Automatic processing method and system for sound barrier post groove based on three-dimensional detection

Abstract

The present application relates to the technical field of automatic processing of metal structural parts, and particularly relates to a sound barrier stand column groove automatic processing method and system based on three-dimensional detection, which comprises the following steps: clamping and positioning a square tube stand column and determining a processing section; using a line laser scanning method to scan the processing section along the axial direction to obtain three-dimensional point cloud data; extracting a cross section contour based on the three-dimensional point cloud data, calculating a wall thickness value and a surface angle, performing outlier rejection, filtering processing and cubic spline curve fitting on the wall thickness value to obtain a wall thickness continuous distribution function; calculating a feed depth, an axial feed speed and a indexing angle according to the continuous distribution function, generating a processing program and performing milling processing to form a groove. According to the present application, the whole section of the stand column processing section is scanned in three dimensions, the continuous distribution function of the wall thickness along the axial direction is established, and the feed depth is adjusted point by point according to the wall thickness changes at different positions, so that the groove processing parameters can be compensated according to the actual appearance changes of the pipe.

Description

Technical Field

[0001] This invention relates to the field of automated processing technology for metal structural components, and in particular to an automated processing method and system for the bevel of sound barrier columns based on three-dimensional detection. Background Technology

[0002] With the development of automated processing technology in the metal structural component manufacturing industry, metal frame columns in products such as sound barriers, guardrails, and steel fences are increasingly being mass-produced using CNC machining. These metal frames are typically formed by welding square tube columns to a mounting base plate. The ends of the square tube columns require beveling to meet the requirements of subsequent welding connections. Currently, for beveling the ends of square tube columns, CNC milling equipment is typically used to cut the column ends according to a preset machining program. This program is designed based on the product specifications, wall thickness, and preset beveling angle. With the development of 3D inspection technology, some processing equipment further incorporates line laser scanning technology to acquire workpiece surface contour data and adjusts processing parameters based on the inspection data to adapt to the processing needs of different specifications of metal tubing.

[0003] The existing method for beveling thin-walled square tubes mostly involves scanning a few sections at the tube end and using a single wall thickness value to represent the entire tube for uniform compensation. Because the periodic fluctuation of the wall thickness along the axial direction of the high-frequency welded pipe is ignored, this results in the problem that the blunt edge is excessively cut off in areas with thinner wall thickness and the blunt edge exceeds the standard in areas with thicker wall thickness. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides an automatic processing method and system for beveling sound barrier columns based on three-dimensional detection. It aims to improve the existing practice of scanning a few sections at the pipe end and using a single wall thickness value to represent the entire pipe section for uniform compensation, which easily leads to the problem of excessive cutting of the blunt edge in areas with thin wall thickness and the blunt edge exceeding the standard in areas with thick wall thickness.

[0005] In a first aspect, the present invention provides the following technical solution: an automatic processing method for the bevel of a sound barrier column based on three-dimensional detection, comprising the following steps:

[0006] S1. Clamp and position the square tube column so that the column axis is parallel to the feed direction during processing, and determine the processing section starting from the end face of the column.

[0007] S2. Using a line laser scanning method, the entire processing section is continuously scanned along the axial direction to obtain three-dimensional point cloud data reflecting the contours of each outer surface of the processing section;

[0008] S3. Based on the three-dimensional point cloud data, extract the cross-sectional contours at different locations along the axial direction, calculate the wall thickness value of each surface at the corresponding position of each cross-sectional contour and the included angle between adjacent surfaces; perform outlier removal and filtering on each wall thickness value, and perform cubic spline curve fitting on the processed wall thickness value to obtain the continuous distribution function of the wall thickness of each surface along the axial direction.

[0009] S4. Calculate the depth of cut point by point along the axial direction based on the continuous distribution function, wherein the depth of cut at any axial position is equal to the difference between the function value of the continuous distribution function at that position and the preset blunt edge width; adjust the axial feed rate according to the derivative of the continuous distribution function, wherein the axial feed rate is inversely related to the derivative; determine the indexing angle for machining different surfaces based on the included angle; generate a machining program by combining the calculated depth of cut, axial feed rate, and indexing angle.

[0010] S5. Execute the machining program to mill the machining section along the axial direction and form a bevel at the end of the square tube column.

