阅读说明：本技术 钢轨廓形及磨耗检测系统的标定方法及检定方法 (Calibration method and verification method of steel rail profile and abrasion detection system ) 是由 王昊 陈春雷 赵延峰 王胜春 王宁 赵鑫欣 方玥 周谦 甄宇峰 郝晋斐 夏承亮 于 2019-10-11 设计创作，主要内容包括：本发明公开了一种钢轨廓形及磨耗检测系统的标定方法及检定方法,该标定方法包括：在棋盘格板置放于承载平台上时,将提取的棋盘格板图像的亚像素角点的世界坐标与图像坐标进行关联标定,确定各相机的标定参数；取下棋盘格板将标定板置放于承载平台上时,根据各相机的标定参数将提取到的标定板的激光光条图像中激光光条中心的亚像素坐标转换为世界坐标后,与预设标准世界坐标对齐确定两相机的拼接关系参数。本发明利用棋盘格板及标定板分别实现对单侧相机及两相机拼接关系的标定,能够提高标定精度。(The invention discloses a calibration method and a calibration method of a steel rail profile and wear detection system, wherein the calibration method comprises the following steps: when the checkerboard is placed on the bearing platform, the world coordinates of the sub-pixel angular points of the extracted checkerboard image and the image coordinates are subjected to associated calibration, and calibration parameters of each camera are determined; when the checkerboard is taken down and the calibration board is placed on the bearing platform, the sub-pixel coordinates of the centers of the laser light stripes in the laser light stripe images of the extracted calibration board are converted into world coordinates according to the calibration parameters of the cameras, and then the world coordinates are aligned with the preset standard world coordinates to determine the splicing relation parameters of the two cameras. The invention utilizes the checkerboard and the calibration board to respectively realize the calibration of the splicing relation of the single-side camera and the two cameras, and can improve the calibration precision.)

1. A calibration method of a steel rail profile and abrasion detection system is characterized in that the steel rail profile and abrasion detection system comprises a steel rail profile beam, two cameras and two lasers; the rail profile beam is provided with an installation surface, two first accommodating parts are concavely arranged on the installation surface, the two phase machines can be respectively detachably positioned and installed in the two first accommodating parts, and the two phase machines are symmetrically arranged by a longitudinal reference plane vertical to the installation surface; two second accommodating parts are further concavely arranged on the mounting surface, the two lasers can be detachably positioned and mounted in the two second accommodating parts respectively, the two lasers are symmetrically arranged on a longitudinal reference plane, and laser light emitted by the two lasers can intersect in the longitudinal reference plane; a mark calibrating installation for demarcating rail profile and wearing and tearing detecting system includes:

the device comprises a bearing platform and a calibration piece, wherein the calibration piece comprises a checkerboard and a calibration board, and the bearing platform can be detachably connected with a track profile beam; the mounting surface is vertically arranged during calibration, the longitudinal center line of the bearing platform is parallel to the mounting surface and is positioned in a longitudinal reference plane, and the lenses of the two cameras face the bearing platform; the calibration method comprises the following steps that a calibration piece is placed on a bearing platform in the calibration process, and laser lines emitted by two lasers are collinear with a positioning groove of the calibration piece;

the calibration method of the steel rail profile and abrasion detection system comprises the following steps:

placing the checkerboard on a bearing platform, and extracting sub-pixel angular points of checkerboard images acquired by a camera; in the calibration process, laser lines emitted by the two lasers are collinear with the positioning grooves of the checkerboard;

performing correlation calibration on world coordinates of sub-pixel angular points of the checkerboard image and image coordinates of the checkerboard image, and determining calibration parameters of each camera;

taking down the checkerboard, placing the calibration board on the bearing platform, and extracting sub-pixel coordinates of the center of the laser light stripe in the laser light stripe image of the calibration board collected by the camera; laser lines emitted by the two lasers are collinear with the positioning groove of the calibration plate in the calibration process;

converting the sub-pixel coordinates of the center of the laser light bar into world coordinates according to the calibration parameters of each camera;

and determining splicing relation parameters of the two cameras by aligning the world coordinate of the center of the laser light bar with a preset standard world coordinate.

2. The calibration method according to claim 1, wherein the sub-pixel coordinates of the center of the laser light bar in the laser light bar image are extracted by:

dividing the laser light bar into a plurality of sections based on a deep learning network according to the gray scale and gradient direction characteristics of the laser light bar;

determining a normal direction for each segment based on gradient histogram statistics; each section corresponds to a normal direction;

and determining a maximum light intensity point in the normal direction of each section, and taking the maximum light intensity point as a center to extract the sub-pixel coordinates of the center of the laser light bar along the normal direction according to a gray scale gravity center method.

