Method, device, equipment, medium and program for correcting embankment mileage data

Abstract

The present disclosure relates to a method, device, equipment, medium and program for correcting subgrade mileage data, comprising: acquiring a target radar image, performing feature extraction on the target radar image to obtain target features; performing image labeling on the target radar image according to the target features to obtain position data of a corresponding single complete bridge in the target radar image; performing regression calculation on the position data by using a preset linear regression model to obtain benchmark mileage data; and performing mileage correction on preset bridge sequence information according to the benchmark mileage data to obtain corrected mileage data.

Description

Technical Field

[0001] This disclosure relates to the field of deep learning technology, and in particular to a method, apparatus, device, medium and program for correcting roadbed mileage data. Background Technology

[0002] When using vehicle-mounted ground-penetrating radar (GPR) technology for roadbed inspection, the antenna is suspended and fixedly mounted at the bottom of the inspection vehicle, maintaining a relatively fixed distance from the top surface of the sleepers for rapid inspection. Since vehicle-mounted radar can detect at speeds up to 100 km / h, long-distance detection leads to significant cumulative errors in the radar data's mileage, which substantially affects the spatial accuracy of subsequent processing. Therefore, mileage correction is a necessary step in processing railway roadbed radar inspection data.

[0003] Current railway subgrade ground-penetrating radar (GPR) detection largely relies on single-period data acquisition, only obtaining the structural morphology of the current strata and failing to determine whether the subgrade deformation has reached a stable state, thus reducing the accuracy of subgrade location identification. Some railways conduct periodic inspections, inspecting the same section of the line at regular intervals and determining the subgrade location by comparing strata changes across multiple periods. However, periodic inspections can generate massive amounts of data, making automated and efficient processing and interpretation of radar data difficult. Furthermore, multi-period data requires precise registration and accurate positioning, and current manual interpretation methods are inefficient and time-consuming. Therefore, improving the accuracy and efficiency of mileage correction for railway subgrade GPR monitoring data is a pressing issue. Summary of the Invention

[0004] This disclosure provides a method, apparatus, equipment, medium, and program for correcting roadbed mileage data, the main purpose of which is to solve the problems of insufficient accuracy and low efficiency in correcting roadbed mileage data.

[0005] Firstly, this disclosure provides a method for correcting roadbed mileage data, including: Acquire a target radar image, and extract features from the target radar image to obtain target features; Based on the target features, the target radar image is annotated to obtain the location data of a single complete bridge in the target radar image; The location data is regressed using a preset linear regression model to obtain baseline mileage data; The preset bridge sequence information is corrected based on the reference mileage data to obtain corrected mileage data.

[0006] In some embodiments, the step of extracting features from the target radar image to obtain target features includes: The target radar image is subjected to multi-layer convolution to obtain a convolution feature matrix; Multi-level self-attention calculation is performed on the convolutional feature matrix to obtain a single feature map at each level; The target features are obtained by performing feature fusion on the single feature map.

[0007] In some embodiments, the feature fusion of the single feature map to obtain the target feature includes: The single feature map is traversed through a sliding window using a pre-constructed convolutional kernel to obtain the pixel value corresponding to the single feature map. The pixel values ​​within the convolution kernel are multiplied by the pixel values ​​corresponding to the single feature map, and then summed to obtain the pixel values ​​corresponding to the target feature. The pixel values ​​corresponding to the target features are mapped to a preset target feature map to obtain the target features.

[0008] In some embodiments, the step of annotating the target radar image according to the target features to obtain the location data of a single complete bridge in the target radar image includes: Based on the target features and the preset bridge features, feature matching is performed to obtain the single complete bridge corresponding to the target radar image; Edge detection is performed on the single complete bridge to obtain the region contour of the single complete bridge; Based on the region contour, the single complete bridge is image-annotated to obtain the location data of the corresponding single complete bridge in the target radar image.

[0009] In some embodiments, the step of image annotation of the single complete bridge based on the region contour to obtain the location data of the corresponding single complete bridge in the target radar image includes: Mark the starting and ending points of the single complete bridge according to the outline of the region; The mileage difference is obtained by calculating the mileage difference between the two ends of the single complete bridge. The mileage difference is used as the location data of the corresponding single complete bridge in the target radar image.

