A method and device for vectorizing a road skeleton raster image

Abstract

The application discloses a road skeleton grid image vectorization method and device, and the method comprises the following steps: acquiring a road skeleton grid image; classifying pixel points in the road skeleton grid image according to the aggregation condition of the pixel points with a value of 1 in the eight-neighborhood value of each pixel point, to obtain a plurality of types of pixel points; initializing the state of all non-isolated pixel points with a value of 1 as untracked, and deleting the remaining pixel points; tracking the plurality of types of pixel points according to the category to which the pixel points belong, and marking the tracked pixel points as tracked; connecting the pixel points marked as tracked into line segments, and constructing a vector geometric road model according to the line segments. The application solves the problems of road repetition and topological relationship errors that may exist in the road skeleton grid image extracted by using an indirect method in the prior art, thereby improving the automation degree and robustness of the construction of the open-pit mine road.

Description

Technical Field

[0001] This application relates to the field of road image processing, and more specifically, to a method and apparatus for vectorizing road skeleton raster images. Background Technology

[0002] Open-pit mine road models and topological relationships are crucial foundations for mine information technology construction, intelligent scheduling, and unmanned driving. Methods for constructing mine road geometric models can be categorized into direct and indirect methods. Direct methods involve using surveying tools such as total stations and RTK to measure points at regular intervals along the mine road, thereby constructing the geometric model. However, for the constantly changing open-pit mine roads, the direct method suffers from a long construction cycle and a large workload. Therefore, indirect methods are generally used when constructing open-pit mine road models. Indirect methods involve extracting road skeleton raster images from open-pit mine truck GPS trajectory lines or mine area imagery to further trace the open-pit mine road geometric model. The indirect method faces challenges such as potential road duplication and topological relationship errors during the raster model vectorization process. Summary of the Invention

[0003] This application provides a method and apparatus for vectorizing road skeleton raster images, which at least solves the problems of road duplication and topological relationship errors that may exist in road skeleton raster images extracted by indirect methods in the prior art.

[0004] According to one aspect of this application, a method for vectorizing a road skeleton raster image is provided, comprising: acquiring a road skeleton raster image; classifying the pixels in the road skeleton raster image according to the clustering of pixels with a value of 1 in the eight neighborhoods of each pixel to obtain multiple types of pixels; initializing all non-isolated pixels with a value of 1 to untracked, and deleting the remaining pixels, wherein an isolated pixel is a pixel whose eight neighborhood values ​​are all 0, and a pixel with a value of 1 indicates that the pixel is a pixel constituting the road; tracking the pixels of the multiple types according to their respective categories, and marking the tracked pixels as tracked; connecting the pixels marked as tracked into line segments, and constructing a vector geometric road model based on the connected line segments.

[0005] Furthermore, tracking the pixels of the multiple types according to their respective types and marking the tracked pixels as tracked includes: marking four-connected pixels as tracked, and starting from each four-connected pixel, tracking the adjacent pixels of the four-connected pixel and marking the adjacent pixels as "tracked" until the type of the adjacent pixels changes to single-connected, triple-connected or quad-connected.

[0006] Furthermore, tracking the pixels of the multiple types according to their respective types and marking the tracked pixels as tracked includes: marking the three-connected pixels as tracked, and starting from each three-connected pixel, tracking the adjacent pixels of the three-connected pixel and marking the adjacent pixels as "tracked" until the type of the adjacent pixels changes to single-connected, three-connected or four-connected.

[0007] Furthermore, tracking the pixels of the multiple types according to their respective types and marking the tracked pixels as tracked includes: marking single-connected pixels as tracked, and starting from each single-connected pixel, tracking the neighboring pixels of the single-connected pixel and marking the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single-connected, triple-connected or quad-connected.

[0008] Furthermore, tracking neighboring pixels of a quad-connected, triple-connected, or single-connected pixel includes: preferentially tracking neighboring pixels from the four directions of east, west, south, and north of the quad-connected, triple-connected, or single-connected pixel.