3. The calibration method as claimed in claim 1, wherein the checkerboard plate is a plate body having a predetermined thickness and having a checkerboard formed on an upper surface thereof, and the calibration plate is a rectangular plate body having a predetermined thickness.

4. The calibration method according to claim 1, wherein the calibration device further comprises a base and an adjusting mechanism, the base is detachably connected to the track profile beam, the adjusting mechanism is detachably disposed on the base, and the supporting platform is detachably mounted on the adjusting mechanism, the adjusting mechanism is capable of lifting the supporting platform along a longitudinal axis, swinging along a first axis horizontally disposed and swinging along a second axis horizontally disposed, the first axis is located in a longitudinal reference plane, the second axis is perpendicular to the first axis, and an intersection point of the first axis and the second axis is located on the longitudinal axis.

5. The calibration method according to claim 4, wherein a handle is provided on the base.

6. The calibration method according to claim 4, wherein the adjustment mechanism comprises a lifting table, a first rotating table and a second rotating table sequentially connected from bottom to top, the lifting table is connected to the base, the supporting platform is disposed on the second rotating table, the lifting table can drive the first rotating table, the second rotating table and the supporting platform to lift, the first rotating table can drive the second rotating table and the supporting platform to rotate and swing along a first axis, and the second rotating table can drive the supporting platform to rotate and swing along a second axis.

7. The calibration method of claim 1, wherein the calibration verification device further comprises an illumination assembly, the illumination assembly comprising:

the lamp and the supporting and adjusting piece, the lower end of the supporting and adjusting piece is arranged on the base, and the lamp is connected to the upper end of the supporting and adjusting piece and is positioned above the bearing platform.

8. A verification method based on the calibration method according to any one of claims 1 to 7, wherein the calibration verification device is further used for verifying the rail profile and wear detection system, the calibration verification device further comprises a verification piece, the verification piece is taken off from the bearing platform and placed on the bearing platform during verification, and the verification method comprises:

extracting sub-pixel coordinates of the center of a laser light bar in a laser image of the detection piece collected by a camera; in the verification process, laser lines emitted by the two lasers are collinear with the positioning groove of the detection piece;

determining actually measured profile data of the verification piece according to sub-pixel coordinates of the centers of laser light bars in the laser image, calibration parameters of each camera and splicing parameters of the two cameras;

and matching and comparing the actually measured profile data of the detecting piece with the standard profile data to determine the profile detection precision and the abrasion detection precision.

9. The certification method according to claim 8, wherein the certification part comprises a standard rail block, the standard profile data comprises standard rail profile data, and the profile detection accuracy is determined by using the standard rail block through the following steps:

extracting sub-pixel coordinates of the center of a laser light bar in a laser image of a standard rail test block acquired by a camera; laser lines emitted by the two lasers are collinear with the positioning groove of the standard rail test block in the verification process;

determining actually measured profile data of the standard rail test block according to the sub-pixel coordinates of the center of the laser light strip in the laser image, the calibration parameters of each camera and the splicing parameters of the two cameras;

and matching and comparing the actually measured profile data of the standard track test block with the standard track profile data to determine the profile detection precision.

10. The certification method according to claim 9, wherein the certification part comprises a wear rail test block, the edge of the wear rail test block is concavely provided with the simulated defects, the standard profile data comprises standard wear rail profile data, and the wear rail test block is used for determining the wear detection precision by the following steps:

extracting sub-pixel coordinates of the center of a laser light strip in a laser image of the wear rail test block acquired by a camera; laser lines emitted by the two lasers are collinear with the positioning groove of the abrasion rail test block in the verification process;

determining actual measurement profile data of the abrasion rail test block according to the sub-pixel coordinates of the center of the laser light strip in the laser image, the calibration parameters of each camera and the splicing parameters of the two cameras;

and matching and comparing the actually measured profile data of the wear rail test block with the standard wear rail profile data to determine the wear detection precision.

Technical Field

The invention relates to the technical field of railway infrastructure detection, in particular to a calibration method and a verification method of a steel rail profile and abrasion detection system.

Background

This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.

Rails are important components of railroad tracks, directly bearing train loads and guiding the operation of wheels. Whether the steel rail is in good condition is directly related to whether the train runs safely, stably and uninterruptedly at the specified speed. The steel rail is inevitably worn during the running of the train, and particularly the side grinding of the outer strand of the small-radius curve is serious. The degree of rail wear, which affects the service life of the rail, is a basic parameter for making large, medium and small maintenance work plans of the line. Therefore, it is necessary to accurately measure the parameters of the rail full-section profile, the rail abrasion loss, the rail bottom slope and the like, and efficiently and accurately predict the rail abrasion tendency.

