Unmanned aerial vehicle remote sensing image seamline network generation method and system based on grid transformation

Abstract

This invention discloses a method and system for generating mosaic network of UAV remote sensing images based on raster transformation. Applied to the field of high-resolution UAV remote sensing image technology, the method includes: using a seed point region growing method to obtain the effective region range of the orthophoto in each UAV remote sensing image; using the boundary pixels of the effective region range as distance seed points, performing a raster distance transformation to obtain a raster distance map; extracting the nearest point set boundaries based on the raster distance map to complete the extraction of the bisecting lines between overlapping images; and generating a seam network using the bisecting lines between overlapping images. This invention uses a fast and accurate Euclidean distance transformation on overlapping images and extracts the bisecting lines between overlapping images based on this distance transformation. It can adapt to overlapping polygons of arbitrary shapes, overcoming the limitation of existing algorithms that can only handle overlapping polygons that are convex.

Description

Technical Field

[0001] This invention relates to the field of high-resolution remote sensing imagery technology for unmanned aerial vehicles (UAVs), and particularly to a method and system for generating mosaic line networks from UAV remote sensing images based on raster transformation. Background Technology

[0002] Digital orthophotos, as an important component of space infrastructure, are a crucial part of basic data products for photogrammetry and remote sensing. Automated production of digital orthophoto products has always been a goal pursued in the development of photogrammetry and remote sensing. The traditional orthophoto production process mainly includes three stages: orthorectification, homogenization, and image mosaicking. Image mosaicking, as the final stage in the production and manufacturing of digital orthophoto maps, has not yet been adequately resolved.

[0003] Traditional serial processing based on pairwise mosaicking is no longer suitable for the challenges brought about by the development of remote sensing data acquisition technology, such as large data volume, high overlap and small coverage area, because its results are related to the processing order of images and the processing efficiency is low.

[0004] The mosaicking method based on seam line networks can determine the range of pixel regions that contribute to mosaicking in each image, greatly improving the production efficiency of large-scale image mosaicking processing. There are mainly seam line network generation methods based on traditional Voronoi diagrams and those that take into account overlap. However, the seam line network generated based on traditional Voronoi diagrams cannot guarantee that the seam lines are located in the overlapping areas between images. The seam line network generation method that takes into account overlap can obtain better results when the overlapping area is a general convex polygon or a combination thereof, but it cannot handle the case where the overlapping surface is a concave polygon.

[0005] Therefore, there is an urgent need to develop a method and system for generating mosaic lines from UAV remote sensing images based on raster transformation to solve the above problems. Summary of the Invention

[0006] This invention provides a method and system for generating mosaic line networks from UAV remote sensing images based on raster transformation, which solves the problem that existing mosaic methods cannot guarantee that the seam lines are located in the overlapping areas between images and cannot be applied to overlapping areas of arbitrary shapes.

[0007] According to a first aspect of the present invention, a method for generating mosaic line networks from UAV remote sensing images based on raster transformation is provided. The method includes: The effective area of ​​orthophoto in each UAV remote sensing image is obtained by using the seed point region growing method. Using the boundary pixels of the effective area as distance seed points, perform grid distance transformation to obtain a grid distance map; Based on the raster distance map, the nearest point set boundary is extracted to complete the extraction of the bisecting line between overlapping area images; A seam line network is generated using the bisecting lines between the overlapping images.

[0008] Furthermore, the effective area of ​​the orthophoto in each UAV remote sensing image is obtained using the seed point region growing method, including the following steps: The outermost pixels of the UAV remote sensing image are scanned sequentially. It is determined whether the pixel value corresponding to the outermost pixel is 0. If it is, it is pushed onto the stack and marked as an invalid region; otherwise, it is marked as a valid region. Take the first element in the stack as the target pixel. Starting from the initial seed point, scan all the adjacent pixels of the target pixel in sequence. Determine whether the pixel value of the adjacent pixel is 0. If it is, mark it as an invalid region and push it onto the stack. At the same time, pop the target pixel from the stack and mark it as a valid region. Repeat the previous step until there are no more elements in the stack, and then stop to obtain the effective area of ​​the orthophoto.

