Automatic detection method for key manufacturing size of bogie frame

Abstract

The present application relates to a kind of bogie frame key manufacturing size automatic detection method, belong to laser detection field.It includes: obtaining the coordinate of bogie frame under the coordinate system of robot, coordinate under the coordinate system of robot is converted into coordinate under the coordinate system of world by hand-eye matrix, coordinate data is imported into drawing software to form point cloud, the fitting of three plane direction point clouds in the vicinity of corner point is carried out to realize the acquisition of corner point, the origin of reference coordinate is determined by corner point, the reference coordinate system is constructed by passing through reference origin, the overall point cloud is reconstructed to curved surface, feature point on boundary plane is identified, bogie frame point cloud is segmented, the key size of bogie frame, including total length, total width, total height, height difference of four corners and the torsion of side beam, is calculated by using "plane approximation method" and "section method".Its advantages are to reduce subjective error when artificial detection, improve detection precision, realize the automation of bogie frame detection and the quantification of size.

Description

Technical Field

[0001] This invention relates to the field of laser inspection, and particularly to an automatic detection method for key manufacturing dimensions of bogie frames. This method is based on the automatic dimensional identification of bogie structural features. It primarily addresses the complex structural design of bogie frames by utilizing planes and cross-sections to automatically detect key manufacturing dimensions of the bogie frame. These key manufacturing dimensions refer to the bogie frame's total length, total width, total height, torsion, and the height difference between its four corners. Background Technology

[0002] With the rapid economic development, accelerated industrialization, and continuous expansion of cities in recent years, rail transit systems have effectively alleviated transportation pressure. As a key component of rail vehicles, the bogie plays a crucial role in load-bearing, force transmission, cushioning, and guidance. The frame, as the foundation for bogie installation, determines the safety, stability, and reliability of the rail vehicle through its welding quality. Welding is a major and critical forming and manufacturing process for bogies; its welding quality directly affects the overall manufacturing quality. Furthermore, because welding is a thermoforming process, welding deformation is unavoidable during bogie manufacturing. Effective control of welding deformation is necessary to ensure the required structural dimensions of the bogie.

[0003] Destructive testing is the most commonly used method for inspecting the quality of welded joints in enterprise production. This method not only wastes raw materials and indirectly increases production costs, but also suffers from low efficiency and long inspection cycles. With increasingly stringent requirements for weld quality, inspection frequency, and efficiency, non-destructive testing (NDT) has become the preferred choice. Currently, traditional visual quality inspection relies mainly on visual inspection by personnel or the use of measuring rulers. These traditional methods are inefficient, and their accuracy is often affected by the subjective factors of the inspectors, inevitably reducing inspection efficiency and product pass rates, which is detrimental to the current trend of automated production inspection.

[0004] Currently, in the field of appearance quality inspection, complex structures such as bogie frames cannot be effectively inspected using existing methods due to their intricate shapes and structures. Therefore, it is essential to develop an automatic dimensional identification method based on bogie structures. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic detection method for key manufacturing dimensions of bogie frames, solving the problems of limited detection methods, low accuracy, and long detection time in industrial production, and filling the gap in automatic detection technology for the dimensional appearance quality of bogie frames. This method automates and intelligently detects the quality of key manufacturing dimensions of bogie frames. By fitting point cloud coordinate data, it identifies feature points on the point cloud boundary, constructs a reference coordinate system, and uses the "planar approximation method" and the "section method" to detect the key manufacturing dimensions of the bogie frame, thus achieving the quantification of key manufacturing dimensions of the bogie frame.

