A method for detecting train appearance by laser three-dimensional scanning

Abstract

The present invention relates to the field of visual inspection, and discloses a method for laser three-dimensional scanning inspection of train appearance. The method comprises the following steps: S1. Scanning a train using a laser radar and obtaining a three-dimensional point data set of the train appearance; S2. Dividing the three-dimensional point data set into a plurality of columns of first data blocks i according to a first spacing along the scanning direction; S3. Dividing each column of first data blocks i into a plurality of rows of second data blocks ij according to a second spacing along the vertical direction of the scanning direction; S4. Performing three-dimensional point distance distribution statistics on each second data block ij; S5. Obtaining a plurality of second data blocks Aij by the first scanning, and obtaining a plurality of second data blocks Bij by the second scanning. If the change amplitude of the three-dimensional point distance distribution between the second data block Bij and the second data block Aij is less than a threshold value, the appearance of the train area corresponding to the second data block ij does not change. This scheme is used to solve the problem of low accuracy in identifying abnormal situations during train appearance inspection.

Description

Technical Field

[0001] The present invention relates to the field of visual inspection, and more specifically, to a method for laser three-dimensional scanning inspection of train appearance. Background Art

[0002] As a common means of transportation in modern society, trains provide convenience for modern people's travel and play an important role in social development. As a common means of transportation, the safe operation of trains is extremely important, so trains need to undergo routine safety inspections every trip or every day. Routine safety inspections have complicated steps and high inspection frequency, so automated safety inspections are the focus of research in related fields.

[0003] Trains have a complex structure and many key components (pantographs, bogies). During the operation of a train, the environment, aging of parts, human operation and other conditions may cause abnormalities in the train body, such as foreign objects floating on the roof, missing key components, deformation or changes in the position of components. The interference of foreign objects, missing key components and position changes will affect the safe operation of the train. Therefore, in routine safety inspections, the inspection of the train appearance is an indispensable and important link.

[0004] In the prior art, the more common method of train appearance detection is to use a camera to capture a passing vehicle, obtain an image, and then identify the image manually or through computer software. When using manual detection, it is easy to miss abnormal features of the vehicle body due to visual fatigue or incomplete patterns; when using computer software for automatic recognition, it is necessary to meet high standards of image capture requirements and use accurate recognition algorithms to identify abnormal features of the vehicle body, and the complex on-site environment and train appearance often lead to image recognition errors, and cannot provide accurate detection results.

[0005] LiDAR is a sensor that uses lasers to achieve precise distance measurement. LiDAR emits laser pulses, which are reflected by surrounding objects. By measuring the time it takes for the laser to reach each object and return to the object, the precise distance of the object can be calculated. LiDAR emits thousands of pulses per second, and by collecting these distance measurements (distance points), a model of the appearance of the scanned object can be constructed. Summary of the invention

[0006] The present invention aims to overcome at least one of the deficiencies of the above-mentioned prior art and provide a method for three-dimensional laser scanning detection of train appearance, which is used to solve the problem of low accuracy in identifying abnormal situations during train appearance detection.

[0007] The technical solution adopted by the present invention is a method for detecting the appearance of a train by three-dimensional laser scanning, comprising the following steps:

[0008] S1. Scan the train using a laser radar and obtain a set of three-dimensional point data of the train appearance;

[0009] S2. Divide the three-dimensional point data set into a plurality of columns of first data blocks i (i is a positive integer) according to a first spacing along the scanning direction;

[0010] S3. In the vertical direction along the scanning direction, each column of the first data block i is further divided into a plurality of rows of the second data block ij (i, j are positive integers) according to the second spacing;

[0011] S4. Perform distance distribution statistics of three-dimensional points on each second data block ij;

[0012] S5. The first scan obtains a number of second data blocks Aij, and the second scan obtains a number of second data blocks Bij. If the change in the three-dimensional point distance distribution between the second data block Bij and the second data block Aij is less than a threshold, the appearance of the train area corresponding to the second data block ij remains unchanged.