The rail profile and wear detection system is a rail profile measuring system based on laser camera structured light measurement and image processing, comprises an image acquisition module, an image processing data analysis module, a processing result display module and the like (waveform and data editing function), and has the capability of performing high-precision detection on the parameters. In order to meet the requirement of high-precision measurement, the precision and the accuracy of the calibration method of the image acquisition module are crucial. At present, a needle plate target calibration method is usually used for the system, but on one hand, because the coplanar adjustment error of a needle plate target surface and a line laser projection surface is large, the object space light spot coordinate value and the actual coordinate value have deviation; on the other hand, the diameter of the needle plate target steel needle is too small, the manufacturing process is difficult, the array position of the steel needle is deviated from the ideal position, and ghost images may appear on part of the steel needle during line laser projection, so that the light spot coordinate value of the planar light spot array is deviated from the actual coordinate value, namely, the pixel point is deviated, and the calibration precision is limited.

Therefore, the existing calibration for the steel rail profile and wear detection system has the problem of poor calibration precision.

Disclosure of Invention

The embodiment of the invention provides a calibration method of a steel rail profile and abrasion detection system, which is used for improving the calibration precision of the steel rail profile and abrasion detection system, wherein the steel rail profile and abrasion detection system comprises a steel rail profile beam, two cameras and two lasers; the rail profile beam is provided with an installation surface, two first accommodating parts are concavely arranged on the installation surface, the two phase machines can be respectively detachably positioned and installed in the two first accommodating parts, and the two phase machines are symmetrically arranged by a longitudinal reference plane vertical to the installation surface; two second accommodating parts are further concavely arranged on the mounting surface, the two lasers can be detachably positioned and mounted in the two second accommodating parts respectively, the two lasers are symmetrically arranged on a longitudinal reference plane, and laser light emitted by the two lasers can intersect in the longitudinal reference plane; a mark calibrating installation for demarcating rail profile and wearing and tearing detecting system includes:

the device comprises a bearing platform and a calibration piece, wherein the calibration piece comprises a checkerboard and a calibration board, and the bearing platform can be detachably connected with a track profile beam; the mounting surface is vertically arranged during calibration, the longitudinal center line of the bearing platform is parallel to the mounting surface and is positioned in a longitudinal reference plane, and the lenses of the two cameras face the bearing platform; the calibration method comprises the following steps that a calibration piece is placed on a bearing platform in the calibration process, and laser lines emitted by two lasers are collinear with a positioning groove of the calibration piece;

the calibration method of the steel rail profile and abrasion detection system comprises the following steps:

placing the checkerboard on a bearing platform, and extracting sub-pixel angular points of checkerboard images acquired by a camera; in the calibration process, laser lines emitted by the two lasers are collinear with the positioning grooves of the checkerboard;

performing correlation calibration on world coordinates of sub-pixel angular points of the checkerboard image and image coordinates of the checkerboard image, and determining calibration parameters of each camera;

taking down the checkerboard, placing the calibration board on the bearing platform, and extracting sub-pixel coordinates of the center of the laser light stripe in the laser light stripe image of the calibration board collected by the camera; laser lines emitted by the two lasers are collinear with the positioning groove of the calibration plate in the calibration process;

converting the sub-pixel coordinates of the center of the laser light bar into world coordinates according to the calibration parameters of each camera;

and determining splicing relation parameters of the two cameras by aligning the world coordinate of the center of the laser light bar with a preset standard world coordinate.

The embodiment of the invention also provides a calibration method based on the calibration method, which is used for calibrating the detection precision of the steel rail profile and wear detection system, the calibration device also comprises a calibration piece, the calibration piece is taken down from the bearing platform and placed on the bearing platform during calibration, and the calibration method comprises the following steps:

extracting sub-pixel coordinates of the center of a laser light bar in a laser image of the detection piece collected by a camera; in the verification process, laser lines emitted by the two lasers are collinear with the positioning groove of the detection piece;

determining actually measured profile data of the verification piece according to sub-pixel coordinates of the centers of laser light bars in the laser image, calibration parameters of each camera and splicing parameters of the two cameras;

and matching and comparing the actually measured profile data of the detecting piece with the standard profile data to determine the profile detection precision and the abrasion detection precision.

In the embodiment of the invention, laser lines emitted by the two lasers during calibration are collinear with the positioning grooves of the checkerboard or the calibration board; the calibration parameters of each camera are determined by extracting sub-pixel angular points of checkerboard images acquired by the cameras, carrying out association calibration on world coordinates and image coordinates of the sub-pixel angular points, and realizing calibration of the single-side camera; and then, on the basis of the calibration parameters of each camera, converting the sub-pixel coordinates of the laser light strip center in the laser light strip image collected by the extracted camera into world coordinates, aligning the world coordinates with preset standard world coordinates to determine the splicing relation parameters of the two cameras, and calibrating the splicing relation of the two cameras. Therefore, the embodiment of the invention respectively realizes the calibration of the single-side camera and the calibration of the splicing relation of the two cameras by utilizing the checkerboard and the calibration board, and can improve the calibration precision of the steel rail profile and wear detection system.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts. In the drawings:

fig. 1 is a schematic structural diagram of a rail profile and wear detection system according to an embodiment of the present invention;

fig. 2 is a schematic perspective view of a calibration and calibration apparatus according to an embodiment of the present invention;

fig. 3 is a schematic side view of a calibration verification apparatus according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of a checkerboard in the calibration verification apparatus provided in the embodiment of the present invention;

fig. 5 is a schematic structural diagram of a calibration plate in the calibration and calibration device provided in the embodiment of the present invention;