[0010] In some embodiments, the step of using a preset linear regression model to perform regression calculations on the location data to obtain baseline mileage data includes: The start and end mileages of the location data are marked to obtain mileage calibration points; The calibration mileage is obtained by performing calibration calculations on the mileage calibration point and the preset actual calibration point; The calibration mileage is input into a preset linear regression model for regression calculation to obtain baseline mileage data.

[0011] Secondly, this disclosure provides a correction device for roadbed mileage data, characterized in that it includes: The feature extraction module is used to acquire target radar images and extract features from the target radar images to obtain target features; An image annotation module is used to annotate the target radar image according to the target features to obtain the location data of a single complete bridge in the target radar image; The regression calculation module is used to perform regression calculations on the location data using a preset linear regression model to obtain baseline mileage data. The mileage correction module is used to correct the mileage of the preset bridge sequence information based on the reference mileage data to obtain corrected mileage data.

[0012] Thirdly, this disclosure provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the foregoing aspects.

[0013] Fourthly, this disclosure provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the methods described in the above aspects.

[0014] Fifthly, this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the methods described in the above aspects.

[0015] This invention utilizes ground-penetrating radar (GPR) technology to preprocess target radar images, improving image quality and clarity. Feature extraction from the preprocessed images allows for more accurate target identification and reduces false recognition rates. Image recognition and annotation technology enables bridge target identification, enhancing accuracy and efficiency. Compared to traditional manual interpretation methods, this significantly simplifies the workflow, reduces labor costs, and facilitates subsequent location data analysis and application. A pre-defined linear regression model is used to perform regression calculations on the location data, resolving the issue of significant discrepancies between identified mileage and actual distance in existing technologies, thereby improving correction accuracy and reducing errors. Furthermore, this invention first acquires bridge sequence information for each subsequent radar data period; then, mileage data correction is achieved simply by registering the bridge sequence information with reference data, greatly improving registration accuracy and mileage correction efficiency. Attached Figure Description

[0016] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings: Figure 1This is a flowchart illustrating a method for correcting roadbed mileage data provided in an embodiment of the present disclosure.

[0017] Figure 2 Mileage data of a single complete bridge in a radar image provided in an embodiment of this disclosure.

[0018] Figure 3 A linear relationship curve is provided for embodiments of this disclosure.

[0019] Figure 4 This is a schematic diagram of a roadbed mileage data correction device provided in an embodiment of the present disclosure.

[0020] Figure 5 This is a schematic diagram of the structure of an electronic device for a roadbed mileage data correction device provided in an embodiment of this disclosure.

[0021] In the accompanying drawings, the same parts are referred to by the same reference numerals, and the drawings are not drawn to scale. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solutions of this disclosure, and to fully understand and implement the process of how this disclosure applies technical means to solve technical problems and achieve corresponding technical effects, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, not all embodiments. The embodiments of this disclosure and the various features within them can be combined with each other without conflict, and the resulting technical solutions are all within the protection scope of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort should fall within the protection scope of this disclosure.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] This application provides a method for correcting roadbed mileage data. The execution subject of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the system provided in this application: a server, a terminal, etc. In other words, the method for correcting roadbed mileage data can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks, and big data and artificial intelligence platforms.

[0025] Example 1 Reference Figure 1 The diagram shown is a flowchart illustrating a method for correcting roadbed mileage data according to an embodiment of the present invention.

[0026] In this embodiment, the method for correcting roadbed mileage data includes: S1. Acquire a target radar image, and extract features from the target radar image to obtain target features.

[0027] In this embodiment of the invention, different ground-penetrating radars are selected to acquire relevant target radar images based on target characteristics and requirements. For example, for long-range, high-resolution imaging requirements, high-performance synthetic aperture radar can be selected to ensure that high-quality radar images can be acquired.

