A method and system for aligning detection data waveforms based on data feature matching

Abstract

This invention discloses a method and system for aligning detection data waveforms based on data feature matching, comprising: constructing a standard pull-out value data waveform for the entire track based on the catenary detection data of the track line; generating a first image contour as a standard data image template based on the standard pull-out value data in the standard pull-out value data waveform; generating a second image contour for the pull-out value data to be aligned; cutting the second image contour to obtain image slices; using a morphology-based template matching method, performing image matching between each image slice and the standard data image template to obtain a matching result for each image slice; filtering out valid matching sequences based on the matching results; performing an initial matching based on the feature point positions of the standard pull-out value data and the pull-out value data to be aligned, and the valid matching sequences; calibrating the initial matching results to obtain a final matching result. This invention achieves accurate alignment of detection data solely based on the inherent features of the catenary detection data itself.

Description

Technical Field

[0001] This invention relates to the field of overhead contact line inspection technology, and specifically to a method and system for aligning inspection data waveforms based on data feature matching. Background Technology

[0002] Existing technical solutions for achieving waveform alignment in track detection mainly rely on correlation functions (such as Pearson correlation coefficient) to evaluate the correlation of data and on some pre-known positional information to achieve waveform alignment.

[0003] In the field of overhead contact line inspection, due to its installation specifications, similar inspection data may exist between multiple stations. However, existing technical solutions may match the same data at multiple locations simultaneously. This is because the correlation coefficient method is based on the linear correlation of data, and the magnitude of the correlation coefficient depends on the approximation of the data to the local mean. This calculation method can lead to a very high linear correlation between the pull-out values ​​at different locations. Therefore, this calculation method needs to limit the matching range of the data. Matching within a specific range will not lead to the problem of matching multiple different results. Thus, this calculation method must limit the registration range and can only achieve registration locally.

[0004] Therefore, the existing technical solutions mentioned above cannot achieve accurate alignment of contact network detection data based solely on the correlation coefficient method, resulting in poor alignment performance. Furthermore, these solutions cannot adapt to situations where the data to be aligned itself lacks usable location information. Summary of the Invention

[0005] The technical problem to be solved by the present invention is that the existing methods for aligning waveforms of overhead contact line detection data mainly rely on the correlation coefficient method, which has a linear relationship based on the detection data and relies on some pre-known position information to achieve local registration, thus failing to achieve accurate alignment of the overhead contact line detection data and resulting in poor alignment effect.

[0006] The purpose of this invention is to provide a method and system for aligning detection data waveforms based on data feature matching. This invention relies solely on the inherent features of the contact wire detection data and utilizes image contour template matching technology to generate image contours using numerical values. By matching the contours of the data slices to be aligned onto standard data contours, this invention makes greater use of the nonlinear relationships of the detection data compared to the correlation coefficient method. It can perform matching based on the similarity of the contours and regress to the position where the Euclidean distance between the two contours is minimized, thus improving the registration accuracy. Subsequently, by searching the longest ordered subsequence to remove a few mismatched positions, preliminary positioning can be achieved. Then, through the identification of inflection points and special regions, accurate alignment of the detection data can be achieved.

[0007] This invention is achieved through the following technical solution:

[0008] In a first aspect, the present invention provides a detection data waveform alignment method based on data feature matching, the method comprising:

[0009] Based on the overhead contact line inspection data, construct the standard pull-out value data waveform for the entire line;

[0010] Based on the standard pull-out value data in the standard pull-out value data waveform, a first image contour is generated as a standard data image template; a second image contour is generated from the pull-out value data to be aligned; the second image contour is cut to obtain image slices;

[0011] Using a morphology-based template matching method, each image slice is matched with a standard data image template to obtain the matching result for each image slice; and based on the matching result, valid matching sequences are selected.

[0012] Based on the feature point positions and valid matching sequences of the standard pull-out value data and the pull-out value data to be aligned, an initial matching is performed to obtain the initial matching result; and the initial matching result is calibrated to obtain the final matching result.

[0013] Furthermore, the standard pull-out value data waveforms for the entire line are constructed by using linear interpolation to supplement the contact network inspection data, including:

[0014] Based on the overhead contact line inspection data, the pull-out value of the positioning points and the span between adjacent positioning points are obtained for the entire line.

