On-line measurement method for inner side distance of train wheel set

Abstract

The invention provides an on-line measurement method for the inner side distance of a train wheel set, and the method comprises the steps: respectively installing structured light visual sensors at the inner side of a track, and enabling the sensors to project structured light to the inner side surfaces of a left wheel and a right wheel; when the train travels to an area to be measured, the two structured light vision sensors synchronously acquire a structured light actual measurement image; respectively acquiring a left wheel actual measurement inner side surface and a right wheel actual measurement inner side surface by utilizing the two structured light actual measurement images; the deviation value generated by the left wheel actual measurement inner side face relative to the left wheel reference inner side face and the deviation value generated by the right wheel actual measurement inner side face relative to the right wheel reference inner side face are calculated respectively; the two deviation values are used for correcting d0, and the corrected distance value is recorded as the train wheel pair inner side distance. The invention aims to provide an inner side distance measuring method without external parameter calibration / a unified coordinate system, which is convenient to implement in an outdoor environment and is suitable for online detection of inner side distances of various types of trains.

Description

Technical Field

[0001] The invention relates to the field of rail transit automated detection, and in particular to an online measurement method for the inner distance of a train wheel pair. Background Art

[0002] Wheelsets are important components of railway trains. During the operation of the train, the wheelset bears a large load, which will inevitably cause wear and defects. When the wear is severe, it may affect the running performance of the train and may even cause driving safety problems. Therefore, it is necessary to measure the profile dimensions of the train wheelset regularly and repair the profile defects found in time. The inner distance of the wheelset is one of the measurement items of the wheelset profile dimensions, such as Figure 5 As shown in the figure, the distance between backs of wheel flanges refers to the distance between the backs of the wheel flanges. According to regulations, the distance between backs of wheelsets for subways and EMUs is generally guaranteed to be within the range of 1353±2 mm, and for passenger trains is generally guaranteed to be within the range of 1353±3 mm.

[0003] At present, the measurement of the inner distance includes contact measurement and non-contact measurement methods. Traditional inner distance rulers and wheel-free lathes need to stop and check the wheels one by one, which is inefficient. In recent years, a type of non-contact online measurement method based on optical technology has emerged. However, such methods often require external parameter calibration of multiple sensors first, and unify the coordinate system. The steps are cumbersome and time-consuming. For example: Chinese patent CN104251667A proposes a dynamic measurement method and system for the inner distance of train wheels. In this technical solution, it is necessary to unify the measurement points of each laser rangefinder in the same coordinate system before the two inner sides can be plane-fitted based on at least 6 laser rangefinders. Chinese patent CN113776446A proposes a wheel inner distance measurement device, system and measurement method, which also requires converting the coordinate systems of all three sensors to the coordinate system of another sensor before performing optical measurement of the wheelset.

[0004] In addition, Chinese patent CN105774840A proposes a dynamic detection system and method for the inner distance of a train wheel set, and discloses a simple inner distance measurement method that only requires two laser rangefinders. Although this method does not require a coordinate system, it requires measuring the horizontal distance between the two sensors, has high requirements for installation and distance measurement, and is difficult to implement. Summary of the invention

[0005] In order to solve the above technical problems, this paper proposes an online measurement method for the inner distance of train wheelsets. This method aims to ensure the measurement accuracy and proposes an inner distance measurement method that is easy to implement and does not require external parameter calibration / unified coordinate system. Since there is no need for complicated calibration process, this method is easy to implement in outdoor environments, takes a short time, has high accuracy, and is suitable for online detection of the inner distance of various types of trains.

