node.

[0055] Calculate the image four-corner vertex data based on the matrix of each child node obtained by recalculation, so as to obtain the area and origin position of the second total image outer frame.

[0056] Preferably, select the first image as the root node image, which serves as the zero mapping matrix and also the reference plane.

[0057] Preferably, calculate the offset according to the origin position obtained from the first total image outer frame, recalculate the mapping matrix of the root node through the offset, and perform the operation of multiplying the matrix of the child node image on the right by the matrix of the parent node image again, calculate the image four corners vertices, so as to obtain the final second total image outer frame.

[0058] Step 105: Create a blank bottom according to the second total image outer frame, traverse the pixels of the captured images respectively according to the image sorting of the tree structure, and draw them on the blank bottom to obtain the total image; wherein, the pixels are obtained from the image information.

[0059] In this embodiment, step 105 is specifically: obtain the pixel and position information of each captured image according to the tree structure, and draw them on the blank bottom to obtain a preliminary total image.

[0060] Identify the pixels in the overlapping area of the preliminary total image, calculate the mixing ratio by using the sin function for the distances from the pixels to the edges of the area, and complete the color mixing.

[0061] See Figure 2 , Figure 2 is a schematic structural diagram of an embodiment of the non-linear image stitching device based on a drone provided by the present invention. As Figure 2 shown, the device includes: an extraction module 201, a construction module 202, a calculation module 203, a framing module 204, and a stitching module 205.

[0062] Among them, the extraction module 201 is used to obtain the captured images and obtain the stitching order of the captured images according to the image information of the captured images.

[0063] The construction module 202 is used to obtain a tree structure according to the stitching order, and obtain child node images, parent node images, and root node images.

[0064] The calculation module 203 is configured to calculate the image mapping matrix from all child node images to all parent node images according to the tree structure.

[0065] The bounding box module 204 is configured to set the root node image as a zero mapping matrix, calculate the matrices of all child node images according to the image mapping matrix, and obtain the first total image bounding box; perform a secondary operation on the mapping matrix of the root node image and the matrices of the child node images according to the first total image bounding box, so as to obtain the second total image bounding box.

[0066] The stitching module 205 is configured to create a blank bottom according to the image bounding box, traverse the pixels of the captured images respectively according to the image sorting of the tree structure, and draw them on the blank bottom to obtain the total image; wherein, the pixels are obtained from the image information.

[0067] Preferably, the construction module 202 is specifically: in the tree structure obtained from the stitching order, each of the captured images has a connection relationship; wherein, each of the captured images is a child node image, the previous image of each captured image is a parent node image, and the captured image without a previous image is a root node image.

[0068] See Figure 3 , Figure 3 is a schematic flowchart of another embodiment of the non-linear image stitching method based on an unmanned aerial vehicle provided by the present invention. As Figure 3 shown, the method includes steps 301 to 304, and the specific steps are as follows:

[0069] In this embodiment, step 301: Preprocessing process, input the tree stitching relationship of all captured images, construct a tree structure according to the stitching order to obtain child node images, parent node images, and root node images, and calculate the image mapping matrix from all child node images to parent node images; wherein the previous image of the child node image is the parent node image, and the image without a previous image is the root node image.

[0070] Step 302: First projection relationship generation, set the root node image as a zero mapping matrix, right-multiply the matrix of each child node image by the matrix of the parent node image to update the matrices of all child node images, and finally use the linear characteristics of projective transformation to calculate the total size of the image bounding box using the four corner vertices of the image.

[0071] Step 303: Second projection relationship generation, recalculate the mapping matrix of the root node through the total size, right-multiply the matrix of each child node image by the matrix of the parent node image again to update the matrices of all child node images, and again use the linear characteristics of projective transformation to calculate the final total size of the image bounding box using the four corner vertices of the image.

[0072] Step 304: Drawing process. Create a blank map according to the total size of the final image outer frame, traverse all the images in a tree structure and draw them on the blank base map to finally obtain the stitched total image.

[0073] See Figure 4 , Figure 4 which is a schematic diagram of the result of an embodiment of the non - linear image stitching method based on an unmanned aerial vehicle provided by the present invention. As Figure 4 shown, the black quadrilateral frame is the stitched image after mapping. Each origin point is the center point of an image, the thick black dot is the root node of the stitching tree (which is also the center point of the reference image of the projection plane), the thin black dots are each sub - node of the stitching tree, and the lines between the points are the tree - like relationships between the sub - nodes.

[0074] See Figure 5 , Figure 5 which is a schematic diagram of the structure of a terminal device provided by an embodiment of the present invention.

[0075] A terminal device in this embodiment includes: a processor 501, a memory 502, and a computer program stored in the memory 502 and executable on the processor 501. When the processor 501 executes the computer program, it implements the steps in the above - mentioned embodiments of the non - linear image stitching method based on an unmanned aerial vehicle, such as Figure 1 all the steps of the non - linear image stitching method based on an unmanned aerial vehicle shown. Or, when the processor executes the computer program, it implements the functions of each module in the above - mentioned device embodiments, such as: Figure 2 all the modules of a non - linear image stitching device based on an unmanned aerial vehicle shown.

[0076] In addition, an embodiment of the present invention also provides a computer - readable storage medium. The computer - readable storage medium includes a stored computer program. When the computer program runs, it controls the device where the computer - readable storage medium is located to execute the non - linear image stitching method based on an unmanned aerial vehicle described in any of the above embodiments.

[0077] Those skilled in the art can understand that the schematic diagram is only an example of the terminal device, and does not constitute a limitation on the terminal device. It may include more or fewer components than shown, or combine some components, or different components. For example, the terminal device may further include input / output devices, network access devices, buses, etc.