[0015] Calculate the standard pull-out value between any two positioning points based on the pull-out value of the positioning point and the span between adjacent positioning points.

[0016] The linear interpolation method is used to set sampling points at preset distances to generate standard pull-out value data waveforms for the entire line.

[0017] Furthermore, the formula for calculating the standard pull-out value data is as follows:

[0018]

[0019] In the formula, s i s1 is the standard pull-out value data; s2 is the pull-out value of the first positioning point; s2 is the pull-out value of the second positioning point; and D is the span between the first positioning point and the second positioning point.

[0020] Furthermore, the second image contour is generated after preprocessing the data to be aligned and pulled out. The preprocessing includes:

[0021] The outlier data to be aligned is filtered out, and abnormal jump values ​​are removed to obtain the preprocessed data to be aligned.

[0022] Furthermore, the method of using morphology-based template matching to match each image slice with a standard data image template to obtain the matching result for each image slice includes:

[0023] Each image slice template is matched against a standard data image template:

[0024] If a match is successful, the slice data is expanded until the number of expansions is less than a set threshold, and the image slice is retained. The expansion of the slice data involves increasing the length of the selected slice data by M points before and after, and then judging whether the two regions are similar. If they are similar, the expansion is performed again.

[0025] If a match fails, the slice data is reduced, and the matching results before and after the slice data are used to supplement the image, and the image matching is performed again.

[0026] Furthermore, the effective matching sequence is the longest ordered subsequence selected from the matching results.

[0027] Furthermore, the feature point positions of the standard pull-out value data and the pull-out value data to be aligned are determined by using the feature points of the data waveform, taking the inflection points of the standard pull-out value data waveform along the entire line as the feature point positions.

[0028] Furthermore, based on the feature point positions and valid matching sequences of the standard pull-out value data and the pull-out value data to be aligned, an initial matching is performed to obtain the initial matching result; and the initial matching result is calibrated to obtain the final matching result, including:

[0029] Using a valid matching sequence, the mileage information of the pull-out value data to be aligned is rewritten to complete the initial matching and obtain the initial matching result.

[0030] Find matching sequences that are similar to the feature point mileage information in the standard pull-out data in the initial matching result, and perform calibration to obtain the final matching result.

[0031] Secondly, the present invention provides a detection data waveform alignment system based on data feature matching, which is used to implement the aforementioned detection data waveform alignment method based on data feature matching; the system includes:

[0032] Construct standard waveform units for the entire line, which are used to construct standard pull-out value data waveforms for the entire line based on the catenary inspection data of the track line;

[0033] The template generation unit is used to generate a first image contour as a standard data image template based on the standard pull-out value data in the standard pull-out value data waveform; and to generate a second image contour for the pull-out value data to be aligned.

[0034] The image cutting unit is used to cut the second image contour to obtain image slices;

[0035] The waveform matching and filtering unit is used to perform image matching between each image slice and a standard data image template using a morphology-based template matching method to obtain the matching result of each image slice; and to filter out the valid matching sequences based on the matching results.

[0036] The precise matching unit is used to perform an initial match based on the feature point positions and valid matching sequences of the standard pull-out value data and the pull-out value data to be aligned, and to obtain the initial matching result; and to calibrate the initial matching result to obtain the final matching result.

[0037] Furthermore, the waveform matching and filtering unit includes a waveform matching subunit and a filtering subunit;

[0038] The waveform matching and filtering unit includes a waveform matching subunit and a filtering subunit;

[0039] The waveform matching subunit is used to perform image matching between each image slice template and the standard data image template. If the match is successful, the slice data is expanded until the number of expansions is less than a set threshold, and the image slice is retained. The expansion of the slice data involves increasing the length of the selected slice data by M points before and after, and then judging whether the two regions are similar. If they are similar, the slice data is expanded again. If the match fails, the slice data is reduced, and the matching results before and after the slice data are used to supplement the data. The image matching is then performed again.

[0040] The filtering sub-unit is used to select the longest ordered subsequence from the matching results as the valid matching sequence, based on the matching results.