FIG. 6 is a flowchart illustrating an implementation of a calibration method according to an embodiment of the present invention;

fig. 7 is a flowchart illustrating an implementation of extracting sub-pixel coordinates of the center of a laser light bar according to an embodiment of the present invention;

FIG. 8 is a flowchart illustrating an implementation of a verification method according to an embodiment of the present invention;

fig. 9 is a schematic structural diagram of a standard rail test block in the calibration and calibration apparatus according to the embodiment of the present invention;

FIG. 10 is a flowchart illustrating an implementation of determining the profile detection accuracy using a standard rail test block according to an embodiment of the present invention;

fig. 11 is a schematic structural diagram of a wear rail test block in the calibration and calibration device provided in the embodiment of the present invention;

fig. 12 is a flowchart for determining the wear detection accuracy by using a wear rail test block according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.

Fig. 1 shows a structural schematic diagram of a rail profile and wear detection system according to an embodiment of the present invention, fig. 2 shows a perspective structural schematic diagram of a calibration verification apparatus according to an embodiment of the present invention, fig. 3 shows a side view structural schematic diagram of the calibration verification apparatus according to an embodiment of the present invention, and for convenience of description, only the parts related to the embodiment of the present invention are shown, and the following details are described:

as shown in fig. 1 to 3, the rail profile and wear detection system includes a rail profile beam 1, a two-phase machine 2 and two lasers 3; the track profile beam 1 is provided with a mounting surface 11, two first accommodating parts 111 are concavely arranged on the mounting surface 11, the two cameras 2 can be respectively detachably positioned and mounted in the two first accommodating parts 111, and the two cameras 2 are symmetrically arranged by a longitudinal reference plane vertical to the mounting surface 11; two second accommodating parts 112 are further concavely arranged on the mounting surface 11, the two lasers 3 can be detachably positioned and mounted in the two second accommodating parts 112 respectively, the two lasers 3 are symmetrically arranged on a longitudinal reference plane, and laser light emitted by the two lasers 3 can intersect in the longitudinal reference plane.

The calibration device 4 for calibrating the rail profile and wear detection system comprises a bearing platform 41 and a calibration piece 42, wherein the calibration piece 42 comprises a checkerboard 421 and a calibration board 422, and the bearing platform 41 can be detachably connected with the rail profile beam 1; the mounting surface 11 is vertically arranged during calibration, the longitudinal center line of the bearing platform 41 is parallel to the mounting surface 11 and is positioned in a longitudinal reference plane, and the lenses of the two cameras 2 face the bearing platform 41; during the calibration process, the calibration piece 42 is placed on the bearing platform 41, and the laser lines emitted by the two lasers 3 are collinear with the positioning groove of the calibration piece 42.

The rail profile and wear detection system serves as a mounting base for the cameras 2, and the two cameras 2 are detachably and fixedly mounted in the two first accommodating portions 111. Meanwhile, the two cameras 2 are symmetrically arranged on a longitudinal reference plane perpendicular to the mounting surface 11, that is, the two cameras 2 are symmetrically arranged on two sides of the longitudinal reference plane perpendicular to the mounting surface 11, and a connecting line between central points of lenses of the two cameras 2 is horizontally arranged, so that the cameras 2 can shoot from the same highly symmetrical position, and accurate splicing calibration can be conveniently carried out on the splicing relation of the two cameras 2 in the later period, thereby ensuring the calibration precision. In addition, after the calibration piece 42 is placed on the bearing platform 41, the posture of the calibration piece 42 needs to be adjusted, the center of the calibration piece 42 should coincide with the center of the bearing platform 41 in the vertical direction, and the laser lines emitted by the two lasers 3 are collinear with the positioning groove of the calibration piece 42, so that the calibration accuracy is ensured.

Fig. 4 shows a structural schematic diagram of a checkerboard in a calibration verification apparatus provided by an embodiment of the present invention, and fig. 5 shows a structural schematic diagram of a checkerboard in a calibration verification apparatus provided by an embodiment of the present invention, and for convenience of description, only the parts related to the embodiment of the present invention are shown, and detailed descriptions are as follows:

as shown in fig. 4, the checkerboard 421 is a plate body having a predetermined thickness and having a checkerboard formed on an upper surface thereof. The checkerboard 421 is photographed by the camera 2, and the checkerboard image is collected, so as to realize the calibration of the single-side camera 2.

