Track steel rail full-section point cloud matching method

Abstract

The invention discloses a rail steel rail full-section point cloud matching method. The method comprises the following steps: step 1, data scanning; step 2, data screening; step 3, acquiring a rail head coordinate point; step 4, determining a rail head and rail waist segmentation point; 5, dividing the rail head and the rail waist; 6, rail head data abnormity is judged; 7, dividing rail webs; step 8, acquiring coordinate data of track feature points; step 9, determining a fitting circle; step 10, carrying out rotation fitting; step 11, matching; and 12, acquiring parameters. By designing the accurate steel rail full-section point cloud matching method, point cloud matching is carried out according to the rail full section and the rail design standard, so that whether the rail is different from the standard rail in shape after being used or not is obtained, and the geometric parameters of the rail are calculated; the method is high in response speed, high in detection precision, low in equipment installation requirement condition and good in robustness; the method is high in practicability and has high popularization significance.

Description

Technical Field

[0001] The present invention relates to the technical field of train tracks, and in particular to a method for matching full-section point clouds of railway tracks. Background Art

[0002] With the continuous development of my country's transportation industry, railway transportation has become an integral and important part of it. While ensuring the development of transportation, safety issues cannot be ignored and are placed in front of technicians. Tracks, as the "highways" for trains to run, play a very important role in railways. The state of the tracks directly affects whether the trains can run safely. In the traditional track maintenance process, people can only rely on the naked eye to observe whether the track state is normal. Not only is the workload huge, but it is also easy to make mistakes during the maintenance process. For this reason, people began to use automated equipment and designed automatic measurement methods to replace manual measurement and maintenance of train tracks.

[0003] Most existing measurement methods use non-contact measurement methods, which mainly use laser displacement sensors and structured light vision, and are based on the principle of triangulation. The position of the track profile can be determined based on the geometric relationship between the laser and the camera and the imaging position on the laser linear array CCD. Some people also use the inertial reference method to measure, using the instantaneous acceleration change of the contact point between the train wheel and the track to determine the height change of the left and right tracks, and then combine the inner side profile of the track obtained by the laser linear array camera with the inertial navigation data, and calculate other parameters of the track based on the geometric relationship of the installation position.

[0004] However, the track cross-section profile obtained by the non-contact measurement method may have sparse sampling points, and the measurement data may be affected by ambient light interference; the acquisition speed is slow and not suitable for longer lines, it consumes too much manpower, the measurement cost is too high, and maintenance is inconvenient. The inertial reference rule has high requirements on train speed and equipment installation, and the accuracy of determining track abnormalities is easily affected by speed and produces errors. Summary of the invention

[0005] Based on this, it is necessary to provide a point cloud matching method for the entire cross-section of rails to address the shortcomings of the existing technology.

[0006] A method for matching point clouds of a full section of a rail, comprising the following steps:

[0007] Step 1: Data scanning: Use a laser profiler to scan the left and right rails of the track to obtain point cloud data of its cross section;

[0008] Step 2: Data screening, counting the number of scan point coordinates in the point cloud data and performing judgment screening. When the number of scan point coordinates is less than the set parameter value, it is judged as scan missing and no processing is performed. When the number of scan point coordinates is greater than or equal to the set parameter value, the next step of processing is performed.

[0009] Step 3: Get the coordinate point of the rail head, arrange the original data in descending order in the Y-axis direction, find n consecutive points in the point cloud data whose vertical distance is less than the judgment value a, determine them as the top point of the rail head and get the coordinates of the corresponding point, where the judgment value a can determine that the point cloud is not a noise point on a horizontal line;

[0010] Step 4: Determine the rail head and rail waist segmentation point. Due to the installation angle of the profile scanning device, the distance between the rail head and the rail waist will be greater than the set value. That is, if the distance between two adjacent scanning points is greater than the set value, it can be said that the point with a larger Y coordinate is the segmentation point between the rail head and the rail waist. The set value is related to the specific installation angle of the device and is not a fixed value. The obtained point coordinates and data are compared with the track design standard for judgment. If the collected data status is normal, the coordinates of the rail head and rail waist segmentation point are calculated values. If the data status is abnormal, the 36mm below the rail head in the track design standard is used as the rail head and rail waist segmentation point.

[0011] Step 5: Segment the rail head and rail waist. Segment the rail head and rail waist according to the coordinates of the rail head and rail waist segmentation point. The point whose Y-axis coordinate value in the point cloud data is greater than or equal to the Y-axis coordinate value of the rail head and rail waist segmentation point is the rail head point, and the point whose Y-axis coordinate value is less than the Y-axis coordinate value of the rail head and rail waist segmentation point is the rail waist point. Remove the noise points in the rail head.