[0013] In this solution, laser radar is used to scan the train's exterior, obtain its 3D point data, and further identify abnormal situations based on the 3D point data. Since 3D point data is distance data, compared with traditional image recognition technology, color changes, pattern changes, surface stains, and on-site lighting conditions of the train's exterior will not interfere with the laser radar's acquisition of 3D point data on the train's exterior, thereby ensuring the accuracy of abnormal situation identification from the perspective of data source.

[0014] Furthermore, this scheme realizes the identification of abnormal conditions on the train appearance by performing distance value distribution statistics on the three-dimensional points within a certain spatial range and comparing the three-dimensional point distance distributions before and after. Within a certain spatial range, if there are shape and position changes such as structural deformation, missing and relative position changes of parts on the train appearance, and foreign objects with a certain volume attached to the surface, the distance value of the three-dimensional point scanned at the corresponding position will change, and the three-dimensional point distance distribution within the spatial range will also change accordingly. That is, the three-dimensional point distance distribution corresponds to the train appearance. If the three-dimensional point distance distribution changes, it indicates that the train appearance has an abnormal situation. The three-dimensional point distance distribution can be a quantity distribution or a probability distribution. Compared with traditional image recognition technology, this scheme does not need to first perform feature recognition on the three-dimensional point data set, and then compare and judge the features before and after; instead, it directly compares the three-dimensional point distance distribution within a certain spatial range, and then determines whether the spatial range has an abnormal situation. Since this scheme only performs statistics and comparison calculations on three-dimensional points, it simplifies the calculation and judgment process, reduces interference factors, and thus improves the speed and accuracy of abnormal situation recognition from the perspective of judgment method. In addition, when considering the interference of other external factors, when the change amplitude of the three-dimensional point distance distribution is greater than the threshold, it indicates that there is an abnormality in the appearance of the train.

[0015] Secondly, this scheme divides the scanned three-dimensional point data set into several second data blocks ij according to the gridding (first spacing and second spacing), and then performs the distance distribution statistics of the three-dimensional points. On the one hand, the three-dimensional point distance distribution can only determine whether an abnormal situation occurs in a certain spatial range, and cannot give a specific location; the gridding of the overall appearance of the train can just confirm the area where the abnormal situation occurs. Taking the scanning start end of the train as the reference, the train appearance area where the abnormal situation occurs can be located by the number ij, the first spacing and the second spacing of the second data block ij. On the other hand, based on the distance distribution statistics of the three-dimensional points, the smaller the statistical spatial range, the higher its sensitivity to changes and the higher the resolution of detecting foreign objects. For example, a 20mm×10mm foreign object accounts for 0.01% in the range of 2000mm×10000mm, and accounts for 1% in the range of 200mm×100mm. Obviously, when the same foreign object changes, the change range of the three-dimensional point distance distribution in the range of 200mm×100mm is larger, and the abnormal situation is easier to be discovered. Therefore, the second data block ij with a smaller range can improve the sensitivity or resolution of abnormal situation identification.

[0016] Preferably, before step S4, the first data block i or the second data block ij is subjected to three-dimensional point interpolation or fitting processing to make the three-dimensional point density the same. In step S1, when using a laser radar to scan a train, since the relative movement speed of the two is difficult to remain constant, the three-dimensional point density obtained at different positions along the scanning direction is different, and the three-dimensional point density obtained at the same position twice is also different. If the three-dimensional point density of the second data block ij is the same, the standard of distance value distribution statistics can be unified, which is conducive to improving the accuracy of the obtained three-dimensional point distance distribution and further improving the accuracy of abnormal situation identification. Secondly, this scheme uses a set fixed value to perform interpolation point processing on positions with sparse three-dimensional point density and performs fitting point processing on positions with dense three-dimensional point density. Interpolation or fitting processing can be performed after step S1, step S2, or step S3.

[0017] In this solution, there are multiple scanning methods according to the relative motion relationship between the laser radar and the train. It can be that the train is fixed and the 2D laser radar or 3D laser radar scans the surface of the train. It can also be that the 2D laser radar or 3D laser radar is fixed and the train passes through its scanning area.