As shown in fig. 5, the calibration plate 422 is a rectangular plate body having a predetermined thickness. The camera 2 is used for collecting the laser light strip image of the standard block, converting the sub-pixel coordinates of the laser light strip center in the laser light strip image into world coordinates according to the calibration parameters of each single-side camera 2, and aligning the world coordinates with the preset standard world coordinates to determine the splicing relation parameters of the two cameras 2 so as to realize calibration of the splicing relation of the two cameras 2.

Fig. 6 illustrates an implementation flow of the calibration method provided by the embodiment of the present invention, and for convenience of description, only the portions related to the embodiment of the present invention are illustrated, and the details are as follows:

as shown in fig. 6, the method for calibrating the rail profile and wear detection system includes:

601, placing the checkerboard 421 on the bearing platform 41, and extracting sub-pixel corners of the checkerboard image acquired by the camera 2; in the calibration process, laser lines emitted by the two lasers 3 are collinear with the positioning grooves of the checkerboard 421;

step 602, performing correlation calibration on world coordinates of sub-pixel angular points of the checkerboard image and image coordinates of the checkerboard image to determine calibration parameters of each camera 2;

step 603, taking down the checkerboard 421, placing the calibration board 422 on the bearing platform 41, and extracting the sub-pixel coordinates of the laser light stripe center in the laser light stripe image of the calibration board 422 collected by the camera 2; in the calibration process, laser lines emitted by the two lasers 3 are collinear with the positioning groove of the calibration plate 422;

step 604, converting the sub-pixel coordinates of the center of the laser light bar into world coordinates according to the calibration parameters of each camera 2;

step 605, determining the stitching relationship parameter of the two cameras 2 by aligning the world coordinate of the center of the laser light bar with the preset standard world coordinate.

One-side camera calibration (step 601 to step 602)

The single-sided camera 2 is calibrated first when calibration is performed. At this time, the checkerboard 421 is first placed on the supporting platform 41, positioning slots are respectively disposed on two sides of the checkerboard 421, and the posture of the checkerboard 421 is adjusted so that the laser lines emitted by the two lasers 3 are collinear with the positioning slots of the checkerboard 421.

Then, the depth of field of the laser 3 is adjusted to make the laser line at the railhead position be the thinnest, camera parameters such as the exposure time, the lens focal length and the light source position of the camera 2 are adjusted, the checkerboard image is collected through the camera 2, the collected checkerboard image is required to be clear, and the black and white connection angular points in the checkerboard 421 cannot be shielded.

And after the checkerboard image is obtained, extracting sub-pixel angular points in the checkerboard image. In an embodiment of the invention, the edge of the checkerboard image can be extracted through a Canny operator, the straight line in the edge is extracted through a Hough algorithm, and all sub-pixel angular points in the checkerboard image are extracted through filtering by using the straight line intersection points.

For camera calibration, three coordinate systems are mainly involved: world coordinate system, image coordinate system and camera coordinate system. Here, we select a reference coordinate system in the environment to describe the position of the camera 2, and describe the position of any object in the environment through the reference coordinate system, which is called a world coordinate system. The image coordinate system is generally a coordinate system established by using an optical axis of the camera and an angular point of an image plane (generally, an image center) as an origin and a plane where the image is located. The camera coordinate system is a coordinate system which is constructed by taking the optical center of the camera as an origin and enabling the optical axis of the camera to be perpendicular to the image plane. There is a specific transformation relationship between the three. The calibration of the camera 2 is to determine internal and external parameters of the camera 2. Camera internal parameters refer to parameters that are related only to the internal structure of the camera, and camera external parameters are determined entirely by the orientation of the camera with respect to the world coordinate system.

After the sub-pixel angular points of the checkerboard image are extracted, the world coordinates of the sub-pixel angular points of the checkerboard image and the image coordinates of the checkerboard image are subjected to correlation calibration, so that the calibration parameters of each single-side camera can be determined, and the calibration of each single-side camera 2 is realized.

Two-phase and two-phase machine splicing relation calibration (step 603 to step 605)

After the single-side camera 2 is calibrated, the splicing relationship of the two cameras 2 needs to be calibrated. At this time, the checkerboard 421 is removed from the supporting platform 41, and the calibration board 422 is placed on the supporting platform 41. Wherein the calibration plate 422 is a precision standard block (see fig. 5) of known shape and size. And positioning grooves are respectively arranged at two sides of the calibration plate 422, and the posture of the calibration plate 422 is adjusted so that laser lines emitted by the two lasers 3 are collinear with the positioning grooves of the calibration plate 422.