[0012] Step 6: Determine the abnormality of the rail head data, obtain the feature point data of the rail head, and arrange the maximum and minimum coordinate values ​​in descending order according to the Y-axis coordinate and the X-axis coordinate; set an area value. If the rail head is wider than the maximum value in the area value, the rail head is too wide; if it is lower than the minimum value in the area value, the rail head is too narrow;

[0013] Step 7: Rail waist division: first, take the two points with the largest and smallest X-axis coordinate values ​​among the rail head points as references, delete the redundant data points on both sides of the rail waist, and then use the clustering method according to the X-axis direction points to divide the rail waist into the left side of the rail waist and the right side of the rail waist;

[0014] Step 8: Obtain the coordinate data of the track feature points, and update the rail waist width according to the number of scanning points on the left side and the right side of the rail waist. When the number of scanning points on the left side and the right side of the rail waist is greater than the set parameter value, take the minimum value of the left side of the rail waist minus the right side of the rail waist and take the center of the two points as the center of the minimum rail waist spacing. If only one side is greater than the set parameter value, take the extreme value of the rail waist depression greater than one side minus 8.25mm as the center of the minimum rail waist spacing. After removing the noise data and the data with compensation collection problems, obtain the coordinate data of the track feature points and save them;

[0015] Step 9: Determine the fitting circle. According to the coordinates of the rail head corner points in the track feature point coordinate data and the radius parameters in the track design standard, determine the fitting circle at the rail head, and obtain the fitting circle at the rail waist and the fitting circle of the rail waist and rail bottom curve in the same way;

[0016] Step 10: Rotation fitting: Rotation fitting is performed on the entire track scanning section based on the rail waist and rail bottom curve fitting circle and the track design standard. The coordinate data of the track feature points are obtained again after each rotation.

[0017] Step 11: Matching: when the center of the fitting circle of the rail waist and rail bottom curve is less than the set parameter value of the track design standard, the track matching is completed, and the coordinate data of the track feature points are updated;

[0018] Step 12: Obtain parameters, and obtain the required orbital geometric parameters based on the saved orbital feature point coordinate data.

[0019] Furthermore, in step 2, when the number of scanning point coordinates is greater than the set parameter value, interpolation method can be used to remove points for downsampling to reduce the computing pressure and improve the computing speed.

[0020] Furthermore, in step 4, if there are multiple groups of two adjacent scanning points whose distances are greater than the set parameter value in the point cloud data, these points are arranged in descending order in the Y-axis direction, and the scanning point with the largest Y-axis coordinate value is used as the rail head and rail waist segmentation point and the coordinates of the corresponding points are obtained.

[0021] Furthermore, in step 6, when the track turns, the track will have a rail bottom slope, and the entire track will tilt to the left or right, so that the height distances from the highest point of the rail head to the maximum and minimum points on the X-axis will be different. In this case, the laser profiler will have a partial missing scan of the track interface; the data anomaly caused by the missing scan can be compensated according to the track design standard, that is, the maximum and minimum values ​​of the X-axis of the rail head part are given according to the corresponding data in the track design standard.

[0022] Furthermore, in step 9, if the rail waist data is not good, the fitting circle at the rail waist is obtained according to the given value in the track design standard; if the rail waist and rail bottom data are problematic, the rail waist and rail bottom curve fitting circle is obtained according to the ordinate of the fitting point, which should be one-third of the distance from the far end of the rail bottom to the abscissa of the fitting point plus the ordinate of the far end of the rail bottom.

[0023] Furthermore, after step 10, a vertical calibration step is provided, in which a fitting circle at the rail waist is used for vertical calibration. Since the distance from the fitting circle to the rail bottom is fixed, the distance is compared with the track design standard in the vertical state to avoid abnormal rail bottom data. When the difference is too large after comparison, a secondary vertical adjustment is performed to make the rail vertical. Then the characteristic value of the rail is recalculated.