[0078] The so-called processor 501 may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor 501 is the control center of the terminal device, connecting various parts of the entire terminal device through various interfaces and lines.

[0079] The memory 502 can be used to store the computer programs and / or modules. The processor 501 realizes various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory 502. The memory 502 may mainly include a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0080] Among them, if the modules / units integrated in the terminal device are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0081] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationships between the modules indicate that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0082] As can be seen from the above, the present invention provides a non-linear image stitching method, device, and terminal device based on an unmanned aerial vehicle. This method constructs a tree-like bifurcation structure for the connection relationships between images, projects images with complex stitching relationships onto the same projection plane and stitches them together, realizing the transformation from one-dimensional stitching to two-dimensional stitching. By setting the root node image as the reference plane for the entire stitching process, and calculating the projection transformation relationship and the occupied spatial position of each child node image on the plane one by one along the trunk of the tree-like structure, the present invention completes the remapping of all images to the reference plane, and well solves the problem of image stitching with complex association relationships.

[0083] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A non - linear image stitching method based on an unmanned aerial vehicle, characterized in that, Including: Obtain a captured image, and according to the image information of the captured image, obtain the splicing order of the captured images; Obtain a tree structure according to the splicing order, and obtain child node images, parent node images, and a root node image; Calculate the image mapping matrix from all child node images to all parent node images according to the tree structure; Set the root node image as a zero mapping matrix, calculate the matrices of all child node images according to the image mapping matrix, and obtain the first total image outer frame; Perform a secondary operation on the mapping matrix of the root node image and the matrices of the child node images according to the first total image outer frame, so as to obtain the second total image outer frame; Create a blank bottom according to the second total image outer frame, traverse the pixels of the captured images respectively according to the image sorting of the tree structure, and draw them on the blank bottom to obtain a total image; wherein, the pixels are obtained from the image information.

2. The non - linear image stitching method based on an unmanned aerial vehicle according to claim 1, characterized in that, The step of traversing the pixels of the captured images respectively according to the image sorting of the tree structure and drawing them on the blank bottom to obtain a total image is specifically: Obtain the pixels and position information of each captured image according to the tree structure, and draw them on the blank bottom to obtain a preliminary total image; Identify the pixels in the overlapping area of the preliminary total image, calculate the mixing ratio by using the sin function for the distance from the pixels to the edges of the area, and complete color mixing.

3. The non - linear image stitching method based on an unmanned aerial vehicle according to claim 1, characterized in that, The step of obtaining a tree structure according to the splicing order, and obtaining child node images, parent node images, and a root node image is specifically: In the tree structure obtained from the splicing order, there is a connection relationship between each captured image; wherein, each captured image is a child node image, the previous image of each captured image is a parent node image, and the captured image without a previous image is a root node image.

4. The non - linear image stitching method based on an unmanned aerial vehicle according to claim 1, characterized in that, The step of setting the root node image as a zero mapping matrix, calculating the matrices of all child node images according to the image mapping matrix, and obtaining the first total image outer frame is specifically: After setting the root node image as a zero mapping matrix, according to the image mapping matrix, right-multiply the matrix of each child node image by the matrix of the parent node image to calculate the matrix of each child node; Calculate the four corner vertex data of the image according to the matrix of each child node, so as to obtain the area and origin position of the first total image outer frame.

5. The non - linear image stitching method based on an unmanned aerial vehicle according to claim 1, characterized in that, The step of performing a secondary operation on the mapping matrix of the root node image and the matrices of the child node images according to the first total image outer frame, so as to obtain the second total image outer frame is specifically: According to the origin position of the second total image outer frame, recalculate the mapping matrix of the root node image, right-multiply the matrix of each child node image by the matrix of the parent node image again, and recalculate the matrix of each child node; Calculate the four corner vertex data of the image according to the recalculated matrix of each child node, so as to obtain the area and origin position of the second total image outer frame.

6. The non - linear image stitching method based on an unmanned aerial vehicle according to claim 1, characterized in that, The step of calculating the image mapping matrix from all child node images to all parent node images according to the tree structure is specifically: The image mapping matrix from all child node images to all parent node images is calculated using the AKAZE feature point extraction method and the RANSAC feature point matching method.

7. A non - linear image stitching device based on an unmanned aerial vehicle, characterized in that, It includes: an extraction module, a construction module, a calculation module, a framing module, and a stitching module; Among them, the extraction module is used to obtain the captured image and, according to the image information of the captured image, obtain the stitching order of the captured images; The construction module is used to obtain a tree structure according to the stitching order, and obtain child node images, parent node images, and a root node image; The calculation module is used to calculate the image mapping matrix from all child node images to all parent node images according to the tree structure; The framing module is used to set the root node image as a zero mapping matrix, calculate the matrices of all child node images according to the image mapping matrix, and obtain the first total image outer frame; perform a secondary operation on the mapping matrix of the root node image and the matrices of the child node images according to the first total image outer frame, so as to obtain the second total image outer frame; The stitching module is used to create a blank bottom according to the image outer frame, traverse the pixels of the captured images respectively according to the image sorting of the tree structure, and draw them on the blank bottom to obtain a total image; wherein, the pixels are obtained from the image information.

8. The non - linear image stitching device based on an unmanned aerial vehicle according to claim 7, characterized in that, The construction module is used to obtain a tree structure according to the stitching order, and obtain child node images, parent node images, and a root node image. Specifically: In the tree structure obtained from the stitching order, each of the captured images is connected; among them, each of the captured images is a child node image, the previous image of each of the captured images is a parent node image, and the captured image without a previous image is a root node image.

9. A computer terminal device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the drone-based non-linear image stitching method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the drone-based non-linear image stitching method according to any one of claims 1 to 6.

Images