Then, the camera 2 collects the laser light stripe image of the calibration plate 422, so as to extract the sub-pixel coordinates of the laser light stripe center in the laser light stripe image, the sub-pixel coordinates of the laser light stripe center are converted into world coordinates according to the calibrated calibration parameters of the single-side camera 2, and then the world coordinates of the laser light stripe center are aligned with the preset standard world coordinates to determine the splicing relation parameters of the two cameras 2, so as to realize the calibration of the splicing relation of the two cameras 2.

In the embodiment of the invention, the world coordinates of the sub-pixel angular points and the image coordinates of the checkerboard image are subjected to associated calibration by extracting the sub-pixel angular points of the checkerboard image acquired by the camera 2, and the calibration parameters of each camera 2 are determined to realize the calibration of each camera 2; then, on the basis of the calibration parameters of each camera 2, the extracted sub-pixel coordinates of the laser light strip center in the laser light strip image collected by the camera 2 are converted into world coordinates, and then the world coordinates are aligned with preset standard world coordinates to determine the splicing relation parameters of the two cameras 2, so that the splicing relation of the two cameras 2 is calibrated. Therefore, in the embodiment of the present invention, the checkerboard 421 and the calibration board 422 are used to respectively calibrate the single-side camera 2 and calibrate the splicing relationship of the two cameras 2, so that the calibration accuracy of the rail profile and wear detection system can be improved.

Fig. 7 illustrates an implementation flow of extracting sub-pixel coordinates of the center of the laser light bar in step 603 provided by the embodiment of the present invention, and for convenience of description, only the relevant portions of the embodiment of the present invention are shown, and detailed below:

as shown in fig. 7, extracting sub-pixel coordinates of the center of the laser light stripe includes:

701, dividing a laser light bar into a plurality of sections based on a deep learning network according to the gray scale and gradient direction characteristics of the laser light bar;

step 702, determining the normal direction of each section based on the gradient histogram statistics; each section corresponds to a normal direction;

step 703, determining the maximum light intensity point in the normal direction of each segment, and extracting the sub-pixel coordinates of the center of the laser light bar along the normal direction by taking the maximum light intensity point as the center according to the gray scale gravity center method.

The method comprises the steps of realizing fine segmentation of a steel rail outline light bar image based on a deep learning network, dividing a light bar into a plurality of sections according to the gray scale and gradient direction characteristics of the light bar, and independently extracting the center of the light bar in each section. This operation removes background noise and interference caused by non-uniform diffuse reflection and specular reflection on light bar extraction, and improves the robustness of the algorithm.

The normal main direction of each segmented sub-region is determined based on gradient histogram statistics, each sub-region corresponds to one normal main direction, complex template matching and time-consuming gradient calculation are avoided, and timeliness of the algorithm is improved.

And constructing a corresponding direction template for each divided sub-region to obtain a maximum light intensity point in the normal direction, and extracting the sub-pixel coordinates of the light strip center along the normal direction by taking the point as the center according to a gray scale gravity center method, thereby ensuring the extraction precision of the algorithm.

In the embodiment of the invention, the laser light bar is divided into a plurality of sections based on the deep learning network, the normal direction of each section is determined based on the gradient histogram statistics, the maximum light intensity point of the normal direction of each section is further determined, the sub-pixel coordinates of the center of the laser light bar are extracted along the normal direction by taking the maximum light intensity point as the center according to the gray scale gravity center method, and the extraction precision of the center of the laser light bar can be ensured.

In an embodiment of the present invention, as shown in fig. 1 to 3, the calibration calibrating apparatus 4 further includes a base 43 and an adjusting mechanism 44, the base 43 can be detachably connected to the track profile beam 1, the adjusting mechanism 44 can be detachably disposed on the base 43, and the supporting platform 41 can be detachably mounted on the adjusting mechanism 44, the adjusting mechanism 44 can realize that the supporting platform 41 can be lifted along a longitudinal axis, can swing along a first axis disposed horizontally, and can swing along a second axis disposed horizontally, the first axis is located in a longitudinal reference plane, the second axis is perpendicular to the first axis, and an intersection point of the first axis and the second axis is located on the longitudinal axis.

The base 43 and the track profile beam 1 can be detachably connected, the adjusting mechanism 44 can be detachably arranged on the base 43, the bearing platform 41 can be detachably installed on the adjusting mechanism 44, the adjusting mechanism 44 can realize that the bearing platform 41 can lift along a longitudinal axis, swing along a first axis which is horizontally arranged and swing along a second axis which is horizontally arranged, wherein the first axis is positioned in a longitudinal reference plane, the second axis is mutually vertical to the first axis, an intersection point (generated by space intersection) of the first axis and the second axis is positioned on the longitudinal axis, the bearing platform 41 can be adjusted from multiple directions through the adjusting mechanism 44, so that each scaling piece 42 positioned on the bearing platform 41 can be strictly coplanar with laser emitted by the two lasers 3, and detection precision is guaranteed.