[0006] The above-mentioned objective of the present invention is achieved through the following technical solution:

[0007] An automatic detection method for key manufacturing dimensions of the bogie frame is proposed. Based on the coordinate data of the bogie frame obtained by a 2D laser sensor, a point cloud is created in drawing software. Corner points are obtained by fitting small planes to points in three plane directions surrounding them. A reference origin and reference coordinate system are constructed based on the four corner points on one side of the side beam. The point cloud is then reconstructed as a whole surface. The reconstructed bogie frame is segmented by extracting segment points. The key manufacturing dimensions of the bogie frame are measured using a "planar approximation method" and a "section method." The method includes the following steps:

[0008] Step 1: Acquiring point cloud data:

[0009] A two-dimensional laser sensor is mounted on the end effector of an industrial robot and follows its movement in three-dimensional space to complete a non-contact measurement of the bogie frame under test, thereby obtaining point cloud data of the bogie frame.

[0010] Step 2: Extraction of side beam corner points:

[0011] The point cloud within the U-domain (small spatial domain) near the corner of the side beam is fitted in three planar directions using the least squares method;

[0012] The equation of the plane is expressed as:

[0013] Mx + Ny + Pz + Q = 0

[0014] Transform it into:

[0015]

[0016] make:

[0017]

[0018] but:

[0019] z = a0x + a1y + a2

[0020] The coordinate of each point is (x i , y i , z i ), where i=0, 1, 2, …, n-1;

[0021] To minimize the distance from the point to the plane, that is, to minimize S, where S is:

[0022]

[0023] To minimize S, the following conditions shall be satisfied:

[0024]

[0025] a0, a1 and a2 are calculated through the above formula to obtain the plane equation expression;

[0026] Three small planes fitted within a U-domain near corner points of side beams are extended, and the intersection point of the three small planes is determined as the corner point; corner points of the two side beams are extracted through plane fitting within said U-domain;

[0027] Step 3: Establishment of reference origin and reference coordinate system:

[0028] Four corner points on the end face at the same side of one of the side beams are determined, three of the corner points are selected to form a reference plane M0, the distance from the fourth corner point to the reference plane M0 is determined, a distance threshold d0 is set, the distance d from the fourth corner point to the reference plane M0 is determined, d is compared with the distance threshold d0, if d<d0, it is determined that the fourth corner point is also on the reference plane M0, conversely, if d>d0, it is determined that the fourth corner point is not on the reference plane M0;

[0029] The fourth corner point is projected onto the reference plane M0, diagonal connecting lines are drawn respectively for the four points on the reference plane M0, the intersection of the two straight lines is defined as the reference origin O; a reference coordinate system (u, v, w) is established, the v-axis of the coordinate system is a straight line starting from the reference origin O and perpendicular to the plane M0, and the u-axis and the w-axis are mutually orthogonal and perpendicular to the v-axis;

[0030] Step 4: Surface reconstruction:

[0031] the overall contour shape of the bogie frame is reproduced through overall surface reconstruction of the point cloud of the bogie frame;

[0032] Step 5: Extraction of segmentation points and segmentation of bogie frame:

[0033] since the side beam has a bent and descending portion, the side beam needs to be divided into three sections; after surface reconstruction, the boundary contour line of the side beam is extracted, and the boundary straight line of the side beam edge in the v-axis direction is extracted; a boundary straight line l1 and a segmentation point E are set.

[0034] Expression of the straight line equation:

[0035] y = b0 + b1x

[0036] The coordinates of each point on the boundary surface are (x i y i , z i ), where i = 0, 1, 2, ..., n-1;

[0037] This minimizes the distance from each point on the boundary surface to the line, i.e., minimizes L, where L is:

[0038]

[0039] To minimize L, the following conditions must be met:

[0040]

[0041] Since the point cloud around the right boundary point E of the defined line l1 shows a downward trend, calculate the distance from the points near the right end of line l1 to line l1, and select the point with the smallest distance as the right boundary point E of line l1.

[0042] Step Six: Inspection of total length, total width, and total height:

[0043] In the space where the bogie frame point cloud is located, planes parallel to plane M0 are continuously constructed along the v-axis of the reference coordinate system. The constructed planes continuously approach the bogie frame point cloud until the bogie frame point cloud appears on the constructed plane. A threshold m0 is set for the number of point clouds on the constructed plane. When the number of point clouds on the plane reaches m0, this plane is taken as the detection starting plane M1. As the plane is continuously constructed, when the plane reaches the other end face of the side beam and the number of point clouds on the plane reaches the set threshold m0, this plane is considered as the detection ending plane M2. The distance between the two mutually parallel planes M1 and M2 is calculated as the detection value of the total length of the bogie frame.