[0041] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0042] This invention discloses a method and system for aligning detection data waveforms based on data feature matching. This invention relies solely on the inherent features of the contact network detection data, without introducing mileage information. Utilizing image contour template matching technology, it can generate image contours numerically. By matching the contours of the data slices to be aligned onto standard data contours, it leverages the nonlinear relationships of the detection data more effectively than the correlation coefficient method. It can match based on contour similarity and regress to the position with the minimum Euclidean distance between two contours, improving registration accuracy. Subsequently, by searching the longest ordered subsequence to remove a few mismatched positions, preliminary positioning can be achieved. Finally, through inflection point and special region identification, precise alignment of the detection data is achieved. Attached Figure Description

[0043] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0044] Figure 1 This is a flowchart of a detection data waveform alignment method based on data feature matching according to the present invention;

[0045] Figure 2 This is a schematic diagram of local matching using the existing correlation coefficient method;

[0046] Figure 3 This is a schematic diagram illustrating the existence of multiple similar regions when using the existing correlation coefficient method for calculation.

[0047] Figure 4 This is the image slice template to be aligned in Embodiment 1 of the present invention;

[0048] Figure 5 This is the standard data image template of Embodiment 1 of the present invention;

[0049] Figure 6 This is a schematic diagram of the matching results in Embodiment 1 of the present invention;

[0050] Figure 7 This is a detailed flowchart of steps 2 and 3 in Embodiment 1 of the present invention;

[0051] Figure 8 This is a diagram showing the effect of finding inflection point feature points in Embodiment 1 of the present invention.

[0052] Figure 9 This is a schematic diagram of the initial matching result in Embodiment 1 of the present invention;

[0053] Figure 10 This is a schematic diagram of the final matching result in Embodiment 1 of the present invention;

[0054] Figure 11 This is a block diagram of a detection data waveform alignment system based on data feature matching according to the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of this invention are only for explaining this invention and are not intended to limit this invention.

[0056] In the field of overhead contact line inspection, due to its installation specifications, similar inspection data (i.e., extremely high similarity of pull-out value inspection data) may exist between multiple stations. However, existing inspection data alignment methods based on Pearson correlation coefficient may result in the same data segment matching multiple locations simultaneously, such as... Figure 3As shown; because the correlation coefficient method is based on the linear correlation of data, the magnitude of the correlation coefficient depends on the degree of approximation of the data to the local mean. This calculation method can lead to very high linear correlation between the pulled values ​​at different locations. Therefore, this calculation method needs to limit the matching range of the data. Matching within a specific range will not lead to multiple different matching results. Therefore, this calculation method must limit the registration range and can only achieve registration locally, such as... Figure 2 As shown. The formula for calculating the correlation coefficient in the detection data alignment method based on the Pearson correlation coefficient is:

[0057]

[0058] In the formula, r is the correlation coefficient, n is the number of standard data and test data, and X i Y is the i-th detection value of the standard data. i Let i be the i-th detection value of the data to be matched. The standard data mean. This represents the mean of the data to be matched.

[0059] In addition, although existing related technologies have achieved the function of aligning detection data waveforms using track geometry parameters, they require input of detection data and related mileage information, specify the maximum range of data mileage error, and mainly use similarity calculation functions. These functions are poorly applicable in the field of catenary detection with high repeatability. In the detection of catenary geometry parameters, there are a large number of similar waveforms, and this type of method is not very effective in catenary detection alignment.

[0060] Therefore, existing methods for aligning waveforms of overhead contact line inspection data mainly rely on the correlation coefficient method. These methods suffer from limitations due to their linear relationship with the inspection data and reliance on partially known positional information for local registration, resulting in inaccurate alignment and poor alignment quality. To address these issues, this invention proposes a waveform alignment method and system based on data feature matching. This invention utilizes only the inherent characteristics of the overhead contact line inspection data, eliminating the need for mileage information or initial / endpoint location information. Employing image contour template matching technology, it generates image contours numerically. By matching the contours of the data slices to be aligned onto standard data contours, it leverages the nonlinear relationship of the inspection data more effectively than the correlation coefficient method. It matches contours based on their similarity, regressing to the position with the minimum Euclidean distance between the two contours, thus improving registration accuracy. Subsequently, the longest ordered subsequence is used to remove a few mismatched positions for initial localization. Finally, through inflection point and special region identification, accurate alignment of the inspection data is achieved.

[0061] Example 1

[0062] like Figure 1As shown, the present invention provides a detection data waveform alignment method based on data feature matching, the method comprising:

[0063] Step 1: Based on the overhead contact line inspection data, construct the standard pull-out value waveform for the entire line; specifically, use linear interpolation to complete the overhead contact line inspection data, including:

[0064] Based on the overhead contact line inspection data, the pull-out values ​​(manually measured values) of the positioning points along the entire line and the span between adjacent positioning points are obtained.