[0018] Preferably, in step S1, the train is running on the track, and the laser radar is fixed on the outside of the track; the train passes through the scanning area of ​​the laser radar, thereby obtaining a three-dimensional point data set of the train appearance. The method of fixing the laser radar and moving the train can save moving parts of the laser radar and realize online detection of the train.

[0019] In this solution, the 3D point data set obtained in step S1 can be a constructed 3D model of the train appearance, and steps S2 to S5 perform statistical judgment on the 3D points of the 3D model. The laser radar scans the train and transmits the original scanned frame data and speed or position data to the modeling module, thereby realizing the construction of the 3D model.

[0020] Preferably, the 3D point data set obtained in step S1 can also be the original scanned frame data, and steps S2 to S5 directly perform statistical judgment on the 3D points of the frame data. Since the modeling process of the train appearance is eliminated, the acquisition speed of the 3D point data set is accelerated and the data interference factors are reduced, which is conducive to further improving the speed and accuracy of abnormal situation identification.

[0021] When using frame data, the actual length corresponding to the three-dimensional point data cannot be obtained because the frame data itself lacks reference coordinates. Therefore, it specifically also includes a sensor, and when the train passes through the scanning area of ​​the laser radar, the sensor is used to trigger the acquisition of a set of timestamps containing the time Ti (i is a positive integer) corresponding to each first spacing distance of the train. The length of the three-dimensional point data between any two adjacent timestamps is the first spacing. Then in step S2, the frame data is divided by a number of timestamps, so that the three-dimensional point data set can be divided into a number of columns of first data blocks i according to the first spacing.

[0022] Furthermore, the sensor is a wheel sensor, and a plurality of wheel sensors are arranged in sequence on the track; the plurality of wheel sensors are used to generate a set of trigger signals including a front-to-back sequence when the train passes by. The plurality of wheel sensors are arranged in sequence on the track approximately at a first spacing, and when the wheels of the train pass the wheel sensors in sequence, the trigger signals generated by the plurality of vehicle sensors include information of the first spacing. A set of trigger signals is recorded by a clock to obtain a set of timestamps required above.

[0023] Preferably, the first spacing or the second spacing is 300mm to 900mm. When the three-dimensional point data set is a three-dimensional model, the first spacing and the second spacing can be selected within the above range according to the detection position and requirements of the train appearance, and the computing and processing capabilities of the equipment. The smaller the first spacing and the second spacing, the greater the amount of computation in the gridding and judgment process during the detection process, and the higher the sensitivity or resolution of abnormal situation identification. When the three-dimensional point data set is frame data, the first spacing is selected within the above range, which can effectively avoid scanning distortion caused by train speed fluctuations, and further ensure the accuracy of abnormal situation identification from the perspective of data source. When the train passes through the laser radar scan, the speed will inevitably fluctuate, but due to the inertia of the train and the limitation of the acceleration and deceleration, the speed of the train remains basically unchanged when passing through a small range. That is, the travel of the train within a small range can be regarded as uniform motion. Through repeated experimental tests, it can be obtained that the range of the first spacing should be between 300mm and 900mm.

[0024] Preferably, the scanning frequency of the laser radar is not less than 500 Hz.

[0025] Preferably, several of the laser radars are arranged at the upper left corner, upper right corner, lower left corner, lower right corner and directly below the cross section of the train, and the scanning area of ​​the laser radar covers the roof and running parts of the train. The laser radars at the upper left corner and upper right corner complete the appearance scan of the roof part of the train from the left and right sides, including key components such as pantographs and refrigeration equipment; the laser radars at the lower left corner, lower right corner and directly below complete the appearance scan of the bottom of the train from the left and right sides and below, including key components such as bogies.

[0026] Optionally, a second camera is further included, which is arranged in parallel on one side of the laser radar, and the second camera is used to synchronously obtain a set of images of the train appearance when the train passes by.

[0027] Optionally, the method further includes step S6. If the appearance of the train area corresponding to the second data block ij changes, an alarm message is sent to the user, and an image of the corresponding position is pushed for manual confirmation.