[0024] To sum up, the beneficial effect of the full-section point cloud matching method of the track steel rail of the present invention is that: by designing an accurate full-section point cloud matching method of the rail, the pain points of the current track measurement in terms of portability, accuracy, using less equipment, low cost, and adaptability to various environments are solved, and point cloud matching is performed according to the full section of the track and the track design standard, so as to obtain whether the shape of the track is different from the standard track after use, and calculate the geometric parameters of the track. This method not only has the advantages of fast response speed and high detection accuracy, but also has low requirements for equipment installation and good robustness; the present invention is highly practical and has a strong promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the laser profiler scanning the track profile in the present invention;

[0026] Figure 2 A scanned image of the point cloud profile of the track section in the present invention;

[0027] Figure 3 A parameter diagram of the track design standard in the present invention;

[0028] Figure 4 This is the effect diagram of the track section point cloud matching in the present invention. DETAILED DESCRIPTION

[0029] In order to further understand the features, technical means and specific objectives and functions achieved by the present invention, and analyze the advantages and spirit of the present invention, a further understanding is obtained through the following detailed description of the present invention in combination with the accompanying drawings and specific implementation methods.

[0030] like Figures 1 to 4 As shown, it is a method for matching the full cross-section point cloud of a rail track of the present invention. It should be noted that the marked parameters in the drawings of the specification are only exemplary and reference data and do not represent the precise values ​​of the parameters in this embodiment. The method includes the following steps:

[0031] Step 1: Data scanning: Use a laser profiler to scan the left and right rails of the track to obtain point cloud data of its cross section;

[0032] Step 2: Data screening, counting the number of coordinates of the scanned points in the point cloud data and performing judgment and screening. When the number of coordinates of the scanned points is less than the set parameter value, it is judged as scan missing and no processing is performed. When the number of coordinates of the scanned points is greater than or equal to the set parameter value, the next step of processing is performed. When the number of coordinates of the scanned points is greater than the set parameter value, interpolation method can be used to remove points for downsampling to reduce the computing pressure and improve the computing speed. In this embodiment, the set parameter value is based on the specific parameters of the installed equipment, and the set parameter value will be different for different equipment;

[0033] Step 3: Get the coordinate point of the rail head, arrange the original data in descending order in the Y-axis direction, find n consecutive points in the point cloud data whose vertical distance is less than the judgment value a, determine them as the top points of the rail head and get the coordinates of the corresponding points. The judgment value a can determine that the point cloud is not a noise point on a horizontal line. During the selection of continuous points, if the vertical distance is less than the judgment value a, it will make it difficult to find the rail head. If it is greater than the judgment value a, it will result in too many noise points.

[0034] Step 4: Determine the dividing point between the rail head and the rail waist. After obtaining the coordinate point of the rail head, due to the installation angle of the profile scanning device, the distance between the rail head and the rail waist will be greater than the set value. That is, if the distance between two adjacent scanning points is greater than the set value, it can be said that the point with a larger Y coordinate is the dividing point between the rail head and the rail waist. The set value is related to the specific installation angle of the device and is not a fixed value.

[0035] If there are multiple sets of two adjacent scanning points whose distance is greater than the set parameter value in the point cloud data, these points are arranged in descending order in the Y-axis direction, and the scanning point with the largest Y-axis coordinate value is used as the rail head and rail waist segmentation point and the coordinates of the corresponding points are obtained;

[0036] The obtained point coordinates and data are compared with the track design standards for judgment. If the collected data is in normal state, the coordinates of the rail head and rail waist division point are calculated values. If the data is abnormal, the point 36mm below the rail head in the track design standard is used as the rail head and rail waist division point.

[0037] Step 5: Segment the rail head and rail waist. Segment the rail head and rail waist according to the coordinates of the rail head and rail waist segmentation point. The point whose Y-axis coordinate value in the point cloud data is greater than or equal to the Y-axis coordinate value of the rail head and rail waist segmentation point is the rail head point, and the point whose Y-axis coordinate value is less than the Y-axis coordinate value of the rail head and rail waist segmentation point is the rail waist point. Remove the noise points in the rail head.

[0038] Step 6: Determine the abnormality of the rail head data, obtain the feature point data of the rail head, and arrange the maximum and minimum coordinate values ​​in descending order according to the Y-axis coordinate and the X-axis coordinate; set an area value. If the rail head is wider than the maximum value in the area value, the rail head is too wide; if it is lower than the minimum value in the area value, the rail head is too narrow;

[0039] When the track turns, there will be a rail bottom slope, and the entire track will tilt to the left or right, so that the height distance from the highest point of the rail head to the maximum and minimum points of the X-axis will be different. In this case, the laser profiler will scan part of the track interface. The data anomaly caused by this scan deficiency can be compensated according to the track design standard, that is, the maximum and minimum values ​​of the X-axis of the rail head part are given according to the corresponding data in the track design standard.

