CN101576375A - Fast processing method of laser vision image of steel rail wear - Google Patents

Abstract

The invention discloses a fast processing method for laser visual images of rail abrasion, which includes: projecting two or more light planes parallel to each other and perpendicular to the axial direction of the rail, and selecting one light plane to project an image of the light bar on the rail waist and the rail head , calculate the coordinates of the center of the light strip at the rail waist in the light plane coordinate system, and calculate the center coordinates of the large circle and small circle at the rail waist in the light plane coordinate system; calculate the selected light plane coordinate system and the standard rail section profile coordinate system transformation matrix between them; in the image plane of the visual sensor, respectively determine the intersection of the line segment between the vertical wear constraint point and the horizontal wear constraint point and the light bar of the rail head; reverse transform the image coordinates of the vertical wear point and the horizontal wear point to In the coordinate system of the standard rail section profile, determine the coordinates of the vertical wear point and the horizontal wear point of the rail in the standard rail section profile coordinate system, and the vertical and horizontal distances between the vertical wear point and the horizontal wear point of the rail and the standard rail profile, namely is the wear value of the measured rail. The invention greatly improves the detection efficiency of the rail wear value.

Description

A rapid processing method for laser vision images of rail wear

technical field

The invention relates to a fast processing method for rail wear laser visual images.

Background technique

The condition of railways is closely related to safety, especially for high-speed trains. Regular inspection of rails is very important to plan properly and keep maintenance costs low. Measuring rail wear and deformation at an early stage can help create better maintenance schedules. In recent years, with the rapid increase of my country's railway speed, passenger and freight volume and traffic density, the wear and tear of rails has become increasingly serious. In order to meet the high-speed development of railway transportation, fast and high-precision measurement of rail wear is an important topic that railway departments at home and abroad have been studying deeply.

According to different detection methods, the existing rail wear measurement devices can be roughly divided into contact type and non-contact type measurement devices. Among them, the measurement accuracy of the contact measurement device is high, but the operation is complicated and the measurement efficiency is low, so it is not suitable for online measurement. Non-contact measuring devices are suitable for dynamic measurements. Currently representative products include the Laserrail 3000 track measurement system and comprehensive inspection vehicle from ImageMap in the United States, the Roger2000 comprehensive inspection vehicle from Italy, and the East-i comprehensive inspection vehicle from Japan. However, the existing detection methods have problems such as high equipment cost, complex structure, and complicated calculation results.

In the Chinese patent application with the publication number CN101144714A and the title of the invention "A Vehicle-mounted Dynamic Measuring Device and Method for Comprehensive Parameters of Rail Wear", a device for measuring rail vertical wear, side wear and wave wear with a grating structured light vision sensor is disclosed and methods. The device and method described in the invention can increase the acquisition density without improving the performance of the image acquisition hardware, thereby meeting the online dynamic measurement requirements of wave wear. However, this method needs to extract the center of all light strips at the rail head and rail waist, so the calculation speed is slow. At the same time, because some light strips at the rail waist will be blocked by the bolts fixing the rail during measurement, these light planes cannot measure rail wear value. Therefore, it is very urgent and necessary to find a rail wear measurement method that can solve the problems of slow speed and easy occlusion of the light plane in the multi-light rail wear measurement.

Contents of the invention

In view of this, the main purpose of the present invention is to provide a rapid processing method for laser visual images of rail wear, which can quickly determine the wear value of the rail.

In order to achieve the above object, technical solution of the present invention is achieved in that way:

A method for rapidly processing laser visual images of rail wear, comprising:

The grating structured light vision sensor projects more than two light planes that are parallel to each other and perpendicular to the axis of the rail. In the image of the grating structured light vision sensor, one of the light planes is arbitrarily selected and projected onto the light bar image of the rail waist and rail head, and the extracted The image coordinates of the center point of the light strip at the rail waist, calculate the coordinate data of the selected light strip center point in the light plane coordinate system according to the mathematical model of the grating structured light visual sensor, and fit the large circle and small circle of the rail waist The coordinates of the center of the circle;

Calculate the conversion matrix between the selected light plane coordinate system and the standard rail section profile coordinate system; determine the vertical wear and horizontal wear constraint points of the rail in the standard rail section profile coordinate system, so that the vertical wear value and horizontal wear of the rail The values are all on the line between the constraint points;