[0028] In this solution, the second camera is used as a supplement to the laser radar to achieve manual assisted diagnosis. The second camera is parallel to one side of the laser radar, and the field of view of the second camera covers the appearance of the train scanned by the laser radar. The second camera can use the above-mentioned timestamp as a mark to confirm the location of the shooting part. When an abnormality is identified in the area corresponding to a second data block ij, the corresponding timestamp is first found by its number i, and then the image involving the first data block i area is found from the image set by the timestamp, and finally the image is sent to the user to allow the user to confirm whether an abnormality has occurred.

[0029] Optionally, a third camera is further included, which is arranged on both sides of the train and is used to synchronously obtain a set of images of the side appearance of the train when the train passes by. For the side appearance of the train, the laser radar will not work properly due to the presence of glass windows, so traditional image recognition technology is used instead. The third camera and the laser radar complement each other to jointly realize the detection of the 360° appearance of the train.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] This solution uses LiDAR to scan the exterior of the train, avoiding the interference caused by color changes, pattern changes, surface stains, and on-site lighting conditions, and ensures the accuracy of abnormal situation identification from the perspective of data source. This solution simplifies the calculation and judgment process by comparing the three-dimensional point distance distribution within a certain spatial range, reduces interference factors, and thus improves the speed and accuracy of abnormal situation identification from the perspective of judgment method. This solution grids the three-dimensional point data set obtained by scanning to improve the sensitivity or resolution of abnormal situation identification. This solution obtains a timestamp set through sensor triggering, and then can directly use the original scanned frame data as a three-dimensional point data set for statistical judgment, further improving the speed and accuracy of abnormal situation identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic top view of embodiment 1 of the present invention.

[0033] Figure 2This is a schematic cross-sectional view of a train according to Embodiment 1 of the present invention.

[0034] Figure 3 This is a schematic diagram of gridding processing of a three-dimensional point data set according to Example 1 of the present invention.

[0035] Figure 4 This is a schematic diagram of the second data block A11 according to Embodiment 1 of the present invention.

[0036] Figure 5 This is a three-dimensional point distance distribution diagram of the second data block A11 according to Example 1 of the present invention.

[0037] Figure 6 This is a schematic diagram of the second data block B11 of Embodiment 1 of the present invention.

[0038] Figure 7 This is a three-dimensional point distance distribution diagram of the second data block B11 according to Example 1 of the present invention.

[0039] Figure 8 This is a schematic diagram of the principle of dividing frame data by timestamp in Example 1 of the present invention.

[0040] Explanation of reference numerals: pantograph 11 , bogie 12 , lidar 30 , second camera 42 , third camera 43 , wheel sensor 50 . DETAILED DESCRIPTION

[0041] The drawings of the present invention are only for illustrative purposes and should not be construed as limiting the present invention. In order to better illustrate the following embodiments, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; it is understandable to those skilled in the art that some well-known structures and their descriptions in the drawings may be omitted.

[0042] Example 1

[0043] like Figure 1 , Figure 2 As shown, this embodiment is a method for detecting the appearance of a train by three-dimensional laser scanning, comprising the following steps:

[0044] S1. Scan the train using a laser radar and obtain a set of three-dimensional point data of the train appearance;

[0045] S2. Divide the three-dimensional point data set into a plurality of columns of first data blocks i (i is a positive integer) according to a first spacing along the scanning direction;

[0046] S3. In the vertical direction along the scanning direction, each column of the first data block i is further divided into a plurality of rows of the second data block ij (i, j are positive integers) according to the second spacing;

[0047] S4. Perform distance distribution statistics of three-dimensional points on each second data block ij;

[0048] S5. The first scan obtains a number of second data blocks Aij, and the second scan obtains a number of second data blocks Bij. If the change in the three-dimensional point distance distribution between the second data block Bij and the second data block Aij is less than a threshold, the appearance of the train area corresponding to the second data block ij remains unchanged.

[0049] In this solution, laser radar is used to scan the train's exterior, obtain its 3D point data, and further identify abnormal situations based on the 3D point data. Since 3D point data is distance data, compared with traditional image recognition technology, color changes, pattern changes, surface stains, and on-site lighting conditions of the train's exterior will not interfere with the laser radar's acquisition of 3D point data on the train's exterior, thereby ensuring the accuracy of abnormal situation identification from the perspective of data source.