[0040] Step 7: Rail waist division: first, take the two points with the largest and smallest X-axis coordinate values ​​among the rail head points as references, delete the redundant data points on both sides of the rail waist, and then use the clustering method according to the X-axis direction points to divide the rail waist into the left side of the rail waist and the right side of the rail waist;

[0041] Step 8: Obtain the coordinate data of the track feature points, and update the rail waist width according to the number of scanning points on the left and right sides of the rail waist. When the number of scanning points on the left and right sides of the rail waist is greater than the set parameter value, the minimum value of the left side of the rail waist minus the right side of the rail waist is taken as the center of the minimum rail waist spacing. If only one side is greater than the set parameter value, the center of the minimum rail waist spacing is taken as the extreme value of the rail waist depression greater than one side minus 8.25mm. After removing the noise data and the data with compensation collection problems, obtain the coordinate data of the track feature points such as rail amount, rail angle, rail head key value, rail waist key point and save them;

[0042] Step 9: Determine the fitting circle. According to the coordinates of the rail head corner points in the track feature point coordinate data and the radius parameters in the track design standard, determine the fitting circle at the rail head, and obtain the fitting circle at the rail waist and the fitting circle of the rail waist and rail bottom curve in the same way;

[0043] If the rail waist data is not good, the fitting circle at the rail waist is obtained according to the given value in the track design standard; if the rail waist and rail bottom data are problematic, the fitting circle of the rail waist and rail bottom curve is obtained according to the ordinate of the fitting point, which should be one-third of the distance from the far end of the rail bottom to the abscissa of the fitting point plus the ordinate of the far end of the rail bottom;

[0044] Step 10: Rotation fitting: Rotation fitting is performed on the entire track scanning section based on the rail waist and rail bottom curve fitting circle and the track design standard. The coordinate data of the track feature points are obtained again after each rotation.

[0045] Step 11: Vertical calibration: Use the fitting circle at the rail waist for vertical calibration. Since the distance from the fitting circle to the rail bottom is fixed, compare it with the track design standard in the vertical state to avoid abnormal rail bottom data. If the difference is too large after comparison, perform secondary vertical adjustment to make the rail vertical; then recalculate the characteristic value of the rail.

[0046] Step 12: Matching: when the center of the fitting circle of the rail waist and rail bottom curve is less than the set parameter value of the track design standard, the track matching is completed, and the coordinate data of the track feature points are updated;

[0047] Step 13: Obtain parameters, and obtain the required orbital geometric parameters based on the saved orbital feature point coordinate data.

[0048] To sum up, the beneficial effect of the full-section point cloud matching method of the track steel rail of the present invention is that: by designing an accurate full-section point cloud matching method of the rail, the pain points of the current track measurement in terms of portability, accuracy, using less equipment, low cost, and adaptability to various environments are solved, and point cloud matching is performed according to the full section of the track and the track design standard, so as to obtain whether the shape of the track is different from the standard track after use, and calculate the geometric parameters of the track. This method not only has the advantages of fast response speed and high detection accuracy, but also has low requirements for equipment installation and good robustness; the present invention is highly practical and has a strong promotion significance.

[0049] The above-mentioned embodiment only expresses one implementation method of the invention, and its description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the inventive concept, which all belong to the protection scope of the invention. Therefore, the protection scope of the invention patent shall be based on the attached claims.