Determine the vertical wear and The coordinates of the constraint point of horizontal abrasion in the image coordinate system;

In the image plane where the light bars of the selected rail waist and rail head are located, respectively determine the intersection points of the line segments between the vertical wear constraint points and the line segments between the horizontal wear constraint points and the rail head light bars, and the intersection points are vertical wear points and horizontal wear points, determining the image coordinates of the vertical wear points and the horizontal wear points; and

Invert the image coordinates of the vertical wear points and horizontal wear points into the standard rail section profile coordinate system, determine the coordinates of the rail vertical wear points and horizontal wear points in the standard rail section profile coordinate system, and the rail vertical wear points and horizontal wear points The vertical distance and horizontal distance from the standard rail profile are the wear value of the measured rail.

Preferably, in the image of the grating structured light vision sensor, the determination of the wear value of the rail where the remaining light planes are projected includes:

Select one of the remaining light planes, use the conversion matrix between the aforementioned light plane coordinate system and the standard rail section outline coordinate system, and the single coordinate system between the currently selected light plane coordinate system and the image coordinate system of the grating structured light vision sensor Response matrix to determine the coordinates of the constraint points of vertical wear and horizontal wear in the image coordinate system;

Respectively determine the intersection points of the line segment between the vertical wear constraint points and the line segment between the horizontal wear constraint points and the currently selected rail head light strip, the intersection points are the vertical wear point and the horizontal wear point respectively, determine the vertical wear point and the horizontal wear point the image coordinates of ; and

Invert the image coordinates of the vertical wear points and horizontal wear points into the standard rail section profile coordinate system, determine the coordinates of the rail vertical wear points and horizontal wear points in the standard rail section profile coordinate system, and the rail vertical wear points and horizontal wear points The vertical distance and horizontal distance from the standard rail profile are the wear value of the currently measured rail.

Preferably, the conversion matrix between the light plane coordinate system where each light plane is located and the image coordinate system of the grating structured light vision sensor is determined in advance by means of calibration;

The conversion matrix between the light plane coordinate system and the standard rail section profile coordinate system is determined by the following method: the center of the large circle and the small circle of the image coordinates of the center point of the light strip at the rail waist determined in the image of the grating structured light vision sensor The coordinates are transformed into the optical plane coordinate system where the optical plane is located, and the coordinates of the center of the large circle and the small circle of the rail waist of the standard rail are determined in the coordinate system of the standard rail section outline, and the center of the large circle and the small circle of the rail waist of the rail is used The coordinates in the light plane coordinate system and the standard rail section outline coordinate system are respectively calculated to calculate the conversion matrix between the light plane coordinate system and the standard rail section outline coordinate system.

Preferably, in the image plane where the light bar of the selected rail waist and rail head is located, determine the intersection of the line segment between the vertical wear constraint point and the horizontal wear constraint point and the rail head light bar, specifically:

Find the pixel point with the largest gray scale on the line segment between the vertical wear constraint point and the horizontal wear constraint point, take the pixel point with the largest gray scale as the center, select a 3×3 pixel point area, and extract the selected rail head The center point of the light strip, after image distortion correction, select two points closest to the center point of the selected rail head light strip, and the line segment between the straight line passing through the two points and the vertical wear constraint point and the horizontal wear constraint point The intersection point is called the intersection point of the line segment between the vertical wear constraint point and the horizontal wear constraint point and the rail head light bar.