In an embodiment of the present invention, as shown in fig. 1 to 3, a handle 431 is further provided on the base 43 to facilitate the taking movement of the calibration device 4.

In an embodiment of the present invention, as shown in fig. 1 to 3, the adjusting mechanism 44 includes a lifting platform 441, a first rotating platform 442, and a second rotating platform 443 sequentially connected from bottom to top, the lifting platform 441 is connected to the base 43, the supporting platform 41 is disposed on the second rotating platform, the lifting platform 441 can drive the first rotating platform 442, the second rotating platform 443, and the supporting platform 41 to lift, the first rotating platform 442 can drive the second rotating platform 443 and the supporting platform 41 to rotate and swing along a first axis, and the second rotating platform 443 can drive the supporting platform 41 to rotate and swing along a second axis. In this way, the position of the bearing platform 41 and the calibration piece 42 disposed thereon can be flexibly adjusted, it should be noted that the specific structure of the adjusting mechanism is only a preferred embodiment of the present invention, and other structures capable of adjusting the bearing platform 41 and the calibration piece 42 thereon can be selected, and the present invention is not limited thereto.

In one embodiment of the present invention, as shown in fig. 1 to 3, the calibration verification apparatus 4 further includes an illumination assembly 45, and the illumination assembly 45 includes: the lamp 451 and the supporting adjuster 452, the lower end of the supporting adjuster 452 is disposed on the base 43, and the lamp 451 is connected to the upper end of the supporting adjuster 452 and is located above the supporting platform 41. The light fitting 451 is used to illuminate each of the targets 42 to facilitate clear imaging of the two cameras 2. Preferably, the longitudinal center line of the lamp 451 should coincide with the longitudinal center line of the carrying platform 41 to ensure that the light emitted from the lamp 451 can be uniformly irradiated on the index member 42.

Fig. 8 illustrates an implementation flow of the calibration method provided by the embodiment of the present invention, and for convenience of description, only the portions related to the embodiment of the present invention are illustrated, and the details are as follows:

in the embodiment of the present invention, the calibration and verification device 4 is further used for verifying the rail profile and wear detection system, and the calibration and verification device 4 further includes a verification piece, during verification, the calibration piece 42 is taken down from the bearing platform 41, the verification piece is placed on the bearing platform 41, and during verification, the laser lines emitted by the two lasers 3 are collinear with the positioning grooves of the verification piece.

As shown in fig. 8, the calibration method based on the calibration method according to the above embodiment includes:

step 801, extracting sub-pixel coordinates of the center of a laser light bar in a laser image of the detection part acquired by a camera 2; in the verification process, laser lines emitted by the two lasers 3 are collinear with the positioning groove of the detection piece;

step 802, determining actually measured profile data of the calibration piece according to the sub-pixel coordinates of the center of the laser light strip in the laser image, the calibration parameters of each camera 2 and the splicing parameters of the two cameras 2;

and step 803, matching and comparing the actually measured profile data of the detection piece with the standard profile data to determine the profile and abrasion detection precision.

Firstly, detecting the profile detection precision through a standard rail test block

Fig. 9 shows a schematic structure of a standard rail test block in the calibration verification apparatus according to an embodiment of the present invention, in the embodiment of the present invention, the verification object includes a standard rail test block 461, the standard profile data includes standard profile data, and the standard profile data is preset ideal profile data.

The standard rail test block 461 is a rail simulation test block with a lossless surface, and the detection precision of the rail profile and the profile of the wear detection system can be determined through the standard rail test block 461.

Fig. 10 shows a flow of implementing the determination of the profile detection accuracy by using the standard rail test block according to the embodiment of the present invention, and for convenience of description, only the relevant portions of the embodiment of the present invention are shown, and the following details are described below:

as shown in fig. 10, determining the profile detection accuracy using the standard rail block 461 includes:

step 1001, extracting sub-pixel coordinates of the center of a laser light bar in a laser image of the standard rail test block 461 acquired by the camera 2; in the verification process, laser lines emitted by the two lasers 3 are collinear with the positioning groove of the standard rail test block 461;

step 1002, determining actually measured profile data of the standard orbit test block 461 according to sub-pixel coordinates of the centers of laser light stripes in the laser image, calibration parameters of each camera 2 and splicing parameters of the two cameras 2;

step 1003, matching and comparing the actually measured profile data of the standard track test block 461 with the standard track profile data, and determining the profile detection precision.

When the standard rail test block 461 is used to determine the profile detection accuracy, the standard rail test block 461 is placed on the carrying platform 41, positioning grooves are respectively arranged at two sides of the standard rail test block 461, and the posture of the standard rail test block 461 is adjusted so that the laser lines emitted by the two lasers 3 are collinear with the positioning grooves of the standard rail test block 461.