[0044] In the space where the bogie frame point cloud is located, planes perpendicular to the u-axis and w-axis are constructed respectively. By continuously constructing the planes and setting the threshold m0, the detection start plane M1 and the detection end plane M2 are selected to complete the detection of the total width and total height of the bogie frame.

[0045] Step 7: Torsion test:

[0046] Select a side beam of the bogie frame after surface reconstruction, and use the "section method". That is, draw a plane parallel to the reference plane M0 and tangent to the contour of the bogie side beam to form a section. By continuously drawing sections starting from the reference plane M0, a series of sections are obtained, which constitute a section set {S1, S2, ..., S...}. nThe pose of each section in this section set can be determined. By projecting the entire section set onto the reference plane M0, the torsion angle of each section relative to the other sections can be analyzed.

[0047] Step 8: Detecting the height difference at the four corners:

[0048] Since a reference plane M0 is constructed, the extracted four corner vertices are projected onto the reference plane M0 to obtain the positions of the four corner vertices on the reference plane M0. The coordinates of the four corner vertices in the u-axis direction in the reference coordinate system are obtained as the height difference between the four corners.

[0049] The two-dimensional laser sensor mentioned in step one is a linear array high-precision two-dimensional laser rangefinder sensor, which is mounted on the end effector of the robot.

[0050] The extraction of segment points in step five involves determining the distance from each point near a segment point on the boundary surface to the fitted line l. i distance δ i If δ i >δ i+1 Then the smaller δ value is retained. i+1 Continue to determine the next point δ i+2 With δ i+1 This process continues until the point with the smallest distance is selected as the segmentation point.

[0051] The torsion calculation described in step seven involves dividing the bogie frame into three segments. A "section method" is used to obtain a set of sections for each of the three segments. All sections in the second segment's section set are translated so that their centers lie on the v-axis. At this point, the centers of all sections are on the same axis. The translated section set of the second segment, together with the section sets of the first and third segments, constitutes a large section set I = {S1, S2, ..., S...}. n Starting from S1, each profile will have a torsion angle with the reference plane M0. A set of torsional values ​​J = {J1, J2, ..., J...} is obtained from the poses of all profiles in the profile set I. n}; From this, we can obtain the maximum and minimum values ​​of the torsion.

[0052] To calculate the torsion angle of the two sections, project the sections onto the reference plane M0. The angle between the boundary lines of the two sections is the torsion angle α. Calculate the average of the torsion angles α between the four boundary lines of the two sections as the amount of torsion of the two sections, as shown in the following formula:

[0053]

[0054] The torsion of the bogie frame is calculated using the "section method" to obtain the changes in torsion on the side beams of the bogie frame, as well as the maximum and minimum values ​​of the torsion.

[0055] The beneficial effects of this invention are as follows: This invention utilizes a two-dimensional laser sensor mounted on the end effector of a robot to acquire coordinate data of the bogie frame. This coordinate data is imported into drawing software to form a point cloud. Small plane fitting is performed on the point cloud in three planar directions near the corner points to obtain the corner points. A reference coordinate origin is determined through the corner points, and a reference coordinate system is constructed through the reference origin. The entire point cloud is reconstructed into a curved surface, and feature points on the boundary plane are identified. The bogie frame point cloud is segmented, and the key dimensions of the bogie frame, including total length, total width, total height, height difference at the four corners, and torsion of the side beams, are calculated using a "planar approximation method" and a "section method." This reduces subjective errors during manual inspection, improves inspection accuracy, and achieves automation and quantification of bogie frame inspection dimensions. Attached Figure Description

[0056] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate the invention and are used to explain it, but do not constitute an undue limitation of the invention.

