Method for detecting welding position of axle workpiece

Abstract

The invention provides an axle workpiece welding position detection method, and belongs to the field of workpiece welding position detection.The axle workpiece welding position detection method comprises the steps that an initial axle covering circle is obtained according to an axle covering image, n subdivision points are obtained through uniform subdivision of the initial axle covering circle, and then an accurate axle covering circle is obtained; 3D scanning equipment is adopted to perform left-right scanning with the position of the workpiece as the center to obtain workpiece scanning data, multiple pieces of line data are selected up and down with the feature points as reference points in the workpiece scanning data, jumping platforms are searched for in all the line data, feature positions are obtained according to the jumping platforms, and feature information is recorded; the accurate position of the workpiece relative to the circle center is obtained by combining the feature positions and the accurate bridge wrapping circle; during detection, the set length is scanned left and right with each detection workpiece as the center to obtain the scanning data of the detection workpieces, then the feature information is obtained, and if the feature information of a certain detection workpiece can be matched with one of the recorded feature information, it is judged that the welding position is qualified. The detection efficiency and the detection accuracy can be improved.

Description

Technical Field

[0001] The present invention belongs to the field of workpiece welding position detection, and particularly relates to a method for detecting the welding position of axle components. Background Art

[0002] For the components welded on the axle, due to reasons such as assembly errors of the components or deformation of the components caused by heat during welding, there may be a deviation between the actual welding position of the components and the set welding position. This deviation has an adverse effect on the quality of the axle. Therefore, it is necessary to detect the welding position of the components. In the prior art, it is usually detected manually, which not only has low detection efficiency, but also strongly depends on manual experience for detection accuracy, has poor consistency, and cannot meet the requirements. Summary of the Invention

[0003] The object of the present invention is to propose a method for detecting the welding position of axle components, which can improve the detection efficiency and detection accuracy.

[0004] The present invention is realized through the following technical solutions: A method for detecting the welding position of axle components, comprising the following steps: Step S1: Use a 2D camera to photograph the bridge package of the template axle, obtain the bridge package image, perform edge detection on the bridge package image to obtain edge information, use the edge information for Hough transform circle detection to obtain the initial bridge package circle, uniformly subdivide the initial bridge package circle to obtain n subdivision points, respectively obtain m data points in the bridge package image centered on each subdivision point and along the direction of the line connecting the subdivision point and the center of the initial bridge package circle, and obtain the jump positions where jumps occur in the data array composed of m data points, and use n jump positions for Hough transform circle detection to obtain the precise bridge package circle; Step S2: For each component welded on the bridge package of the template axle, obtain the position of the component, use a 3D scanning device to scan left and right a set length centered on the position of the component to obtain the component scan data, manually select feature points from the component scan data, and select several line data up and down with the feature points as reference points in the component scan data. Search for jump platforms in each line data respectively, and take the positions in the head and end positions of each jump platform whose distances from the feature points are within the distance threshold and closer to the feature points as the feature positions. Record each feature position, as well as the jump method and jump height corresponding to the feature position as a feature information, and combine the normalized data of each feature position with the center of the precise bridge package circle to obtain the accurate position of the component relative to the center of the circle; Step S3: During detection, based on the precise position of the workpiece relative to the center of the circle obtained in Step S2, with the center of the precise bridge package as the reference, adjust the 3D scanning device so that it takes each workpiece to be detected for the axle welding as the center, scan left and right by a set length to obtain the scanned data of the workpiece to be detected, and obtain the feature information of each workpiece according to Step S2. If the feature information of a certain workpiece to be detected can match one of the feature information recorded in Step S2, it is determined that the welding position of the workpiece to be detected is qualified.

[0005] Further, in Step S1, before obtaining the bridge package image, obtain the diameter of the bridge package circle of the template axle according to the CAD file of the template axle, and convert the diameter of the bridge package circle into pixel diameter information.