[0009] According to another aspect of this application, a road skeleton grid image vectorization device is also provided, comprising: an acquisition module for acquiring a road skeleton grid image; a classification module for classifying pixels in the road skeleton grid image according to the clustering of pixels with a value of 1 in the eight neighborhoods of each pixel, to obtain multiple types of pixels; an initialization module for initializing all non-isolated pixels with a value of 1 to untracked and deleting the remaining pixels, wherein an isolated pixel is a pixel whose eight neighborhood values ​​are all 0, and a pixel with a value of 1 indicates that the pixel is a pixel constituting a road; a tracking module for tracking the multiple types of pixels according to their respective categories, and marking the tracked pixels as tracked; and a construction module for connecting the marked tracked pixels into line segments, and constructing a vector geometric road model based on the connected line segments.

[0010] Furthermore, the tracking module is used to: mark the four connected pixels as tracked, and starting from each four connected pixel, track the adjacent pixels of the four connected pixels and mark the adjacent pixels as "tracked" until the type of the adjacent pixels changes to single connected, triple connected or four connected.

[0011] Furthermore, the tracking module is used to: mark the three-connected pixels as tracked, and starting from each three-connected pixel, track the adjacent pixels of the three-connected pixel and mark the adjacent pixels as "tracked" until the type of the adjacent pixels changes to single-connected, three-connected or four-connected.

[0012] Furthermore, the tracking module is used to: mark single-connected pixels as tracked, and starting from each single-connected pixel, track the neighboring pixels of the single-connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single-connected, triple-connected or quadruple-connected.

[0013] Furthermore, the tracking module is used to: prioritize tracking adjacent pixels from the four directions of east, west, south, and north of quad-connected, triple-connected, or single-connected pixels.

[0014] In this embodiment, a road skeleton grid image is acquired; the pixels in the road skeleton grid image are classified according to the clustering of pixels with a value of 1 in their eight neighborhoods, resulting in multiple types of pixels; all non-isolated pixels with a value of 1 are initialized to untracked, and the remaining pixels are deleted. Isolated pixels are those whose eight neighborhood values ​​are all 0, and pixels with a value of 1 indicate that the pixel constitutes a road. The pixels of the multiple types are tracked according to their category, and the tracked pixels are marked as tracked. The tracked pixels are connected into line segments, and a vector geometric road model is constructed based on the connected line segments. This application solves the problems of road duplication and topological relationship errors that may exist in road skeleton grid images extracted using indirect methods in the prior art, thereby improving the automation and robustness of open-pit mine road construction. Attached Figure Description

[0015] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0016] Figure 1 This is a flowchart of a method for vectorizing a raster image of an open-pit mine road skeleton according to an embodiment of this application;

[0017] Figure 2 This is a schematic diagram of a grid image of the skeleton of an open-pit mine road according to an embodiment of this application;

[0018] Figure 3 This is a schematic diagram of a single connected pixel according to an embodiment of this application;

[0019] Figure 4 This is a schematic diagram of dual-connected pixels according to an embodiment of this application;

[0020] Figure 5 This is a schematic diagram of three connected pixels according to an embodiment of this application;

[0021] Figure 6This is a schematic diagram of four connected pixels according to an embodiment of this application;

[0022] Figure 7 This is a vectorized road diagram of an open-pit mine constructed according to an embodiment of this application. Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0025] In this embodiment, a method for vectorizing a road skeleton raster image is provided, which includes the following steps:

[0026] Step A: Obtain the road skeleton grid image.

[0027] Step B involves classifying the pixels in the road skeleton grid image according to the clustering of pixels with a value of 1 in the eight neighborhoods of each pixel, resulting in multiple types of pixels.

[0028] Step C: Initialize the state of all non-isolated pixels with a value of 1 to untracked, and delete the remaining pixels. Isolated pixels are those whose eight neighboring values ​​are all 0, and pixels with a value of 1 are those that constitute the road.

[0029] Step D: Track the pixels of the multiple types according to their category, and mark the tracked pixels as tracked.