[0050] Furthermore, this scheme realizes the identification of abnormal conditions on the train appearance by performing distance value distribution statistics on the three-dimensional points within a certain spatial range and comparing the three-dimensional point distance distributions before and after. Within a certain spatial range, if there are shape and position changes such as structural deformation, missing and relative position changes of parts on the train appearance, and foreign objects with a certain volume attached to the surface, the distance value of the three-dimensional point scanned at the corresponding position will change, and the three-dimensional point distance distribution within the spatial range will also change accordingly. That is, the three-dimensional point distance distribution corresponds to the train appearance. If the three-dimensional point distance distribution changes, it indicates that the train appearance has an abnormal situation. The three-dimensional point distance distribution can be a quantity distribution or a probability distribution. Compared with traditional image recognition technology, this scheme does not need to first perform feature recognition on the three-dimensional point data set, and then compare and judge the features before and after; instead, it directly compares the three-dimensional point distance distribution within a certain spatial range, and then determines whether the spatial range has an abnormal situation. Since this scheme only performs statistics and comparison calculations on three-dimensional points, it simplifies the calculation and judgment process, reduces interference factors, and thus improves the speed and accuracy of abnormal situation recognition from the perspective of judgment method. In addition, when considering the interference of other external factors, when the change amplitude of the three-dimensional point distance distribution is greater than the threshold, it indicates that there is an abnormality in the appearance of the train.

[0051] like Figure 3As shown, this scheme divides the scanned 3D point data set into several second data blocks ij according to the gridding (first spacing and second spacing), and then performs the distance distribution statistics of the 3D points. On the one hand, the 3D point distance distribution can only determine whether an abnormal situation occurs in a certain spatial range, and cannot give a specific location; the gridding of the overall appearance of the train can just confirm the area where the abnormal situation occurs. Taking the scanning start end of the train as the reference, the train appearance area where the abnormal situation occurs can be located by the number ij, the first spacing and the second spacing of the second data block ij. On the other hand, based on the distance distribution statistics of the 3D points, the smaller the statistical spatial range, the higher its sensitivity to changes and the higher the resolution of detecting foreign objects. For example, a 20mm×10mm foreign object accounts for 0.01% in the range of 2000mm×10000mm, and accounts for 1% in the range of 200mm×100mm. Obviously, when the same foreign object changes, the change range of the 3D point distance distribution in the range of 200mm×100mm is larger, and the abnormal situation is easier to be discovered. Therefore, the second data block ij with a smaller range can improve the sensitivity or resolution of abnormal situation identification.

[0052] In this embodiment, Figures 4 to 7 As shown, the above judgment process is further explained by taking the second data block 11 as an example. Figure 4 and Figure 6 These are visualization images of the 3D point data sets obtained in the local area of ​​the train exterior for the first and second time, respectively. The line segments in the figure are composed of several 3D points with distance values. Figure 5 and Figure 7 The three-dimensional point distance distribution of the second data block 11 obtained for the first and second times is shown in Figure 1. When a certain volume of foreign matter (cube) is attached to the local area, the three-dimensional point distance distribution of the second data block changes accordingly. By judging the change range, it can be identified that an abnormal situation has occurred in the local area.

[0053] Preferably, before step S4, the first data block i or the second data block ij is subjected to three-dimensional point interpolation or fitting processing to make the three-dimensional point density the same. In step S1, when using a laser radar to scan a train, since the relative movement speed of the two is difficult to remain constant, the three-dimensional point density obtained at different positions along the scanning direction is different, and the three-dimensional point density obtained at the same position twice is also different. If the three-dimensional point density of the second data block ij is the same, the standard of distance value distribution statistics can be unified, which is conducive to improving the accuracy of the obtained three-dimensional point distance distribution and further improving the accuracy of abnormal situation identification. Secondly, this scheme uses a set fixed value to perform interpolation point processing on positions with sparse three-dimensional point density and performs fitting point processing on positions with dense three-dimensional point density. Interpolation or fitting processing can be performed after step S1, step S2, or step S3.