[0057] Figure 1 This is a schematic diagram of the logic flow of the present invention;

[0058] Figure 2 This is a schematic diagram of corner point extraction according to the present invention;

[0059] Figure 3 This is a schematic diagram illustrating the construction of the reference coordinate system of the present invention;

[0060] Figure 4 This is a schematic diagram of the segmentation point extraction method of the present invention;

[0061] Figure 5 This is a schematic diagram of the segmented steering frame of the present invention;

[0062] Figure 6 This is a schematic diagram showing the overall length, width, and height dimensions of the bogie frame of the present invention.

[0063] Figure 7 This is a cross-sectional view of the present invention;

[0064] Figure 8 This is a schematic diagram of the torsion calculation of the present invention. Detailed Implementation

[0065] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] See Figures 1 to 8 As shown, the automatic detection method for key manufacturing dimensions of the bogie frame of the present invention includes the following steps: acquiring the coordinates of the bogie frame in the robot coordinate system; converting the coordinates in the robot coordinate system to coordinates in the world coordinate system using a hand-eye matrix; importing the coordinate data into drawing software to form a point cloud; fitting small planes to the point cloud in three planar directions near the corner point to obtain the corner point; determining the origin of the reference coordinate system through the corner point; constructing a reference coordinate system through the origin; reconstructing the surface of the entire point cloud; identifying feature points on the boundary plane; segmenting the point cloud of the bogie frame; and calculating the key dimensions of the bogie frame, including the total length, total width, total height, height difference at the four corners, and torsion of the side beams, using a "planar approximation method" and a "section method". Its advantages include reducing subjective errors during manual inspection, improving inspection accuracy, and achieving automation and quantification of bogie frame inspection dimensions.

[0067] See Figures 1 to 8 As shown, the automatic detection method for key manufacturing dimensions of the bogie frame of the present invention, based on the coordinate data of the bogie frame obtained by a two-dimensional laser sensor, establishes a point cloud in drawing software, obtains corner points by fitting small planes to points in three planar directions around the corner points, constructs a reference origin and reference coordinate system based on the four corner points on one side of the side beam, reconstructs the overall surface of the point cloud, extracts segment points and segments the reconstructed bogie frame, and measures the key manufacturing dimensions of the bogie frame using the "planar approximation method" and the "section method". The method includes the following steps:

[0068] Step 1: Acquiring point cloud data:

[0069] A two-dimensional laser sensor is mounted on the end effector of an industrial robot and follows its movement in three-dimensional space to complete a non-contact measurement of the bogie frame under test. The acquired two-dimensional coordinate data of the bogie frame and the robot's pose data are transmitted to a computer. Through the conversion of the hand-eye matrix, the three-dimensional coordinate data of the bogie frame in the base coordinate system is obtained, and the three-dimensional coordinate data is imported into drawing software to form a point cloud.

[0070] Step 2: Extraction of side beam corner points:

[0071] As Figure 2 shown, plane fitting is performed on the point cloud in the U-domain near the corner point of the side beam in three planar directions by adopting the least square method.

[0072] The expression of the plane equation is:

[0073] Mx+Ny+PZ+Q=0

[0074] Transform it into:

[0075]

[0076] Let:

[0077]

[0078] Then:

[0079] z=a0x+a1y+a2

[0080] The coordinate of each point is (x i , y i , z i ), wherein i=0, 1, 2, …,n-1.

[0081] To make the distance from the point to the plane the shortest, that is, to minimize S, S is:

[0082]

[0083] To minimize S, the following conditions shall be met:

[0084]

[0085] a0, a1 and a2 are calculated through the above formula, and the expression of the plane equation can be obtained.

[0086] The three small planes fitted in the U-domain near the corner points of the side beam are extended, and the intersection point of the three small planes is determined as the corner point. The corner points of the two side beams are extracted through plane fitting in the U-domain.