[0011] The above technical solution involves using line laser scanning to continuously scan the entire processing section at the end of the column, acquiring three-dimensional point cloud data reflecting the contours of each outer surface of the section. Based on the three-dimensional point cloud data, the wall thickness values ​​at each position along the axial direction are extracted and fitted to obtain a continuous distribution function of the wall thickness along the axial direction for each surface. Then, the depth of cut is calculated point by point according to the continuous distribution function along the axial position, so that the depth of cut is automatically increased in areas with thicker wall thickness and automatically decreased in areas with thinner wall thickness. This improves the existing practice of scanning a few sections at the pipe end and using a single wall thickness value to represent the entire pipe for uniform compensation, which ignores the periodic fluctuation of the wall thickness along the axial direction of high-frequency welded pipes, resulting in excessive removal of the blunt edge in areas with thinner wall thickness and the blunt edge exceeding the standard in areas with thicker wall thickness.

[0012] Furthermore, following S5, it also includes:

[0013] S6. Perform a full-segment continuous scan on the processed section after processing to obtain detection point cloud data reflecting the bevel morphology after processing; based on the detection point cloud data, extract multiple detection points along the axial and circumferential directions, and calculate the bevel angle and blunt edge width of each detection point; compare the measured values ​​of the bevel angle and blunt edge width of all detection points with the preset bevel angle tolerance and the preset blunt edge width tolerance, respectively, and output the judgment result of qualified or unqualified.

[0014] Further, in S2, the step of acquiring three-dimensional point cloud data reflecting the contours of each outer surface of the processing section includes:

[0015] Acquire the axial position corresponding to each acquisition frame during the line laser scanning process;

[0016] Based on the axial position corresponding to each acquisition frame, the two-dimensional contour data acquired in each frame are transformed to the same three-dimensional coordinate system;

[0017] The three-dimensional point cloud data is formed by sequentially stitching together the converted two-dimensional contour data according to their axial positions.

[0018] Furthermore, in S3, the step of outlier removal and filtering for each of the wall thickness values ​​includes:

[0019] The wall thickness values ​​of each surface are sorted according to their axial position to form a wall thickness data sequence for each surface.

[0020] Calculate the local outlier factor for each wall thickness value in each wall thickness data sequence;

[0021] Remove the wall thickness values ​​corresponding to local outliers that are greater than a preset threshold to obtain the wall thickness data sequence after removing outliers;

[0022] The wall thickness data sequence after removing outliers is filtered to obtain the processed wall thickness value.

[0023] Further, in S3, the step of performing cubic spline curve fitting on the processed wall thickness values ​​to obtain the continuous distribution function of the wall thickness along the axial direction for each surface includes:

[0024] Establish a one-to-one correspondence between the treated wall thickness of each surface and its corresponding axial position to form fitting data points;

[0025] Cubic spline curves were fitted to the fitted data points of each surface to obtain the cubic spline curves corresponding to each surface.

[0026] Generate a continuous axial distribution function of the wall thickness of each surface based on the cubic spline curves corresponding to each surface.

[0027] Further in S4, the step of calculating the depth of cut point by point according to the axial position includes:

[0028] Obtain the function values ​​of the continuous distribution function at each axial position;

[0029] The function value corresponding to each axial position is used as the wall thickness value corresponding to each axial position.

[0030] Calculate the difference between the wall thickness value and the preset blunt edge width at each axial position to obtain the depth of cut at each axial position.

[0031] Further, in S4, the step of adjusting the axial feed rate based on the derivative of the continuous distribution function includes:

[0032] Calculate the derivative values ​​of the continuous distribution function at each axial position;

[0033] Calculate the axial feed rate corresponding to each axial position based on the derivative value corresponding to each axial position;

[0034] Establish the correspondence between each axial position and the corresponding axial feed rate to form axial feed rate data.

[0035] Furthermore, in S4, the step of determining the rotation angle when machining different surfaces based on the included angle includes:

[0036] Obtain the included angle between each adjacent surface;

[0037] Establish the angular correspondence between each surface to be processed and each adjacent surface;

[0038] The rotation angle corresponding to each surface to be processed is determined based on the included angle of each surface to be processed.

[0039] Furthermore, in S4, the step of generating the machining program by comprehensively calculating the depth of cut, axial feed rate, and indexing angle includes:

[0040] Acquire the depth of cut, axial feed rate data corresponding to each axial position, and the indexing angle corresponding to each surface to be machined;

[0041] The tool path data for each surface to be machined is generated based on the depth of cut corresponding to each axial position.

[0042] The machining program is generated based on the tool path data, axial feed rate data, and the indexing angles of each surface to be machined.

