Unmanned aerial vehicle-based road disease point pile number positioning method and system

Abstract

The application discloses a road disease point pile number positioning method and system based on a UAV, and comprises the following steps: constructing a road center line based on a road orthographic image and a design file; taking each whole hundred pile number point as an initial reference point, constructing a plurality of vertical offset lines along the road center line, and distributing pile number data of each vertical offset line; taking each vertical offset line as a central axis to construct a rectangular buffer zone, and merging all the rectangular buffer zones into a strip-shaped buffer zone surface; performing grid division on the strip-shaped buffer zone surface according to the resolution of the orthographic image, and distributing pile number data of each grid; aligning each grid of the strip-shaped buffer zone surface with the orthographic image pixel by pixel, and fusing the pile number data of each grid with image data of corresponding pixels to generate corresponding fusion grid data; combining image data of the road disease point and the fusion grid data of each grid, searching for a grid closest to the road disease point by using a KD tree, and accurately positioning the pile number of the road disease point according to corresponding pile number data.

Description

Technical Field

[0001] This invention relates to the field of image analysis for determining the location of objects, and specifically to a method and system for locating road defect points based on unmanned aerial vehicles (UAVs). Background Technology

[0002] With the continuous expansion of highway networks and the increase in service life, highway slope defects have become a significant factor affecting road safety and operational efficiency. Traditional defect detection methods largely rely on manual inspections, which are not only inefficient but also lack comprehensive coverage. The development of modern computer vision and drone technology has provided new avenues for automatic road defect identification. However, current systems primarily focus on the acquisition and recognition of defect images, lacking the ability to efficiently convert the recognition results into the station information required for actual maintenance work. This leads to inaccurate defect location, impacting subsequent maintenance efficiency and cost control. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention proposes a method and system for locating road defect points using unmanned aerial vehicles (UAVs), which can accurately pinpoint the station number information of the defect location. The specific technical solution is as follows:

[0004] In a first aspect, a method for locating road defect points based on unmanned aerial vehicles (UAVs) is provided. In a first implementable method of this first aspect, the method includes:

[0005] Construct the road centerline based on orthophotos and design documents;

[0006] Using each hundred-station point on the road as an initial reference point, multiple vertical offset lines are constructed along the centerline of the road, and station number data for each vertical offset line is assigned according to the corresponding hundred-station number.

[0007] Construct corresponding rectangular buffers with each vertical offset line as the central axis, and merge all rectangular buffers into a continuous strip buffer surface;

[0008] The strip buffer area is divided into grids according to the resolution of the orthophoto, and the station data of each grid is assigned according to the vertical offset line corresponding to the rectangular buffer.

[0009] Align each grid of the strip buffer surface with the orthophoto pixel by pixel, and fuse the station data of each grid with the image data of the corresponding pixel to generate the corresponding fused grid data;

[0010] By combining the image data of road defect points with the fused raster data of each raster, a KD tree is used to search for the raster closest to the road defect point, and the station number of the road defect point is determined based on its corresponding station number data.

[0011] In conjunction with the first feasible method of the first aspect, in the second feasible method of the first aspect, the road centerline is constructed based on the orthophoto of the road and design documents, including:

[0012] Using orthophotos as the base layer, the design file is overlaid and registered with the orthophotos through multiple registration points in the orthophotos, and the road centerline is constructed based on the registered road registration data.

[0013] In conjunction with the first possible implementation of the first aspect, in the third possible implementation of the first aspect, each hundredth-mile marker on the road is used as an initial reference point, including:

[0014] The position coordinates of each hundred-station point were corrected based on the actual on-site measurement data.

[0015] In conjunction with the first possible implementation of the first aspect, in the fourth possible implementation of the first aspect, multiple vertical offset lines are constructed along the centerline of the road, including:

[0016] Using each of the aforementioned hundredth-station points as initial reference points, draw perpendicular lines to the road centerline and determine the intersection point of each perpendicular line with the road centerline;

[0017] Multiple incremental marker points are evenly distributed along the road centerline between two adjacent intersections;

[0018] A vertical offset line is created perpendicular to the road centerline, using each incremental marker as a reference point. This vertical offset line covers the road width.

