JP7449191B2 - How to measure the inclination of a joint plate - Google Patents

Description

The present invention relates to a method for measuring the inclination of a joint plate installed on the abdomen of a rail laid on a railway track, and in particular, the present invention relates to a method for measuring the inclination of a joint plate provided on the abdomen of a rail laid on a railway track. Concerning techniques for measuring the slope of.

On railway tracks, a joint plate is installed on the abdomen of the rail to connect the rails to each other, and if the joint plate comes off, it may cause transportation problems, so detection of abnormalities in the joint plate is necessary. is extremely important in track maintenance.
Generally, the joint plate is attached to both flanks of the rail by bolts, and the most common cause of the joint plate falling off is the bolts breaking. Therefore, the current situation is to conduct maintenance management in which broken joint plate bolts are detected visually or by images taken with a camera and replaced. However, visual inspection of broken joint plate bolts requires a lot of manpower and labor, and therefore labor saving is desired. Detecting breakage of a seam plate bolt using images can save labor, but it is difficult to predict breakage of a seam plate bolt.

Therefore, there is a need to quickly detect abnormalities such as loosening of the joint plate bolts before they are damaged, without requiring human labor. Note that the pair of joint plates J1 and J2, which are normally attached to the left and right sides of the rail R, are installed parallel to each other as shown by the dotted lines in FIG. It has been reported that if this occurs, the parallelism of the left and right seam plates J1 and J2 is impaired, as shown by the solid line, and the seam plate bolt B is likely to break.
Therefore, if the inclination or parallelism of the joint plates J1 and J2 could be measured, it would be effective because preventive maintenance could be carried out before the bolt B breaks, but a device that measures the inclination or parallelism of the joint plates is not practical. This method is not standardized, and each step must be performed manually, which requires an extremely large amount of labor. Therefore, it has not been possible to detect and manage the inclination (parallelism) of the joint plates.

Note that an invention related to a joint plate detection device for easily and reliably detecting a rail joint plate has been proposed (see Patent Document 1). In addition, in order to detect abnormalities in rail fastening devices, joint plates, and track pads, by installing them on commercial vehicles and acquiring 3D shape data of track components, we can observe deterioration trends and accurately predict maintenance timing. In addition, an invention has been proposed in which an abnormality in a track component can be discovered by transforming three-dimensional shape data into a distance image and displaying it (see Patent Document 2).

Japanese Patent Application Publication No. 10-332324 JP2011-214933A

However, although the detection device of the invention disclosed in Patent Document 1 can detect the presence or absence of a rail joint plate, it cannot measure the inclination of the joint plate. Further, in the invention of Patent Document 2, although distance images regarding track components can be displayed, the inclination of the joint plate cannot be measured. Even if it were possible to display a distance image in which the joint plate is tilted, sufficient labor savings would not be achieved because a person would have to visually judge whether or not the tilt would lead to loosening of the joint plate bolts. The problem is that there is no.

Therefore, the present inventors focused on the fact that the track material monitoring device installed on trains is equipped with a profile camera (distance meter using a laser) that detects the distance to the track material. We considered calculating the inclination of the joint plate based on the measured distance to the plate. However, there is a height difference called cant on the track at curved sections, and the inclination of the laser distance meter (distance sensor) on the vehicle body side with respect to the rail is different from the inclination at flat points. It became clear that there was a problem in that it was difficult to accurately measure the inclination of the joint plate at curved parts of the track.

The present invention has been made in view of the above-mentioned problems, and provides a seam plate that can measure the inclination of the seam plate, which may lead to damage of the seam plate bolt, based on the measured value of a distance sensor mounted on a vehicle. The object of the present invention is to provide a method for measuring the slope of
Another object of the present invention is to provide a method for measuring the inclination of a seam plate that can accurately measure the inclination of a seam plate even at curved locations on a track.

In order to achieve the above-mentioned problems, this invention
A method for measuring the inclination of a joint plate attached to the abdomen of a rail laid on a railway track based on distance image data of track material acquired while a vehicle equipped with a distance sensor is running. In,
Based on the distance image data, calculate the slope of the upper surface of the lower edge of the joint plate and the slope of the upper surface of the bottom of the rail,
The difference between the calculated inclination of the upper surface of the lower edge of the joint plate and the upper surface of the bottom of the rail is set as the inclination of the joint plate with respect to the rail.