[0043] Secondly, the present invention provides an automatic processing system for the bevel of sound barrier columns based on three-dimensional detection, comprising:

[0044] A line laser profile sensor is installed on the side of the spindle of a CNC machining device. After the square tube column is clamped and positioned, it is used to continuously scan the entire machining section at the end of the column along the axial direction to obtain three-dimensional point cloud data reflecting the profile of each outer surface of the machining section.

[0045] The industrial control computer receives the 3D point cloud data and performs the following operations:

[0046] Based on the three-dimensional point cloud data, the cross-sectional profiles at different locations along the axial direction are extracted, and the wall thickness value of each surface at the corresponding position of each cross-sectional profile and the included angle between adjacent surfaces are calculated.

[0047] Outlier removal and filtering are performed on each of the wall thickness values, and cubic spline curve fitting is performed on the processed wall thickness values ​​to obtain the continuous distribution function of the wall thickness along the axial direction for each surface.

[0048] According to the continuous distribution function, the depth of cut is calculated point by point according to the axial position, wherein the depth of cut at any axial position is equal to the difference between the function value of the continuous distribution function at that position and the preset blunt edge width;

[0049] The axial feed rate is adjusted according to the derivative of the continuous distribution function, and the axial feed rate is inversely related to the derivative.

[0050] The rotation angle is determined based on the included angle when machining different surfaces;

[0051] A machining program is generated by combining the calculated depth of cut, axial feed rate, and indexing angle, and the machining program is sent to the CNC machining device.

[0052] The CNC machining device receives and executes the machining program, drives the milling cutter to mill the machining section along the axial direction, and forms a bevel at the end of the square tube column.

[0053] The present invention has the following beneficial effects:

[0054] 1. In this invention, a continuous full-section scanning of the processing section at the end of the column is performed using line laser scanning to obtain three-dimensional point cloud data reflecting the contours of each outer surface of the section. Based on the three-dimensional point cloud data, the wall thickness values ​​at each position along the axial direction are extracted and fitted to obtain a continuous distribution function of the wall thickness along the axial direction for each surface. Then, the depth of cut is calculated point by point according to the continuous distribution function along the axial position, so that the depth of cut is automatically increased in areas with thicker wall thickness and automatically decreased in areas with thinner wall thickness. This improves the existing practice of scanning a few sections at the pipe end and using a single wall thickness value to represent the entire pipe for uniform compensation in the beveling of thin-walled square tubes. Because the periodic fluctuation of the wall thickness along the axial direction of the high-frequency welded pipe is ignored, the blunt edge in areas with thinner wall thickness is excessively cut off, and the blunt edge in areas with thicker wall thickness exceeds the standard.

[0055] 2. In this invention, the axial feed rate is adjusted according to the derivative of the continuous distribution function, so that the axial feed rate and the derivative are inversely related. The feed rate is automatically reduced when the wall thickness changes drastically and restored to the normal feed rate when the wall thickness changes gently. This improves the problem that most existing beveling methods perform variable depth cutting with a constant feed rate. Due to the lag in the response of the servo system to the depth change, the accuracy of the blunt edge in the wall thickness change section is difficult to control stably.

[0056] 3. In this invention, the included angle between adjacent surfaces is calculated based on three-dimensional point cloud data, and the rotation angle is determined according to the included angle when processing different surfaces. This improves the problem that most existing square tube beveling processes are indexed and rotated according to fixed theoretical angles. Since the cross-section of thin-walled square tubes generally has a non-square diagonal difference, the milling cutter axis deviates from the normal direction of the surface to be processed and the beveling angle is distorted. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the overall process of an automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to the present invention.

[0058] Figure 2 This is a schematic diagram of the process for obtaining three-dimensional point cloud data of the processing section based on line laser scanning in an automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to the present invention.

[0059] Figure 3 This is a schematic diagram of the data processing flow for establishing a continuous axial distribution function of the wall thickness of each surface based on three-dimensional point cloud data in the automatic processing method for bevel of sound barrier columns based on three-dimensional detection of the present invention.

[0060] Figure 4 This is a flowchart illustrating the automatic processing method for the bevel of a sound barrier column based on three-dimensional detection, which calculates processing parameters and generates a processing program according to a continuous distribution function.

[0061] Figure 5 This is a schematic diagram illustrating the process of performing three-dimensional detection and verification of the bevel area after processing in an automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to the present invention.