The present invention only needs to determine the image coordinates of the center of the light bar at the waist of the rail measured by a certain light plane in the image of the grating structured light vision sensor, and does not need to calculate the relevant data at the rail head and the other light planes at the waist (except wear and tear) Points outside), through the coordinates of the rail waist big circle and small circle in the light plane coordinate system and the standard rail section profile coordinate system, calculate the transformation matrix between the light plane coordinate system and the standard rail section profile coordinate system, and the wear constraint The coordinates of the point in the coordinate system of the standard rail section profile are transformed into the image of the grating structured light visual sensor, thereby determining the coordinate point of the wear point in the image coordinate system of the grating structured light visual sensor, and converting the coordinates of the wear point to the standard In the rail section contour coordinate system, the distance between the wear point and the standard rail cross-section contour can be determined, and the distance is the wear value. Since each light plane is parallel and perpendicular to the rail, and at the same time, when the visual sensor is calibrated, it is ensured that the connection line of the origin of each light plane coordinate system is perpendicular to the light plane and parallel to the direction of the rail axis, so each light plane coordinate system is consistent with the standard The transformation matrices between the coordinate systems of the rail section outline are the same. When calculating the wear points in the other rail images in the image of the grating structured light vision sensor, the coordinates of the wear constraint points in the standard rail section outline coordinate system are directly converted to the grating It can be determined from the image of the structured light vision sensor, therefore, the required calculation amount will be greatly reduced, especially in the detection system with more light planes, the efficiency of the present invention to determine the wear value is higher, and at the same time Because this patented method utilizes the parallelism of the light planes, it only needs the data of the center point of the rail waist light strip measured by one light plane to realize the calculation of the wear value of multiple light planes, so this patent method also solves the problem of The problem that the part of the light plane cannot measure the rail wear value caused by the partial light plane being blocked in the middle.

Description of drawings

Fig. 1 is the flow chart of the rapid determination method of rail wear value of the present invention;

Fig. 2 is a schematic diagram of the positions of the large circle and the small circle of the standard rail section profile;

Fig. 3 is the schematic diagram of the light strip formed on the rail by the light plane photographed by the visual sensor;

Fig. 4 is a schematic diagram of the position of the wear constraint point and the wear point in the coordinate system of the standard rail section profile;

Fig. 5 is a schematic diagram of the conversion relationship between the standard rail cross-section profile coordinate system, the light plane coordinate system and the visual sensor image coordinate system.

Detailed ways

The present invention is described in further detail below.

Fig. 1 is the flow chart of the rapid determination method of rail wear value of the present invention, as shown in Fig. 1, the rapid determination method of rail wear value of the present invention comprises the following steps:

Step 101: Randomly select a light plane A among the multiple light planes projected by the grating structured light vision sensor, and use Zhou Fuqiang et al. in the article "A Hybrid Image Processing Method for Extracting Structured Light Strips [J], Optoelectronics·Laser, 2008, 19(11): 1534-1537", the light strip extraction method mentioned in "extracts the image coordinates of the center point of the rail waist light strip.

According to the calibration method mentioned in the article "Field Calibration Method of Line Structured Light Vision Sensor [J], Chinese Journal of Mechanical Engineering, 2004, 40(6): 169-173" by Zhou Fuqiang et al. The homography matrix H _s,im of the light plane coordinate system and the visual sensor image coordinate system. For the light plane coordinate system where each light plane projected by the grating laser is located, the homography matrix of the image coordinate system of the visual sensor is respectively calibrated.

The main process of determining the origin of each light plane coordinate system is to first calculate the intersection point of the straight line where the optical axis of the camera is located on the light plane A, and use the intersection point as the coordinate origin of the light plane A coordinate system, and calculate the perpendicular to the light plane A through the origin. , determine the intersection points of the line and each light plane, and set these intersection points as the coordinate origin of the coordinate system of the light plane. Since each light plane is parallel and perpendicular to the rail, the conversion matrix from each light plane coordinate system to the standard rail section profile coordinate system is guaranteed to be the same.

According to the determined H _s,im , calculate the coordinates of the center point of the rail waist in the light plane A coordinate system.

Among the extracted central points of the light strips of the rail waist, they are respectively distributed on the large circle and the small circle of the rail waist. The more center points are extracted on the large circle and the small circle of the rail waist, the higher the accuracy of the determined coordinates of the center of the large circle and the small circle is.

Among them, the grating structured light vision sensor is a system dedicated to measuring the wear value of the rail, including a grating laser and a camera. Then shoot the light strip projected on the rail by the light plane at an angle set with the light plane and at a certain distance, and complete the extraction of the image coordinates of the center point on the light strip through the image processing module of the system software in the computer; the taken light strip As shown in Figure 3, the light strip in the figure is the light strip image formed by the grating laser projected onto the rail. Since the grating structured light vision sensor is an existing technology, its structure and its image processing methods are roughly the same, and the present invention will not repeat its structural details and how to perform image processing.

Step 102: Calculate the coordinates of the centers of the large circle and the small circle of the rail waist in the light plane A coordinate system.