[0006] The technical solution is as follows:

[0007] An online measurement method for the inner distance of a train wheel set, wherein structured light vision sensors are respectively installed on the inner side of a track, and a line between the installation positions of two structured light vision sensors is approximately perpendicular to the extension direction of the track;

[0008] Two structured light vision sensors are used to project multi-point structured light / multi-line structured light / grating structured light onto the inner sides of the left and right wheels respectively;

[0009] Obtain the inner distance of the train wheelset by following the steps below:

[0010] ① When the train travels to the area to be tested, the wheel triggers the sensor, the sensor sends out a signal of an approaching vehicle, and the two structured light vision sensors are turned on synchronously, projecting structured light onto the inner side of the left and right wheels respectively and acquiring structured light measured images; the sensors are fixed around the structured light vision sensors;

[0011] ② Use the three-dimensional point information in the two structured light measured images to fit planes respectively, which are recorded as the measured inner side of the left wheel and the measured inner side of the right wheel;

[0012] ③ Retrieve the pre-stored reference inner side of the left wheel, the reference inner side of the right wheel and the reference distance d0 between the two wheels;

[0013] Calculate the deviation of the measured inner side surface of the left wheel relative to the reference inner side surface of the left wheel and the deviation of the measured inner side surface of the right wheel relative to the reference inner side surface of the right wheel respectively;

[0014] The two deviations are used to correct d0, and the corrected distance value is recorded as the inner distance of the train wheelset.

[0015] Further, before step ①, the following steps are used in advance to store the reference inner side surface of the left wheel, the reference inner side surface of the right wheel, and the distance d0 between the two reference inner side surfaces:

[0016] A normal wheel set or mold without wear is recorded as a standard part; the mold includes a left side surface, a right side surface and a rigid connecting rod to keep the relative positions of the two fixed, the left side surface and the right side surface are parallel to each other, and the distance between the inner side surfaces of the two is equal to the inner side distance between normal wheels without wear;

[0017] The standard part is placed within the field of view of the structured light vision sensor; the distance between the left and right inner sides of the standard part is manually measured and recorded as the reference distance d0;

[0018] The structured light vision sensors are turned on synchronously to project structured light onto the left and right inner sides of the standard component and obtain structured light standard images;

[0019] The three-dimensional point information in the two structured light standard images is used to fit planes respectively, which are recorded as the left wheel reference inner side surface and the right wheel reference inner side surface.

[0020] Furthermore, the two deviations are used to correct d0, and the corrected distance value d is calculated. P =d0+d1-d2

[0021] Among them, d1 is the deviation of the actually measured inner side surface of the left wheel relative to the reference inner side surface of the left wheel, and d2 is the deviation of the actually measured inner side surface of the right wheel relative to the reference inner side surface of the right wheel.

[0022] Preferably, the plane equation of the reference inner side surface of the left wheel is A1x+B1y+C1z+D1=0, and the plane equation of the reference inner side surface of the right wheel is A2x+B2y+C2z+D2=0;

[0023] In step ②, the plane A1x+B1y+C1z+D1′=0 is fitted using the three-dimensional point information in the structured light image on the inner side of the left wheel, and is recorded as the measured inner side of the left wheel;

[0024] The plane A2x+B2y+C2z+D2'=0 is fitted using the three-dimensional point information in the structured light image on the inner side of the right wheel, and is recorded as the measured inner side of the right wheel;

[0025]

[0026] Preferably, the method further comprises the following steps:

[0027] The left auxiliary plane A3x+B3y+C3z+D3=0 is fitted using the three-dimensional point information in the structured light measured image on the inner side of the left wheel;

[0028] The right auxiliary plane A4x+B4y+C4z+D4=0 is fitted using the three-dimensional point information in the structured light measured image on the inner side of the right wheel;

[0029] Calculate the cosine of the angle between the normal vector of the left auxiliary plane and the normal vector of the left wheel reference inner surface:

[0030]

[0031] Calculate the cosine of the angle between the normal vector of the right auxiliary plane and the normal vector of the right wheel reference inner surface:

[0032]

[0033] Use the following method to d p Make another correction to obtain the final train wheel set inner distance d c :

[0034] When the difference between cosθ1 and cosθ2 is less than the threshold, or C =d P cosθ1 or d C =d P cosθ2;

[0035] When the difference between cosθ1 and cosθ2 is not less than the threshold, Among them, d D is the theoretical / measured diameter of the wheel.

[0036] Preferably, the threshold value is between 0.1 and 0.5.