[0028] Radar images can also be obtained through professional radar data providers or research institutions. For example, for remote sensing data such as synthetic aperture radar, it is possible to obtain them from official channels such as the National Meteorological Administration and the Satellite Remote Sensing Center. A series of image preprocessing steps are performed on the obtained radar images, such as geometric correction and denoising using median filtering, to ensure that high-quality and accurate radar images are obtained.

[0029] In this embodiment of the invention, the step of extracting features from the target radar image to obtain target features includes: The target radar image is subjected to multi-layer convolution to obtain a convolution feature matrix; Multi-level self-attention calculation is performed on the convolutional feature matrix to obtain a single feature map at each level; The target features are obtained by performing feature fusion on the single feature map.

[0030] In detail, this invention utilizes a convolutional neural network to perform convolution operations on target radar images. Convolution increases the receptive field of the convolution kernel, effectively extracting features from the target image. Through a series of convolutional layers and the ReLU activation function, the abstraction level of the features is gradually improved. After convolution, average pooling is performed on the convolution matrix, averaging the convolution matrices within each pooling window to obtain a pooling matrix. This reduces the spatial dimension of the convolution matrix while preserving the most important shape features. Furthermore, multi-layer self-attention calculations are performed using two independent LSTM units to obtain feature matrices at different levels. These feature matrices are then fused to obtain a fused feature map, enabling more accurate extraction of target features from the radar image. This helps improve the accuracy of railway roadbed radar image recognition. The target features include texture features, shape features, etc.

[0031] In this embodiment of the invention, the step of performing feature fusion on the single feature map to obtain the target feature includes: The single feature map is traversed through a sliding window using a pre-constructed convolutional kernel to obtain the pixel value corresponding to the single feature map. The pixel values ​​within the convolution kernel are multiplied by the pixel values ​​corresponding to the single feature map, and then summed to obtain the pixel values ​​corresponding to the target feature. The pixel values ​​corresponding to the target features are mapped to a preset target feature map to obtain the target features.

[0032] In detail, the convolution kernel is a matrix used for multi-level convolution with a single feature map. The size of the convolution kernel, the element values, and the stride all affect the result of multi-level convolution. The sliding window traversal is a way of implementing convolution. The convolution kernel moves on the single feature map like a sliding window. Each time it moves, it multiplies with the feature values ​​in the window and sums them to obtain a new pixel value. This new pixel value corresponds to a point on the target feature map. All pixel values ​​obtained through convolution are mapped to the preset target feature map according to their positions on the single feature map to obtain the target feature.

[0033] In this embodiment of the invention, the target radar image obtained by ground penetrating radar technology is preprocessed to improve the quality and clarity of the radar image. Feature extraction is then performed on the preprocessed radar image to more accurately identify the target and reduce the false identification rate. Feature extraction helps to distinguish subtle differences between different targets, thereby improving the accuracy and reliability of subsequent identification.

[0034] S2. Based on the target features, perform image annotation on the target radar image to obtain the location data of a single complete bridge in the target radar image.

[0035] In this embodiment of the invention, the prior art makes it difficult to completely label or identify a bridge in a single image due to the variable length of the bridge and the limited width of the image slice. The present invention uses image labeling technology, which combines the target features of the bridge in the radar image, to label the left and right parts of the bridge. The mileage difference between the right and left sides is used to determine the bridge of different lengths. The location data of a single complete bridge includes basic information such as the coordinates of the bridge's center point, the bridge's length, and its width.

[0036] In this embodiment of the invention, the step of annotating the target radar image according to the target features to obtain the location data of a single complete bridge in the target radar image includes: Based on the target features and the preset bridge features, feature matching is performed to obtain the single complete bridge corresponding to the target radar image; Edge detection is performed on the single complete bridge to obtain the region contour of the single complete bridge; Based on the region contour, the single complete bridge is image-annotated to obtain the location data of the corresponding single complete bridge in the target radar image.

[0037] In detail, this invention compares target features in a target radar image with preset bridge features, including texture features and shape features. Then, a preset target detection model is used to identify the bridge in the target bridge image, and a single complete bridge matching the preset bridge features is identified in the target radar image. Edge detection techniques such as the Canny edge detector and the Sobel operator are used to identify areas with drastic changes in grayscale values ​​in the image, thereby outlining the edges of the target bridge. Based on this, the single complete bridge is labeled in the image, thereby extracting the location data of the target bridge.