[0087] Step 3: Establishment of reference origin and reference coordinate system:

[0088] As Figure 3 shown, four corner points on the end surface of the same side of one side beam are determined, three of the corner points are selected to form a reference plane M0, the distance from the fourth corner point to the reference plane M0 is determined, a distance threshold d0 is set, the distance d from the fourth corner point to the reference plane M0 is determined, d is compared with the distance threshold d0, if d<d0, it can be considered that the fourth corner point is also on the reference plane M0; conversely, if d>d0, it is considered that the fourth corner point is not on the reference plane M0.

[0089] Project the fourth corner point onto the reference plane M0. Connect the four points diagonally on the reference plane M0, and define the intersection of the two lines as the reference origin O. Establish a reference coordinate system (u, v, w). The v-axis of this coordinate system is drawn from the reference origin O and perpendicular to the plane M0. The u-axis is orthogonal to the w-axis and perpendicular to the v-axis.

[0090] Step 4: Surface Reconstruction

[0091] By reconstructing the overall surface of the bogie frame point cloud, the overall outline of the bogie frame is reproduced.

[0092] Step 5: Extraction of segmentation points and segmentation of the steering framework:

[0093] Because the side beam has a curved and descending section, it needs to be divided into three segments. After surface reconstruction, the boundary contour line of the side beam can be extracted, and the boundary line of the side beam edge in the v-axis direction can be extracted. The boundary line l1 and the segment point E are set.

[0094] The equation of a straight line is:

[0095] y = b0 + b1x

[0096] The coordinates of each point on the boundary surface are (x i y i , z i ), where i = 0, 1, 2, ..., n-1.

[0097] The goal is to minimize the distance from every point on the boundary surface to the line, i.e., to minimize L, where L is:

[0098]

[0099] To minimize L, the following conditions must be met:

[0100]

[0101] Because the point cloud around the right boundary point E of the defined straight line l1 shows a descending trend, such as Figure 4 Calculate the distance from points near the right end of line l1 to line l1, and select the point with the smallest distance as the right boundary point E of line l1.

[0102] Based on the extracted feature points at the four corner vertices A, A', B, and B', their coordinate values ​​along the z-axis are compared. The point with the smallest coordinate value along the z-axis is selected as the reference point, and the difference between the coordinate values ​​of the other three points and the reference point is taken as the height difference between the four corner points.

[0103] Step Six: Inspection of total length, total width, and total height:

[0104] In the space where the bogie frame point cloud is located, planes parallel to plane M0 are continuously constructed along the v-axis of the reference coordinate system. These constructed planes continuously approach the bogie frame point cloud until the point cloud appears on the constructed plane. A threshold m0 is set for the number of point clouds on the constructed plane. When the number of point clouds on the plane reaches m0, this plane is designated as the detection starting plane M1. As planes are continuously constructed, when the plane reaches the other end face of the side beam and the number of point clouds on the plane reaches the set threshold m0, this plane is considered the detection ending plane M2. The distance between the two mutually parallel planes M1 and M2 is calculated as the detection value for the total length of the bogie frame.

[0105] Based on the above principle, planes perpendicular to the u-axis and w-axis are constructed in the space where the bogie frame point cloud is located. By continuously constructing the planes and setting the threshold m0, the detection start plane M1 and the detection end plane M2 are selected to complete the detection of the total width and total height of the bogie frame.

[0106] Step 7: Torsion test:

[0107] Select a side beam of the bogie frame after surface reconstruction, and use the "section method". That is, draw a plane parallel to the reference plane M0 and tangent to the contour of the bogie side beam to form a section. By continuously drawing sections starting from the reference plane M0, a series of sections are obtained, which constitute a section set {S1, S2, ..., S...}. n The pose of each section in this section set can be determined. By projecting the entire section set onto the reference plane M0, the torsion angle of each section relative to the other sections can be analyzed.

[0108] Step 8: Detecting the height difference at the four corners:

[0109] Since a reference plane M0 is constructed, the extracted four corner vertices are projected onto the reference plane M0 to obtain the positions of the four corner vertices on the reference plane M0. The coordinates of the four corner vertices in the u-axis direction in the reference coordinate system are obtained as the height difference between the four corners.

