A method for laser cladding point position detection and automatic equipment programming

Abstract

This invention discloses a method for detecting laser cladding points and programming automated equipment, comprising the following steps: industrial camera intrinsic parameter calibration, laser processing head focal length and spot image feature calibration—work site preparation—affixing a calibration plate to the surface to be processed—laser processing head tilt angle calibration—laser processing head focal length calibration—generating a processing program. This invention uses machine vision calibration, where the indicator light spot image features serve as the focal length reference, eliminating the need for manual determination of the focal length reference and ensuring accurate detection. Simultaneously, the laser processing head tilt angle calibration further guarantees focal length accuracy during processing, improving cladding quality.

Description

Technical Field

[0001] This invention relates to the field of laser processing, and more specifically to a method for detecting laser cladding points and programming automated equipment. Background Technology

[0002] Currently, laser cladding has become one of the common methods for surface treatment of parts. Common laser cladding processing methods include: manually teaching robots or CNC machine tools to edit processing programs, and editing processing programs based on the dimensions of part design drawings or actual measured dimensions.

[0003] The manual standard calibration bar method for fixed-point programming specifically involves: positioning the processing equipment—installing the processing head—the operator holding the standard calibration bar—visually observing and adjusting the focal length of the processing head—completing the corner point calibration of the processing program point by point—compiling the processing program. Its disadvantages are: 1. The standard length calibration bar held during manual calibration serves as the focal length reference. When manually adjusting the focal length of the processing head visually, the operator's experience and working conditions can affect the accuracy of the focal length, resulting in poor accuracy and a long processing time. 2. It does not consider the impact of the workpiece itself or the processing head itself on the cladding quality, and therefore cannot adequately guarantee the quality of the cladding. Summary of the Invention

[0004] The problem this invention aims to solve is to provide a method for detecting laser cladding points and programming automated equipment. This method automatically generates a processing program by comparing optical image features and performing data calculations. The method uses machine vision calibration, with the indicator light spot image features serving as the focal length reference, eliminating the need for manual determination of the focal length reference and ensuring accurate detection. Simultaneously, the tilt angle calibration of the laser processing head further guarantees accurate focal length during processing, improving cladding quality.

[0005] This invention discloses a method for detecting laser cladding points and programming automated equipment, comprising the following steps: calibration of industrial camera intrinsic parameters, calibration of laser processing head focal length and spot image features—work site preparation—affixing calibration plate to the surface to be processed—calibration of laser processing head tilt angle—calibration of laser processing head focal length—generation of processing program.

[0006] The steps "Industrial camera intrinsic parameter calibration and laser processing head focal length and spot image feature calibration" specifically involve: placing a planar two-dimensional calibration plate on the front of the industrial camera, adjusting the distance and angle between the industrial camera and the calibration plate, taking multiple images of the calibration plate from different angles with the industrial camera, integrating them into image set a1, and processing them using image processing software to obtain camera intrinsic parameter data; moving the laser processing head along the Z-axis of the tool coordinate system, measuring the distance between the laser processing head and the calibration plate, and measuring the laser spot size on the calibration plate, adjusting the position of the laser processing head to meet the focal length requirements of the cladding process, taking and storing the laser processing head indicator red light spot image features through the industrial camera, and repeatedly building a laser processing head focal length and spot image feature dataset.

[0007] The specific steps of "Work Site Preparation" are as follows: cleaning the surface of the parts to be processed to the processing state; processing equipment; installing the laser processing head onto the tool end of the industrial robot, installing the industrial camera onto the side outer wall of the laser processing head, and setting up low-angle side lighting on site to achieve diffuse dark field.

[0008] The step "laser processing head tilt angle calibration" specifically involves: the robot reciprocating along the Z-axis of the tool coordinate system, the industrial camera acquiring the motion trajectory of the red light spot indicating the laser processing head relative to the calibration point, calculating the motion trajectory of the laser processing head, obtaining the tilt angle of the laser processing head relative to the surface to be processed, and feeding back to the robot to correct the tilt angle so that the light output direction of the laser processing head is perpendicular to the surface to be processed.

[0009] Specifically, the step "laser processing head focal length calibration" involves: the laser processing head moving along the Z-axis of the tool coordinate system, the industrial camera capturing the image features of the red light spot indicating the laser processing head, comparing it with the pre-stored image features of the red light spot indicating the laser head, and adjusting the robot position so that the focus of the laser head is located at the predetermined focal length on the surface to be processed of the part.