[0006] Further, in Step S1, after obtaining the bridge package image, perform adaptive histogram equalization on the bridge package image and filter it to obtain a preprocessed image, and perform edge detection on the preprocessed image to obtain the edge information.

[0007] Further, in Step S1, evenly divide the initial bridge package circle by 1° to obtain 360 subdivision points, and obtain the m / 2 data points along the direction close to the center of the initial bridge package circle and the m / 2 data points along the direction away from the center of the initial bridge package circle in the bridge package image with the subdivision points as the center. Search for the jump platform in the data array in the direction away from the center of the initial bridge package circle, and record the position of the start end of the jump platform as the jump position.

[0008] Further, in Step S2, obtain the positions of each workpiece according to the CAD file of the template axle.

[0009] Further, in Step S2, when searching for the jump platform in the line data, if the difference between the depth information of the i +1th data and the depth information of the i th data is greater than the set depth threshold, several consecutive data before the i th data are all less than the set depth threshold compared with the depth information of the i th data, and the sum of the distances of the i th data and several consecutive data before it is greater than the set length threshold, it is determined that a jump occurs. The jump methods include depth increase and depth decrease. The jump platform includes the data between the depth increase point and the depth decrease point, and the jump height is the difference between the average depth information of the data in two adjacent jump platforms.

[0010] Further, in Step S2, the coordinates of the manually selected feature point are expressed as ( x , y , z ), withy When selecting 3-line data respectively above and below the center of the coordinate value to determine the characteristic position, compare the x coordinate value at the start position of the jump platform and the x coordinate value at the end position with the x coordinate value of the characteristic point to see if the distance therebetween is within the distance threshold.

[0011] Furthermore, in the step S2, the obtaining of the precise position of the workpiece relative to the center of the circle refers to the position of the workpiece relative to the center of the precise bridge package circle in the horizontal direction of the axle housing and the welding height of the workpiece on the axle housing.

[0012] Furthermore, both the 2D camera and the 3D scanning device are arranged on the bracket, the bracket is arranged to slide in the horizontal direction of the axle housing, the 2D camera and the 3D scanning device are arranged at intervals, and before obtaining the bridge package image, rotate the axle housing so that the bridge package can be located within the shooting fields of the 2D camera and the 3D scanning device.

[0013] The present invention has the following beneficial effects: 1. The present invention first obtains the precise bridge package circle according to the obtained bridge package image, then scans a set length left and right with the position of the workpiece as the center to obtain the workpiece scanning data, manually selects the characteristic points from the workpiece scanning data, and selects several line data above and below the characteristic points as the center in the workpiece scanning data, searches for the jump platforms in each line of data respectively, and takes the positions in the start positions and end positions of each jump platform whose distances from the characteristic points are within the distance threshold and are closer to the characteristic points as the characteristic positions, records the jump modes and jump heights corresponding to each characteristic position as the characteristic information, and combines the normalized data corresponding to each characteristic position with the center of the precise bridge package circle to obtain the precise position of the workpiece relative to the center of the circle. During actual detection, with the precise bridge package center as the reference, adjust the 3D scanning device so that it takes each workpiece to be welded on the axle housing to be detected as the center, scans a set length left and right to obtain the scanning data of the workpiece to be detected, and obtains the characteristic information of each workpiece to be detected according to the scanning data of the workpiece to be detected. If the characteristic information of a certain workpiece to be detected can match one of the recorded characteristic information, it is determined that the welding position of the workpiece to be detected is qualified, thereby realizing the detection of the welding position of the axle housing workpiece. Compared with the prior art, the detection efficiency is higher, and it does not need to rely on manual experience. Since the detection standards are the same, the detection accuracy and consistency are both improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The following further describes the present invention in detail with reference to the drawings.

[0015] Figure 1 is the flowchart of the present invention.

[0016] Figure 2 is the structural schematic diagram of the axle housing of the present invention.