[0009] Further, using the boundary pixels of the effective area as distance seed points, a raster distance transformation is performed to obtain a raster distance map, including: The scanned foreground pixel is taken as the current pixel P. The eight neighborhoods of the current pixel P are divided into neighborhood set N1 and neighborhood set N2. Neighborhood set N1 includes the left neighborhood, top-left neighborhood, top neighborhood, and top-right neighborhood of the current pixel in sequence. Neighborhood set N2 includes the right neighborhood, bottom-right neighborhood, bottom neighborhood, and bottom-left neighborhood of the current pixel in sequence. The grid distance transformation includes: a sequential full access based on the neighborhood set N1, specifically including the following steps: The binary digital image value f(p) of the current pixel P is initialized to infinity. Then, the neighboring pixels in the neighborhood set N1 are traversed in order. The sum of the binary digital image value f(q) of each neighboring pixel and the square Euclidean distance h(p,q) from the current pixel P to the neighboring pixel is calculated in turn. The sum is compared with the corresponding binary digital image value f(p). The minimum value between the two is taken to update the binary digital image value f(p) of the current pixel. Determine if there exists a minimum value for the summation result f(q) + h(p, q) of adjacent pixels, and this minimum value is less than the updated binary digital image value f(p) of the current pixel; If so, then take the sum of the relative coordinate vector R(q) of the adjacent pixel and the relative coordinate difference G(p, q) between the current pixel P and the adjacent pixel, and update the relative coordinate vector R(p) of the current pixel. Continue iterating through the remaining foreground pixels until all pixels have been processed.

[0010] Furthermore, the grid distance transformation also includes a reverse full access based on the neighborhood set N2; A reverse full access includes the following steps: Iterate through the neighboring pixels in the neighborhood set N2 of the current pixel P in order, calculate the sum of the binary digital image value f(q) of each neighboring pixel and the square Euclidean distance h(p,q) from the current pixel P to the neighboring pixel, compare the sum with the corresponding binary digital image value f(p), and take the minimum value between the two to update the binary digital image value f(p) of the current pixel. Determine if there exists a minimum value for the summation result f(q) + h(p, q) of adjacent pixels, and this minimum value is less than the updated binary digital image value f(p) of the current pixel; If so, take the sum of the relative coordinate vector R(q) of the neighboring pixel and the relative coordinate difference G(p,q) between the current pixel P and the neighboring pixel, update the relative coordinate vector R(p) of the current pixel, and assign the square root of the updated binary digital image value f(p) of the current pixel as the Euclidean distance transformation matrix E(p). Continue iterating through the remaining foreground pixels until all pixels have been processed.

[0011] Furthermore, based on the raster distance map, the nearest point set boundary is extracted to complete the extraction of the bisecting lines between overlapping area images, including: The start and end pixels are defined based on the intersection of the boundaries of the two raster distance maps; The starting pixel is used as the seed point. The effective region boundary point is found by using the seed point region expansion method. The process ends when the termination pixel is found. The found points are stored in an array to obtain the distance generator point set A. Using the termination pixel as the seed point, the effective region boundary point is found by using the seed point region expansion method. The process ends when the starting pixel is found. The found points are stored in another array to obtain the distance generator point set B. Push the points in the distance generator point set A onto the stack one by one, and pop the top element of the stack one by one as the pop point. P i Using eight-neighbor pairs for pop points P i Iterate through the adjacent pixels and determine the distance from the adjacent pixels to the pop point. P i If the distance is equal to the Euclidean distance recorded in the Euclidean distance transformation matrix, then pop the stack point. P i Push the marker onto the stack, indicating that it has been traversed and belongs to region A, until the stack is empty; Push the points in the generator point set B into the stack one by one, and pop the top element of the stack one by one as the pop point.P i Using eight-neighbor pairs for pop points P i Iterate through the adjacent pixels and determine the distance from the adjacent pixels to the pop point. P i If the distance is equal to the Euclidean distance recorded in the Euclidean distance transformation matrix, then pop the stack point. P i Push the item onto the stack, marking that it has been traversed and belongs to region B, until the stack is empty; Traverse the markers and extract the overlapping points that belong to both region A and region B to obtain the bisecting line between the overlapping images.