[0054] Preferably, in step S1, the train is running on the track, and the laser radar is fixed on the outside of the track; the train passes through the scanning area of ​​the laser radar, thereby obtaining a three-dimensional point data set of the train appearance. The method of fixing the laser radar and moving the train can save moving parts of the laser radar and realize online detection of the train.

[0055] In this solution, the 3D point data set obtained in step S1 can be a constructed 3D model of the train appearance, and steps S2 to S5 perform statistical judgment on the 3D points of the 3D model. The laser radar scans the train and transmits the original scanned frame data and speed or position data to the modeling module, thereby realizing the construction of the 3D model.

[0056] Preferably, the 3D point data set obtained in step S1 can also be the original scanned frame data, and steps S2 to S5 directly perform statistical judgment on the 3D points of the frame data. Since the modeling process of the train appearance is eliminated, the acquisition speed of the 3D point data set is accelerated and the data interference factors are reduced, which is conducive to further improving the speed and accuracy of abnormal situation identification.

[0057] like Figure 8 As shown, when using frame data, the real length corresponding to the three-dimensional point data cannot be obtained because the frame data itself lacks the reference coordinates. Therefore, it also specifically includes a sensor, and when the train passes through the scanning area of ​​the laser radar, the sensor is used to trigger the acquisition of a set of timestamps containing the time Ti (i is a positive integer) corresponding to each first spacing distance of the train. The length of the three-dimensional point data between any two adjacent timestamps is the first spacing. Then in step S2, the frame data is divided by a number of timestamps, so that the three-dimensional point data set can be divided into a number of columns of first data blocks i according to the first spacing.

[0058] Furthermore, the sensor is a wheel sensor, and a plurality of wheel sensors are arranged in sequence on the track; the plurality of wheel sensors are used to generate a set of trigger signals including a front-to-back sequence when the train passes by. The plurality of wheel sensors are arranged in sequence on the track approximately at a first spacing, and when the wheels of the train pass the wheel sensors in sequence, the trigger signals generated by the plurality of vehicle sensors include information of the first spacing. A set of trigger signals is recorded by a clock to obtain a set of timestamps required above.

[0059] Preferably, the first spacing or the second spacing is 300mm to 900mm. When the three-dimensional point data set is a three-dimensional model, the first spacing and the second spacing can be selected within the above range according to the detection position and requirements of the train appearance, and the computing and processing capabilities of the equipment. The smaller the first spacing and the second spacing, the greater the amount of computation in the gridding and judgment process during the detection process, and the higher the sensitivity or resolution of abnormal situation identification. When the three-dimensional point data set is frame data, the first spacing is selected within the above range, which can effectively avoid scanning distortion caused by train speed fluctuations, and further ensure the accuracy of abnormal situation identification from the perspective of data source. When the train passes through the laser radar scan, the speed will inevitably fluctuate, but due to the inertia of the train and the limitation of the acceleration and deceleration, the speed of the train remains basically unchanged when passing through a small range. That is, the travel of the train within a small range can be regarded as uniform motion. Through repeated experimental tests, it can be obtained that the range of the first spacing should be between 300mm and 900mm.

[0060] Preferably, the scanning frequency of the laser radar is not less than 500 Hz.

[0061] Preferably, several of the laser radars are arranged at the upper left corner, upper right corner, lower left corner, lower right corner and directly below the cross section of the train, and the scanning area of ​​the laser radar covers the roof and running parts of the train. The laser radars at the upper left corner and upper right corner complete the appearance scan of the roof part of the train from the left and right sides, including key components such as pantographs and refrigeration equipment; the laser radars at the lower left corner, lower right corner and directly below complete the appearance scan of the bottom of the train from the left and right sides and below, including key components such as bogies.

[0062] Optionally, a second camera is further included, which is arranged in parallel on one side of the laser radar, and the second camera is used to synchronously obtain a set of images of the train appearance when the train passes by.

[0063] Optionally, the method further includes step S6. If the appearance of the train area corresponding to the second data block ij changes, an alarm message is sent to the user, and an image of the corresponding position is pushed for manual confirmation.