[0037] When this method is applied, only two optical sensors and standard parts are needed. The deviation between the measured inner side and the reference inner side is used to correct the reference distance d0, and the measured inner side distance is obtained. Therefore, this method does not require the operator to perform complicated calibration steps, and performs external parameter calibration / unified coordinate system on the sensor, which simplifies the operation process, saves detection time and cost, and reduces the proficiency requirements for operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a schematic diagram of the top view of the installation position of the structured light vision sensor in a specific implementation manner;

[0039] Figure 2 It is a schematic diagram of the side structure of the installation position of the structured light vision sensor in a specific implementation manner;

[0040] Figure 3 A schematic diagram showing the placement of the mold within the field of view of a structured light vision sensor;

[0041] Figure 4 To solve d C Process diagram;

[0042] Figure 5 Schematic diagram of the inner side and inner distance of the wheelset. DETAILED DESCRIPTION

[0043] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0044] An online measurement method for the inner distance of a train wheel set, such as Figure 1 As shown, structured light vision sensors 1 are respectively installed on the inner side of the track 3, and the line between the installation positions of the two structured light vision sensors is approximately perpendicular to the extension direction of the track (the angle is less than 10°);

[0045] In actual application, two structured light vision sensors are installed side by side under the track, and the camera lens shoots the inner side of the wheel from below diagonally upward;

[0046] The structured light vision sensor 1 includes a structured light projector 11 and a camera 12;

[0047] The structured light projectors 11 of the two structured light vision sensors are used to project multi-point structured light / multi-line structured light / grating structured light onto the inner side surfaces of the left and right wheels 2 respectively, and the camera collects the structured light image; the internal parameters of each structured light vision sensor 1 are pre-calibrated;

[0048] In order to ensure the effect of plane fitting, when a point structured light vision sensor is used, the number of point structured lights projected on the inner side surface is greater than 3; when a line structured light vision sensor is used, the number of line structured lights projected on the inner side surface is greater than 3;

[0049] Obtain the inner distance of the train wheelset by following the steps below:

[0050] ① When the train reaches the area to be tested, Figure 2 As shown, the wheel triggers the sensor 4, the sensor sends out a vehicle-oncoming signal, and the two structured light vision sensors are turned on synchronously to project structured light onto the inner sides of the left and right wheels respectively and obtain structured light measured images; the sensor is fixed around the structured light vision sensor; photoelectric beam switches, magnetic steel sensors, etc. can be selected.

[0051] ② Use the three-dimensional point information in the two structured light measured images to fit planes respectively, which are recorded as the measured inner side of the left wheel and the measured inner side of the right wheel;

[0052] Specifically, the three-dimensional coordinate point cloud of the inner surface in each coordinate system is reconstructed through the pre-calibrated sensor internal system parameters (structured light sensor internal parameters) and the two-dimensional feature points on the image, and the three-dimensional coordinate point cloud is used to fit the plane (such as the least squares method).

[0053] ③ Retrieve the pre-stored reference inner side of the left wheel, the reference inner side of the right wheel and the reference distance d0 between the two wheels;

[0054] Calculate the deviation of the measured inner side surface of the left wheel relative to the reference inner side surface of the left wheel and the deviation of the measured inner side surface of the right wheel relative to the reference inner side surface of the right wheel respectively;

[0055] The two deviations are used to correct d0, and the corrected distance value is recorded as the inner distance of the train wheelset.

[0056] Specifically, before step ①, the following steps are used to store the left wheel reference inner side surface, the right wheel reference inner side surface, and the distance d0 between the two reference inner side surfaces:

[0057] The normal wheelsets or dies without wear shall be recorded as standard parts; Figure 3 As shown, the mold includes a left side surface 5, a right side surface 6 and a rigid connecting rod 7 for keeping the relative positions of the two fixed. The left side surface and the right side surface are parallel to each other, and the distance between the inner side surfaces of the two is equal to the inner side distance between normal wheels without wear (the standard inner side distance is within the range of 1353±2mm);

[0058] Place the standard part within the field of view of the structured light vision sensor (on the track); manually measure the distance between the left and right inner sides of the standard part and record it as the reference distance d0 (such as using an inner distance ruler, a wheel-less lathe, or other tools for precise measurement);

[0059] The structured light vision sensor is turned on synchronously to project structured light onto the left and right inner sides of the standard component and obtain a structured light standard image;

[0060] The three-dimensional point information in the two structured light standard images is used to fit planes respectively, which are recorded as the left wheel reference inner side surface and the right wheel reference inner side surface.