Claims

1. A method for matching point clouds of the entire cross section of a rail, characterized in that: The steps include: Step 1: Data scanning: Use a laser profiler to scan the left and right rails of the track to obtain point cloud data of its cross section; Step 2: Data screening, counting the number of scan point coordinates in the point cloud data and performing judgment screening. When the number of scan point coordinates is less than the set parameter value, it is judged as scan missing and no processing is performed. When the number of scan point coordinates is greater than or equal to the set parameter value, the next step of processing is performed. Step 3: Get the coordinate point of the rail head, arrange the original data in descending order in the Y-axis direction, find n consecutive points in the point cloud data whose vertical distance is less than the judgment value a, determine them as the top point of the rail head and get the coordinates of the corresponding point, where the judgment value a can determine that the point cloud is not a noise point on a horizontal line; Step 4: Determine the rail head and rail waist segmentation point. Due to the installation angle of the profile scanning device, the distance between the rail head and the rail waist will be greater than the set value. That is, if the distance between two adjacent scanning points is greater than the set value, it can be said that the point with a larger Y coordinate is the segmentation point between the rail head and the rail waist. The set value is related to the specific installation angle of the device and is not a fixed value. The obtained point coordinates and data are compared with the track design standard for judgment. If the collected data status is normal, the coordinates of the rail head and rail waist segmentation point are calculated values. If the data status is abnormal, the 36mm below the rail head in the track design standard is used as the rail head and rail waist segmentation point. Step 5: Segment the rail head and rail waist. Segment the rail head and rail waist according to the coordinates of the rail head and rail waist segmentation point. The point whose Y-axis coordinate value in the point cloud data is greater than or equal to the Y-axis coordinate value of the rail head and rail waist segmentation point is the rail head point, and the point whose Y-axis coordinate value is less than the Y-axis coordinate value of the rail head and rail waist segmentation point is the rail waist point. Remove the noise points in the rail head. Step 6: Determine the abnormality of the rail head data, obtain the feature point data of the rail head, and arrange the maximum and minimum coordinate values ​​in descending order according to the Y-axis coordinate and the X-axis coordinate; set an area value. If the rail head is wider than the maximum value in the area value, the rail head is too wide; if it is lower than the minimum value in the area value, the rail head is too narrow; Step 7: Rail waist division: first, take the two points with the largest and smallest X-axis coordinate values ​​among the rail head points as references, delete the redundant data points on both sides of the rail waist, and then use the clustering method according to the X-axis direction points to divide the rail waist into the left side of the rail waist and the right side of the rail waist; Step 8: Obtain the coordinate data of the track feature points, and update the rail waist width according to the number of scanning points on the left side and the right side of the rail waist. When the number of scanning points on the left side and the right side of the rail waist is greater than the set parameter value, take the minimum value of the left side of the rail waist minus the right side of the rail waist and take the center of the two points as the center of the minimum rail waist spacing. If only one side is greater than the set parameter value, take the extreme value of the rail waist depression greater than one side minus 8.25mm as the center of the minimum rail waist spacing. After removing the noise data and the data with compensation collection problems, obtain the coordinate data of the track feature points and save them; Step 9: Determine the fitting circle. According to the coordinates of the rail head corner points in the track feature point coordinate data and the radius parameters in the track design standard, determine the fitting circle at the rail head, and obtain the fitting circle at the rail waist and the fitting circle of the rail waist and rail bottom curve in the same way; Step 10: Rotation fitting: Rotation fitting is performed on the entire track scanning section based on the rail waist and rail bottom curve fitting circle and the track design standard. The coordinate data of the track feature points are obtained again after each rotation. Step 11: Matching: when the center of the fitting circle of the rail waist and rail bottom curve is less than the set parameter value of the track design standard, the track matching is completed, and the coordinate data of the track feature points are updated; Step 12: Obtain parameters, obtain the required orbital geometric parameters based on the saved orbital feature point coordinate data.

2. The method for matching the full-section point cloud of a track rail according to claim 1, characterized in that: In step 2, when the number of scanning point coordinates is greater than the set parameter value, interpolation method can be used to remove points for downsampling to reduce computing pressure and improve computing speed.

3. The method for matching the full-section point cloud of a track rail according to claim 1, characterized in that: In step 4, if there are multiple sets of two adjacent scanning points whose distance is greater than the set parameter value in the point cloud data, these points are arranged in descending order in the Y-axis direction, and the scanning point with the largest Y-axis coordinate value is used as the rail head and rail waist segmentation point and the coordinates of the corresponding points are obtained.

4. The method for matching the full-section point cloud of a track rail according to claim 1, characterized in that: In step 6, when the track turns, the track will have a rail bottom slope, and the entire track will tilt to the left or right, so that the height distance from the highest point of the rail head to the maximum and minimum points of the X-axis will be different. In this case, the laser profiler will have a partial missing scan of the track interface; the data anomaly caused by the missing scan can be compensated according to the track design standard, that is, the maximum and minimum values ​​of the X-axis of the rail head part are given according to the corresponding data in the track design standard.

5. The method for matching the full-section point cloud of a track rail according to claim 1, characterized in that: In step 9, if the rail waist data is not good, the fitting circle at the rail waist is obtained according to the given value in the track design standard; if the rail waist and rail bottom data are problematic, the rail waist and rail bottom curve fitting circle is obtained according to the ordinate of the fitting point, which should be one-third of the distance from the far end of the rail bottom to the abscissa of the fitting point plus the ordinate of the far end of the rail bottom.

6. The method for matching the full-section point cloud of a track rail according to claim 1, characterized in that: After step 10, a vertical calibration step is further provided, in which a fitting circle at the rail waist is used for vertical calibration. Since the distance from the fitting circle to the rail bottom is fixed, the distance is compared with the track design standard in a vertical state to avoid abnormal rail bottom data. When the difference is too large after comparison, a secondary vertical adjustment is performed to make the rail vertical. Then the characteristic value of the rail is recalculated.

Images