[0019] In conjunction with the first implementable method of the first aspect, in the fifth implementable method of the first aspect, all rectangular buffers are merged into a continuous strip buffer surface, including:

[0020] The system detects gaps and overlaps between rectangular buffers and processes these gaps and overlaps through automatic merging.

[0021] In conjunction with the first possible implementation of the first aspect, in the sixth possible implementation of the first aspect, corresponding fused raster data is generated, including:

[0022] The station data of the raster is fused with the image data of the corresponding pixels of the raster by band combination.

[0023] In conjunction with the first implementable method of the first aspect, the seventh implementable method of the first aspect also includes:

[0024] A training set is constructed based on the fused raster data corresponding to each raster in the strip buffer surface;

[0025] Using the image data as input and the station number data as output, the constructed data regression model is trained using the training set;

[0026] When the KD tree fails to find the nearest grid cell to the road defect point, the station number of the road defect point is inferred from the image data of the road defect point using a trained data regression model.

[0027] In conjunction with the seventh possible implementation method of the first aspect, in the eighth possible implementation method of the first aspect, the data regression model is a random forest regression model.

[0028] In conjunction with the seventh feasible method of the first aspect, and in the ninth feasible method of the first aspect, the data regression model is constructed by combining the random forest regression model and the XGBoost model.

[0029] Secondly, a road defect point location system is provided, including:

[0030] The centerline construction module is configured to construct road centerlines based on orthophotos and design files.

[0031] The offset line construction module is configured to use each hundred-station point on the road as an initial reference point, construct multiple vertical offset lines along the center line of the road, and assign station number data for each vertical offset line according to the corresponding hundred-station number.

[0032] The buffer surface construction module is configured to construct corresponding rectangular buffers with each vertical offset line as the central axis, and merge all rectangular buffers into a continuous strip buffer surface;

[0033] The grid division module is configured to divide the strip buffer surface into grids according to the resolution of the orthophoto, and to assign station data of each grid according to the vertical offset line corresponding to the rectangular buffer in which it is located.

[0034] The data fusion module is configured to align each grid of the strip buffer surface with the orthophoto pixel by pixel, and fuse the station data of each grid with the image data of the corresponding pixel to generate corresponding fused grid data.

[0035] The stationing module is configured to combine the image data of road defect points with the fused grid data of each grid, use a KD tree to search for the grid closest to the road defect point, and determine the station number of the road defect point based on its corresponding station number data.

[0036] Beneficial Effects: The road defect point location method and system based on UAVs of this invention fuses on-site measured orthophotos of the road with road design data to obtain accurate and reliable fused raster data for a predefined buffer zone, providing a foundation for precise location of defect point station information. Based on the obtained fused raster data and defect image data, KD-tree search can be used to accurately locate the station of defect points within the predefined buffer zone, thereby reducing the time spent locating road defect points, enabling faster data-driven intervention, and minimizing maintenance time. Attached Figure Description

[0037] To more clearly illustrate the specific embodiments of the present invention, the accompanying drawings used in the specific embodiments will be briefly described below. In all the drawings, the elements or parts are not necessarily drawn to scale.

[0038] Figure 1 A flowchart illustrating a method for locating road defect points based on unmanned aerial vehicles (UAVs) according to an embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of a strip buffer surface provided in an embodiment of the present invention. Detailed Implementation

[0040] The embodiments of the technical solution of the present invention will now be described in detail with reference to the accompanying drawings. These embodiments are merely illustrative of the technical solution of the present invention and are therefore intended to limit the scope of protection of the present invention.

[0041] like Figure 1 The flowchart shown is a method for locating road defect points based on unmanned aerial vehicles (UAVs). This method includes:

[0042] Step 1: Construct the road centerline based on the orthophoto of the road and the design files;

[0043] Step 2: Using each hundred-station point on the road as an initial reference point, construct multiple vertical offset lines along the centerline of the road, and assign station number data for each vertical offset line according to the corresponding hundred-station number.

[0044] Step 3: Construct corresponding rectangular buffers with each vertical offset line as the central axis, and merge all rectangular buffers into a continuous strip buffer surface;

[0045] Step 4: Divide the strip buffer area into grids according to the resolution of the orthophoto, and assign station data to each grid according to the vertical offset line corresponding to the rectangular buffer.