According to the above method, the inclination of the seam plate can be calculated based on the brightness data obtained by the distance sensor, so the inclination of the seam plate that can lead to damage of the seam plate bolt can be discovered without much effort. be able to. In addition, since the relative inclination of the joint plate to the rail can be determined, it is possible to accurately measure the inclination of the joint plate even at curved places on the track where the inclination of the detection axis of the distance sensor with respect to the rail is different from flat places. Can be done.

More specifically, the inclination of a joint plate attached to the abdomen of a rail laid on a railway track is measured based on distance image data of track material acquired while a vehicle equipped with a distance sensor is running. In the method of measuring the inclination of the joint plate,
A first step of detecting an image area of the joint plate from the distance image data and extracting distance image data of a predetermined range including the joint plate and the bottom of the rail;
a second step of calculating a difference in brightness between adjacent pixels for all rows from one end of the joint plate to the other end based on the distance image data extracted in the first step;
a third step of detecting the boundary of the upper surface area of the lower edge of the joint plate by determining whether the difference in brightness is within a preset threshold range;
a fourth step of detecting a boundary of an upper surface area of the bottom of the rail by determining whether the difference in brightness is within a preset threshold range;
Based on the boundary of the upper surface area of the lower edge of the joint plate detected in the third step, distance image data of the upper surface of the lower edge of the joint plate is specified from the extracted distance image data, and a fifth step of calculating the slope of the upper surface of the lower edge of the board;
Based on the boundary of the top surface area of the bottom of the rail detected in the fourth step, distance image data of the top surface of the bottom of the rail is specified from the extracted distance image data, and an inclination of the top surface of the bottom of the rail is calculated. The sixth step is to
By taking the difference between the slope of the upper surface of the lower edge of the joint plate calculated in the fifth step and the slope of the upper surface of the bottom part of the rail calculated in the sixth step, the slope of the joint plate with respect to the rail can be determined. and a seventh step of calculating.

According to this method, the boundary information of the upper edge of the joint plate and the bottom of the rail is acquired, and the luminance data of the upper surface of the upper edge of the joint plate and the rail are selected from the distance image data acquired by the distance sensor. It is possible to extract the brightness data of the bottom and top surfaces of the . Therefore, by calculating the inclination of the top surface of the upper edge of the joint plate and the slope of the top surface of the bottom of the rail from each luminance data, it is possible to accurately determine the relative inclination of the joint plate with respect to the rail.

Preferably, in the fifth step, based on the boundary of the upper surface area of the lower edge of the joint plate detected in the third step, the distance image data extracted in the first step is Distance image data of the upper surface of the lower edge of the joint plate is specified , the average value of the luminance of the distance image data along the longitudinal direction of the joint plate is calculated, and the distance image data of the upper surface of the lower edge of the joint plate is calculated. Calculate the slope of the top surface of the bottom edge of the board,
In the sixth step, a distance image of the top surface of the bottom of the rail is calculated from the distance image data extracted in the first step based on the boundary of the top surface area of the bottom of the rail detected in the fourth step. The data is specified , the average value of the brightness of the distance image data along the longitudinal direction of the rail is calculated, and the slope of the upper surface of the bottom of the rail is calculated based on the calculated average value of the brightness.

According to this method, even if the resolution in the height direction of the brightness data obtained from the distance sensor used is coarse, the resolution can be improved by taking the average value, and more accurate seams can be created. The inclination of the board can be calculated.

Furthermore, desirably, in the fifth step, a difference is calculated in the row direction based on the average value of the luminance of the upper surface of the lower edge of the seam plate along the longitudinal direction of the seam plate, and determining an order statistic such as the median or quartile of the difference between the top surfaces of the edges as a representative value, and calculating the slope of the top surface of the bottom edge of the joint plate based on the determined representative value;
In the sixth step, differences are calculated in the row direction based on the average value of the brightness of the bottom top surface of the rail along the rail length direction, and the median value of the differences of the bottom top surface of the rail is determined as a representative value. Then, the inclination of the bottom upper surface of the rail is calculated based on the determined representative value.