[0061] In specific implementation, two deviations are used to correct d0, and the corrected distance value d is obtained. P =d0+d1-d2

[0062] Wherein, d1 is the deviation of the measured inner side of the left wheel relative to the reference inner side of the left wheel. When the measured inner side of the left wheel deviates toward the right wheel relative to the reference inner side of the left wheel, the generated deviation is a negative value, otherwise it is a positive value;

[0063] d2 is the deviation of the actually measured inner side surface of the right wheel relative to the reference inner side surface of the right wheel. When the actually measured inner side surface of the right wheel is offset toward the left wheel relative to the reference inner side surface of the right wheel, the offset is a negative value, otherwise it is a positive value.

[0064] Specifically, Figure 4 As shown, the plane equation of the left wheel reference inner side is A1x+B1y+C1z+D1=0(P 左基 ), the plane equation of the right wheel reference inner side is A2x+B2y+C2z+D2=0(P 右基 );

[0065] In step ②, the preferred implementation is: using the three-dimensional point information in the structured light image on the inner side of the left wheel to fit the plane A1x+B1y+C1z+D1'=0, recorded as the measured inner side of the left wheel ( Figure 4 The dashed line in the 左实 );

[0066] The plane A2x+B2y+C2z+D2'=0 is fitted using the three-dimensional point information in the structured light image on the inner side of the right wheel, and is recorded as the measured inner side of the right wheel ( Figure 4 The dashed line in the 右实 ).

[0067] calculate:

[0068]

[0069] In order to further reduce the solution error caused by the non-parallel installation position of the wheelset itself, the serpentine motion of the train, the reference plane placement not parallel to the track and other problems, as a preferred implementation, the following steps are also performed:

[0070] like Figure 4 As shown, the left auxiliary plane A3x+B3y+C3z+D3=0 (P 左辅 );

[0071] The right auxiliary plane A4x+B4y+C4z+D4=0(P 右辅 );

[0072] Calculate the cosine of the angle between the normal vector of the left auxiliary plane and the normal vector of the left wheel reference inner surface:

[0073]

[0074] Calculate the cosine of the angle between the normal vector of the right auxiliary plane and the normal vector of the right wheel reference inner surface:

[0075]

[0076] Use the following method to p Make another correction to obtain the final train wheel set inner distance d c :

[0077] When the difference between cosθ1 and cosθ2 is less than a threshold value (e.g. 0.2), or C =d P cosθ1 or d C =d P cosθ2;

[0078] When the difference between cosθ1 and cosθ2 is not less than a threshold value (e.g. 0.2), Among them, dD is the theoretical / measured diameter of the wheel.

[0079] The threshold value is between 0.1 and 0.5.

[0080] The foregoing description of the specific exemplary embodiments of the present invention is presented for the purpose of illustration and description. The foregoing description is not intended to be exhaustive, nor is it intended to limit the present invention to the precise form disclosed, and it is apparent that many changes and variations are possible in light of the above teachings. The exemplary embodiments are selected and described to explain the specific principles of the present invention and its practical application, so that other technicians in the field can realize and utilize the various exemplary embodiments of the present invention and its different selected forms and modified forms. The scope of the present invention is intended to be limited by the appended claims and their equivalents.