[0065] Calculate the standard pull-out value between any two positioning points based on the pull-out value of the positioning point and the span between adjacent positioning points.

[0066] A sampling point is set according to a preset T meters using linear interpolation to generate the standard pull-out value waveform for the entire line.

[0067] Assume the pull-out value of the first positioning point is s1, the pull-out value of the second positioning point is s2, the span between the two positioning points is D, and the standard pull-out value between these two positioning points is S. i The calculation formula is:

[0068]

[0069] In the formula, s i s1 is the standard pull-out value data; s2 is the pull-out value of the first positioning point; s2 is the pull-out value of the second positioning point; and D is the span between the first positioning point and the second positioning point.

[0070] Step 2: Based on the standard pull-out value data in the standard pull-out value waveform, generate a first image contour as a standard data image template, such as... Figure 5 As shown; generate a second image contour from the data to be aligned and pulled out; cut the second image contour into multiple parts to obtain multiple image slices, as shown. Figure 4 As shown;

[0071] Specifically, the second image contour is generated after preprocessing the data to be aligned and pulled out. The preprocessing includes:

[0072] To align the pull-out value data, it is necessary to first filter out outliers and remove abnormal jump values ​​to obtain the preprocessed pull-out value data to be aligned, denoted as K. Then, an image contour is generated from it, denoted as the second image contour.

[0073] Step 3: Using a morphology-based template matching method, each image slice is matched against a standard data image template to obtain the matching result for each image slice, such as... Figure 6 As shown; and based on the matching results, select the valid matching sequences;

[0074] Specifically, the method of using morphology-based template matching to match each image slice with a standard data image template to obtain the matching result for each image slice includes:

[0075] Each image slice template is matched against a standard data image template:

[0076] If a match is successful, the slice data is expanded until the number of expansions is less than a set threshold, and the image slice is retained. The expansion of the slice data involves increasing the length of the selected slice data by M points before and after, and then judging whether the two regions are similar. If they are similar, the expansion is performed again.

[0077] If a match fails, the slice data is reduced, and the matching results before and after the slice data are used to supplement the image, and the image matching is performed again.

[0078] The above technical solution takes into account that due to the overhead contact line construction, the same slice may correspond to multiple locations in the original data. Therefore, this invention designs the above matching method to ensure the correctness of the slice corresponding to the data. When an image slice data can be matched in the standard data image, the length of the selected slice data is expanded by M points before and after it, and the similarity between the two regions is judged again. If they are still similar, they are expanded again. When the number of expansions exceeds a set threshold N, the match is considered valid. For data slices that fail to match, they can be supplemented by subsequent matching results before and after the data.

[0079] The flowcharts for steps 2 and 3 are as follows: Figure 7 As shown.

[0080] As a further implementation, the effective matching sequence is the longest ordered subsequence selected from the matching results.

[0081] A second verification is performed on the image matching results from step 3. Since the image templates for the slices to be aligned are matched sequentially in step 3, the correct matching sequences are also arranged in order. The longest sequential subsequence is selected from the obtained matching results as the valid matching sequence U. n .

[0082] Step 4: Perform initial matching based on the feature point positions and valid matching sequences of the standard pull-out value data and the pull-out value data to be aligned, and obtain the initial matching result; then calibrate the initial matching result to obtain the final matching result.

[0083] Specifically, the feature point positions of the standard pull-out value data and the pull-out value data to be aligned are determined by using the feature points of the data waveform, taking the inflection points of the standard pull-out value data waveform along the entire line as the feature point positions.

[0084] This is because overhead contact lines are typically erected in a zigzag pattern, resulting in strong regularity in their data. For the pull-out value, the positioning point often represents an inflection point in the waveform. Utilizing these characteristic points of the data waveform improves waveform alignment accuracy. Taking the contact wire pull-out value as an example: find the "0" point of the derivative of its detection data as the characteristic point; the effect is shown in the image. Figure 8 As shown, the feature point positions of the standard data and the data to be aligned are recorded.