[0062] Figure 6 This is a schematic diagram of the module architecture and data flow of an automatic processing system for the bevel of a sound barrier column based on three-dimensional detection, according to the present invention. Detailed Implementation

[0063] 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.

[0064] Example 1: In the first embodiment of the present invention, the present invention provides an automatic processing method for the bevel of sound barrier columns based on three-dimensional detection, such as... Figure 1 As shown, it includes the following steps:

[0065] S1. Clamp and position the square tube column so that the column axis is parallel to the feed direction during processing, and determine the processing section starting from the end face of the column.

[0066] Specifically, the thin-walled square tube column is placed on the pneumatic fixture and pushed against the fixed axial positioning surface and lateral positioning surface. After confirming that it is firmly attached, the cylinder drives the gripper to press it. After clamping, the column axis is parallel to the X-axis of the machine tool. The length of the column end extending out of the front end face of the fixture is the processing section. This length L is set according to the product process requirements. In this solution, L is preferably 60mm.

[0067] like Figure 2 As shown, S2 uses a line laser scanning method to continuously scan the entire processing section along the axial direction to obtain three-dimensional point cloud data reflecting the contours of each outer surface of the processing section.

[0068] Furthermore, in S2, the step of acquiring three-dimensional point cloud data reflecting the contours of each outer surface of the processing section includes:

[0069] Acquire the axial position corresponding to each acquisition frame during the line laser scanning process;

[0070] Based on the axial position corresponding to each acquisition frame, the two-dimensional contour data acquired in each frame are transformed to the same three-dimensional coordinate system;

[0071] The converted two-dimensional contour data are stitched together sequentially according to their axial positions to form three-dimensional point cloud data.

[0072] Specifically, a line laser contour sensor is installed on the side of the spindle. The spindle moves at a constant speed along the X-axis, driving the sensor to scan from the end face of the column towards the fixture. The scanning stroke covers the entire machining section length L. During the scanning process, the column remains stationary, and the sensor continuously collects contour data at a fixed frequency. The axial spacing between adjacent acquisition frames is set to no more than 1 mm. During the scanning process, the spindle X-axis coordinate value at the acquisition time of each frame is recorded, and this coordinate value is used as the axial position of the contour data of that frame. A three-dimensional coordinate system is established with the center of the end face of the column as the origin and the X-axis coinciding with the feed direction. In this scheme, setting this origin is a preferred method. The two-dimensional contour points acquired in each frame are translated into the three-dimensional coordinate system according to their corresponding axial positions. The contour data of each frame are arranged in ascending order of axial position to form three-dimensional point cloud data used to describe the geometry of the outer surface of the entire machining section.

[0073] like Figure 3 As shown in Figure S3, based on the three-dimensional point cloud data, the cross-sectional contours at different locations are extracted along the axial direction, and the wall thickness value of each surface at the corresponding position of each cross-sectional contour and the included angle between adjacent surfaces are calculated; outlier removal and filtering are performed on each wall thickness value, and cubic spline curve fitting is performed on the processed wall thickness value to obtain the continuous distribution function of the wall thickness of each surface along the axial direction.

[0074] Furthermore, in S3, the steps for outlier removal and filtering of each wall thickness value include:

[0075] The wall thickness values ​​of each surface are sorted according to their axial position to form a wall thickness data sequence for each surface.

[0076] Calculate the local outlier factor for each wall thickness value in each wall thickness data sequence;

[0077] Remove the wall thickness values ​​corresponding to local outliers that are greater than a preset threshold to obtain the wall thickness data sequence after removing outliers;

[0078] The wall thickness data sequence after removing outliers is filtered to obtain the processed wall thickness value.

[0079] Specifically, the processing section is divided into N sections at equal intervals along the axial direction, with a spacing of 2mm to 3mm. Taking the end face of the column as the axial zero point, the axial coordinate of the i-th section is... ; in each section At this point, the outer surface contour segments of the four faces are extracted from the 3D point cloud data, and the straight line equations of each face are fitted. Simultaneously, the end face contour line at this cross-section is obtained. For each face, the perpendicular distance from the fitted outer surface line segment to the fitted end face line segment is calculated, serving as the preliminary measured wall thickness value for that face at this cross-section. The angle between adjacent surfaces is obtained by calculating the angle between the normal directions of the fitted outer surface lines of two adjacent faces. For each face j, the preliminary wall thickness values ​​extracted from N cross-sections are then calculated according to the axial coordinates. Arranged in ascending order, they form a sequence of wall thickness data, in which... The four sides of the corresponding square tube are scanned using a local outlier algorithm. The local density deviation within the neighborhood of each data point is calculated. When the local outlier value of a point exceeds a preset threshold, it is identified as measurement noise and removed. In this scheme, the preset threshold is preferably 2.0, which can be adjusted according to the actual surface quality of the pipe and the measurement noise level. For the valid data points retained after noise removal, Gaussian filtering is used for smoothing. The filter window width is selected based on the axial sampling interval to obtain the wall thickness values ​​of each section reflecting the true wall thickness trend of the pipe. .