[0012] Furthermore, a seam line network is generated using the bisectors between the overlapping area images, including: The effective area of ​​the orthophoto is divided using the Voronoi polygon generation algorithm to obtain the Voronoi polygons of each orthophoto. After generating Voronoi polygons, the common edge of two adjacent Voronoi polygons is the seam line, and the intersection of three adjacent Voronoi polygons is the common vertex. All common edges and common vertices constitute the seam line network.

[0013] According to a second aspect of the present invention, a system for generating mosaic line networks from UAV remote sensing images based on raster transformation is provided, comprising: The region acquisition module is used to acquire the effective region range of orthophotos in each UAV remote sensing image using the seed point region growing method. The raster transformation module, connected to the region acquisition module, is used to perform raster distance transformation using the boundary pixels of the effective region as distance seed points to obtain a raster distance map. The image bisector module, connected to the raster transformation module, is used to extract the nearest point set boundary based on the raster distance map, and to extract the bisector line between overlapping images. The network generation module, connected to the image bisector module, is used to generate a seam network using the bisectors between the overlapping images.

[0014] According to a third aspect of the present invention, an electronic device is provided, including a memory, a processor, and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement a method for generating mosaic line networks of UAV remote sensing images based on raster transformation.

[0015] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon; the computer program is executed by a processor to implement a method for generating mosaic line networks of UAV remote sensing images based on raster transformation.

[0016] The beneficial effects of this invention are: This invention uses a fast and accurate Euclidean distance transformation on overlapping area images, and extracts the bisecting lines between overlapping images based on this distance transformation. It can adapt to overlapping area polygons of arbitrary shapes, overcome the shortcomings of existing algorithms that can only handle overlapping area polygons that are convex, meet the requirements for generating seam lines, and improve the production efficiency of large-scale image mosaic processing.

[0017] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of the present invention, nor is it intended to restrict the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0018] The above and other features, advantages, and aspects of the various embodiments of the present invention will become more apparent from the accompanying drawings and the following detailed description. The drawings are provided for a better understanding of the invention and are not intended to limit the invention. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein: Figure 1 A flowchart of the method for generating mosaic line networks of UAV remote sensing images based on raster transformation provided in an embodiment of the present invention is shown; Figure 2 This diagram illustrates the neighboring pixel neighborhood numbering provided in an embodiment of the present invention. Figure 3 This invention provides schematic diagrams illustrating four types of relationships between adjacent pixels according to embodiments of the present invention. Figure 4 This diagram illustrates two adjacent UAV remote sensing images provided in an embodiment of the present invention. Figure 5 A schematic diagram of the grid distance map provided in an embodiment of the present invention is shown; Figure 6 A schematic diagram of a region-marked image provided in an embodiment of the present invention is shown; Figure 7 This diagram illustrates the bisecting line between overlapping area images provided in an embodiment of the present invention. Figure 8 A schematic diagram of the bisector between the widened overlapping area images provided in an embodiment of the present invention is shown; Figure 9 This diagram illustrates the framework of a UAV remote sensing image mosaic network generation system based on raster transformation, as provided in an embodiment of the present invention. Figure 10 A block diagram of an electronic device according to an embodiment of the present invention is shown. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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.

[0020] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0021] First, let me explain the following: Orthophotos are remote sensing images that have undergone orthorectification, eliminating distortions caused by terrain undulations and sensor orientation. Each pixel corresponds to fixed geographic coordinate information.

[0022] UAV remote sensing images are often rectangular, that is, regular rectangles, and belong to the original images.

[0023] A rectangular image includes the effective area of ​​the orthophoto and the surrounding ineffective area. During image mosaicking, the effective area of ​​the orthophoto must first be extracted from the rectangular image, and then the seam lines of the effective area are generated and mosaicked.