Then, the camera 2 acquires the laser light strip image of the standard track test block 461, further extracts the sub-pixel coordinates of the center of the laser light strip in the laser light strip image of the standard track test block 461, determines the actually measured profile data of the standard track test block 461 according to the calibrated parameters of the single-side camera 2 and the splicing relation parameters of the two cameras 2, and finally matches and compares the actually measured profile data of the standard track test block 461 with the standard track profile data to determine the profile detection precision.

In the embodiment of the present invention, the sub-pixel coordinates of the center of the laser light stripe in the laser image of the standard rail test block 461 acquired by the camera 2 are extracted, the actually measured profile data of the standard rail test block 461 is determined according to the sub-pixel coordinates of the center of the laser light stripe in the laser image, the calibration parameters of each camera 2 and the splicing parameters of the two cameras 2, and finally, the actually measured profile data of the standard rail test block 461 is matched and compared with the standard rail profile data to determine the profile detection precision, thereby realizing the detection precision verification of the rail profile and wear detection system and improving the verification precision.

Secondly, detecting the abrasion detection precision through the abrasion rail test block

Fig. 11 shows a structural schematic diagram of a wear rail test block according to an embodiment of the present invention, in the embodiment of the present invention, a verification piece includes a wear rail test block 462, a simulated defect 4621 is formed at an edge of the wear rail test block 462 in a recessed manner, standard profile data includes standard wear profile data, the standard wear profile data is preset theoretical ideal wear profile data, and a rail profile and wear detection accuracy of a wear detection system can be determined by the wear rail test block 462.

Fig. 12 shows a flow of implementing the determination of the wear detection accuracy by using the wear rail test block according to the embodiment of the present invention, and for convenience of description, only the relevant parts of the embodiment of the present invention are shown, and the following details are described below:

as shown in fig. 12, determining the wear detection accuracy using the wear rail test block 462 includes:

step 1201, extracting the sub-pixel coordinates of the laser light strip center in the laser image of the wear rail test block 462 collected by the camera 2; in the verification process, laser lines emitted by the two lasers 3 are collinear with the positioning groove of the wear rail test block 462;

step 1202, determining actual measurement profile data of the wear rail test block 462 according to sub-pixel coordinates of the center of the laser light stripe in the laser image, calibration parameters of each camera 2 and splicing parameters of the two cameras 2;

step 1203, matching and comparing the actually measured profile data of the wear rail test block 462 with the standard wear rail profile data, and determining the wear detection precision.

When the wear detection accuracy is determined by using the wear rail test block, the wear rail test block 462 is placed on the bearing platform 41, positioning grooves are respectively arranged on two sides of the wear rail test block 462, and the posture of the wear rail test block 462 is adjusted so that laser lines emitted by the two lasers 3 are collinear with the positioning grooves of the wear rail test block 462.

Then, the camera 2 collects the laser light strip image of the wear rail test block 462, and further extracts the sub-pixel coordinates of the center of the laser light strip in the laser light strip image of the wear rail test block 462, the actual measurement profile data of the wear rail test block 462 is determined according to the calibrated parameters of the single-side camera 2 and the splicing relation parameters of the two cameras 2, and finally the actual measurement profile data of the wear rail test block 462 is matched and compared with the standard wear rail profile data to determine the wear detection precision.

In the embodiment of the present invention, the sub-pixel coordinates of the center of the laser light stripe in the laser image of the wear rail test block 462 collected by the camera 2 are extracted, the actual measurement profile data of the wear rail test block 462 is determined according to the sub-pixel coordinates of the center of the laser light stripe in the laser image, the calibration parameters of each camera 2 and the splicing parameters of the two cameras 2, and finally the actual measurement profile data of the wear rail test block 462 is matched and compared with the standard wear rail profile data to determine the wear detection precision, thereby realizing the detection precision verification of the steel rail wear and wear detection system and improving the verification precision.

In summary, in the embodiment of the present invention, the sub-pixel angular points of the checkerboard image collected by the camera 2 are extracted, the world coordinates of the sub-pixel angular points and the image coordinates of the checkerboard image are calibrated in an associated manner, and the calibration parameters of each camera 2 are determined to calibrate each camera 2; then, on the basis of the calibration parameters of each camera 2, the extracted sub-pixel coordinates of the laser light strip center in the laser light strip image collected by the camera 2 are converted into world coordinates, and then the world coordinates are aligned with preset standard world coordinates to determine the splicing relation parameters of the two cameras 2, so that the splicing relation of the two cameras 2 is calibrated. Therefore, in the embodiment of the invention, the checkerboard and the calibration board 422 are used for respectively realizing the calibration of the single-side camera 2 and the calibration of the splicing relation of the two cameras 2, so that the calibration precision of the steel rail profile and wear detection system can be improved.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.