[0085] Specifically, based on the feature point positions and valid matching sequences of the standard pull-out value data and the pull-out value data to be aligned, an initial matching is performed to obtain the initial matching result; and the initial matching result is then calibrated to obtain the final matching result, including:

[0086] Using a valid matching sequence, the mileage information of the data to be aligned is rewritten to complete the initial matching, and the initial matching result is obtained, such as... Figure 9 As shown;

[0087] Find matching sequences that are similar to the feature point mileage information in the standard pull-out data from the initial matching results, perform calibration to obtain the final matching results, such as... Figure 10 As shown.

[0088] The method of this invention only utilizes the consistency of the waveform images before and after the data itself, without introducing mileage information of the detection data, and automatically finds the position of the current detection data, which reduces the error dimension and eliminates mileage error; at the same time, the method of this invention utilizes the characteristics of the catenary construction, and achieves accurate alignment of the detection data by identifying inflection points and special areas.

[0089] Example 2

[0090] like Figure 11 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment provides a detection data waveform alignment system based on data feature matching. This system is used to implement the detection data waveform alignment method based on data feature matching in Embodiment 1. The system includes:

[0091] Construct standard waveform units for the entire line, which are used to construct standard pull-out value data waveforms for the entire line based on the catenary inspection data of the track line;

[0092] The template generation unit is used to generate a first image contour as a standard data image template based on the standard pull-out value data in the standard pull-out value data waveform; and to generate a second image contour for the pull-out value data to be aligned.

[0093] The image cutting unit is used to cut the second image contour to obtain image slices;

[0094] The waveform matching and filtering unit is used to perform image matching between each image slice and a standard data image template using a morphology-based template matching method to obtain the matching result of each image slice; and to filter out the valid matching sequences based on the matching results.

[0095] The precise matching unit is used to perform an initial match based on the feature point positions and valid matching sequences of the standard pull-out value data and the pull-out value data to be aligned, and to obtain the initial matching result; and to calibrate the initial matching result to obtain the final matching result.

[0096] As a further implementation, the waveform matching and filtering unit includes a waveform matching subunit and a filtering subunit;

[0097] The waveform matching and filtering unit includes a waveform matching subunit and a filtering subunit;

[0098] The waveform matching subunit is used to perform image matching between each image slice template and the standard data image template. If the match is successful, the slice data is expanded until the number of expansions is less than a set threshold, and the image slice is retained. The expansion of the slice data involves increasing the length of the selected slice data by M points before and after, and then judging whether the two regions are similar. If they are similar, the slice data is expanded again. If the match fails, the slice data is reduced, and the matching results before and after the slice data are used to supplement the data. The image matching is then performed again.

[0099] The filtering sub-unit is used to select the longest ordered subsequence from the matching results as the valid matching sequence, based on the matching results.

[0100] The execution process of each unit can be carried out according to the steps of the detection data waveform alignment method based on data feature matching in Embodiment 1, and will not be described in detail in this embodiment.

[0101] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0102] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0103] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0104] These computer program instructions 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 of the function specified in one or more boxes.

[0105] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for aligning detection data waveforms based on data feature matching, characterized in that, The method includes: Based on the overhead contact line inspection data, construct the standard pull-out value data waveform for the entire line; Based on the standard pull-out value data in the standard pull-out value data waveform, a first image contour is generated as a standard data image template; a second image contour is generated from the pull-out value data to be aligned; the second image contour is cut to obtain image slices; Using a morphology-based template matching method, each image slice is matched with a standard data image template to obtain the matching result for each image slice; and based on the matching result, valid matching sequences are selected. Based on the feature point positions of the standard pull-out value data and the pull-out value data to be aligned, and the effective matching sequence, an initial matching is performed to obtain an initial matching result; and the initial matching result is calibrated to obtain a final matching result. The standard pull-out value data waveform for the entire line is constructed by using linear interpolation to complete the contact network inspection data, including: Based on the overhead contact line inspection data, the pull-out value of the positioning points and the span between adjacent positioning points are obtained for the entire line. Based on the pull-out value of the positioning point and the span between adjacent positioning points, calculate the standard pull-out value data between any two positioning points; Using linear interpolation, sampling points are set at preset distances to generate standard pull-out value data waveforms for the entire line; The formula for calculating the standard pull-out value data is as follows: In the formula, s i s1 is the standard pull-out value data; s2 is the pull-out value of the first positioning point; s2 is the pull-out value of the second positioning point; and D is the span between the first positioning point and the second positioning point.