[0017] Among them, 1. Axle; 2. Axle housing; 3. Workpiece; 4. 2D camera; 5. 3D scanning device. Specific implementation mode

[0018] As Figure 1 shown, the welding position detection method of the axle workpiece includes the following steps: Step S1: Use the 2D camera 4 to photograph the axle housing 2 of the template axle 1 to obtain an axle housing image, perform edge detection on the axle housing image to obtain edge information, use the edge information to perform Hough transform circle detection to obtain an initial axle housing circle, and uniformly subdivide the initial axle housing circle to obtain n subdivision points, respectively obtain m data points on the axle housing image centered on each subdivision point and along the direction of the line connecting the subdivision point and the center of the initial axle housing circle, and obtain the jump positions where jumps occur in the data array composed of m data points, and use n jump positions to perform Hough transform circle detection to obtain an accurate axle housing circle; As Figure 2 shown, both the 2D camera 4 and the 3D scanning device 5 are arranged on the bracket, the bracket is arranged to slide horizontally along the axle 1, the bracket is arranged at an interval from the axle 1, the 2D camera 4 and the 3D scanning device 5 are arranged at an interval, before obtaining the axle housing image, rotate the axle 1 so that the axle housing 2 can be located within the shooting fields of the 2D camera 4 and the 3D scanning device 5. In this embodiment, the axle 1 is rotated 90° relative to the 2D camera 4 and the 3D scanning device 5.

[0019] Before obtaining the axle housing image, import the CAD file of the template axle 1 into the server, so as to obtain the diameter of the axle housing circle of the template axle 1, and convert the diameter of the axle housing circle into pixel diameter information. On the premise of camera calibration, use the 2D camera 4 to obtain the axle housing image. After obtaining the axle housing image, perform adaptive histogram equalization on the axle housing image and perform filtering to obtain a preprocessed image, and perform edge detection on the preprocessed image to obtain edge information. The filtering method adopted in this embodiment is Gaussian and Kalman filtering. The process of performing edge detection to obtain edge information is a prior art.

[0020] The initial axle housing circle is uniformly subdivided at 1° to obtain 360 subdivision points. For each subdivision point a , take from the axle housing image a as the center and along the subdivision point a and the direction of the line connecting the center of the initial axle housing circle are distributed m ∈[60, 80] data points, that is, take m / 2 data points along the direction close to the center of the initial axle housing circle and take m / 2 data points along the direction away from the center of the initial axle housing circle. From these mA data group is composed of data points. Starting from the data point closest to the center of the circle, search for the jump platform in the data array in the direction away from the center of the initial bridge package circle, and record the position of the start end of the jump platform as the jump position.

[0021] Step S2: For each workpiece 3 welded on the bridge package 2 of the template axle 1, obtain the position of the workpiece 3, and use a 3D scanning device 5 to scan left and right at a set length centered on the position of the workpiece 3 Length / 2 to obtain the scanning data of the workpiece 3. Manually select feature points from the scanning data of the workpiece 3, and select several line data up and down with the feature points as reference points in the scanning data of the workpiece 3. Search for jump platforms in each line of data respectively, and take the position where the distance between the start end position and the end position of each jump platform and the feature point is within the distance threshold and closer to the feature point as the feature position. Record each feature position, as well as the jump method and jump height corresponding to the feature position, as a feature information, and combine the normalized data of each feature position with the center of the precise bridge package circle to obtain the precise position of the workpiece 3 relative to the center of the circle, where, Length Set according to the width of the workpiece 3 to be scanned, generally 80 - 120 mm; Specifically, obtain the positions of the workpieces 3 according to the CAD file of the template axle 1 imported into the server.