[0110] The acquisition of point cloud data described in step one specifically involves:

[0111] Start the detection system and initialize it. Adjust the spatial position of the two-dimensional laser rangefinder sensor so that the laser line emitted by the sensor is exactly at the starting section position of the bogie frame under test, and keep the emitted laser line perpendicular to the bogie frame under test.

[0112] The outline of the initial section of the bogie frame under test is obtained by a two-dimensional laser rangefinder sensor. It is then determined whether the laser line has covered the surface of the area to be measured. The height of the two-dimensional laser sensor is adjusted so that the bogie frame under test is exactly within its detection range, and this position is determined as the starting position of the detection. At the same time, the motion trajectory of the industrial robot is set according to the shape and characteristics of the bogie frame, so that the two-dimensional laser sensor can move according to the set motion trajectory. Due to the extremely complex shape and structure of the bogie frame, multi-segment scanning is required when scanning it.

[0113] The industrial robot controls a two-dimensional laser sensor to simultaneously transmit and receive laser signals. Based on the principle of laser triangulation, the sensor automatically calculates the coordinates of points on each contour line of the measured steering frame along the laser line and feeds back the detection results in coordinate form to the industrial computer.

[0114] While receiving the detection data, the industrial computer controls the industrial robot to move according to the pre-set motion trajectory and provides real-time feedback on the robot's pose information; the industrial robot saves the acquired bogie frame surface contour data and the robot's pose information.

[0115] The two-dimensional laser sensor mentioned in step one is a linear array high-precision two-dimensional laser rangefinder sensor, which is mounted on the end effector of an industrial robot. It can emit and receive linear array lasers, and can acquire all the information on the laser line in a single detection. It also converts photoelectric signals into digital signals and finally transmits them to an industrial computer.

[0116] Step 5 describes the extraction of segmentation points, such as... Figure 4 Each point near a segment point on the boundary surface is checked to determine its distance from the fitted line l. i distance δ i If δ i >δ i+1 Then the smaller δ value is retained. i+1 Continue to determine the next point δ i+2 With δ i+1 This process continues until the point with the smallest distance is selected as the segmentation point.

[0117] The torsion calculation described in step seven involves segmenting the bogie frame during torsion testing due to the significant bending and descent of the side beams. As shown in Figure 5, the bogie frame is divided into three segments. A "section method" is used to obtain a set of sections for each of the three segments. All sections in the second segment's section set are translated so that their centers lie on the v-axis. At this point, the centers of all sections are on the same axis. The translated section set of the second segment, together with the section sets of the first and third segments, constitutes a large section set I = {S1, S2, ..., S...}. nStarting from S1, each profile will have a torsion angle with the reference plane M0. From the poses of all profiles in profile set I, a set of torsion amounts J = {J1, J2, ..., J...} can be obtained. n From this, we can obtain the maximum and minimum values ​​of the torsion.

[0118] like Figure 8 As shown, the torsion angle of the two sections is calculated by projecting the sections onto the reference plane M0. The angle between the boundary lines of the two sections is the torsion angle α. The average value of the torsion angle α between the four boundary lines of the two sections is calculated as the torsion amount of the two sections, as shown in the following formula:

[0119]

[0120] By calculating the torsion of the bogie frame using the "section method", the changes in torsion on the side beams of the bogie frame, as well as the maximum and minimum values ​​of the torsion, can be obtained.