[0046] Step 5: Align each grid of the strip buffer surface with the orthophoto pixel by pixel, and fuse the station number data of each grid with the image data of the corresponding pixel to generate the corresponding fused grid data.

[0047] Step 6: Combining the image data of the road defect points with the fused raster data of each raster, use a KD tree to search for the raster closest to the road defect point, and determine the station number of the road defect point based on its corresponding station number data.

[0048] Specifically, firstly, the road centerline can be created by combining orthophotos of the road taken on-site by drones with the road design documents. Then, using each 100-point station as an initial reference point, multiple vertical offset lines perpendicular to the road centerline are constructed along the road centerline, and the station information of the reference 100-point stations is assigned to each vertical offset line.

[0049] Next, rectangular buffers of equal size are created using each vertical offset line as the central axis. All these rectangular buffers are then merged into a strip-shaped buffer surface distributed along the road surface. The strip-shaped buffer surface is then rasterized according to the image resolution of the orthophoto, and the stationing information of the vertical offset line corresponding to the rectangular buffer of each raster is assigned to each raster. Finally, by using the set capture raster, all rasteres are aligned one-to-one with each pixel of the orthophoto, thereby minimizing misalignment during data fusion and ensuring the accuracy of the fused raster data obtained by fusing the stationing data of each raster with the image data corresponding to the aligned pixels.

[0050] Finally, existing image defect identification methods are used to identify defect points in orthophotos. Based on the image data of the defect points and the fused raster data of each raster, a KD tree is used to search for the raster closest to the road defect point. The station number of the road defect point is determined according to its corresponding station number data, thereby achieving high-precision positioning of the station number of the defect point at the pixel level. This reduces the time spent locating road defect points, enables faster data-driven intervention, and minimizes maintenance time.

[0051] In this embodiment, optionally, in step 1, constructing the road centerline based on the orthophoto of the road and the design file includes:

[0052] Using orthophotos as the base layer, the design file is overlaid and registered with the orthophotos through multiple registration points in the orthophotos, and the road centerline is constructed based on the registered road registration data.

[0053] Specifically, when creating a road centerline, the orthophoto of the road can first be imported into the GIS software as the primary reference layer. Simultaneously, the road design file and all station numbers (in whole hundred points) should also be imported into the GIS software. The orthophoto and station numbers from the design file should then be overlaid in the GIS software. Next, using the orthophoto as the base layer, multiple pre-set registration points within the orthophoto are used to align the orthophoto, design file, and station numbers to ensure the accuracy of the subsequent creation of the road centerline and vertical offset lines.

[0054] After the layers are registered and aligned, the road centerline can be created based on the registered and aligned data. In this embodiment, the road centerline can be manually drawn directly using GIS software based on the registered and aligned data. Alternatively, the deviation between the road centerline in the orthophoto and the road centerline in the design file can be compared. If the deviation is within one pixel, the road centerline in the design file can be directly extracted and used, reducing the creation time of the road centerline and improving efficiency.

[0055] In this embodiment, registration points can be selected from important road features such as curves and intersections, and these registration points should be distributed as evenly as possible along the road. Simultaneously, the acquired orthophoto image can be appropriately cropped to exclude irrelevant areas far from the road and slopes.

[0056] In this embodiment, optionally, each hundredth station on the road is used as an initial reference point, including: correcting the position coordinates of each hundredth station based on on-site measured data.

[0057] It should be understood that there may be differences between the station numbers in the road design documents and those used during actual road operation. Therefore, the station number information for each hundredth station can be corrected based on the accurate location of the station numbers obtained on-site using high-precision RTK equipment. This ensures that the subsequently created vertical offset lines match the actual road conditions and guarantee the accuracy of station numbering for road defects.

[0058] In this embodiment, optionally, in step 2, multiple vertical offset lines are constructed along the road centerline, including:

[0059] Using each of the aforementioned hundredth-station points as initial reference points, draw perpendicular lines to the road centerline and determine the intersection point of each perpendicular line with the road centerline;

[0060] Multiple incremental marker points are evenly distributed along the road centerline between two adjacent intersections;

[0061] A vertical offset line is created perpendicular to the road centerline, using each incremental marker as a reference point. This vertical offset line covers the road width.