2. The detection data waveform alignment method based on data feature matching according to claim 1, characterized in that, The second image contour is generated after preprocessing the data to be aligned and pulled out. The preprocessing includes: The outlier data to be aligned is filtered out, and abnormal jump values ​​are removed to obtain the preprocessed data to be aligned.

3. The detection data waveform alignment method based on data feature matching according to claim 1, characterized in that, The aforementioned morphology-based template matching method performs image matching between each image slice and a standard data image template to obtain the matching result for each image slice, including: Each image slice template is matched against a standard data image template: If a match is successful, the slice data is expanded until the number of expansions is less than a set threshold, and the image slice is retained. The expansion of the slice data involves increasing the length of the selected slice data by M points before and after, and then judging whether the two regions are similar. If they are similar, the expansion is performed again. If a match fails, the slice data is reduced, and the matching results before and after the slice data are used to supplement the image, and the image matching is performed again.

4. The detection data waveform alignment method based on data feature matching according to claim 1, characterized in that, The effective matching sequence is the longest ordered subsequence selected from the matching results.

5. The detection data waveform alignment method based on data feature matching according to claim 1, characterized in that, The feature point positions of the standard pull-out value data and the pull-out value data to be aligned are determined by using the feature points of the data waveform, taking the inflection points of the standard pull-out value data waveform along the entire line as the feature point positions.

6. The detection data waveform alignment method based on data feature matching according to claim 1, characterized in that, Based on the feature point positions of the standard pull-out value data and the pull-out value data to be aligned, and the effective matching sequence, an initial matching is performed to obtain the initial matching result. The initial matching result is then calibrated to obtain the final matching result, including: Using a valid matching sequence, the mileage information of the pull-out value data to be aligned is rewritten to complete the initial matching and obtain the initial matching result. Find matching sequences that are similar to the feature point mileage information in the standard pull-out data in the initial matching result, and perform calibration to obtain the final matching result.

7. A detection data waveform alignment system based on data feature matching, characterized in that, The system includes: Construct standard waveform units for the entire line, which are used to construct standard pull-out value data waveforms for the entire line based on the catenary inspection data of the track line; The template generation unit is used to generate a first image contour as a standard data image template based on the standard pull-out value data in the standard pull-out value data waveform; and to generate a second image contour for the pull-out value data to be aligned. An image cutting unit is used to cut the second image contour to obtain image slices; The waveform matching and filtering unit is used to perform image matching between each image slice and a standard data image template using a morphology-based template matching method to obtain the matching result of each image slice; and to filter out the valid matching sequence based on the matching result. The precise matching unit is used to perform an initial match based on the feature point positions of the standard pull-out value data and the pull-out value data to be aligned, as well as the effective matching sequence, to obtain an initial matching result; and to calibrate the initial matching result to obtain a final matching result. The standard pull-out value data waveform for the entire line is constructed by using linear interpolation to complete the contact network inspection data, including: Based on the overhead contact line inspection data, the pull-out value of the positioning points and the span between adjacent positioning points are obtained for the entire line. Based on the pull-out value of the positioning point and the span between adjacent positioning points, calculate the standard pull-out value data between any two positioning points; Using linear interpolation, sampling points are set at preset distances to generate standard pull-out value data waveforms for the entire line; The formula for calculating the standard pull-out value data is as follows: In the formula, s i s1 is the standard pull-out value data; s2 is the pull-out value of the first positioning point; s2 is the pull-out value of the second positioning point; and D is the span between the first positioning point and the second positioning point.

8. The detection data waveform alignment system based on data feature matching according to claim 7, characterized in that, The waveform matching and filtering unit includes a waveform matching subunit and a filtering subunit; The waveform matching and filtering unit includes a waveform matching subunit and a filtering subunit; The waveform matching subunit is used to perform image matching between each image slice template and the standard data image template. If the matching is successful, the slice data is expanded until the number of expansions is less than a set threshold, and the image slice is retained. The expansion of the slice data involves increasing the length of the selected slice data by M points before and after, and then judging whether the two regions are similar. If they are similar, the slice data is expanded again. If the matching fails, the slice data is reduced, and the matching results before and after the slice data are used to supplement the data. Image matching is then performed again. A filtering subunit is used to select the longest ordered subsequence from the matching results as a valid matching sequence, based on the matching results.

Images