[0038] In this embodiment of the invention, the step of annotating the single complete bridge according to the region contour to obtain the location data of the corresponding single complete bridge in the target radar image includes: Mark the starting and ending points of the single complete bridge according to the outline of the region; The mileage difference is obtained by calculating the mileage difference between the two ends of the single complete bridge. The mileage difference is used as the location data of the corresponding single complete bridge in the target radar image.

[0039] In detail, based on edge detection, this invention can perform image annotation on a single complete bridge in a target radar image. Image annotation typically involves drawing rectangles, polygons, or other shapes on the image to mark the starting and ending points of a single complete bridge. For example, this invention uses polygons (such as quadrilaterals or more complex shapes) to more accurately mark the starting and ending points of a single complete bridge. By performing a subtraction operation on the mileage data corresponding to the starting and ending points of the marked single complete bridge, the absolute value of the result is converted into a mileage difference, and the mileage difference is used as the position data of the corresponding single complete bridge in the target radar image.

[0040] In this embodiment of the invention, image recognition and annotation technology can be used to identify bridge targets, thereby improving the accuracy and efficiency of identification. Compared with traditional manual interpretation methods, it greatly simplifies the workflow and reduces labor costs. Based on the identification results, the bridge targets are assigned corresponding location information, which facilitates the subsequent analysis and application of location data.

[0041] S3. Use a preset linear regression model to perform regression calculations on the location data to obtain baseline mileage data.

[0042] In this embodiment of the invention, the regression calculation refers to marking the start and end mileages of the target location data to obtain mileage calibration points, and inputting the mileage calibration points and preset actual calibration points into a linear regression model for regression calculation to obtain baseline mileage data. The baseline mileage data includes the center mileage of the bridge, the start and end mileages of the bridge, and the length of the bridge.

[0043] In this embodiment of the invention, the step of using a preset linear regression model to perform regression calculations on the location data to obtain baseline mileage data includes: The start and end mileages of the location data are marked to obtain mileage calibration points; The calibration mileage is obtained by performing calibration calculations on the mileage calibration point and the preset actual calibration point; The calibration mileage is input into a preset linear regression model for regression calculation to obtain baseline mileage data.

[0044] In detail, the baseline mileage data can be obtained by performing regression calculations using the following formula:

[0045]

[0046] in, The reference mileage data, The slope of the linear regression formula. The intercept of the linear regression formula is... For the calibration mileage, To preset the actual mileage, The number of location data. The average value of the calibration mileage. The average value of the preset actual mileage.

[0047] In detail, the mileage calibration point refers to the midpoint of the calibration mileage between the start mileage and the end mileage of the location data, the actual calibration point refers to the midpoint of the actual mileage from the start mileage to the end mileage, the calibration calculation refers to the subtraction operation between the mileage calibration point and the preset actual calibration point, and the calibration mileage refers to the current mileage minus the difference between the midpoint of the actual mileage and the midpoint of the calibration mileage.

[0048] Furthermore, after obtaining the mileage calibration points and actual calibration points, they are used as input data and fed into a preset linear regression model. The linear regression model will learn and fit an optimal linear relationship curve based on these input data. This curve can be as close as possible to the actual mileage data. Through this linear relationship curve, the mileage data corresponding to any given location can be predicted, thereby obtaining the benchmark mileage data.

[0049] For example, such as Figure 2 The radar image shown has an original mileage range of 0-100. Three target bridges were found. The calibrated mileage and the actual mileage were input into a preset linear regression model for regression calculation, resulting in a regression formula. This allows for the fitting of an optimal linear relationship curve. Substituting x=0 and x=100 into the regression formula respectively, the mileage data in the radar image can be corrected to the new mileage, i.e., the baseline mileage data. The linear relationship curve is shown below. Figure 3 As shown.

[0050] In this embodiment of the invention, a preset linear regression model is used to perform regression calculations on the location data, which solves the problem in the prior art that the identified mileage differs greatly from the actual distance, thereby improving the correction accuracy and reducing errors.