[0024] This invention provides a method for generating mosaic line networks from UAV remote sensing images based on raster transformation. See [link to related documentation]. Figures 1 to 8 The method specifically includes the following steps: S1. Use the seed point region growing method to obtain the effective area range of the orthophoto in each UAV remote sensing image. This invention solves the problem of determining the effective area of ​​an orthophoto when it is an irregular or non-convex polygon by using a seed point region growing method.

[0025] Seed point region growing is a clustering method based on the similarity of pixels within the same type of land cover. It starts with several seed points and, according to certain similarity metrics (such as average gray value, texture, color, etc.), merges pixels with the same properties as the seed points into the seed point region, thus gradually growing the region until no more pixels meet the criteria, at which point the process terminates.

[0026] In practical applications, the seed point region growing method has two key influencing factors: the selection of seed points and the choice of similarity metric. These two key factors directly affect the final result of the overall method. In this invention, since the invalid region of the orthophoto is always located on the periphery of the rectangular image, the pixels with the highest or lowest gray level on the outermost ring of the rectangular image can be used as seed points. This invention defaults to assigning the minimum gray level, i.e., 0, to the pixels in the invalid region. As for the selection of the similarity metric, this invention determines whether the pixel value is 0; if it is, growth is performed.

[0027] Specifically, it includes the following steps: S11. Sequentially scan the outermost pixels of the UAV remote sensing image, determine whether the pixel value corresponding to the outermost pixel is 0. If it is, push it onto the stack and mark it as an invalid area; otherwise, mark it as a valid area. S12. Take the first element in the stack QSeed as the target pixel. Starting from the initial seed point, scan all the adjacent pixels of the target pixel in order. Determine whether the pixel value of the adjacent pixel is 0. If it is, mark it as an invalid region and push it onto the stack QSeed. At the same time, pop the target pixel from the stack and mark it as a valid region. S13. Repeat S12 until there are no more elements in the stack QSeed, and then obtain the effective area range of the orthophoto.

[0028] For example, see Figure 2 The P in the middle represents the target pixel, and there are 8 dots around it. ~ , representing adjacent pixels, can be As the initial seed point, according to... Start to End, then visit adjacent pixels clockwise and determine the validity of the target pixel.

[0029] In summary, the seed point region growing method can accurately identify the effective area of ​​an image.

[0030] S2. Using the boundary pixels of the effective area as distance seed points, perform grid distance transformation to obtain a grid distance map.

[0031] This invention relates to a fast Euclidean distance transformation based on a 3×3 neighborhood range for raster distance transformation, which requires two full scans of UAV remote sensing images.

[0032] Specifically, see Figure 2 ,make , , For binary digital images, F represents the foreground pixel, i.e., the pixel within the effective area. Let represent the squared Euclidean distance between pixels p and q in the orthophoto, where q represents adjacent pixels. G(p, q) represents the relative coordinate difference between pixels p and q. Represents a relative coordinate vector, where and These represent the pixel distances in the horizontal and vertical directions, respectively.

[0033] The specific calculation is performed using the following formula: ; in, Indicates the horizontal pixel distance. G(p, q), representing the vertical pixel distance, is calculated using the following formula: .

[0034] The fast Euclidean distance transformation designed in this invention requires two full image scans. First, the scanned foreground pixel is taken as the current pixel, and the eight-neighborhood of the current pixel is divided into a neighborhood set N1 and a neighborhood set N2. The neighborhood set N1 includes the left neighbor, upper left neighbor, upper neighbor, and upper right neighbor of the current pixel in sequence; the neighborhood set N2 includes the right neighbor, lower right neighbor, lower neighbor, and lower left neighbor of the current pixel in sequence. This includes a sequential full access based on the neighborhood set N1 and a reverse full access based on the neighborhood set N2. Through two scans, the shortest Euclidean distance is approximated by dynamic programming.