[0030] In this step, tracing can begin with any type of pixel, such as a four-connected, three-connected, or one-connected pixel. For example, four-connected pixels are marked as traced, and starting from each four-connected pixel, its neighboring pixels are traced and marked as "traced" until the type of the neighboring pixels changes to one-connected, three-connected, or four-connected. Similarly, three-connected pixels are marked as traced, and starting from each three-connected pixel, its neighboring pixels are traced and marked as "traced" until the type of the neighboring pixels changes to one-connected, three-connected, or four-connected. Likewise, one-connected pixels are marked as traced, and starting from each one-connected pixel, its neighboring pixels are traced and marked as "traced" until the type of the neighboring pixels changes to one-connected, three-connected, or four-connected. Preferably, neighboring pixels can be traced primarily in the east, west, south, and north directions of the four-connected, three-connected, or one-connected pixels.

[0031] Step E: Connect the pixels marked as tracked into line segments, and construct a vector geometric road model based on the connected line segments.

[0032] The road raster image was further processed in the above steps. Therefore, the above steps solve the problems of road duplication and topological relationship errors that may exist in the road skeleton raster image extracted by the indirect method in the prior art, thereby improving the automation and robustness of open mine road construction.

[0033] The following describes an optional embodiment with reference to the accompanying drawings. In this optional embodiment, a method for vectorizing a raster image of an open-pit mine road skeleton is provided, including the following steps: S1, pixel feature classification; S2, pixel tracking state initialization; S3, starting and tracking of four connected pixels; S4, starting and tracking of three connected pixels; S5, starting and tracking of a single connected pixel.

[0034] In step S1, the basic characteristic of a standard binary skeleton raster image (which can be obtained through existing methods and will not be elaborated here) is that in the eight-neighborhood of each pixel, at most two pixels with a value of 1 are clustered together. The pixels of the open-pit mine road skeleton raster image are categorized into isolated pixels, single-connected pixels, double-connected pixels, triple-connected pixels, and quadruple-connected pixels based on the clustering distribution of their eight-neighborhoods. An isolated pixel is defined as one whose eight-neighborhood values ​​are all 0; a single-connected pixel is defined as one whose eight-neighborhood contains exactly one cluster of pixels with a value of 1; a double-connected pixel is defined as one whose eight-neighborhood contains exactly one cluster of pixels with a value of 1; and a quadruple-connected pixel is defined as one whose eight-neighborhood contains exactly one cluster of pixels with a value of 1. A pixel with only two clustered values ​​of 1 has its eight-neighbor clustering distribution in seven forms: "1+5", "2+4", "3+3", "1+4", "2+3", "1+3", and "2+2". A triple-connected pixel has exactly three clustered values ​​of 1 in its eight-neighbor clustering distribution in four forms: "1+1+3", "1+2+2", "1+1+2", and "1+1+1". A quadruple-connected pixel has exactly four clustered values ​​of 1 in its eight-neighbor clustering distribution.

[0035] In step S2, the tracking status of all non-isolated pixels with a value of 1 is initialized to "untracked", and the remaining pixels are deleted.

[0036] In step S3, the eight neighborhoods of a pixel are defined as follows: the neighborhoods in the east, west, south, and north directions are strong neighbors, and the neighborhoods in the other four directions are weak neighbors. It is determined whether all four connected pixels are marked as "tracked". If yes, proceed to the next step; if no, it is further determined whether all pixels adjacent to the four connected pixels are marked as "tracked". If yes, it jumps back and re-executes this step; if no, it starts from the four connected pixels and prioritizes tracking along the directions of its strong adjacent pixels, tracking and marking pixels as "tracked" until the type changes to single connected, triple connected, or quad connected.

[0037] In step S4, it is determined whether all three connected pixels are marked as "tracked". If yes, proceed to the next step; if no, it is further determined whether all pixels adjacent to the three connected pixels are marked as "tracked". If yes, it is jumped back to re-execute this step; if no, starting from the three connected pixels, it is prioritized to track and mark pixels as "tracked" along the direction of its strong adjacent pixels until the type changes to single connected or three connected.

[0038] In step S5, it is determined whether all single-connected pixels are marked as "tracked". If so, the execution ends and the open-pit mine vector road model is successfully constructed. If not, starting from the single-connected pixel, the pixel is tracked and marked as "tracked" along the direction of its strong adjacent pixel until the type becomes single-connected.