[0062] Specifically, such as Figure 2 As shown, when creating vertical offset lines, firstly, each 100-meter station point is used as an initial reference point to draw a perpendicular line to the road centerline. The intersection of this perpendicular line and the road centerline is then determined using GIS software as vertical marker points. These vertical marker points will mark precise 100-meter intervals on the road centerline, and errors will not propagate, thus preventing the typical error propagation characteristic of long linear road features and maintaining the accuracy of road segment stationing. Furthermore, these intersection points inherit the stationing information for their respective stations, which includes a prefix letter, kilometers, and meters, such as K001+900.

[0063] Then, 99 incremental markers can be created on the road centerline between two adjacent vertical markers, each representing a 1-meter increment. These 99 incremental markers are then assigned sequentially along the direction of increasing road station numbers, each given the smaller prefix letter and kilometer value from the two vertical markers. For example, the 99 incremental markers generated between the two vertical markers K001+900 and K002+000 are K001+901, K001+902, ..., K001+999.

[0064] Finally, using each incremental marker point as a reference point, create a vertical offset line perpendicular to the road centerline. The two ends of each vertical offset line extend to both sides of the road, forming a vertical segment covering the road width. After creating the vertical offset lines, determine the attribute value of each vertical offset line based on the assigned values ​​of the incremental marker points and the driving direction. This attribute value includes the station number and driving direction, such as K001+901 for uphill.

[0065] In this embodiment, optionally, in step 3, merging all rectangular buffers into a continuous strip-shaped buffer surface includes:

[0066] The system detects gaps and overlaps between rectangular buffers and processes these gaps and overlaps through automatic merging.

[0067] Specifically, after creating all the vertical offset lines, rectangular buffer zones of the same size and perpendicular to the road centerline can be created, using each vertical offset line as the center axis. The length of the rectangular buffer zone can be set to the road width, and the width can be set to 1 meter. Finally, all the created rectangular buffer zones are merged to generate the entire strip buffer surface.

[0068] During the creation of adjacent rectangular buffer zones, overlaps and gaps may occur due to road curvature or misalignment. Therefore, topology tools in GIS software can be used to locate gaps and overlaps between adjacent rectangular buffer zones. The width of these gaps or overlaps can be compared to a set tolerance threshold. If the width exceeds the threshold, a joint automatic merging process is used to process the gaps and overlaps, resulting in a strip-shaped buffer surface without redundant overlaps and with high consistency with the road route. This ensures the accuracy of subsequent raster data fusion.

[0069] In this embodiment, optionally, in step 5, generating the corresponding fused raster data includes:

[0070] The station data of the raster is fused with the image data of the corresponding pixels of the raster by band combination.

[0071] Specifically, after constructing the strip buffer area, the strip buffer area can be converted into a strip buffer raster map according to the resolution of the road orthophoto, so as to realize pixel-level fusion of image data and station data, thereby improving the accuracy of fused raster data.

[0072] After converting the strip buffer area into a strip buffer raster map, capture gratings can be set within it. These capture gratings are then used to align each grid in the strip buffer raster map with each cell in the orthophoto, minimizing misalignment during data fusion and further improving the accuracy of the fused raster data. After aligning the strip buffer raster map with the orthophoto pixel by pixel, the band combination function in GIS software can be used to fuse the station information of the grids in the strip buffer raster map with the image data of the aligned cells in the orthophoto, generating accurate and reliable fused raster data. This provides a foundation for precisely locating station signals of disease points.

[0073] The fused raster data includes not only the geospatial information and R, G, B information of the rasters, but also station information. Existing image-based road defect identification and localization methods can be used to locate the geospatial coordinates of road defect points in the orthophoto. Based on these geospatial coordinates, a KD-tree search algorithm can be used to find the raster closest to the road defect point from among numerous rasters. The station information in the fused raster data of this raster point can then be used to accurately locate the station number of the road defect point. This achieves high-precision localization of the station number of the defect point, thereby reducing the time spent locating road defects, enabling faster data-driven intervention, and minimizing maintenance time.