[0035] The sequential full access is the first scan of the fast Euclidean distance transform, and the specific process is described in the following algorithm: Sequential full access can be summarized in the following steps: The binary digital image value f(p) of the current pixel P is initialized to infinity. Then, the neighboring pixels in the neighborhood set N1 are traversed in order. The sum of the binary digital image value f(q) of each neighboring pixel and the square Euclidean distance h(p,q) from the current pixel P to the neighboring pixel is calculated in turn. The sum is compared with the corresponding binary digital image value f(p). The minimum value between the two is taken to update the binary digital image value f(p) of the current pixel. Determine if there exists a minimum value for the summation result f(q) + h(p, q) of adjacent pixels, and this minimum value is less than the updated binary digital image value f(p) of the current pixel; If so, take the sum of the relative coordinate vector R(q) of the adjacent pixel and the relative coordinate difference G(p,q) between the current pixel P and the adjacent pixel, R(q)+G(p,q), and update the relative coordinate vector R(p) of the current pixel. Continue iterating through the remaining foreground pixels until all pixels have been processed.

[0036] For the detailed process of reverse full access, please refer to the following algorithm: A full reverse access can be summarized in the following steps: Iterate through the neighboring pixels in the neighborhood set N2 of the current pixel P in order, calculate the sum of the binary digital image value f(q) of each neighboring pixel and the square Euclidean distance h(p,q) from the current pixel P to the neighboring pixel, compare the sum with the corresponding binary digital image value f(p), and take the minimum value between the two to update the binary digital image value f(p) of the current pixel. Determine if there exists a minimum value for the summation result f(q) + h(p, q) of adjacent pixels, and this minimum value is less than the updated binary digital image value f(p) of the current pixel; If so, take the sum of the relative coordinate vector R(q) of the neighboring pixel and the relative coordinate difference G(p,q) between the current pixel P and the neighboring pixel, update the relative coordinate vector R(p) of the current pixel, and assign the square root of the updated binary digital image value f(p) of the current pixel as the Euclidean distance transformation matrix E(p). Continue iterating through the remaining foreground pixels until all pixels have been processed.

[0037] The raster distance map can be obtained by sequential and reverse access. The value of each pixel represents the shortest distance from that pixel to the boundary of the effective area. It can be displayed in grayscale or pseudo-color, with a value of 0 at the boundary and gradually increasing as it moves away from the boundary.

[0038] In a raster distance map, there are four types of relationships between adjacent pixels, such as... Figure 3 As shown, (a) represents a 0-degree relationship, (b) represents a 135-degree relationship, (c) represents a 90-degree relationship, and (d) represents a 45-degree relationship. Figure 4 (a) and 4(b) are two different UAV remote sensing image data. Figure 5 The extracted raster distance map shows that the white areas around the image are non-overlapping areas, while the other areas are overlapping areas. The grayscale values ​​of the pixels represent the distance values, with larger grayscale values ​​indicating greater distances.

[0039] In summary, for each UAV remote sensing image, the minimum distance between each pixel and the boundary of the effective area is calculated. That is, the straight-line distance from each pixel to the effective area is calculated by using fast Euclidean distance transformation, and a raster distance map is obtained. The brighter the pixel in the raster distance map, the farther away the pixel is from the effective area.

[0040] S3. Extract the nearest point set boundary based on the raster distance map to complete the extraction of the bisecting line between overlapping images.

[0041] Specifically, it includes the following steps: S31. Define the start and end pixels based on the intersection of the boundaries of the two raster distance maps; S32. Using the starting pixel as the seed point, the effective region boundary point is found by using the seed point region expansion method. The process ends when the termination pixel is found. The found points are stored in an array to obtain the distance generator point set A. The seed point region expansion method is a region growing algorithm that starts from a certain pixel and expands according to preset rules until a stopping condition is met.