[0022] The coordinates of the feature points manually selected are expressed as ( x , y , z ), select y 3 lines of data with coordinate values greater than the coordinate values of the feature point y , and y 3 lines of data with coordinate values less than the coordinate values of the feature point y . The y coordinate values of the same line of data are the same. When determining the feature position, compare the distance between the x coordinate value of the start end position of the jump platform and the x coordinate value of the feature point, and the distance between the x coordinate value of the end position and the x coordinate value of the feature point respectively to see if they are within the distance threshold (the distance threshold is set in the range of 1 mm). If both are within the distance threshold, select the position with the closer distance as the feature position. If only one is within the distance threshold, it is the feature position. If neither is within the distance threshold, neither of them is used as the feature position. Among them, when manually selecting feature points, select the visible jump edge, such as the position where the height of the workpiece 3 changes; When searching for the jump platform in the same line of data, if the difference between the depth information of the i +1th data and the depth information of the i th data is greater than the set depth threshold (the depth threshold is set to 2 mm), thei Before a certain number of consecutive data before a data are all less than a set depth threshold from the depth information of the i data, and when the sum of the distances of the i data and a certain number of consecutive data before it is greater than a set length threshold (the length threshold is set to 3 mm), it is determined that a jump occurs. The jump methods include depth increase and depth decrease. When a jump occurs, if the depth information of the i +1 data is greater than the depth information of the i data, then the i +1 data is a depth increase point (i.e., the starting position of the jump platform). If the depth information of the i +1 data is less than the depth information of the i data, then the i +1 data is a depth decrease point (i.e., the ending position of the jump platform). The jump platform includes the data corresponding between the depth increase point and the depth decrease point, and the jump height is the difference between the average depth information of the data in two adjacent jump platforms.

[0023] Obtaining the exact position of the workpiece 3 relative to the center of the circle refers to the position of the workpiece 3 in the horizontal direction of the axle 1 relative to the center of the accurate axle housing circle and the height at which the workpiece 3 is welded on the axle 1, and this height is the height relative to the 3D scanning device 5.

[0024] Step S3: During detection, according to the exact position of the workpiece 3 relative to the center of the circle obtained in step S2, with the accurate axle housing center as the reference, adjust the 3D scanning device 5 in the horizontal direction so that it takes each detected workpiece 3 welded on the axle 1 to be detected as the center, and scan left and right by a set length Length / 2 to obtain the scanning data of the detected workpiece 3, and process the scanning data of the detected workpiece 3 according to step S2 to obtain the characteristic information of each detected workpiece 3. If the characteristic information of a certain detected workpiece 3 can match one of the characteristic information recorded in step S2 (the match includes that the characteristic position, jump method, and jump height are all the same), it is determined that the welding position of the detected workpiece 3 is qualified.

[0025] The specific matching process is as follows: Corresponding matching errors are set for the characteristic position and the jump height respectively. If the differences between the characteristic position and the jump height of the detected workpiece 3 and the characteristic position and the jump height in the characteristic information recorded in step S2 are within the matching error range, it is considered a match.

[0026] As described above, it is only the preferred embodiment of the present invention, so the scope of implementation of the present invention cannot be limited thereby. That is, equivalent changes and modifications made according to the scope of the patent application of the present invention and the content of the specification should still fall within the scope covered by the patent of the present invention.

Claims

1. A welding position detection method for an axle workpiece, characterized in that: It includes the following steps: Step S1: Use a 2D camera to capture the bridge housing of the template axle, obtain the bridge housing image, perform edge detection on the bridge housing image to obtain edge information, use the edge information to perform Hough transform circle detection to obtain the initial bridge housing circle, and evenly divide the initial bridge housing circle to obtain n subdivision points, respectively obtain m data points in the bridge housing image centered on each subdivision point and along the direction of the line connecting the subdivision point and the center of the initial bridge housing circle, and obtain the jump positions where jumps occur in the data array composed of m data points, and use n jump positions to perform Hough transform circle detection to obtain the precise bridge housing circle; Step S2: For each workpiece welded on the template axle housing, obtain the position of the workpiece. Use a 3D scanning device to scan left and right a set length with the position of the workpiece as the center to obtain workpiece scanning data. Manually select feature points from the workpiece scanning data, and select several line data up and down with the feature points as reference points in the workpiece scanning data. Search for jump platforms in each line data respectively, and take the position in the head position and the end position of each jump platform whose distance from the feature point is within the distance threshold and closer to the feature point as the feature position. Record each feature position, as well as the corresponding jump mode and jump height of the feature position as a feature information, and combine the normalized data of each feature position with the center of the precise axle housing circle to obtain the precise position of the workpiece relative to the center of the circle; Step S3: During detection, according to the precise position of the workpiece relative to the center of the circle obtained in Step S2, with the center of the precise axle housing circle as the reference, adjust the 3D scanning device so that it takes each detected workpiece welded on the axle to be detected as the center, scan left and right a set length to obtain the detected workpiece scanning data, and obtain the feature information of each detected workpiece according to Step S2. If the feature information of a certain detected workpiece can match one of the feature information recorded in Step S2, it is determined that the welding position of the detected workpiece is qualified.