According to this method, order statistics such as the median or quartile of the difference between the top surface of the lower edge of the joint plate and the top surface of the bottom of the rail are used as representative values, and based on the representative values, In order to calculate the inclination of the joint plate with respect to the rail, it is possible to prevent data from the curved part of the lower edge of the joint plate and the hem of the bottom of the rail from affecting the calculation of the inclination, which allows for highly accurate inclination of the joint plate. can be calculated.
Further, in the fifth step and the sixth step, when the pixel size of the distance image data is represented by r and the difference as the representative value is ΔH, the slope of the upper surface of the lower edge of the joint plate is α and the slope β of the top surface of the bottom of the rail can be calculated using the formula tan-1 (ΔH/r).

According to the method for measuring the inclination of a joint plate of the present invention, it is possible to measure the inclination of a joint plate that may lead to damage to a joint plate bolt, based on the measured value of a distance sensor mounted on a vehicle. Further, there is an effect that the inclination of the joint plate can be accurately measured even at curved portions of the track.

FIG. 2 is a schematic front view showing an example of installing a distance sensor in a track material monitoring device mounted on a train. FIG. 3 is a schematic diagram showing the relationship between a joint plate to be measured and a rail. It is a graph which shows an example of the brightness change in the rail width direction acquired by the distance sensor provided in the track material monitoring device, and the change of the difference of the brightness value of adjacent pixels. It is a flowchart which shows an example of the procedure of the inclination measurement method of the joint plate of this invention. (A) is a graph showing the change in brightness in the row direction of a certain row based on distance image data and the change in brightness after column direction averaging processing; (B) is a diagram showing the relationship between the difference in brightness value and the slope. . FIG. 3 is a front view showing normal and abnormal states of a seam plate, which is a target of the seam plate inclination measurement method of the present invention.

Hereinafter, with reference to the drawings, an embodiment in which the present invention is applied to a case where the inclination of a joint plate is measured using data acquired by a distance sensor provided in a track material monitoring device mounted on a train. will be explained in detail.
Note that the distance sensor used when applying the present invention may be of any type, such as a laser distance meter, as long as it can measure the distance to the target object. The following explanation assumes that a sensor such as a profile camera, a depth sensor, or a depth camera that outputs distance image data representing the distance to the object as a brightness value is used.

Regarding the method of installing the distance sensors, as shown in Fig. 1, it is recommended to install a pair of distance sensors S1 and S2 facing downward, corresponding to both sides of a pair of rails R1 and R2 on the track. desirable. In addition, in FIG. 1, what is shown by the symbol RC is a rail fastening device.
Some existing commercial trains are equipped with track material monitoring devices that are equipped with profile cameras that can measure distance and line sensor cameras that can see shading, and that acquire track conditions while running and store them in storage devices. The inclination of the seam plate may be measured using data acquired by a profile camera of such a track material monitoring device.

Before describing the embodiments, the concept of the method for detecting the inclination of a joint plate according to the present invention will be explained.
As mentioned above, there is a cant in the track at curved sections, and the inclination of the detection axis of the distance sensor on the vehicle body with respect to the rail is different from the inclination at flat locations. It is difficult to accurately measure the inclination of the joint plates at curved locations. On the other hand, if the joint plate is correctly attached to the abdomen of the rail, whether in a curved area or a flat area, the angle β of the top surface F1 of the bottom (lower flange) of the rail R and The angle α of the upper surface F2 of the lower edge portion (lower overhang portion) of the joint plate J is approximately equal, and the relative angle of F2 to F1 is constant (approximately 0°).

Therefore, by measuring the distance from the distance sensor to the top surface F1 of the bottom of the rail R and the top surface F2 of the bottom edge of the joint plate J, calculating the respective angles, and taking the difference, the influence of cant can be eliminated and the rail of the joint plate We decided to calculate the relative slope to . To do this, it is necessary to extract distance data between the bottom upper surface F1 of the rail R and the lower edge upper surface F2 of the joint plate J from the measurement data of the distance sensor.
If the distance sensor is a sensor that outputs distance image data expressing distance as brightness, in the acquired distance image data, boundary #1 between the head of the rail R and the upper edge (upper overhang) of the joint plate J , at the boundary #2 between the upper edge and the lower edge of the joint plate J, and at the boundary #3 between the lower edge of the joint plate J and the bottom (lower flange) of the rail R, the brightness changes significantly. It is expected that a peak will appear in the difference between.