[0042] S33. Using the termination pixel as the seed point, the seed point region expansion method is used to find the effective region boundary point. The process ends when the starting pixel is found. The found points are stored in another array to obtain the distance generator point set B. S34. Push the points in the distance generator point set A onto the stack in sequence, and take out the top element of the stack as the pop point Pi in sequence. Use the eight neighborhood to traverse the adjacent pixels of the pop point Pi, and determine whether the distance from the adjacent pixels to the pop point Pi is equal to the Euclidean distance recorded in the previously obtained Euclidean distance transformation matrix E(p). If so, push the pop point Pi onto the stack, mark it as traversed and belonging to region A, and continue until the stack is empty. S35. Repeat the operation of distance generator point set A on distance generator point set B, that is, push the points in distance generator point set B onto the stack in turn, take out the top element of the stack in turn as the pop point Pi, and use the eight neighborhood to traverse the adjacent pixels of the pop point Pi. Determine whether the distance from the adjacent pixel to the pop point Pi is equal to the Euclidean distance recorded in the previously obtained Euclidean distance transformation matrix E(p). If so, push the pop point Pi onto the stack, mark it as traversed and belonging to region B, and continue until the stack is empty.

[0043] Based on the above steps, the points belonging to the two images are marked respectively, so as to obtain the subsequent bisecting line. That is, the bisecting line between the overlapping images is obtained by extracting the repeated points that belong to both region A and region B.

[0044] S36. Traverse the markers and extract the overlapping points that belong to both region A and region B to obtain the bisecting line between the overlapping area images.

[0045] In one specific embodiment, Figure 6 To obtain the region-labeled image using the method described above, see [link to original text]. Figure 6 The area where the two marked colors meet is the bisector that needs to be found. Figure 7The bisecting lines between overlapping images are extracted by the method of this invention. The precision of the bisecting lines is one pixel, meaning the width of the bisecting lines is one pixel. For easier observation, the bisecting lines are all widened to 11 pixels, as shown in the image. Figure 8 As shown.

[0046] In summary, this invention finds the overlapping area of ​​two UAV remote sensing images and defines the starting and ending intersection points. Then, starting from the intersection points, it finds two sets of distance generator point sets. Based on the raster distance map, it marks the pixels in the overlapping area one by one. Thus, a line is drawn at the boundary of the two markings. Each point on this line is equidistant from the effective area range in the two UAV remote sensing images, thus obtaining the bisector of the overlapping area, which is equivalent to drawing a non-biased center line in the middle of the overlapping images.

[0047] S4. Generate a seam line network using the bisecting lines between the overlapping area images.

[0048] Specifically, it includes the following steps: The effective area of ​​the orthophoto is divided using the Voronoi polygon generation algorithm to obtain the Voronoi polygons of each orthophoto. After generating Voronoi polygons, the common edge of two adjacent Voronoi polygons is the seam line, and the intersection of three adjacent Voronoi polygons is the common vertex. All common edges and common vertices constitute the seam line network.

[0049] Based on the above technical solutions, the present invention has the following advantages: This invention uses a fast and accurate Euclidean distance transformation on overlapping area images, and extracts the bisecting lines between overlapping images based on this distance transformation. Its algorithm can adapt to overlapping area polygons of arbitrary shapes, solving the shortcomings of existing algorithms that can only handle overlapping area polygons that are convex polygons. It meets the requirements for generating seam lines and improves the production efficiency of large-scale image mosaic processing.

[0050] It should be noted that, for the sake of simplicity, the foregoing embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0051] The acquisition, storage, and application of user personal information involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0052] This invention also provides a mosaic line network generation system 900 for UAV remote sensing imagery based on raster transformation, see [link to documentation]. Figure 9 Specifically, it includes: The region acquisition module 910 is used to acquire the effective region range of the orthophoto in each UAV remote sensing image using the seed point region growing method. The raster transformation module 920, connected to the region acquisition module 910, is used to perform raster distance transformation using the boundary pixels of the effective region as distance seed points to obtain a raster distance map. The image bisector module 930 is connected to the raster transformation module 920 and is used to extract the nearest point set boundary according to the raster distance map, thereby completing the extraction of the bisector line between overlapping images. The network generation module 940 is connected to the image bisector module 930 and is used to generate a seam line network using the bisectors between the overlapping images.

[0053] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing embodiments, and will not be repeated here.

[0054] Figure 10 A schematic block diagram of an electronic device 1000 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0055] Electronic device 1000 includes a computing unit 1001, which can perform various appropriate actions and processes according to a computer program stored in ROM 1002 or a computer program loaded into RAM 1003 from storage unit 1008. RAM 1003 may also store various programs and data required for the operation of electronic device 1000. The computing unit 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. I / O interface 1005 is also connected to bus 1004.