[0080] Furthermore, in S3, the step of performing cubic spline curve fitting on the processed wall thickness values ​​to obtain the continuous axial distribution function of the wall thickness of each surface includes:

[0081] Establish a one-to-one correspondence between the treated wall thickness of each surface and its corresponding axial position to form fitting data points;

[0082] Cubic spline curves were fitted to the fitted data points of each surface to obtain the cubic spline curves corresponding to each surface.

[0083] Generate a continuous axial distribution function of the wall thickness of each surface based on the cubic spline curves corresponding to each surface.

[0084] Specifically, for each face j, the filtered discrete data points As a type value point, in the interval Constructing a cubic spline interpolation function , In each sub-interval The above is a cubic polynomial The above coefficients The following boundary conditions are simultaneously solved to determine the values ​​at each type point. Satisfying The first derivative is continuous at each internal type value point. and continuity of the second derivative And at the endpoints of the interval and The natural boundary conditions are satisfied. ; income That is, the continuous distribution function of the wall thickness of surface j along the axial direction. ; .

[0085] like Figure 4 As shown in step S4, calculate the depth of cut point by point along the axial position according to the continuous distribution function, where the depth of cut at any axial position is equal to the difference between the function value of the continuous distribution function at that position and the preset blunt edge width; adjust the axial feed rate according to the derivative of the continuous distribution function, the axial feed rate and the derivative are inversely related; determine the indexing angle when machining different surfaces according to the included angle; generate the machining program by combining the calculated depth of cut, axial feed rate and indexing angle.

[0086] Furthermore, in S4, the step of calculating the depth of cut point by point according to the axial position includes:

[0087] Obtain the function values ​​of the continuous distribution function at each axial position;

[0088] The function value corresponding to each axial position is used as the wall thickness value corresponding to each axial position.

[0089] Calculate the difference between the wall thickness value and the preset blunt edge width at each axial position to obtain the depth of cut at each axial position.

[0090] Specifically, the bevel angle α is set to a fixed value according to the welding process specification, and does not change with axial position or wall thickness; for face j, at any position x of the tool along the axial feed, a continuous distribution function is invoked. Obtain the wall thickness value at this location and the target depth of cut. According to the formula Calculation; where To preset the target blunt edge width, a fixed value between 1.0mm and 1.5mm is selected according to welding process requirements. This calculation automatically increases the depth of cut in areas with thicker walls and automatically decreases the depth of cut in areas with thinner walls, ensuring that the remaining blunt edge width after machining at any position is equal to the target width. .

[0091] Furthermore, in S4, the step of adjusting the axial feed rate based on the derivative of the continuous distribution function includes:

[0092] Calculate the derivative values ​​of the continuous distribution function at each axial position;

[0093] Calculate the axial feed rate corresponding to each axial position based on the derivative value corresponding to each axial position;

[0094] Establish the correspondence between each axial position and the corresponding axial feed rate to form axial feed rate data.

[0095] Specifically, calculate the continuous distribution function. first derivative Its physical meaning is the rate of change of wall thickness along the axial direction; axial feed rate. According to the formula Calculation; where This is the reference feed rate during normal cutting. This is a preset feed rate adjustment coefficient, used to adjust the sensitivity of the feed rate to the wall thickness change rate. It is a dimensionless constant in this scheme. Preferably 50; when When the wall thickness is large, the feed rate automatically decreases to ensure the servo system has sufficient response time to accurately execute continuous depth changes. When the wall thickness change is gradual, the feed rate returns to normal. .

[0096] Furthermore, in S4, the step of determining the indexing angle when machining different surfaces based on the included angle includes:

[0097] Obtain the included angle between each adjacent surface;

[0098] Establish the angular correspondence between each surface to be processed and each adjacent surface;

[0099] The rotation angle corresponding to each surface to be processed is determined based on the included angle of each surface to be processed.