[0010] When vision is used as an aid, the laser head focal length is used as a reference, and manual fixed-point programming is used, the specific method is as follows: calibration of the intrinsic parameters of the industrial camera, calibration of the focal length and spot image features of the laser processing laser head — preparation of the work site — affixing the calibration plate to the surface to be processed — calibration of the tilt angle of the laser processing laser head — visual inspection and calibration of the focal length of the laser processing laser head at each corner point of the surface to be processed — generation of the processing program.

[0011] The specific steps for visually inspecting and calibrating the focal length of the laser processing head at each corner of the surface to be processed are as follows: A specific corner of the surface to be processed is manually selected for focal length calibration, which is then automatically completed by the equipment. After the focal length calibration of multiple corners is completed, these points are connected to form a closed shape, which is the outline of the surface to be processed. The calibrated corner robot coordinate values ​​are then input into a pre-written program for data replacement, generating the processing program.

[0012] When vision is the primary factor, machine vision inspection results are the benchmark, and the program is automatically generated, the specific method is as follows: industrial camera intrinsic parameter calibration, laser processing head focal length and spot image feature calibration—work site preparation—affixing calibration plate to the surface to be processed—industrial camera extrinsic parameter calibration, part coordinate system calibration—laser processing head tilt angle calibration—laser processing head focal length calibration—contour extraction, motion trajectory node extraction, motion trajectory generation—generating the processing program.

[0013] Specifically, the calibration of the industrial camera's extrinsic parameters and the calibration of the part coordinate system are as follows: the industrial robot takes pictures of the surface to be processed at different distances and angles, extracts the pixel data of each feature point on the calibration board, and calculates the spatial positional relationship between each feature point on the calibration board; based on the spatial relationship between the camera coordinate system and the robot world coordinate system, and based on the pixel standard size specification and the ratio of image pixel size, a part coordinate system relative to the robot world coordinate system is constructed, and the extrinsic parameters of the camera are calculated.

[0014] Specifically, the extraction of part machining contours, extraction of motion trajectory nodes, and generation of motion trajectories are as follows: a calibration board composed of two-dimensional squares is used for macroscopic dimension calibration; distant images of the robot are collected; a rough vertical contour is extracted from the distant view; the shooting area is divided according to the shooting accuracy and shooting size range of the industrial camera; clear images are taken sequentially according to the shooting area; the position of the camera coordinate system relative to the robot world coordinate system is recorded simultaneously; the images are processed to obtain the contour and corner coordinates of the area to be processed; after obtaining the edge of the image contour, the pixel coordinates of the contour corner points are converted into tool coordinate system coordinates under the robot world coordinate system, that is, the motion trajectory nodes are extracted; the area enclosed by multiple corner points is the machining area, that is, the machining contour of the part, and then the motion trajectory is generated.

[0015] The specific steps of image processing are as follows: input – grayscale conversion – noise reduction – binarization – morphological processing – threshold segmentation – edge extraction – contour feature extraction.

[0016] The advantages of this invention are: 1. During visual calibration, the image features of the indicator light spot serve as the focal length reference, eliminating the need for manual determination of the focal length reference and ensuring accurate detection; 2. The tilt angle calibration of the laser processing head further guarantees accurate focal length during processing, improving processing quality; 3. Method 1 uses manual visual inspection of each corner point of the surface to be processed to calibrate the focal length of the laser processing head, making manual point setting flexible and convenient; 4. Method 2 detects the size and specifications of the surface to be processed and the spatial position of the processing equipment relative to the measured part, automatically plans the cladding processing path trajectory and generates trajectory node coordinate values, generates a motion control program, and controls the laser processing equipment to complete the cladding processing; it achieves vision-driven, machine vision-based detection results, with the machine determining the calibration point instead of manually determining it, thus overcoming the problems inherent in human subjectivity, making the obtained motion trajectory more accurate and comprehensive, and the cladding quality more reliable; the automatic generation of the processing program does not rely on any human experience, greatly improving work efficiency and achieving automation; the industrial camera installed on the outer wall of the processing head can obtain the processing program as soon as it is visible, avoiding the sensitivity of manual inspection and calibration to heights, narrow areas, and dangerous and harsh environments, and ensuring good accessibility. Attached Figure Description

[0017] Figure 1 Schematic diagram for laser processing head tilt angle calibration;

[0018] Figure 2 A schematic diagram illustrating how tilt angles can expand the robot's processing range.

[0019] Figure 3 A schematic diagram for laser processing head focal length calibration. Detailed Implementation

[0020] Example 1

[0021] The method of this invention includes the following steps: calibration of the intrinsic parameters of the industrial camera, calibration of the focal length and spot image features of the laser processing head, preparation of the work site, affixing of the calibration plate to the surface to be processed, calibration of the tilt angle of the laser processing head, calibration of the focal length of the laser processing head, and generation of the processing program.