[0039] Figure 1 This is a flowchart of a method for vectorizing open-pit mine road skeleton raster images according to an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps:

[0040] 101. Pixel Feature Classification. A basic characteristic of standard binary skeleton raster images is that in the eight-neighborhood of each pixel, at most two pixels with a value of 1 cluster together (where a value of 1 represents a black pixel, indicating a road). Figure 3 As shown in the first graph, the eight neighbors of P0 are P1 to P8, and only P7 has a value of 1. Therefore, only one point in the eight neighbors has a value of 1 and they are clustered together. In the second graph, only one point in the eight neighbors has a value of 1 and they are clustered together. In the third graph, P8 and P1 have values ​​of 1, so two points in the eight neighbors have a value of 1 and they are clustered together. For example... Figure 4 As shown in the first image, in the eight-neighborhood, P2 and P8 have a value of 1, but P2 and P8 are not "clustered" together, separated by P1. Therefore, only one point in the eight-neighborhood has a value of 1 clustered together. In the fourth image, in the eight-neighborhood, P1, P3, and P8 have a value of 1, and only P1 and P8 are "clustered" together. Therefore, two points in the eight-neighborhood have a value of 1 clustered together. The pixels in the open-pit mine road skeleton raster image are categorized into isolated pixels, single-connected pixels, double-connected pixels, triple-connected pixels, and quadruple-connected pixels based on their eight-neighborhood clustering. Isolated pixels are those whose eight-neighborhood values ​​are all 0. Single-connected pixels are those whose eight-neighborhood has exactly one clustered value of 1. Double-connected pixels are those whose eight-neighborhood has exactly two clustered values ​​of 1. The eight-neighborhood clustering of double-connected pixels can be further subdivided. There are seven forms: “1+5”, “2+4”, “3+3”, “1+4”, “2+3”, “1+3” and “2+2”. A three-connected pixel is a pixel whose value is 1 in exactly 3 of its eight neighboring regions. The clustering distribution of the eight neighboring regions of a three-connected pixel can be further divided into four forms: “1+1+3”, “1+2+2”, “1+1+2” and “1+1+1”. A four-connected pixel is a pixel whose value is 1 in exactly 4 of its eight neighboring regions.

[0041] 102. Pixel tracking state initialization. Initialize the tracking state of all non-isolated pixels with a value of 1 to "untracked", and delete the remaining pixels.

[0042] 103. Starting and Tracking Four-Connected Pixels. Define the eight neighbors of a pixel: neighbors in the east, west, south, and north directions are strong neighbors, while neighbors in the other four directions are weak neighbors. Determine if all four-connected pixels are marked as "tracked." If yes, proceed to the next step; otherwise, further determine if all pixels adjacent to the four-connected pixel are marked as "tracked." If yes, jump back and re-execute this step; otherwise, starting from the four-connected pixel, prioritize tracking along the directions of its strong neighbors and mark pixels as "tracked" until the type changes to single-connected, triple-connected, or four-connected. That is, if adjacent pixels are double-connected, they are marked as tracked; if they are single-connected, triple-connected, or four-connected, tracking stops. This is because double-connected pixels are ordinary pixels, serving as bridges connecting single-connected, triple-connected, and four-connected pixels.

[0043] 104. Starting from and tracking three connected pixels. Determine if all three connected pixels are marked as "tracked". If yes, proceed to the next step; if no, further determine if all pixels adjacent to the three connected pixels are marked as "tracked". If yes, jump back and re-execute this step; if no, start from the three connected pixels, prioritize the direction of its strongly adjacent pixels, track and mark pixels as "tracked" until the type changes to single or three connected.

[0044] 105. Starting and Tracking Single Connected Pixels. Determine if all single connected pixels are marked as "tracked". If so, the process ends and the open-pit mine vector road model is successfully constructed. If not, start from a single connected pixel, prioritize the direction of its strongly adjacent pixels, track and mark the pixel as "tracked" until the type changes to single connected.

[0045] The following example will illustrate this point.