According to the geometric characteristics of the rail waist section, the extracted data of the center point of the rail waist is divided into the data of the large circle of the rail waist and the data of the small circle of the rail waist. Figure 2 is a schematic diagram of the position of the large circle and the small circle of the standard rail cross-sectional profile. It is the design structure required in the rail waist standard. For details, please refer to the relevant regulations in the rail design standard, and its structure is only schematically given here. The method of radius constraint is used to fit the large circle and the small circle of the rail waist respectively, and the coordinates of the center of the circle are determined. The radii of the large circle and the small circle are calculated and obtained according to the different rail types. For the relevant design standards of the rail waist, please refer to the relevant regulations in the "National Standard of "Hot Rolled Steel Rails for Railways" (GB 2585-2007)".

Step 103: Calculate the conversion matrix T _{b, s} of the coordinate system of the standard rail section profile to the coordinate system of the light plane A of the light plane A.

According to the datum alignment method mentioned in the Chinese patent application whose publication number is CN1776364A, and the invention name is "laser visual dynamic measurement device for rail wear", the conversion between the optical plane coordinate system of optical plane A and the standard rail section profile coordinate system is calculated matrix.

Step 104: Determine rail horizontal and vertical wear constraint points.

According to the definition of rail wear, two horizontal wear constraint points and two vertical wear constraint points are determined in the standard rail section profile coordinate system. Figure 4 is a schematic diagram of the location of wear constraint points _and wear points in the standard rail section profile coordinate system, as shown in Figure 4, the O _b z _b axis of the standard rail section profile coordinate system O b x _by y _b z _b is perpendicular to the standard rail section. Labels 1 and 2 are the constraint points of vertical wear and horizontal wear, respectively. Numbers 3 and 4 are the horizontal and vertical wear points of the rail, respectively. The number 5 is the center point of the light bar extracted from the rail waist after benchmark alignment, the number 6 is the measured profile of the rail head, and the number 7 is the profile of the standard rail section. The "Railway Line Maintenance Rules" stipulates that the horizontal wear is located at a distance of 16 mm from the top surface of the rail on the standard section, and the vertical wear is located at 1/3 of the width of the top surface of the rail. According to the definition of vertical wear, the y coordinate of the vertical wear constraint point is located at 1/3 of the width of the top surface of the rail, which is the position defined by vertical wear. The x coordinate can be determined according to the actual width of the top surface of the rail. The principle is the selected The two wear constraint points can ensure that the actual vertical wear point is between the line segments determined by the two wear constraint points, and the selection of wear constraint points can be determined based on experience. Similarly, the horizontal wear constraint point is also determined according to the aforementioned method of the vertical wear constraint point, which will not be repeated here.

Step 105: Determine the image coordinates p _im of the horizontal and vertical wear constraint points of the rail in the visual sensor image coordinate system.

In the present invention, the image coordinate system of the camera in the grating structured light visual sensor is defined as the visual sensor image coordinate system. Let the coordinates of the rail wear constraint point be P _b = (x _b , y _b , 1) in the standard rail section profile coordinate system, and the image coordinates in the visual sensor image coordinate system be p _im = (x _im , y _im , 1 ). The image coordinates of the rail wear constraint points in the image coordinate system can be calculated by the following formula:

ρ ₁ P _im = H _{s, im} T _{b, s} P _b

In the formula, H _{s, im} is the homography matrix of the light plane coordinate system of the light plane A and the visual sensor image coordinate system, and ρ ₁ is the proportional coefficient.

According to the formula P _im =ρ ₁ H _{s, im} T _{b, s} P _b , the image coordinates of the two constraint points of vertical abrasion and the two constraint points of horizontal abrasion in the visual sensor image coordinate system can be respectively obtained.

Briefly explain the conversion relationship between the aforementioned coordinate systems. Figure 5 is a schematic diagram of the conversion relationship between the standard rail section profile coordinate system, the light plane coordinate system and the visual sensor image coordinate system. As shown in Figure 5, the light plane coordinate system The homography matrix H _s,im to the image coordinate system of the visual sensor is determined through the aforementioned calibration method, wherein the homography matrix between the image coordinate system where each light bar is located and the light plane coordinate system is determined by calibration. The conversion matrix T _{b, s} of the coordinate system of the standard rail cross-section outline and the coordinate system of the light plane is calculated through the method of the aforementioned step 103 . Correspondingly, the conversion matrix between the light plane coordinate system and the standard rail section outline coordinate system is T _{b, s} ^-1 , which is the transposition of T _{b, s} , and the distance between the visual sensor image coordinate system and the light plane coordinate system The homography matrix H ^-1 _s,im of H ^-1 _s,im is the transpose of H _s,im .