2. The welding position detection method of an axle workpiece according to claim 1, characterized in that: In the said Step S1, before obtaining the axle housing image, obtain the axle housing circle diameter of the template axle according to the CAD file of the template axle, and convert the axle housing circle diameter into pixel diameter information.

3. The welding position detection method for an axle workpiece according to claim 2, characterized in that: In the said Step S1, after obtaining the axle housing image, perform adaptive histogram equalization on the axle housing image and filter it to obtain a preprocessed image, and perform edge detection on the preprocessed image to obtain the said edge information.

4. The welding position detection method of an axle workpiece according to claim 3, characterized in that: In the step S1, the initial bridge packet circle is evenly subdivided by 1° to obtain 360 subdivision points, and in the bridge packet image, m / 2 data points along the direction close to the center of the initial bridge packet circle and m / 2 data points along the direction away from the center of the initial bridge packet circle are obtained. The jump platform in the data array is searched in the direction away from the center of the initial bridge packet circle, and the starting position of the jump platform is recorded as the jump position.

5. The welding position detection method for an axle workpiece according to claim 4, wherein: In the said Step S2, obtain the positions of each workpiece according to the CAD file of the template axle.

6. The welding position detection method for an axle workpiece according to claim 5, characterized in that: In step S2, when searching for a jump platform in the line data, if the difference between the depth information of the i +(i + 1)-th data and the depth information of the i i-th data is greater than a set depth threshold, the depth information differences between several consecutive data before the i i-th data and the depth information of the i i-th data are all less than the set depth threshold, and the sum of the distances of the i i-th data and several consecutive data before it is greater than a set length threshold, it is determined that a jump occurs. The jump modes include depth increase and depth decrease. The jump platform includes the data between the depth increase point and the depth decrease point, and the jump height is the difference between the average depth information of the data in two adjacent jump platforms.

7. The welding position detection method for an axle workpiece according to any one of claims 1 to 6, characterized in that: In the step S2, the coordinates of the manually selected feature points are represented as ( x , y , z ). Taking the y coordinate value as the center, 3-line data are selected respectively above and below. When determining the feature position, the x coordinate value at the start position of the jump platform and the x coordinate value at the end position are respectively compared with the x coordinate value of the feature point to check whether the distance therebetween is within the distance threshold.

8. A welding position detection method for an axle workpiece according to any one of claims 1 to 6, characterized in that: In the said Step S2, the obtaining of the precise position of the workpiece relative to the center of the circle refers to the position of the workpiece in the horizontal direction of the axle relative to the center of the precise axle housing circle and the welding height of the workpiece on the axle.

9. The welding position detection method for an axle workpiece according to any one of claims 1 to 6, characterized in that: Both the 2D camera and the 3D scanning device are arranged on a bracket, and the bracket is arranged to slide along the horizontal direction of the axle. The 2D camera and the 3D scanning device are arranged at intervals. Before obtaining the axle housing image, rotate the axle so that the axle housing can be located within the shooting fields of the 2D camera and the 3D scanning device.

Images