[0074] In this embodiment, optionally, it also includes:

[0075] A training set is constructed based on the fused raster data corresponding to each raster in the strip buffer surface;

[0076] Using the image data as input and the station number data as output, the constructed data regression model is trained using the training set;

[0077] When the KD tree fails to find the nearest grid cell to the road defect point, the station number of the road defect point is inferred from the image data of the road defect point using a trained data regression model.

[0078] It should be understood that, due to the limited range of the predefined strip buffer surface, the geographic space of road defect points identified by orthophotos may exceed the range of the strip buffer surface. For example, the identified defect points may not be on the road surface, but on the roadbed slope, or the searched grid may not have corresponding station information.

[0079] To this end, a training set can be constructed based on the fused raster data corresponding to each grid in the strip buffer area. This training set is then used to train a data regression model, which takes geospatial information and R, G, and B information as input and station information as output. During training, a random search cross-validation method can be used to adjust the model parameters to ensure that the model's coefficient of determination exceeds 0.9 and the mean absolute error is less than 1 meter.

[0080] When the distance between the nearest grid cell to a road defect point found using a KD-tree search exceeds a set threshold, the station number of the road defect point can be inferred from the image data of the road defect point using a trained data regression model. In this embodiment, the threshold can be set to 1 pixel.

[0081] In this embodiment, optionally, the data regression model is a random forest regression model. Specifically, the random forest regression model strikes a good balance between interpretability, accuracy, and computational efficiency. Therefore, in this embodiment, the random forest regression model is selected to estimate the station information of road defect points.

[0082] In this embodiment, optionally, the data regression model is constructed by combining the random forest regression model and the XGBoost model.

[0083] Specifically, in order to improve the overall accuracy of the model, the data regression model can be constructed by combining the random forest regression model and the XGBoost model, thereby leveraging the advantages of each model to improve the accuracy and stability of the station information estimation for road defect points.

[0084] Specifically, a stacked ensemble approach can be used to combine random forest regression models and XGBoost models in a hierarchical manner, using the predictions of one model as initial features input into the other. Alternatively, the predictions from random forest regression models and XGBoost models can be mixed, and their predictions can be averaged or weighted based on the performance of the two models.

[0085] A road defect point location system, comprising:

[0086] The centerline construction module is configured to construct road centerlines based on orthophotos and design files.

[0087] The offset line construction module is configured to use each hundred-station point on the road as an initial reference point, construct multiple vertical offset lines along the center line of the road, and assign station number data for each vertical offset line according to the corresponding hundred-station number.

[0088] The buffer surface construction module is configured to construct corresponding rectangular buffers with each vertical offset line as the central axis, and merge all rectangular buffers into a continuous strip buffer surface;

[0089] The grid division module is configured to divide the strip buffer surface into grids according to the resolution of the orthophoto, and to assign station data of each grid according to the vertical offset line corresponding to the rectangular buffer in which it is located.

[0090] The data fusion module is configured to align each grid of the strip buffer surface with the orthophoto pixel by pixel, and fuse the station data of each grid with the image data of the corresponding pixel to generate corresponding fused grid data.

[0091] The stationing module is configured to combine the image data of road defect points with the fused grid data of each grid, use a KD tree to search for the grid closest to the road defect point, and determine the station number of the road defect point based on its corresponding station number data.

[0092] Specifically, the positioning system includes a centerline construction module, an offset line construction module, a buffer surface construction module, a grid division module, a data fusion module, and a stationing module. The centerline construction module creates the road centerline by combining the road design documents and measured orthophotos. The offset line construction module uses each whole-hundred-station point as an initial reference point to construct multiple vertical offset lines perpendicular to the road centerline, and assigns these reference whole-hundred-station points to each vertical offset line based on their stationing information.

[0093] The buffer surface construction module can create rectangular buffers of equal size with each vertical offset line as the central axis, and merge all the created rectangular buffers into a unified strip buffer surface. The raster division module can divide the strip buffer surface into rasteres according to the image resolution of the orthophoto, and assign the station information of the whole hundred station points referenced by the vertical offset line corresponding to the rectangular buffer of the raster to each raster. The data fusion module can align all rasters with each cell of the orthophoto by setting the capture raster, thereby minimizing misalignment during data fusion and ensuring the accuracy of the fused raster data after fusing the station data of each raster with the image data corresponding to the aligned cells.