[0051] S4. Based on the reference mileage data, the preset bridge sequence information is mileage corrected to obtain corrected mileage data.

[0052] In this embodiment of the invention, the preset bridge sequence information includes basic information such as bridge length and bridge spacing.

[0053] In this embodiment of the invention, the reference mileage data is integrated with the preset bridge sequence information to obtain a reference dataset. The mileage data of a single complete bridge corresponding to the target radar image is matched with the reference dataset. By performing regression calculation on the mileage data of the single complete bridge and the reference dataset, the mapping relationship between them is obtained. The mileage data of the single complete bridge is corrected according to the mapping relationship to obtain corrected mileage data, thereby realizing mileage data correction.

[0054] In detail, the baseline mileage data is integrated with the preset bridge sequence information to form a baseline dataset containing bridge identification and location data. The registration of subsequent ground-penetrating radar data for multiple periods is carried out by matching the bridge sequence information extracted from each subsequent period of data with the baseline dataset to obtain the bridges in each period of data that correspond to the bridges in the baseline dataset. For the successfully matched bridges, the correspondence between each subsequent period of data and the baseline data is established to obtain the corrected mileage data and realize mileage correction.

[0055] In this embodiment of the invention, for each subsequent radar data period, after obtaining the bridge sequence information, the mileage data can be corrected simply by registering the bridge sequence information with the reference data, which greatly improves the registration accuracy and the efficiency of mileage correction.

[0056] This invention utilizes ground-penetrating radar (GPR) technology to preprocess target radar images, improving image quality and clarity. Feature extraction from the preprocessed images allows for more accurate target identification and reduces false recognition rates. Image recognition and annotation technology enables bridge target identification, enhancing accuracy and efficiency. Compared to traditional manual interpretation methods, this significantly simplifies the workflow, reduces labor costs, and facilitates subsequent location data analysis and application. A pre-defined linear regression model is used to perform regression calculations on the location data, resolving the issue of significant discrepancies between identified mileage and actual distance in existing technologies, thereby improving correction accuracy and reducing errors. Furthermore, this invention first acquires bridge sequence information for each subsequent radar data period; then, mileage data correction is achieved simply by registering the bridge sequence information with reference data, greatly improving registration accuracy and mileage correction efficiency.

[0057] Example 2 like Figure 4 The diagram shown is a schematic diagram of a roadbed mileage data correction device provided in an embodiment of this application.

[0058] The roadbed mileage data correction device 100 of the present invention can be installed in an electronic device. Depending on the functions implemented, the roadbed mileage data correction device 100 may include a feature extraction module 101, an image annotation module 102, a regression calculation module 103, and a mileage correction module 104. The module described in this invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0059] In this embodiment, the functions of each module / unit are as follows: The feature extraction module 101 is used to acquire a target radar image and extract features from the target radar image to obtain target features; The image annotation module 102 is used to annotate the target radar image according to the target features to obtain the location data of a single complete bridge in the target radar image. The regression calculation module 103 is used to perform regression calculation on the location data using a preset linear regression model to obtain baseline mileage data. The mileage correction module 104 is used to perform mileage correction on the preset bridge sequence information based on the reference mileage data to obtain corrected mileage data.

[0060] Example 3 Figure 5 This is a schematic diagram of the electronic device for a roadbed mileage data correction device provided in an embodiment of this application.

[0061] Based on the above embodiments, this embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the above embodiments.

[0062] In some embodiments of this example, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the steps of the method described in the above embodiments.

[0063] In some embodiments of this example, a computer program product is provided, including a computer program / instructions, characterized in that the computer program, when executed by a processor, implements the steps of the method described in the above embodiments.

[0064] The processor may include, but is not limited to, one or more processors or microprocessors. Each processor may be implemented as an Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic component, for executing the methods in the above embodiments.

[0065] Computer-readable storage media can be implemented by any type of volatile or non-volatile storage device or a combination thereof. Computer-readable storage media may include, but are not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, and computer storage media (e.g., hard disks, floppy disks, solid-state drives, removable disks, CD-ROMs, DVD-ROMs, Blu-ray discs, etc.).