[0100] Specifically, let the measured value of the included angle between adjacent faces calculated from the 3D point cloud data be... When machining the current face and the workpiece or tool needs to be rotated to machine the next face, the rotation axis command angle is adopted. Replace fixed theoretical values This ensures that the milling cutter axis is always aligned with the normal direction of the next surface to be machined.

[0101] Furthermore, in S4, the steps for generating the machining program by combining the calculated depth of cut, axial feed rate, and indexing angle include:

[0102] Acquire the depth of cut, axial feed rate data corresponding to each axial position, and the indexing angle corresponding to each surface to be machined;

[0103] The tool path data for each surface to be machined is generated based on the depth of cut corresponding to each axial position.

[0104] The machining program is generated based on the tool path data, axial feed rate data, and the indexing angles of each surface to be machined.

[0105] Specifically, the depth of cut at each axial position on each face. Axial feed rate and rotation angle The code is compiled into G-code programs executable by the CNC system, first based on... Generate the radial cutting trajectory command of the milling cutter point by point along the X-axis, and then... Map the feed rate words to the corresponding program segments, insert an indexing command at the face switching point, and set the rotation angle to... The final generated program fully describes the motion commands of all axes of motion at any axial position during the machining process; it also specifies the depth of cut for each face and each axial position. Axial feed rate and rotation angle The code is compiled into G-code programs executable by the CNC system, first based on... Generate the radial cutting trajectory command of the milling cutter point by point along the X-axis, and then... Map the feed rate words to the corresponding program segments, insert an indexing command at the face switching point, and set the rotation angle to... The final generated program fully describes the motion commands of all motion axes at any axial position during the machining process.

[0106] S5. Execute the machining program and mill the machining section along the axial direction to form a bevel at the end of the square tube column;

[0107] Specifically, after the CNC system loads the dedicated program, the milling cutter first mills the end face of the pipe to form a flat reference surface. Then, the milling cutter tilts to the bevel angle α and continuously mills from the pipe end toward the fixture according to the axial feed speed and the depth of cut that dynamically changes with the position set by the program, completing the single V-shaped bevel forming of one surface. After the surface is machined, it is rotated to the next surface according to the indexing command in the program. The milling action is repeated until all four surfaces are completed. The entire machining process is completed in one clamping of the column.

[0108] Following S5, it also includes:

[0109] like Figure 5 As shown in step S6, perform a full-segment continuous scan on the processed section after processing to obtain detection point cloud data reflecting the bevel morphology after processing; based on the detection point cloud data, extract multiple detection points along the axial and circumferential directions, and calculate the bevel angle and blunt edge width of each detection point; compare the measured values ​​of the bevel angle and blunt edge width of all detection points with the preset bevel angle tolerance and the preset blunt edge width tolerance respectively, and output the judgment result of qualified or unqualified.

[0110] Specifically, the spindle drives the line laser sensor again to scan the entire bevel area in the same way, with the scanning stroke consistent with the first scan before processing. K detection sections are extracted from the detection point cloud data at equal intervals along the axial direction. For each section, one detection point is taken at the bevel position on each of the four sides, for a total of 4K detection points. For each detection point, the bevel angle is obtained by fitting a straight line to the point cloud on both sides of the bevel and calculating the included angle. At the same time, the remaining thickness from the root of the bevel to the inner side of the pipe wall is measured to obtain the measured blunt edge width. The bevel angle measurement values ​​of all detection points are compared with the preset tolerance zone α±2° and the blunt edge width measurement values ​​are compared with the preset tolerance zone d±0.2mm. If all detection points are within the tolerance zone, they are judged as qualified and the green indicator light is lit. If any point exceeds the tolerance, it is judged as unqualified and an audible and visual alarm is triggered.

[0111] Example 2: In a fence post production workshop, the production line frequently switches between different batches of high-frequency welded square tubes, resulting in significant differences in wall thickness deviation and cross-sectional non-squareness between batches. The original beveling equipment could only execute fixed programs according to theoretical dimensions. Each batch change required repeated manual trial cutting and parameter correction, leading to a high scrap rate and poor consistency in the bevel's blunt edge, becoming a major bottleneck restricting the improvement of subsequent robotic welding pass rates. To solve these problems, an automatic beveling system for sound barrier posts based on three-dimensional detection, as provided in this invention, was adopted. Figure 6 As shown. The specific implementation process of this system is as follows:

[0112] First, the operator loads the square tube column into the pneumatic clamp and pushes it against the positioning surface. After confirming that it is firmly attached, the clamping is started, and the end of the column extends out to a predetermined length as the processing section. Then, the line laser profile sensor scans the entire processing section along the spindle along the axis, and acquires three-dimensional point cloud data covering the four outer surfaces in real time. This eliminates the tedious steps of manually measuring the wall thickness and cross-sectional dimensions, and also avoids the approximate error of inferring the wall thickness of the entire pipe section based on only a local cross-section at the pipe end.