Therefore, the present inventors extracted the data of the joint plate area from the distance image data acquired by the distance sensor installed in the track material monitoring device, and calculated the difference in brightness between pixels along the rail width direction. I took it and expressed it as a graph. Figure 3 shows the graph. Of these, FIG. 3A shows a change in the brightness value of each pixel, and FIG. 3B shows a change in the difference in brightness between adjacent pixels. From FIG. 3, it can be seen that the peak of the difference in brightness appears at each boundary #1, #2, #3, and #4. Note that #4 is the boundary between the bottom of the rail R and parts such as the rail fastening device (P in FIG. 2).

Furthermore, the present inventors applied the same processing to the distance image data of other joint plates acquired by the distance sensor and expressed it in a graph, and found that each boundary #1, #2, # 3. It became clear that the peak of the brightness difference appeared in #4. After performing statistical processing, we found that each boundary #1, #2, and #3 can be detected by setting the threshold values LDth1, LDth2, LDth3, and LDth0, as shown in Figure 3 (B). Ta. Therefore, the bottom upper surface F1 of the rail R and the lower edge upper surface F2 of the joint plate J can be extracted from the measurement data of the distance sensor. It was found that by calculating each angle and taking the difference, it was possible to calculate the inclination of the joint plate with respect to the rail.

Next, an example of a specific procedure of the method for measuring the inclination of a joint plate according to the present invention using the above knowledge will be described using the flowchart of FIG. 4.
In the inclination measurement process of FIG. 4, first, the distance image data acquired by the distance sensor is read out from the storage device to detect the hem position of the rail (step S1). Next, the area of the joint plate is detected from the read distance image data, and an image from the edge (#1) of the rail head to the hem position (#4) of the rail bottom is extracted (step S2). Next, the difference in brightness between adjacent pixels is calculated for all rows from the top end to the bottom end of the joint plate (step S3).

After that, for each row from the top end to the bottom end of the joint plate, find one whose brightness difference calculated in step S3 falls between the thresholds LDth1 and LDth2 shown in FIG. 3(B) (step S4). LDth1 and LDth2 are thresholds for finding boundary #2 between the upper edge and lower edge of the joint plate, and there is no corresponding difference between LDth1 and LDth2 in the row of interest. This is thought to be because it is hidden behind the joint plate bolts, etc., so when it is determined "No" in step S4, the process moves to step S8, excludes the relevant difference data, and returns to step S4. Executes judgment on the next line.

On the other hand, when the determination in step S4 is "Yes", the process proceeds to step S5 to find a value that falls between the threshold values LDth0 and LDth1. LDth0 and LDth1 are thresholds for finding boundary #3 between the lower edge of the joint plate and the bottom of the rail, and there is no corresponding difference between LDth0 and LDth1 in the line of interest. is considered to be data that cannot be determined, so when the determination in step S5 is "No", the process moves to step S8, excludes the difference data, and returns to step S4 to determine the next line. Execute.

Moreover, when it is determined in step S5 that it is "Yes", the process proceeds to step S6, and it is determined whether or not the difference data is larger than the threshold value LDth3. Here, difference data larger than the threshold value LDth3 is impossible due to the structure of the joint plate, so when it is determined as "Yes" in step S6, it is regarded as an abnormal value and the process moves to step S8 to calculate the difference data. The data is excluded and the process returns to step S4 to perform determination on the next line.
When it is determined "No" in step S6, the process proceeds to step S7, and it is determined whether or not the determination has been completed for all the rows of difference data from the top edge to the bottom edge of the joint plate, and if the determination has not been completed (No ), the process returns to step S4 to determine the next line, and when all lines have been processed (Yes), the process moves to step S9.

In step S9, averaging processing is performed to average the luminance in the column direction (longitudinal direction of the seam plate) in the range of the lower edge of the seam plate that satisfies the conditions of steps S4 to S6, and in the next step S10, , the difference in the averaged luminance in the row direction is taken in step S9, and the median value of the differences is taken as a representative value from among the difference data group. The reason for taking the average of the luminance in the column direction is because if the resolution in the height direction of the distance sensor used is coarse, for example 1 mm/1 gradation, the accuracy of tilt measurement will deteriorate if the acquired luminance data is used as is. It is.