Step 106: Determine the image coordinates of the horizontal and vertical wear points of the rail in the visual sensor image coordinate system

Taking the calculation of the vertical wear point as an example, the calculation process of the image coordinates of the wear point is introduced.

It should be noted that the intersection of the line segment formed by the vertical wear constraint points and the image of the light strip formed by the light plane A is the vertical wear point. Specifically: in the image coordinate system of the visual sensor, first search for the pixel with the largest gray scale on the image on the line segment formed by the two vertical wear constraint points, and select a 3×3 pixel area centered on the largest pixel point, Extract the 3×3 region according to the strip extraction method mentioned in Zhou Fuqiang et al.’s article “Hybrid Image Processing Method for Structured Light Strip Extraction [J], Optoelectronics Laser, 2008, 19(11): 1534-1537” The center point of the light bar, the coordinates of the center point of the light bar are corrected for distortion, and then the two center points of the light bar closest to the offline segment are selected, and the intersection point between the line connecting the two points and the line segment is defined as the vertical wear point in the image coordinate system coordinates below.

Similarly, the coordinates of the horizontal wear point in the visual sensor image coordinate system can also be obtained accordingly.

和H_s，im以及T_b，s，求解磨耗点在标准钢轨截面轮廓坐标系下的坐标具体变换关系如下：Step 107: Image coordinates of wear points through the rail

and H _{s, im} and T _{b, s} to solve the coordinates of the wear point in the standard rail section profile coordinate system The specific conversion relationship is as follows:

${\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; P</span>}}}_{\mathrm{<span\; class="notranslate">\; b</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; im</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; p</span>}}}_{\mathrm{<span\; class="notranslate">\; im</span>}}$

In the formula, _ρ2 is the proportionality coefficient.

According to the formula ${\mathrm{\&rho;}}_2{\tilde{P}}_b=T_{b,\mathrm{the\; s}}^{-1}{h}_{\mathrm{the\; s},\mathrm{im}}^{-1}{\tilde{p}}_{\mathrm{im}},$ Calculate the coordinates of the vertical wear point in the standard rail section profile coordinate system. Similarly, the coordinates of the horizontal wear point in the standard rail section profile coordinate system can also be determined according to the above method.

Step 108: Calculate the vertical wear value and the horizontal wear value of the rail.

By comparing the coordinates of the wear point calculated in step 107 in the coordinate system of the standard rail section profile with the coordinates of the defined position of the wear point in the standard rail section profile, calculate the rail wear value, which includes vertical wear value and horizontal wear value .

As shown in Figure 1, the rapid determination method of rail wear value of the present invention also includes the following steps:

Step 109: Calculate the vertical wear value and the horizontal wear value of the steel rail measured on the remaining light planes.

Since the grating laser ensures that the emitted light strips are parallel and perpendicular to the steel rail during installation, and the vision sensor calibration is completed according to the calibration process in step 101, it can be guaranteed that Tb _,s of all light planes are the same, repeat step 105 -107, the rail wear values measured on different light planes can be obtained separately, only in the formula ρ ₁ P _im =H _{s, im} T _{b, s} P _b , ${\mathrm{\&rho;}}_2{\tilde{P}}_b=T_{b,\mathrm{the\; s}}^{-1}{h}_{\mathrm{the\; s},\mathrm{im}}^{-1}{\tilde{p}}_{\mathrm{im}}$ It is enough to replace H _{s and im} with the homography matrix of the light plane coordinate system where the corresponding light plane is located and the visual sensor image coordinate system.