[0056] Multiple components in electronic device 1000 are connected to I / O interface 1005, including: input unit 1006, such as keyboard, mouse, etc.; output unit 1007, such as various types of displays, speakers, etc.; storage unit 1008, such as disk, optical disk, etc.; and communication unit 1009, such as network card, modem, wireless transceiver, etc. Communication unit 1009 allows electronic device 1000 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0057] The computing unit 1001 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1001 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1001 performs the various processes described above. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 1000 via ROM 1002 and / or communication unit 1009.

[0058] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0059] The program code used to implement the embodiments of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0060] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0061] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including voice input, speech input, or tactile input).

[0062] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0063] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0064] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0065] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for generating mosaic line networks from UAV remote sensing images based on raster transformation, characterized in that, Includes the following steps: The effective area of ​​orthophoto in each UAV remote sensing image is obtained by using the seed point region growing method. Using the boundary pixels of the effective area as distance seed points, perform grid distance transformation to obtain a grid distance map; Based on the raster distance map, the nearest point set boundary is extracted to complete the extraction of the bisecting line between overlapping area images; A seam line network is generated using the bisecting lines between the overlapping images.

2. The method for generating mosaic line networks from UAV remote sensing images based on raster transformation according to claim 1, characterized in that, The effective area of ​​the orthophoto in each UAV remote sensing image is obtained using the seed point region growing method, including the following steps: The outermost pixels of the UAV remote sensing image are scanned sequentially. It is determined whether the pixel value corresponding to the outermost pixel is 0. If it is, it is pushed onto the stack and marked as an invalid region; otherwise, it is marked as a valid region. Take the first element in the stack as the target pixel. Starting from the initial seed point, scan all the adjacent pixels of the target pixel in sequence. Determine whether the pixel value of the adjacent pixel is 0. If it is, mark it as an invalid region and push it onto the stack. At the same time, pop the target pixel from the stack and mark it as a valid region. Repeat the previous step until there are no more elements in the stack, and then stop to obtain the effective area of ​​the orthophoto.

3. The method for generating mosaic line networks from UAV remote sensing images based on raster transformation according to claim 1, characterized in that, Using the boundary pixels of the effective area as distance seed points, a raster distance transformation is performed to obtain a raster distance map, including: The scanned foreground pixel is taken as the current pixel P. The eight neighborhoods of the current pixel P are divided into neighborhood set N1 and neighborhood set N2. Neighborhood set N1 includes the left neighborhood, top-left neighborhood, top neighborhood, and top-right neighborhood of the current pixel in sequence. Neighborhood set N2 includes the right neighborhood, bottom-right neighborhood, bottom neighborhood, and bottom-left neighborhood of the current pixel in sequence. The grid distance transformation includes: a sequential full access based on the neighborhood set N1, specifically including the following steps: The binary digital image value f(p) of the current pixel P is initialized to infinity. Then, the neighboring pixels in the neighborhood set N1 are traversed in order. The sum of the binary digital image value f(q) of each neighboring pixel and the square Euclidean distance h(p,q) from the current pixel P to the neighboring pixel is calculated in turn. The sum is compared with the corresponding binary digital image value f(p). The minimum value between the two is taken to update the binary digital image value f(p) of the current pixel. Determine if there exists a minimum value for the summation result f(q) + h(p, q) of adjacent pixels, and this minimum value is less than the updated binary digital image value f(p) of the current pixel; If so, then take the sum of the relative coordinate vector R(q) of the adjacent pixel and the relative coordinate difference G(p, q) between the current pixel P and the adjacent pixel, and update the relative coordinate vector R(p) of the current pixel. Continue iterating through the remaining foreground pixels until all pixels have been processed.