FIG. 5A shows an example of the luminance value in the row direction of a certain row and the average luminance value in the row direction obtained by performing the averaging process in step S9. In FIG. 5A, the ● mark is a plot of the luminance value in the row direction of a certain row, and the □ mark is a plot of the average luminance value in the row direction after averaging processing. From FIG. 5A, it can be seen that the luminance values in the row direction before averaging change by one gradation, whereas the average luminance values change in a width narrower than one gradation.
Furthermore, the reason why the median value of the differences is used as the representative value in step S10 is because, as can be seen from the range #2 to #3 in FIG. 2, the terminal end side of the lower edge of the joint plate is curved. This is because an error will occur if the slope of the upper surface of the lower edge of the joint plate is calculated including the curvature. Instead of taking the median value, you may take quartile values.

After determining the representative value in step S10, the process proceeds to step S11, where the inclination α of the upper surface of the lower edge of the joint plate is calculated using the representative value in accordance with the horizontal resolution of the image. Specifically, if the resolution of the image is rmm/1 pixel and ^the difference in brightness ΔH, then from FIG. I understand.
Next, in the same procedure as steps S9 to S11 above, luminance data in the range of the rail bottom is extracted, a representative value is determined, and the slope β of the upper surface of the rail bottom is calculated (step S12). Then, the difference (α-β) between the inclination α calculated in step S11 and the inclination β calculated in step S12 is calculated to determine the relative inclination of the joint plate with respect to the rail, and the process ends.

As described above, in the embodiment, since the inclination of the seam plate can be calculated based on the brightness image data acquired by the distance sensor, it is possible to calculate the inclination of the seam plate without much effort, which may lead to damage of the seam plate bolt. The slope of can be found. In addition, since the relative inclination of the joint plate to the rail can be determined, it is possible to accurately measure the inclination of the joint plate even at curved places on the track where the inclination of the detection axis of the distance sensor with respect to the rail is different from flat places. Can be done. Furthermore, for joint plates whose inclination as calculated above is greater than a predetermined angle, repair work may be performed to eliminate the inclination, or the parallelism of the joint plate may be manually measured again to ensure that the inclination is greater than the predetermined angle. Repair work may be performed on the joint plates. In other words, the above measurement results can be utilized as screening data.

Incidentally, the measurement of the inclination of the joint plate according to the flowchart shown in FIG. 4 can be carried out using a computer equipped with a microprocessor (MPU) and a storage means such as ROM (read-only memory) or RAM (memory that can be read and written at any time). This can be carried out using a program that describes the steps of the method for measuring the inclination of a joint plate shown in FIG. In that case, a computer equipped with such a program will function as a device for measuring the inclination of the seam plate. This joint plate inclination measuring device measures the inclination of the joint plate by transferring measurement data stored in the storage device of the on-board track material monitoring device via communication or a medium such as a CDROM or USB memory. It can be calculated.

Although the invention made by the present inventor has been specifically explained based on the embodiments above, the present invention is not limited to the embodiments. For example, in the embodiment described above, the inclination of the joint plate is measured based on distance image data acquired by a track material monitoring device already installed on the train, but a train that is not equipped with a track material monitoring device Alternatively, a distance sensor capable of acquiring distance image data may be installed in the measurement vehicle to collect data.

In addition, in the above embodiment, in step S3, the difference in brightness between adjacent pixels is calculated for all rows from the top to the bottom of the seam plate, but for all rows from the top to the bottom of the seam plate, The process of calculating the difference in brightness between adjacent pixels may be performed for, for example, a range of 1/2 to 1/3 of the intermediate portion of the joint plate in the longitudinal direction, instead of the row.
Furthermore, the present invention is not limited to detecting abnormalities such as loosening of joint plate bolts, but can be widely used for detecting falling off, partial damage, etc. of parts of snow melting equipment, etc.