Because in the present invention, when determining the rail wear points measured by multiple optical planes, only the center point coordinates of the large circle and the small circle of the derailed rail waist can be calculated according to the center point of the light bar at the measured rail waist of one optical plane, and it is not necessary to process all the multiple optical planes All light bar images, therefore, the amount of data processed is greatly reduced. After determining the homography matrix of the light plane A coordinate system and the visual sensor image coordinate system, when determining the rail wear values measured on the remaining light planes, directly use the previously determined homography of the light plane A coordinate system and the visual sensor image coordinate system matrix, thus greatly saving the calculation workload and improving the efficiency of determining the rail wear value. At the same time, this method also solves the problem that the rail wear cannot be measured on some light planes due to bolt occlusion. The rail wear value determined by the present invention can also guarantee accuracy.

The technical solutions of the present invention will be further described below through specific examples.

The test uses a CCD camera, three grating lasers and the corresponding image acquisition system to form a rail wear measuring instrument. Among them, the camera uses MINTRON high-sensitivity CCD camera, the image size is 768 pixels × 576 pixels, equipped with ANENIR 12mm CCTV lens. The grating laser is a 5mw line laser. The measured rail is 43-gauge type.

The leftmost light strip plane in Figure 3 is defined as light plane A, and the remaining two are light planes B and C from left to right.

Step 1: Extract the image coordinates of the center point of the light bar at the waist of the track A of the light plane.

Step 2: Determine the coordinates of the center point of the light strip at the waist of track A of the light plane in the light plane coordinate system. The coordinates of the centers of the large and small circles in the light plane coordinate system are determined by fitting. Table 1 gives the coordinates of the center coordinates of the large and small circles in the standard rail outline coordinate system and the light plane coordinate system respectively.

Table 1

Step 3: Calculate the transformation matrix from the standard rail profile coordinate system to the light plane coordinate system. According to the coordinates of the center of the large and small circles in the light plane coordinate system and the standard rail contour coordinate system, determine the conversion matrix T _{b, s} from the standard rail contour coordinate system to the light plane coordinate system.

${\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; 0.879049</span>}& \mathrm{<span\; class="notranslate">\; 0.477488</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 80.148359</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.477488</span>}& \mathrm{<span\; class="notranslate">\; 0.878637</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 54.399825</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]$

Step 4: Determine the coordinates of the constraint points of the rail wear points in the image. First determine the coordinates of the rail wear constraint point in the standard rail contour coordinate system, as shown in Table 2:

x(mm) y(mm) Horizontal Wear Constraint Point 1 50 124 Horizontal Wear Constraint Point 2 20 124 Vertical Wear Constraint Point 1 11.667 140 Vertical wear constraint point 2 11.667 130

Table 2

Then calculate the coordinates of the constraint points of the rail wear point in the visual sensor image coordinate system, as shown in Table 3:

x(pixel) y(pixel) Horizontal Wear Constraint Point 1 480.310 158.168 Horizontal Wear Constraint Point 2 436.837 112.784 Vertical Wear Constraint Point 1 437.958 45.011 Vertical wear constraint point 2 430.187 80.213

table 3

Step 5: Locate the coordinates of the wear point in the image. As shown in Table 4:

x(pixel) y(pixel) horizontal wear point 433.615 63.297 Vertical wear point 460.588 136.053

Table 4

Step 6: Determine the image coordinates of the rail wear points measured on the remaining two optical planes B and C according to the above method. Calculate the coordinates of the rail wear point under the standard rail contour coordinates: according to T _{b, s} and H _{s, im} through the formula ${\mathrm{\&rho;}}_2{\tilde{P}}_b=T_{b,\mathrm{the\; s}}^{-1}{h}_{\mathrm{the\; s},\mathrm{im}}^{-1}{\tilde{p}}_{\mathrm{im}}$ Calculate the coordinates of the rail wear point under the standard rail contour coordinates, and all the results are shown in Table 5:

table 5

Step 7: Calculate the rail wear value. The horizontal and vertical wear values of the rail are determined by comparing with the size of the standard rail section profile at the wear point. The specific values are shown in Table 6:

In the above scenario, the calculation time of the entire rail wear value is 1.4ms.

Those skilled in the art should understand that the above example is only an illustration. In the sixth step, the rail wear value of the light strip corresponding to the light planes B and C can be determined after the rail wear value of the light strip corresponding to the light plane A is determined. calculate.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention.