[0094] The stationing module can use existing image defect recognition methods to identify defect points in orthophotos. Based on the image data of the defect points and the fused grid data of each grid, it uses a KD tree to search for the grid closest to the road defect point and determines the station number of the road defect point according to its corresponding station number data. This achieves high-precision positioning of the station number to which the defect point belongs, thereby reducing the time spent locating road defect points, enabling faster data-driven intervention, and minimizing maintenance time.

[0095] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for locating road defect points using unmanned aerial vehicles (UAVs), characterized in that, include: Construct the road centerline based on orthophotos and design documents; Using each hundred-station point on the road as an initial reference point, multiple vertical offset lines are constructed along the centerline of the road, and station number data for each vertical offset line is assigned according to the corresponding hundred-station number. Construct corresponding rectangular buffers with each vertical offset line as the central axis, and merge all rectangular buffers into a continuous strip buffer surface; The strip buffer surface is divided into grids according to the resolution of the orthophoto, and the station data of each grid is assigned according to the vertical offset line corresponding to the rectangular buffer. Align each grid of the strip buffer surface with the orthophoto pixel by pixel, and fuse the station data of each grid with the image data of the corresponding pixel to generate the corresponding fused grid data; By combining image data of road defect points with fused raster data of each raster, a KD tree is used to search for the raster closest to the road defect point, and the station number of the road defect point is determined based on its corresponding station number data.

2. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, Constructing the road centerline based on orthophotos and design documents, including: Using orthophotos as the base layer, the design file is overlaid and registered with the orthophotos through multiple registration points in the orthophotos, and the road centerline is constructed based on the registered road registration data.

3. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, Using each hundredth-mile marker on the road as an initial reference point, including: The coordinates of each hundred-station point were corrected based on the actual on-site measurement data.

4. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, Multiple vertical offset lines are constructed along the centerline of the road, including: Using each of the aforementioned hundredth-station points as initial reference points, draw perpendicular lines to the road centerline and determine the intersection point of each perpendicular line with the road centerline; Multiple incremental marker points are evenly distributed along the road centerline between two adjacent intersections; A vertical offset line is created perpendicular to the road centerline, using each incremental marker as a reference point. This vertical offset line covers the road width.

5. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, Merge all rectangular buffers into a continuous strip buffer surface, including: The system detects gaps and overlaps between rectangular buffers and processes these gaps and overlaps through automatic merging.

6. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, Generate the corresponding fused raster data, including: The station data of the raster is fused with the image data of the corresponding pixels of the raster by band combination.

7. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, Also includes: A training set is constructed based on the fused raster data corresponding to each raster in the strip buffer surface; Using the image data as input and the station number data as output, the constructed data regression model is trained using the training set; When the KD tree fails to find the nearest grid cell to the road defect point, the station number of the road defect point is inferred from the image data of the road defect point using a trained data regression model.

8. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 7, characterized in that, The data regression model is a random forest regression model.

9. The method for locating road defect points based on unmanned aerial vehicles (UAVs) according to claim 7, characterized in that, The data regression model is constructed by combining the random forest regression model and the XGBoost model.

10. A road defect point location system, characterized in that, include: The centerline construction module is configured to construct road centerlines based on orthophotos and design files. The offset line construction module is configured to use each hundred-station point on the road as an initial reference point, construct multiple vertical offset lines along the center line of the road, and assign station number data for each vertical offset line according to the corresponding hundred-station number. The buffer surface construction module is configured to construct corresponding rectangular buffers with each vertical offset line as the central axis, and merge all rectangular buffers into a continuous strip buffer surface; The grid division module is configured to divide the strip buffer surface into grids according to the resolution of the orthophoto, and to assign station data of each grid according to the vertical offset line corresponding to the rectangular buffer in which it is located. The data fusion module is configured to align each grid of the strip buffer surface with the orthophoto pixel by pixel, and fuse the station data of each grid with the image data of the corresponding pixel to generate corresponding fused grid data. The stationing module is configured to combine the image data of road defect points with the fused grid data of each grid, use a KD tree to search for the grid closest to the road defect point, and determine the station number of the road defect point based on its corresponding station number data.

Images