R, R1, R2 Rail J, J1, J2 Joint plate B Bolt S1, S2 Distance sensor RC Rail fastening device

Claims (5)

A method for measuring the inclination of a joint plate attached to the abdomen of a rail laid on a railway track based on distance image data of track material acquired while a vehicle equipped with a distance sensor is running. And,
Based on the distance image data, calculate the slope of the upper surface of the lower edge of the joint plate and the slope of the upper surface of the bottom of the rail, respectively;
A method for measuring the inclination of a joint plate, characterized in that the difference between the calculated slope of the upper surface of the lower edge of the joint plate and the upper surface of the bottom of the rail is determined as the inclination of the joint plate with respect to the rail. A method for measuring the inclination of a joint plate attached to the abdomen of a rail laid on a railway track based on distance image data of track material acquired while a vehicle equipped with a distance sensor is running. And,
a first step of detecting an image area of the joint plate from the distance image data and extracting distance image data of a predetermined range including the joint plate and the bottom of the rail;
a second step of calculating a difference in brightness between adjacent pixels for all rows from one end of the joint plate to the other end based on the distance image data extracted in the first step;
a third step of detecting the boundary of the upper surface area of the lower edge of the joint plate by determining whether the difference in brightness is within a preset threshold range;
a fourth step of detecting a boundary of an upper surface area of the bottom of the rail by determining whether the difference in brightness is within a preset threshold range;
Based on the boundary of the upper surface area of the lower edge of the joint plate detected in the third step, distance image data of the upper surface of the lower edge of the joint plate is extracted from the distance image data extracted in the first step. a fifth step of determining the slope of the upper surface of the lower edge of the joint plate;
Based on the boundary of the upper surface area of the bottom of the rail detected in the fourth step, distance image data of the upper surface of the bottom of the rail is specified from the distance image data extracted in the first step , and a sixth step of calculating the slope of the top surface of the bottom of the
By taking the difference between the slope of the upper surface of the lower edge of the joint plate calculated in the fifth step and the slope of the upper surface of the bottom of the rail calculated in the sixth step, the rail of the joint plate a seventh step of calculating the slope for
A method for measuring the inclination of a joint plate, the method comprising: In the fifth step, based on the boundary of the upper surface area of the lower edge of the joint plate detected in the third step, the lower edge of the joint plate is extracted from the distance image data extracted in the first step. The distance image data of the upper surface of the joint plate is specified , the average value of the brightness of the distance image data along the longitudinal direction of the joint plate is calculated, and the lower edge of the joint plate is calculated based on the calculated average brightness value. Calculate the slope of the top surface of
In the sixth step, a distance image of the top surface of the bottom of the rail is calculated from the distance image data extracted in the first step based on the boundary of the top surface area of the bottom of the rail detected in the fourth step. data , calculates an average value of brightness of distance image data along the longitudinal direction of the rail, and calculates the slope of the top surface of the bottom of the rail based on the calculated average value of brightness. The method for measuring the inclination of a joint plate according to claim 2. In the fifth step, based on the boundary of the upper surface area of the lower edge of the joint plate detected in the third step, the lower edge of the joint plate is extracted from the distance image data extracted in the first step. The distance image data of the upper surface of the upper surface of the joint plate is specified, and the difference is calculated in the row direction based on the average value of the brightness of the distance image data of the upper surface of the lower edge of the joint plate along the longitudinal direction of the joint plate. determining the order statistic of the difference of the upper surface of the lower edge of the joint plate as a representative value, and calculating the slope of the upper surface of the lower edge of the joint plate based on the determined representative value;
In the sixth step, a distance image of the top surface of the bottom of the rail is calculated from the distance image data extracted in the first step based on the boundary of the top surface area of the bottom of the rail detected in the fourth step. Specify the data, take the difference in the row direction based on the average value of the brightness of the distance image data of the top surface of the bottom of the rail along the longitudinal direction of the rail, and calculate the median value of the difference of the top surface of the bottom of the rail. 3. The method for measuring the inclination of a joint plate according to claim 2 , wherein the inclination of the upper surface of the bottom of the rail is calculated based on the determined representative value. In the fifth step and the sixth step, when the pixel size of the distance image data is represented by r and the difference as the representative value is ΔH, the slope of the upper surface of the lower edge of the joint plate is α 5. The method for measuring the inclination of a joint plate according to claim 4, wherein the inclination β of the upper surface of the bottom of the rail is calculated using the formula tan-1 (ΔH/r).

Images