Claims (4)

1. A rapid processing method for rail wear laser visual images, characterized in that it comprises: The grating structured light vision sensor projects more than two light planes that are parallel to each other and perpendicular to the axis of the rail. In the image of the grating structured light vision sensor, one of the light planes is arbitrarily selected and projected onto the light bar image of the rail waist and rail head, and the extracted The image coordinates of the center point of the light strip at the rail waist, according to the mathematical model of the grating structured light visual sensor, calculate the coordinate data of the selected light strip center point in the light plane coordinate system and fit the large circle and small circle of the rail waist The coordinates of the center of the circle; Calculate the conversion matrix between the selected light plane coordinate system and the standard rail section profile coordinate system; determine the vertical wear and horizontal wear constraint points of the rail in the standard rail section profile coordinate system, so that the vertical wear value and horizontal wear of the rail The values are all on the line segment between the constraint points; Determine the vertical wear and The coordinates of the constraint point of horizontal abrasion in the image coordinate system; In the image plane where the light strips of the selected rail waist and rail head are located, respectively determine the intersection points of the line segment between the vertical wear constraint point and the horizontal wear constraint point and the rail head light strip, and the intersection points are the vertical wear points respectively and the horizontal wear point, determine the image coordinates of the vertical wear point and the horizontal wear point; and Invert the image coordinates of the vertical wear points and horizontal wear points into the standard rail section profile coordinate system, determine the coordinates of the rail vertical wear points and horizontal wear points in the standard rail section profile coordinate system, and the rail vertical wear points and horizontal wear points The vertical distance and horizontal distance from the standard rail profile are the wear value of the measured rail. 2. The method according to claim 1, characterized in that, in the image of the grating structured light visual sensor, the determination of the rail wear value at the projected position of the remaining light planes comprises: Select one of the remaining light planes, use the conversion matrix between the aforementioned light plane coordinate system and the standard rail section outline coordinate system, and the single coordinate system between the currently selected light plane coordinate system and the image coordinate system of the grating structured light vision sensor Response matrix to determine the coordinates of the constraint points of vertical wear and horizontal wear in the image coordinate system; Respectively determine the intersection points of the line segment between the vertical wear constraint points and the line segment between the horizontal wear constraint points and the currently selected rail head light strip, the intersection points are the vertical wear point and the horizontal wear point respectively, determine the vertical wear point and the horizontal wear point the image coordinates of ; and Invert the image coordinates of the vertical wear points and horizontal wear points into the standard rail section profile coordinate system, determine the coordinates of the rail vertical wear points and horizontal wear points in the standard rail section profile coordinate system, and the rail vertical wear points and horizontal wear points The vertical distance and horizontal distance from the standard rail profile are the wear value of the currently measured rail. 3. The method according to claim 1 or 2, characterized in that the conversion matrix between the light plane coordinate system where each light plane is located and the image coordinate system of the grating structured light vision sensor is determined in advance by means of calibration; The conversion matrix between the light plane coordinate system and the standard rail section profile coordinate system is determined by the following method: the center of the large circle and the small circle of the image coordinates of the center point of the light strip at the rail waist determined in the image of the grating structured light vision sensor The coordinates are transformed into the optical plane coordinate system where the optical plane is located, and the coordinates of the center of the large circle and the small circle of the rail waist of the standard rail are determined in the coordinate system of the standard rail section outline, and the center of the large circle and the small circle of the rail waist of the rail is used The coordinates in the light plane coordinate system and the standard rail section outline coordinate system are respectively calculated to calculate the conversion matrix between the light plane coordinate system and the standard rail section outline coordinate system. 4. The method according to claim 1, characterized in that, in the image plane where the light strips of the selected rail waist and rail head are located, determine the line segment and rail distance between the vertical wear constraint point and the horizontal wear constraint point. The intersection point of the head light bar, specifically: Find out the pixel point with the largest gray scale on the line segment between the vertical wear constraint points and the line segment between the horizontal wear constraint points respectively, take the pixel point with the largest gray scale as the center, select a 3×3 pixel point area, and extract the selected The center point of the rail head light strip, after image distortion correction, select two points closest to the center point of the selected rail head light strip, and the line segment and horizontal wear between the straight line passing through the two points and the vertical wear constraint point The intersection of the line segments between the constraint points is called the intersection of the line segments between the vertical wear constraint points and the line segment between the horizontal wear constraint points and the rail head light bar.

Images