[0046] A1. A raster image of the road skeleton in an open-pit mine, as shown below. Figure 2 As shown, the pixels in the open-pit mine road skeleton raster image are categorized into isolated pixels, single-connected pixels, double-connected pixels, triple-connected pixels, and quadruple-connected pixels based on the clustering distribution of their eight neighbors. Isolated pixels are those whose eight neighbors all have a value of 0, while single-connected pixels are those whose eight neighbors have exactly one cluster with a value of 1. Figure 3 As shown. A doubly connected pixel is a pixel whose eight neighborhoods have exactly two clusters with a value of 1. The clustering distribution of the eight neighborhoods of a doubly connected pixel can be further divided into seven forms: "1+5", "2+4", "3+3", "1+4", "2+3", "1+3", and "2+2", as shown. Figure 4 As shown, in Figure 4In the first image, there is one 0 pixel p1 existing alone, and five 0 pixels p3 to p7 clustered together, referred to as "1+5". The remaining "2+4", "3+3", "1+4", "2+3", "1+3", and "2+2" have the same meaning. A three-connected pixel refers to a pixel whose eight-neighborhood has exactly three clusters with a value of 1. The clustering distribution of the eight-neighborhood of a three-connected pixel can be further divided into four forms: "1+1+3", "1+2+2", "1+1+2", and "1+1+1", as shown below. Figure 5 As shown, a four-connected pixel refers to a pixel whose value is 1 in exactly four of its eight neighborhoods, such as... Figure 6 As shown.

[0047] A2. Initialize the tracking status of all non-isolated pixels with a value of 1 to "untracked" and delete the remaining pixels.

[0048] A3, Four-way interconnected pixel point origination and tracking.

[0049] A4. Tri-connection pixel point origination and tracking.

[0050] A5. Starting from and tracking a single connected pixel, after execution, the pixels marked as tracked are connected into line segments. Then, a vector geometric road model is constructed based on these line segments. For example, line segment thinning and smoothing algorithms can be further used to construct a vectorized geometric road model and topological relationships for open-pit mines. Existing algorithms can be used for line segment thinning and smoothing, which will not be elaborated here. Figure 7 As shown.

[0051] In this embodiment, an electronic device is provided, including a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to perform the methods described in the above embodiments.

[0052] The aforementioned program can run on a processor or be stored in memory (or computer-readable medium). Computer-readable medium includes both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable medium does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0053] These computer programs may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes can be implemented using different modules, and different steps can be implemented using different modules.

[0054] This embodiment provides such an apparatus or system. The apparatus, referred to as a road skeleton grid image vectorization apparatus, includes: an acquisition module for acquiring a road skeleton grid image; a classification module for classifying pixels in the road skeleton grid image according to the clustering of pixels with a value of 1 in their eight neighborhoods, resulting in multiple types of pixels; an initialization module for initializing all non-isolated pixels with a value of 1 to untracked and deleting the remaining pixels, wherein an isolated pixel is a pixel whose eight neighborhood values ​​are all 0, and a pixel with a value of 1 indicates that the pixel constitutes a road; a tracking module for tracking the multiple types of pixels according to their category and marking the tracked pixels as tracked; and a construction module for connecting the marked tracked pixels into line segments and constructing a vector geometric road model based on the connected line segments.

[0055] The system or apparatus is used to implement the functions of the methods in the above embodiments. Each module in the system or apparatus corresponds to each step in the method, as has been described in the method and will not be repeated here.

[0056] For example, the tracking module is used to: mark a quad-connected pixel as tracked, and starting from each quad-connected pixel, track the neighboring pixels of that quad-connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single-connected, triple-connected, or quad-connected. Alternatively, the tracking module is used to: mark a triple-connected pixel as tracked, and starting from each triple-connected pixel, track the neighboring pixels of that triple-connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single-connected, triple-connected, or quad-connected. Alternatively, the tracking module is used to: mark a single-connected pixel as tracked, and starting from each single-connected pixel, track the neighboring pixels of that single-connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single-connected, triple-connected, or quad-connected.

[0057] Optionally, the tracking module is used to: preferentially track adjacent pixels from the four directions of east, west, south and north of the four-connected, three-connected or single-connected pixels.

[0058] The above embodiments solve the problems of road duplication and topological relationship errors that may exist in road skeleton raster images extracted by indirect methods in the prior art, thereby improving the automation and robustness of open-pit mine road construction.