[0064] In this solution, the second camera is used as a supplement to the laser radar to achieve manual assisted diagnosis. The second camera is parallel to one side of the laser radar, and the field of view of the second camera covers the appearance of the train scanned by the laser radar. The second camera can use the above-mentioned timestamp as a mark to confirm the location of the shooting part. When an abnormality is identified in the area corresponding to a second data block ij, the corresponding timestamp is first found by its number i, and then the image involving the first data block i area is found from the image set by the timestamp, and finally the image is sent to the user to allow the user to confirm whether an abnormality has occurred.

[0065] Optionally, a third camera is further included, which is arranged on both sides of the train and is used to synchronously obtain a set of images of the side appearance of the train when the train passes by. For the side appearance of the train, the laser radar will not work properly due to the presence of glass windows, so traditional image recognition technology is used instead. The third camera and the laser radar complement each other to jointly realize the detection of the 360° appearance of the train.

[0066] Obviously, the above embodiments of the present invention are only examples for clearly illustrating the technical solutions of the present invention, and are not intended to limit the specific implementation methods of the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the claims of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A method for detecting train appearance by laser three-dimensional scanning, characterized in that: The following steps are involved: S1. Use laser radar to scan the train and obtain a set of 3D point data of the train appearance; S2. Divide the three-dimensional point data set into a plurality of columns of first data blocks i along the scanning direction according to the first spacing, where i is a positive integer; S3 along the vertical direction of the scanning direction, according to the second spacing each column of the first data block i is again divided into a number of rows of second data blocks ij, i, j are positive integers; S4. Perform distance distribution statistics of three-dimensional points on each second data block ij respectively; S5. The first scan obtains several second data blocks A ij , the second scan obtains several second data blocks B ij , if the second data block B ij With the second data block A ij If the change in the three-dimensional point distance distribution is less than the threshold, the appearance of the train area corresponding to the second data block ij does not change; Before step S4, three-dimensional point interpolation or fitting processing is performed on the first data block i or the second data block ij so that the three-dimensional point densities thereof are the same.

2. A train appearance laser three-dimensional scanning detection method according to claim 1, characterized in that: In step S1, the train runs on the track, and the laser radar is fixed on the outside of the track; the train passes through the scanning area of ​​the laser radar, thereby obtaining a three-dimensional point data set of the train appearance.

3. The method for detecting train appearance by laser three-dimensional scanning according to claim 2, characterized in that: It also includes a sensor. When the train passes through the scanning area of ​​the laser radar, the sensor is used to trigger the acquisition of a set of timestamps containing a time Ti corresponding to each first spacing distance of the train, where i is a positive integer.

4. The method for detecting train appearance by three-dimensional laser scanning according to claim 3, characterized in that: The sensors are wheel sensors, and a plurality of wheel sensors are arranged in sequence on the track; the plurality of wheel sensors are used to generate a set of trigger signals including a front-to-back sequence when the train passes by.

5. The method for detecting train appearance by three-dimensional laser scanning according to claim 1, characterized in that: The first spacing or the second spacing is 300 mm to 900 mm.

6. The method for three-dimensional laser scanning of train appearance according to claim 1, characterized in that: The scanning frequency of the laser radar is not less than 500 Hz.

7. The method for detecting train appearance by three-dimensional laser scanning according to claim 2, characterized in that: Several of the laser radars are respectively arranged at the upper left corner, upper right corner, lower left corner, lower right corner and directly below the cross section of the train, and the scanning area of ​​the laser radar covers the roof and running part of the train.

8. A train appearance laser three-dimensional scanning detection method according to any one of claims 2 to 7, characterized in that: It also includes a second camera, which is arranged in parallel on one side of the laser radar, and is used to synchronously obtain a set of images of the appearance of the train when the train passes by.

9. The method for three-dimensional laser scanning of train appearance according to claim 8, characterized in that: The method further comprises step S6. If the appearance of the train area corresponding to the second data block ij changes, an alarm message is sent to the user, and an image of the corresponding position is pushed for manual confirmation.

Images