Claims

1. An online measurement method for the inner distance of a train wheel set, characterized in that: Structured light vision sensors are respectively installed on the inner side of the track, and the line between the installation positions of the two structured light vision sensors is approximately perpendicular to the extension direction of the track; Two structured light vision sensors are used to project multi-point structured light / multi-line structured light / grating structured light onto the inner sides of the left and right wheels respectively; Obtain the inner distance of the train wheelset by following the steps below: ① When the train travels to the area to be tested, the wheel triggers the sensor, the sensor sends out a signal of an approaching vehicle, and the two structured light vision sensors are turned on synchronously, projecting structured light onto the inner sides of the left and right wheels respectively and acquiring structured light measured images; the sensors are fixed around the structured light vision sensors; ② Use the three-dimensional point information in the two structured light measured images to fit planes respectively, which are recorded as the measured inner side of the left wheel and the measured inner side of the right wheel; ③ Retrieve the pre-stored reference inner side of the left wheel, the reference inner side of the right wheel and the reference distance d0 between the two wheels; Calculate the deviation of the measured inner side surface of the left wheel relative to the reference inner side surface of the left wheel and the deviation of the measured inner side surface of the right wheel relative to the reference inner side surface of the right wheel respectively; The two deviations are used to correct d0, and the corrected distance value is recorded as the inner distance of the train wheelset.

2. The method for online measurement of the inner distance of a train wheel pair according to claim 1, characterized in that: Before step ①, the following steps are used to store the left wheel reference inner side surface, the right wheel reference inner side surface, and the distance d0 between the two reference inner side surfaces: A normal wheel set or mold without wear is recorded as a standard part; the mold includes a left side surface, a right side surface and a rigid connecting rod to keep the relative positions of the two fixed, the left side surface and the right side surface are parallel to each other, and the distance between the inner side surfaces of the two is equal to the inner side distance between normal wheels without wear; The standard part is placed within the field of view of the structured light vision sensor; the distance between the left and right inner sides of the standard part is manually measured and recorded as the reference distance d0; The structured light vision sensors are turned on synchronously to project structured light onto the left and right inner sides of the standard component and obtain structured light standard images; The three-dimensional point information in the two structured light standard images is used to fit planes respectively, which are recorded as the left wheel reference inner side surface and the right wheel reference inner side surface.

3. The method for online measurement of the inner distance of a train wheel pair according to claim 1, characterized in that: Use the two deviations to correct d0 and obtain the corrected distance value d P =d0+d1-d2 Among them, d1 is the deviation of the actually measured inner side surface of the left wheel relative to the reference inner side surface of the left wheel, and d2 is the deviation of the actually measured inner side surface of the right wheel relative to the reference inner side surface of the right wheel.

4. The method for online measurement of the inner distance of a train wheel pair as claimed in claim 3, characterized in that: The plane equation of the reference inner side of the left wheel is A1x+B1y+C1z+D1=0, and the plane equation of the reference inner side of the right wheel is A2x+B2y+C2z+D2=0; In step ②, the plane A1x+B1y+C1z+D1′=0 is fitted using the three-dimensional point information in the structured light image on the inner side of the left wheel, and is recorded as the measured inner side of the left wheel; The plane A2x+B2y+C2z+D2'=0 is fitted using the three-dimensional point information in the structured light image on the inner side of the right wheel, and is recorded as the measured inner side of the right wheel; 5. The method for online measurement of the inner distance of a train wheel pair according to claim 4, characterized in that: The following steps are also included: The left auxiliary plane A3x+B3y+C3z+D3=0 is fitted using the three-dimensional point information in the structured light measured image on the inner side of the left wheel; The right auxiliary plane A4x+B4y+C4z+D4=0 is fitted using the three-dimensional point information in the structured light measured image on the inner side of the right wheel; Calculate the cosine of the angle between the normal vector of the left auxiliary plane and the normal vector of the left wheel reference inner surface: Calculate the cosine of the angle between the normal vector of the right auxiliary plane and the normal vector of the right wheel reference inner surface: Use the following method to d p Make another correction to obtain the final train wheel set inner distance d c : When the difference between cosθ1 and cosθ2 is less than the threshold, or C =d P cosθ1 or d C =d P cosθ2; When the difference between cosθ1 and cosθ2 is not less than the threshold, Among them, d D is the theoretical / measured diameter of the wheel.

6. The method for online measurement of the inner distance of a train wheel pair as claimed in claim 5, characterized in that: The threshold value is between 0.1 and 0.5.

Images