[0066] Computer-readable storage media may also store at least one computer-executable program / instruction, such as computer-readable instructions. Computer-readable storage media include, but are not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Computer-readable storage media may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, a non-transitory computer-readable storage medium may be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions stored on the computer-readable storage medium, the various methods described above can be performed.

[0067] In addition, the computer device may include (but is not limited to) a data bus, an input / output (I / O) bus, a display, and input / output devices (e.g., keyboard, mouse, speakers, etc.).

[0068] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product / computer program product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0069] In the embodiments provided in this disclosure, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0070] It should be noted that, in this disclosure, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element limited by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0071] While the embodiments disclosed herein are as described above, the foregoing content is merely for the purpose of facilitating understanding of this disclosure and is not intended to limit this disclosure. Any person skilled in the art to which this disclosure pertains may make any modifications and changes in form and detail of the implementation without departing from the spirit and scope of this disclosure; however, the scope of patent protection of this disclosure shall still be determined by the scope defined in the appended claims.

Claims

1. A method for correcting roadbed mileage data, characterized in that, The method includes: Acquire a target radar image, and extract features from the target radar image to obtain target features; Based on the target features, the target radar image is annotated to obtain the location data of a single complete bridge in the target radar image; The location data is regressed using a preset linear regression model to obtain baseline mileage data; The preset bridge sequence information is corrected based on the reference mileage data to obtain corrected mileage data.

2. The method according to claim 1, characterized in that, The step of extracting features from the target radar image to obtain target features includes: The target radar image is subjected to multi-layer convolution to obtain a convolution feature matrix; Multi-level self-attention calculation is performed on the convolutional feature matrix to obtain a single feature map at each level; The target features are obtained by performing feature fusion on the single feature map.

3. The method according to claim 2, characterized in that, The step of fusing features from the single feature map to obtain target features includes: The single feature map is traversed through a sliding window using a pre-constructed convolutional kernel to obtain the pixel value corresponding to the single feature map. The pixel values ​​within the convolution kernel are multiplied by the pixel values ​​corresponding to the single feature map, and then summed to obtain the pixel values ​​corresponding to the target feature. The pixel values ​​corresponding to the target features are mapped to a preset target feature map to obtain the target features.

4. The method according to claim 1, characterized in that, The step of annotating the target radar image based on the target features to obtain the location data of a single complete bridge in the target radar image includes: Based on the target features and the preset bridge features, feature matching is performed to obtain the single complete bridge corresponding to the target radar image; Edge detection is performed on the single complete bridge to obtain the region contour of the single complete bridge; Based on the region contour, the single complete bridge is image-annotated to obtain the location data of the corresponding single complete bridge in the target radar image.

5. The method according to claim 4, characterized in that, The step of annotating the individual complete bridge according to the region contour to obtain the location data of the corresponding individual complete bridge in the target radar image includes: Mark the starting and ending points of the single complete bridge according to the outline of the region; The mileage difference is obtained by calculating the mileage difference between the two ends of the single complete bridge. The mileage difference is used as the location data of the corresponding single complete bridge in the target radar image.

6. The method according to claim 1, characterized in that, The step of using a preset linear regression model to perform regression calculations on the location data to obtain baseline mileage data includes: The start and end mileages of the location data are marked to obtain mileage calibration points; The calibration mileage is obtained by performing calibration calculations on the mileage calibration point and the preset actual calibration point; The calibration mileage is input into a preset linear regression model for regression calculation to obtain baseline mileage data.

7. A device for correcting roadbed mileage data, characterized in that, include: The feature extraction module is used to acquire target radar images and extract features from the target radar images to obtain target features; An image annotation module is used to annotate the target radar image according to the target features to obtain the location data of a single complete bridge in the target radar image; The regression calculation module is used to perform regression calculations on the location data using a preset linear regression model to obtain baseline mileage data. The mileage correction module is used to correct the mileage of the preset bridge sequence information based on the reference mileage data to obtain corrected mileage data.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 6.

Images