Rapid splicing method for orthoimages of unmanned aerial vehicle

Abstract

The invention discloses a rapid splicing method for an orthoimage of an unmanned aerial vehicle. The method comprises the following steps: S1, importing an aerial image of an unmanned aerial vehicle image into a computer for processing; s2, acquiring image XMP information; s3, carrying out image affine transformation, and obtaining the maximum image size Max (Ix) and Max (Iy) through global traversal; s4, calculating offset and outputting an offset set W; s5, calculating the boundary of the image, and adopting a screen coordinate system to take Max (Dx) + Max (Ix) and Max (Dy) + Max (Iy) as the coordinates of the lower right corner of the boundary; and S6, according to the offset set W, obtaining a spliced image file. According to the method, based on an orthoimage shot by an unmanned aerial vehicle operator in a daily inspection task, a rapid unmanned aerial vehicle orthoimage splicing process is realized through preset parameters and dynamic parameters in an actual shooting process.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to a method for quickly stitching orthophotos of unmanned aerial vehicles. Background Art

[0002] Drone image stitching technology is the process of stitching multiple images taken by drones to generate a panoramic image or map. Although this technology has made a lot of progress, there are still some defects and shortcomings:

[0003] (1) Image registration problem: Since the drone is affected by wind and other environmental factors during flight, the captured images may have a certain degree of offset, rotation or height difference, which may affect the accuracy and quality of image stitching.

[0004] (2) Data processing is large and time-consuming: Since images taken by drones usually have high resolution, the stitching process requires processing a large amount of data and may take a long time to complete, especially for stitching tasks that require high precision and a large range.

[0005] (3) Sensitive to terrain and lighting changes: The undulations of the terrain, obstructions such as trees and buildings, and changes in lighting angles and conditions may all lead to difficulties in image stitching, requiring post-processing corrections or adjustments.

[0006] The present invention relates to a parameter-based rapid unmanned aerial vehicle (UAV) orthophoto stitching method, aiming to find a rapid image stitching method to meet application scenarios with high real-time requirements. The method uses images taken by UAVs for processing and stitching, and rapidly realizes orthophoto stitching through parameterized adjustment, thereby improving image processing efficiency and accuracy. Summary of the invention

[0007] The purpose of the present invention is to solve the shortcomings of the prior art and propose a method for quickly stitching drone orthophotos. The method is based on the orthophotos taken by drone operators during daily inspection tasks, and realizes a fast drone orthophoto stitching process through pre-set parameters and dynamic parameters in the actual shooting process.

[0008] A method for quickly stitching orthophotos of unmanned aerial vehicles comprises the following steps:

[0009] S1. The drone acquires aerial images and imports them into a computer for processing;

[0010] S2. Obtain image XMP information, including latitude, longitude, elevation h, heading angle θ, pitch angle α, roll angle β, and focal length f;

[0011] S3. Image affine transformation: use the heading angle θ, flip angle β and elevation h to calculate the affine matrix M, perform affine transformation according to the affine matrix M, output the image set Z, define the length of each image as Ix and the width as Iy, and obtain the maximum image size Max(Ix) and Max(Iy) through global traversal;

[0012] S4. Calculate the offset. Calculate the coordinate point A of the upper left corner by global traversal. Take the coordinate point A as the reference point. Use the Vincenty algorithm to calculate the longitude distance Lx from the center point of each image to the reference point A. Then use the pixel number calculation formula Dx=f*Lx / h to calculate the longitude pixel offset Dx. Similarly, use the Vincenty algorithm to calculate the latitude distance Ly from the center point of each image to the reference point A. Then use the pixel number calculation formula Dy=f*Ly / h to calculate the latitude pixel offset Dy, and output the offset set W.

[0013] S5. Calculate the image boundary, use the global traversal method to calculate the maximum offset Max(Dx) in the longitude direction and the maximum offset Max(Dy) in the latitude direction, and use the screen coordinate system to take Max(Dx)+Max(Ix) and Max(Dy)+Max(Iy) as the coordinates of the lower right corner of the boundary;

[0014] S6. Draw the image data according to the offset set W, merge the images, and output the results after merging to obtain a spliced ​​image file.

[0015] Preferably, in S1, the shooting requires that the camera suspended by the drone is vertically aimed directly downward to shoot, and the heading overlap rate is required to be greater than or equal to 60%; the lateral overlap rate is required to be greater than or equal to 60%.

[0016] Preferably, the affine transformation matrix M in S3 is expressed as:

[0017]

[0018] Where: α is the scaling ratio, θ is the rotation angle, x c ,y c is the coordinate of the rotation center point; α is the elevation h of each image divided by the average elevation of all images.

[0019] Preferably, the offset set W in S4 also includes corrected offsets of Dx+Max(Ix) / 2 and Dy+Max(Iy) / 2, so that additional offsets are added to each image.

[0020] The advantages and technical effects of the present invention are:

[0021] The invention discloses a rapid stitching method for unmanned aerial vehicle orthophotos, which has a simplified stitching algorithm, is convenient for lightweight and miniaturized deployment, is applicable to a wide range of computing scenarios, and has a stitching effect that can meet the application requirements of real-time image maps.