[0059] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for vectorizing a road skeleton raster image, characterized in that, include: Acquire road skeleton grid images; The pixels in the road skeleton grid image are classified according to the clustering of pixels with a value of 1 in the eight neighborhoods of each pixel, resulting in multiple types of pixels. Initialize the state of all non-isolated pixels with a value of 1 to untracked, and delete the remaining pixels. Isolated pixels are those whose eight neighboring values ​​are all 0, and pixels with a value of 1 are those that constitute the road. The various types of pixels are tracked according to their category, and the tracked pixels are marked as tracked. Each pixel is categorized based on the clustering distribution of its eight neighbors as follows: isolated pixels, single-connected pixels, double-connected pixels, triple-connected pixels, and quadruple-connected pixels. An isolated pixel is defined as one where all eight neighbors have a value of 0; a single-connected pixel is defined as one where only one of its eight neighbors has a value of 1; a double-connected pixel is defined as one where only two of its eight neighbors have a value of 1; a triple-connected pixel is defined as one where only three of its eight neighbors have a value of 1; and a quadruple-connected pixel is defined as one where only four of its eight neighbors have a value of 1. The quadruple-connected pixel is further categorized based on its eight neighbors. The pixels are marked as "tracked", and starting from each four-connected pixel, the neighboring pixels of the four-connected pixel are tracked and marked as "tracked" until the type of the neighboring pixels becomes single-connected, triple-connected or quad-connected; the triple-connected pixels are marked as "tracked", and starting from each triple-connected pixel, the neighboring pixels of the triple-connected pixel are tracked and marked as "tracked" until the type of the neighboring pixels becomes single-connected, triple-connected or quad-connected; the single-connected pixels are marked as "tracked", and starting from each single-connected pixel, the neighboring pixels of the single-connected pixel are tracked and marked as "tracked" until the type of the neighboring pixels becomes single-connected, triple-connected or quad-connected. The pixels marked as tracked are connected into line segments, and a vector geometric road model is constructed based on the connected line segments.

2. The method according to claim 1, characterized in that, Tracking neighboring pixels of quad-connected, triple-connected, or single-connected pixels includes: Prioritize tracing adjacent pixels from the four directions of east, west, south, and north of quad-connected, triple-connected, or single-connected pixels.

3. A road skeleton raster image vectorization device, characterized in that, include: The acquisition module is used to acquire images of the road skeleton grid. The classification module is used to classify the pixels in the road skeleton grid image according to the clustering of pixels with a value of 1 in the eight neighborhoods of each pixel, and obtain multiple types of pixels. The initialization module is used to initialize the state of all non-isolated pixels with a value of 1 to untracked and delete the remaining pixels. An isolated pixel is a pixel whose eight neighboring values ​​are all 0, and a pixel with a value of 1 indicates that the pixel is a pixel that constitutes the road. The tracking module is used to track pixels of the multiple types according to their category and mark the tracked pixels as tracked. Each pixel is categorized based on the clustering distribution of its eight neighbors as follows: isolated pixels, single-connected pixels, double-connected pixels, triple-connected pixels, and quadruple-connected pixels. An isolated pixel is defined as one where all eight neighbors have a value of 0; a single-connected pixel is defined as one where only one of its eight neighbors has a value of 1; a double-connected pixel is defined as one where only two of its eight neighbors have a value of 1; a triple-connected pixel is defined as one where only three of its eight neighbors have a value of 1; and a quadruple-connected pixel is defined as one where only four of its eight neighbors have a value of 1. Mark the four connected pixels as tracked, and starting from each four connected pixel, track the neighboring pixels of that four connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single connected, triple connected, or four connected; mark the three connected pixels as tracked, and starting from each triple connected pixel, track the neighboring pixels of that triple connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single connected, triple connected, or four connected; mark the single connected pixels as tracked, and starting from each single connected pixel, track the neighboring pixels of that single connected pixel and mark the neighboring pixels as "tracked" until the type of the neighboring pixels changes to single connected, triple connected, or four connected; A building module is used to connect the marked and tracked pixels into line segments and construct a vector geometric road model based on the connected line segments.

4. The apparatus according to claim 3, characterized in that, The tracking module is used for: Prioritize tracing adjacent pixels from the four directions of east, west, south, and north of quad-connected, triple-connected, or single-connected pixels.

Images