[0113] Then, the industrial control computer analyzes and processes the three-dimensional point cloud data, extracts the actual wall thickness values ​​of each cross section along the axial direction and removes measurement noise caused by surface oxide scale, etc., and then obtains the continuous distribution function of the wall thickness of each surface along the axial direction by fitting. At the same time, it calculates the actual included angle between adjacent surfaces, thus completely quantifying the axial fluctuation of the wall thickness and the degree of non-squareness of the cross section of each column into a calculable mathematical model, providing an accurate decision basis for adaptive processing.

[0114] Next, the industrial control computer automatically plans the machining path using the aforementioned wall thickness distribution function: at each position along the axial direction, the depth of cut is determined based on the difference between the actual wall thickness and the preset blunt edge width, so that the cutting amount is reduced in areas with thinner wall thickness and increased in areas with thicker wall thickness; at the same time, the axial feed rate is automatically adjusted according to the wall thickness change rate, slowing down the feed in sections with rapid wall thickness changes to ensure that the machining trajectory follows accurately; during rotary indexing, the measured angle between adjacent surfaces is used instead of the fixed theoretical value, so that the milling cutter posture always maintains the correct relative relationship with the surface to be machined; the machining program generated in this way completely includes adaptive compensation instructions for all geometric features of the current column, and the entire planning process does not require manual intervention for debugging;

[0115] Subsequently, the CNC machining device executes the aforementioned dedicated program. The milling cutter first mills the flat end face reference, and then continuously moves along the axial direction to complete the beveling of each face. All machining is completed in one clamping. Since the machining program has automatically adapted to the actual wall thickness and cross-sectional shape of each column, whether it is within the same batch or across batches, it can ensure that the width and angle of the bevel blunt edge produced are stable and consistent in the axial and circumferential directions, which significantly reduces the batch change debugging time and scrap rate.

[0116] Finally, after processing, the sensor scans the bevel area again, automatically measures the bevel angle and blunt edge width at each detection point and compares them with the preset tolerance zone, determines whether it is qualified or not online and outputs the results. Thus, without adding manual inspection, closed-loop monitoring of the bevel quality of each column is achieved, ensuring that unqualified products do not flow into the next welding process.

[0117] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An automatic processing method for the bevel of sound barrier columns based on three-dimensional detection, characterized in that, Includes the following steps: S1. Clamp and position the square tube column so that the column axis is parallel to the feed direction during processing, and determine the processing section starting from the end face of the column. S2. Using a line laser scanning method, the entire processing section is continuously scanned along the axial direction to obtain three-dimensional point cloud data reflecting the contours of each outer surface of the processing section; S3. Based on the three-dimensional point cloud data, extract the cross-sectional contours at different locations along the axial direction, calculate the wall thickness value of each surface at the corresponding position of each cross-sectional contour and the included angle between adjacent surfaces; perform outlier removal and filtering on each wall thickness value, and perform cubic spline curve fitting on the processed wall thickness value to obtain the continuous distribution function of the wall thickness of each surface along the axial direction. S4. Calculate the depth of cut point by point along the axial direction based on the continuous distribution function, wherein the depth of cut at any axial position is equal to the difference between the function value of the continuous distribution function at that position and the preset blunt edge width; adjust the axial feed rate according to the derivative of the continuous distribution function, wherein the axial feed rate is inversely related to the derivative; determine the indexing angle for machining different surfaces based on the included angle; generate a machining program by combining the calculated depth of cut, axial feed rate, and indexing angle. S5. Execute the machining program to mill the machining section along the axial direction and form a bevel at the end of the square tube column.

2. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, Following S5, it also includes: S6. Perform a full-segment continuous scan on the processed section after processing to obtain detection point cloud data reflecting the bevel morphology after processing; based on the detection point cloud data, extract multiple detection points along the axial and circumferential directions, and calculate the bevel angle and blunt edge width of each detection point; compare the measured values ​​of the bevel angle and blunt edge width of all detection points with the preset bevel angle tolerance and the preset blunt edge width tolerance, respectively, and output the judgment result of qualified or unqualified.

3. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, In S2, the step of acquiring three-dimensional point cloud data reflecting the contours of each outer surface of the processing section includes: Acquire the axial position corresponding to each acquisition frame during the line laser scanning process; Based on the axial position corresponding to each acquisition frame, the two-dimensional contour data acquired in each frame are transformed to the same three-dimensional coordinate system; The three-dimensional point cloud data is formed by sequentially stitching together the converted two-dimensional contour data according to their axial positions.

4. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, In S3, the step of outlier removal and filtering for each of the wall thickness values ​​includes: The wall thickness values ​​of each surface are sorted according to their axial position to form a wall thickness data sequence for each surface. Calculate the local outlier factor for each wall thickness value in each wall thickness data sequence; Remove the wall thickness values ​​corresponding to local outliers that are greater than a preset threshold to obtain the wall thickness data sequence after removing outliers; The wall thickness data sequence after removing outliers is filtered to obtain the processed wall thickness value.

5. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, In S3, the step of performing cubic spline curve fitting on the processed wall thickness values ​​to obtain the continuous axial distribution function of the wall thickness of each surface includes: Establish a one-to-one correspondence between the treated wall thickness of each surface and its corresponding axial position to form fitting data points; Cubic spline curves were fitted to the fitted data points of each surface to obtain the cubic spline curves corresponding to each surface. Generate a continuous axial distribution function of the wall thickness of each surface based on the cubic spline curves corresponding to each surface.

6. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, In S4, the step of calculating the depth of cut point by point according to the axial position includes: Obtain the function values ​​of the continuous distribution function at each axial position; The function value corresponding to each axial position is used as the wall thickness value corresponding to each axial position; Calculate the difference between the wall thickness value and the preset blunt edge width at each axial position to obtain the depth of cut at each axial position.

7. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, In S4, the step of adjusting the axial feed rate based on the derivative of the continuous distribution function includes: Calculate the derivative values ​​of the continuous distribution function at each axial position; Calculate the axial feed rate corresponding to each axial position based on the derivative value corresponding to each axial position; Establish the correspondence between each axial position and the corresponding axial feed rate to form axial feed rate data.

8. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 1, characterized in that, In S4, the step of determining the rotation angle when machining different surfaces based on the included angle includes: Obtain the included angle between each adjacent surface; Establish the angular correspondence between each surface to be processed and each adjacent surface; The rotation angle corresponding to each surface to be processed is determined based on the included angle of each surface to be processed.

9. The automatic processing method for the bevel of a sound barrier column based on three-dimensional detection according to claim 7, characterized in that, In S4, the step of generating the machining program by combining the calculated depth of cut, axial feed rate, and indexing angle includes: Acquire the depth of cut, axial feed rate data corresponding to each axial position, and the indexing angle corresponding to each surface to be machined; The tool path data for each surface to be machined is generated based on the depth of cut corresponding to each axial position. The machining program is generated based on the tool path data, axial feed rate data, and the indexing angles of each surface to be machined.

10. An automatic processing system for the bevel of sound barrier columns based on three-dimensional detection, characterized in that, include: A line laser profile sensor is installed on the side of the spindle of a CNC machining device. After the square tube column is clamped and positioned, it is used to continuously scan the entire machining section at the end of the column along the axial direction to obtain three-dimensional point cloud data reflecting the profile of each outer surface of the machining section. The industrial control computer receives the 3D point cloud data and performs the following operations: Based on the three-dimensional point cloud data, the cross-sectional profiles at different locations along the axial direction are extracted, and the wall thickness value of each surface at the corresponding position of each cross-sectional profile and the included angle between adjacent surfaces are calculated. Outlier removal and filtering are performed on each of the wall thickness values, and cubic spline curve fitting is performed on the processed wall thickness values ​​to obtain the continuous distribution function of the wall thickness along the axial direction for each surface. According to the continuous distribution function, the depth of cut is calculated point by point according to the axial position, wherein the depth of cut at any axial position is equal to the difference between the function value of the continuous distribution function at that position and the preset blunt edge width; The axial feed rate is adjusted according to the derivative of the continuous distribution function, and the axial feed rate is inversely related to the derivative. The rotation angle is determined based on the included angle when machining different surfaces; A machining program is generated by combining the calculated depth of cut, axial feed rate, and indexing angle, and the machining program is sent to the CNC machining device. The CNC machining device receives and executes the machining program, drives the milling cutter to mill the machining section along the axial direction, and forms a bevel at the end of the square tube column.

Images