[0022] The invention discloses a rapid stitching method for unmanned aerial vehicle orthophotos, which makes full use of known flight parameters and camera parameters and simplifies the stitching algorithm and processing process of the images by mathematical calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of the method of the present invention;

[0024] Figure 2 is a schematic diagram of an image information list used in the present invention;

[0025] Figure 3 Parameter diagram of the distance calculation method (Vincenty algorithm) used in the present invention;

[0026] Figure 4a , 4b 4c is the original image used before splicing in the present invention;

[0027] Figure 5 This is an image effect diagram completed by splicing using the present invention; DETAILED DESCRIPTION

[0028] The embodiments of the invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the invention, and cannot be understood as limiting the invention.

[0029] A method for quickly stitching orthophotos of unmanned aerial vehicles of the present invention comprises the following steps:

[0030] S1. The drone acquires aerial images and imports them into a computer for processing;

[0031] S2. Obtain image XMP information, including latitude, longitude, elevation h, heading angle θ, pitch angle α, roll angle β, and focal length f;

[0032] S3. Image affine transformation: use the heading angle θ, flip angle β and elevation h to calculate the affine matrix M, perform affine transformation according to the affine matrix M, output the image set Z, define the length of each image as Ix and the width as Iy, and obtain the maximum image size Max(Ix) and Max(Iy) through global traversal;

[0033] S4. Calculate the offset. Calculate the coordinate point A of the upper left corner by global traversal. Take the coordinate point A as the reference point. Use the Vincenty algorithm to calculate the longitude distance Lx from the center point of each image to the reference point A. Then use the pixel number calculation formula Dx=f*Lx / h to calculate the longitude pixel offset Dx. Similarly, use the Vincenty algorithm to calculate the latitude distance Ly from the center point of each image to the reference point A. Then use the pixel number calculation formula Dy=f*Ly / h to calculate the latitude pixel offset Dy, and output the offset set W.

[0034] S5. Calculate the image boundary, use the global traversal method to calculate the maximum offset Max(Dx) in the longitude direction and the maximum offset Max(Dy) in the latitude direction, and use the screen coordinate system to take Max(Dx)+Max(Ix) and Max(Dy)+Max(Iy) as the coordinates of the lower right corner of the boundary;

[0035] S6. Draw the image data according to the offset set W, merge the images, and output the results after merging to obtain a spliced ​​image file.

[0036] Preferably, in S1, the shooting requires that the camera suspended by the drone is vertically aimed directly downward to shoot, and the heading overlap rate is required to be greater than or equal to 60%; the lateral overlap rate is required to be greater than or equal to 60%.

[0037] Preferably, the affine transformation matrix M in S3 is expressed as:

[0038]

[0039] Where: α is the scaling ratio, θ is the rotation angle, x c ,y c is the coordinate of the rotation center point; α is the elevation h of each image divided by the average elevation of all images.

[0040] Preferably, the offset set W in S4 also includes corrected offsets of Dx+Max(Ix) / 2 and Dy+Max(Iy) / 2, so that additional offsets are added to each image.

[0041] Finally, all the parts not described in the present invention adopt mature products and mature technical means in the prior art.

[0042] In the description of this specification, reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in the embodiment or example of the invention.

[0043] Although embodiments of the invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A method for fast stitching of drone orthophotos, characterized in that: The following steps are involved: S1. The drone acquires aerial images and imports them into a computer for processing; S2. Obtain image XMP information, including latitude, longitude, elevation h, heading angle θ, pitch angle α, roll angle β, and focal length f; S3. Image affine transformation: use the heading angle θ, flip angle β and elevation h to calculate the affine matrix M, perform affine transformation according to the affine matrix M, output the image set Z, define the length of each image as Ix and the width as Iy, and obtain the maximum image size Max(Ix) and Max(Iy) through global traversal; S4. Calculate the offset. Calculate the coordinate point A of the upper left corner by global traversal. Take the coordinate point A as the reference point. Use the Vincenty algorithm to calculate the longitude distance Lx from the center point of each image to the reference point A. Then use the pixel number calculation formula Dx=f*Lx / h to calculate the longitude pixel offset Dx. Similarly, use the Vincenty algorithm to calculate the latitude distance Ly from the center point of each image to the reference point A. Then use the pixel number calculation formula Dy=f*Ly / h to calculate the latitude pixel offset Dy, and output the offset set W. S5. Calculate the image boundary, use the global traversal method to calculate the maximum offset Max(Dx) in the longitude direction and the maximum offset Max(Dy) in the latitude direction, and use the screen coordinate system to take Max(Dx)+Max(Ix) and Max(Dy)+Max(Iy) as the coordinates of the lower right corner of the boundary; S6. Draw the image data according to the offset set W, merge the images, and output the results after merging to obtain a spliced ​​image file.

2. The method for rapidly stitching orthophotos from drones according to claim 1, characterized in that: In S1, the shooting requires that the camera suspended by the drone is vertically aimed directly downward to shoot, and the heading overlap rate is required to be greater than or equal to 60%; the lateral overlap rate is required to be greater than or equal to 60%.

3. The method for rapid stitching of drone orthophotos according to claim 1, characterized in that: The expression of the affine transformation matrix M in S3 is: Where: α is the scaling ratio, θ is the rotation angle, x c ,y c is the coordinate of the rotation center point; α is the elevation h of each image divided by the average elevation of all images.

4. The method for rapidly stitching orthophotos from drones according to claim 1, characterized in that: The offset set W in S4 also includes the corrected offsets of Dx+Max(Ix) / 2 and Dy+Max(Iy) / 2, with additional offsets added to each image.

Images