[0121] Example:

[0122] This invention employs the method to automatically inspect a welding simulation component of a high-speed railway bogie frame. The material of the bogie frame welding simulation component is Q235 ordinary carbon steel. The bogie frame welding simulation component is welded from 12 steel plates and two steel pipes using manual arc welding. The wall thickness of both the plates and pipes is 6mm. The automatic inspection method for key manufacturing dimensions of the bogie frame of this invention is used to inspect the welding simulation component of the bogie frame. The specific steps are as follows: A robot drives a two-dimensional laser sensor to acquire the coordinate data of the bogie frame. The coordinate data is imported into drawing software to form a point cloud. The automatic inspection program is started to acquire corner point positions, select corner points, determine the reference coordinate system and reference plane M0, reconstruct the surface, extract the segment point positions and segment the bogie frame, select the dimension inspection direction, and complete the inspection of the total length, total width, and total height of the bogie frame. The side beam is selected, and the torsion inspection in the automatic inspection program is selected to complete the torsion inspection of the bogie frame. The four-corner height difference inspection in the automatic inspection program is selected to complete the inspection of the four-corner height difference of the bogie frame.

[0123] The above method enables automatic detection of key manufacturing dimensions of the bogie frame. Compared with manual detection using measuring rulers, it improves the accuracy of detection and greatly reduces the difficulty of manual detection, thus improving work efficiency, especially considering the complex structure of the bogie frame.

[0124] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the invention by those skilled in the art. Any modifications, equivalent substitutions, or improvements made to the present invention should be included within the scope of protection of the present invention.

Claims

1. An automatic detection method for key manufacturing dimensions of a bogie frame, characterized in that: Comprising the following steps: Step 1: Acquisition of point cloud data: A two-dimensional laser sensor is mounted on the end of an industrial robot, and follows the robot to move in three-dimensional space to complete non-contact measurement of the measured bogie frame, so as to obtain point cloud data of the bogie frame; Step 2: Extraction of corner points of side beams: Plane fitting is performed on the point cloud in the U-domain near the corner points of the side beam in three plane directions, and the least square method is adopted for fitting; The expression of the plane equation is: ； It is transformed into: ； Let: ; ; ； Then: ； The coordinates of each point are (x) i y i , z i ), where i = 0, 1, 2, ..., n-1; To minimize the distance from the point to the plane, that is, to minimize S, where S is: ； To minimize S, it is necessary to satisfy: ，k=0，1，2； Calculated using the above formula , , This yields the equation of the plane. The three small planes fitted in the U-domain near the corner points of the side beam are extended, and the intersection point of the three small planes is determined as the corner point; the corner points of the two side beams are extracted through the plane fitting in the U-domain; Step 3: Establishment of reference origin and reference coordinate system: Four corner points on the same side end face of one of the side beams are determined, three of the corner points are selected to form a reference plane M0, the distance from the fourth corner point to the reference plane M0 is determined, a distance threshold d0 is set, the distance d from the fourth corner point to the reference plane M0 is determined, d and the distance threshold d0 are compared, if d<d0, it is considered that the fourth corner point is also on the reference plane M0, on the contrary, if d>d0, it is considered that the fourth corner point is not on the reference plane M0; The fourth corner point is projected onto the reference plane M0, the four points on the reference plane M0 are connected diagonally respectively, and the intersection point of the two straight lines is defined as the reference origin O; a reference coordinate system (u, v, w) is established, the v-axis of the coordinate system starts from the reference origin O and is a straight line perpendicular to the plane M0, and the u-axis and the w-axis are mutually orthogonal and perpendicular to the v-axis; Step 4: Surface reconstruction: Through overall surface reconstruction of the point cloud of the bogie frame, the overall contour shape of the bogie frame is reproduced; Step 5: Extraction of segmentation points and segmentation of bogie frame: Since the side beam has a bent and lowered part, the side beam needs to be divided into three sections; after surface reconstruction, the boundary contour line of the side beam is extracted, and the boundary straight line of the side beam edge in the v-axis direction is extracted; a boundary straight line l1 and a segmentation point E are set; Expression of the straight line equation: ； The coordinates of each point on the boundary surface are (x i y i , z i ), where i = 0, 1, 2, ..., n-1; such that the distance from each point on the boundary surface to the straight line is minimized, that is, L is minimized, where L is: ； To minimize L, it is necessary to satisfy: ，k=0，1； Since the point cloud around the right-end boundary point E of the defined straight line l1 shows a downward trend, the distances from the points near the right end of the straight line l1 to the straight line l1 are calculated, and the point with the smallest distance is selected as the right-end boundary point E of the straight line l1; Step 6: Detection of total length, total width and total height: In the space where the bogie frame point cloud is located, planes parallel to plane M0 are continuously constructed along the v-axis of the reference coordinate system. The constructed planes continuously approach the bogie frame point cloud until the bogie frame point cloud appears on the constructed plane. A threshold m0 is set for the number of point clouds on the constructed plane. When the number of point clouds on the plane reaches m0, this plane is taken as the detection starting plane M1. As the plane is continuously constructed, when the plane reaches the other end face of the side beam and the number of point clouds on the plane reaches the set threshold m0, this plane is considered as the detection ending plane M2. The distance between the two mutually parallel planes M1 and M2 is calculated as the detection value of the total length of the bogie frame. In the space where the bogie frame point cloud is located, planes perpendicular to the u-axis and w-axis are constructed respectively. By continuously constructing the planes and setting the threshold m0, the detection start plane M1 and the detection end plane M2 are selected to complete the detection of the total width and total height of the bogie frame. Step 7: Torsion test: Select a side beam of the bogie frame after surface reconstruction, and use the "section method". That is, draw a plane parallel to the reference plane M0 and tangent to the contour of the bogie side beam to form a section. By continuously drawing sections starting from the reference plane M0, a series of sections are obtained, which constitute a section set {S1, S2, ..., S...}. n The pose of each section in this section set can be determined. By projecting the entire section set onto the reference plane M0, the torsion angle of each section relative to the other sections can be analyzed. Step 8: Detecting the height difference at the four corners: Since a reference plane M0 is constructed, the extracted four corner vertices are projected onto the reference plane M0 to obtain the positions of the four corner vertices on the reference plane M0. The coordinates of the four corner vertices in the u-axis direction in the reference coordinate system are obtained as the height difference between the four corners.