4. The method for generating mosaic line networks from UAV remote sensing images based on raster transformation according to claim 3, characterized in that, The grid distance transformation also includes a reverse full access based on the neighborhood set N2; A reverse full access includes the following steps: Iterate through the neighboring pixels in the neighborhood set N2 of the current pixel P in order, calculate the sum of the binary digital image value f(q) of each neighboring pixel and the square Euclidean distance h(p,q) from the current pixel P to the neighboring pixel, compare the sum with the corresponding binary digital image value f(p), and take the minimum value between the two to update the binary digital image value f(p) of the current pixel. Determine if there exists a minimum value for the summation result f(q) + h(p, q) of adjacent pixels, and this minimum value is less than the updated binary digital image value f(p) of the current pixel; If so, take the sum of the relative coordinate vector R(q) of the neighboring pixel and the relative coordinate difference G(p,q) between the current pixel P and the neighboring pixel, update the relative coordinate vector R(p) of the current pixel, and assign the square root of the updated binary digital image value f(p) of the current pixel as the Euclidean distance transformation matrix E(p). Continue iterating through the remaining foreground pixels until all pixels have been processed.

5. The method for generating mosaic line networks from UAV remote sensing images based on raster transformation according to claim 4, characterized in that, Based on the raster distance map, the nearest point set boundary is extracted, and the bisectors between overlapping images are extracted, including: The start and end pixels are defined based on the intersection of the boundaries of the two raster distance maps; The starting pixel is used as the seed point. The effective region boundary point is found by using the seed point region expansion method. The process ends when the termination pixel is found. The found points are stored in an array to obtain the distance generator point set A. Using the termination pixel as the seed point, the effective region boundary point is found by using the seed point region expansion method. The process ends when the starting pixel is found. The found points are stored in another array to obtain the distance generator point set B. Push the points in the distance generator point set A onto the stack one by one, and pop the top element of the stack one by one as the pop point. P i Using eight-neighbor pairs for pop points P i Iterate through the adjacent pixels and determine the distance from the adjacent pixels to the pop point. P i If the distance is equal to the Euclidean distance recorded in the Euclidean distance transformation matrix, then pop the stack point. P i Push the marker onto the stack, indicating that it has been traversed and belongs to region A, until the stack is empty; Push the points in the generator point set B into the stack one by one, and pop the top element of the stack one by one as the pop point. P i Using eight-neighbor pairs for pop points P i Iterate through the adjacent pixels and determine the distance from the adjacent pixels to the pop point. P i If the distance is equal to the Euclidean distance recorded in the Euclidean distance transformation matrix, then pop the stack point. P i Push the item onto the stack, marking that it has been traversed and belongs to region B, until the stack is empty; Traverse the markers and extract the overlapping points that belong to both region A and region B to obtain the bisecting line between the overlapping images.

6. The method for generating mosaic line networks from UAV remote sensing images based on raster transformation according to claim 1, characterized in that, Generating a seam line network using the bisecting lines between the overlapping images includes: The effective area of ​​the orthophoto is divided using the Voronoi polygon generation algorithm to obtain the Voronoi polygons of each orthophoto. After generating Voronoi polygons, the common edge of two adjacent Voronoi polygons is the seam line, and the intersection of three adjacent Voronoi polygons is the common vertex. All common edges and common vertices constitute the seam line network.

7. A system for generating mosaic line networks from UAV remote sensing images based on raster transformation, used to implement the method for generating mosaic line networks from UAV remote sensing images based on raster transformation as described in any one of claims 1-6, characterized in that, The system includes: The region acquisition module is used to acquire the effective region range of orthophotos in each UAV remote sensing image using the seed point region growing method. The raster transformation module, connected to the region acquisition module, is used to perform raster distance transformation using the boundary pixels of the effective region as distance seed points to obtain a raster distance map. The image bisector module, connected to the raster transformation module, is used to extract the nearest point set boundary based on the raster distance map, and to extract the bisector line between overlapping images. The network generation module, connected to the image bisector module, is used to generate a seam network using the bisectors between the overlapping images.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method for generating mosaic line networks of UAV remote sensing images based on raster transformation as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores a computer program; the computer program is executed by a processor to implement the method for generating mosaic line networks of UAV remote sensing images based on raster transformation as described in any one of claims 1 to 6.

Images