2. The automatic detection method for key manufacturing dimensions of the bogie frame according to claim 1, characterized in that: The key manufacturing dimensions of the bogie frame refer to the total length, total width, total height, torsion, and height difference at the four corners of the bogie frame.

3. The automatic detection method for key manufacturing dimensions of the bogie frame according to claim 1, characterized in that: The two-dimensional laser sensor mentioned in step one is a linear array high-precision two-dimensional laser ranging sensor.

4. The automatic detection method for key manufacturing dimensions of the bogie frame according to claim 1, characterized in that: The extraction of segment points in step five involves determining the points near each segment point on the boundary surface and their distance to the fitted line l. i distance δ i If δ i >δ i+1 Then the smaller δ value is retained. i+1 Continue to determine the next point δ i+2 With δ i+1 This process continues until the point with the smallest distance is selected as the segmentation point.

5. The automatic detection method for key manufacturing dimensions of bogie frames according to claim 1, characterized in that: The torsion calculation described in step seven involves dividing the bogie frame into three segments. A "section method" is used to obtain a set of sections for each of the three segments. All sections in the second segment's section set are translated so that their centers lie on the v-axis. At this point, the centers of all sections are on the same axis. The translated section set of the second segment, together with the section sets of the first and third segments, constitutes a large section set I = {S1, S2, ..., S...}. n Starting from S1, each profile will have a torsion angle with the reference plane M0. A set of torsion values ​​J = {J1, J2, ..., J...} is obtained from the poses of all profiles in the profile set I. n }; From this, we can obtain the maximum and minimum values ​​of the torsion. To calculate the torsion angle of the two sections, project the sections onto the reference plane M0. The angle between the boundary lines of the two sections is the torsion angle α. Calculate the average of the torsion angles α between the four boundary lines of the two sections as the amount of torsion of the two sections, as shown in the following formula: ； The torsion of the bogie frame is calculated using the "section method" to obtain the changes in torsion on the side beams of the bogie frame, as well as the maximum and minimum values ​​of the torsion.

Images