[0022] The specific steps are described below:

[0023] I. Industrial camera intrinsic parameter calibration, laser processing optical head focal length and spot image feature calibration:

[0024] The intrinsic parameter calibration of the industrial camera and the focal length and spot image feature calibration of the laser processing head are completed in advance in a non-operational field (laboratory). The specific steps are as follows: Industrial camera intrinsic parameter calibration: A two-dimensional planar calibration board is placed on the front of the industrial camera. The calibration board consists of regular squares and marker circles. The size, specifications, and positional relationship of the squares and marker circles are known. The distance and angle between the industrial camera and the calibration board are adjusted so that the field of view of the industrial camera covers the image of the calibration board, and the image of the calibration board is clear and distinct. The industrial camera takes multiple images of the calibration board from different angles and integrates them into an image set a1. Image processing software is used to process the images to obtain the camera's intrinsic parameter data, including focal length, pixel size, and the number of pixels that differ between the center and origin of the image. After the camera intrinsic parameter calibration is completed, the captured images can be automatically corrected so that the images objectively reflect the real world.

[0025] Laser processing head focal length and spot image feature calibration: The laser processing head moves along the Z-axis of the tool coordinate system. The distance between the laser processing head and the calibration plate and the size of the laser spot on the calibration plate are measured to adjust the position of the laser processing head to meet the focal length of the cladding process. The laser processing head indicator red light spot image features (including shape and area) are captured by an industrial camera and stored. The laser processing head focal length and spot image feature dataset is built by repeating this process multiple times.

[0026] Industrial camera intrinsic parameter calibration and laser processing optical head focal length and spot image feature calibration can be completed in advance, without needing to be repeated on the processing site; and once these two calibrations are completed, they can be reused in subsequent processes without needing to be repeated for each processing.

[0027] II. On-site work preparation: Clean the surface of the parts to be processed to the ready-to-process state; install, connect and debug the processing equipment (including CNC machine tools, laser processing industrial robots, etc.) on site, with the robot base equipped with casters and adjustable support feet; install the laser processing head on the tool end of the industrial robot, install the industrial camera on the side outer wall of the laser processing head, and set up low-angle side lighting on site to achieve diffuse dark field.

[0028] 3. Applying a calibration plate to the surface to be processed: Apply a two-dimensional calibration plate to the surface to be processed of the part. The calibration plate consists of regular squares and marking circles. The size and positional relationship of the two-dimensional squares and marking circles are known, and the deviation of the calibration plate thickness from the visual image calculation is known.

[0029] A calibration board is placed at different positions on the surface to be processed as a comparison reference with surrounding images. Using the known information from the two-dimensional calibration board, when it is converted into machine vision image features, a correspondence between the camera's captured pixels and the actual object's dimensions can be established. For example, if the calibration board is a 10mm square, and the camera captures a square with 5 pixels on each side, then each pixel corresponds to 2mm. When the same focal length is used to capture a line segment with 200 pixels, the actual length of this line segment is 400mm. Similarly, if the surface to be processed is at a 45-degree angle to the processing equipment, based on the principle of perspective, the number of pixels on a calibration board of the same specifications in the same image will differ at different positions. The distance and angle between the surface to be processed and the camera, as well as the dimensions of the surface itself, can be calculated based on these proportional relationships.

[0030] To improve accuracy and perform multi-image processing, calibration plates can also be placed around the corners of the surface to be processed (such as the four corners of a rectangle), and the specifications of the multiple two-dimensional calibration plates are the same.

[0031] IV. Laser processing head tilt angle calibration: The robot reciprocates along the Z-axis of the tool coordinate system. The industrial camera acquires the motion trajectory of the red light spot of the laser processing head relative to the calibration point. The motion trajectory of the laser processing head is calculated to obtain the tilt angle of the laser processing head relative to the surface to be processed. The robot is fed back to correct the tilt angle so that the light output direction of the laser processing head is perpendicular to the surface to be processed.

[0032] like Figure 1 As shown: the solid line represents the state without tilt angle, in which the robot moves up and down along the Z-axis, and the light spot only changes in size while the center remains unchanged; the dashed line represents the state with tilt angle, in which the robot moves up and down along the Z-axis, and the light spot moves laterally along the tilt direction. The direction of the light spot's movement is the tilt direction of the optical axis.

[0033] Tilt angle calibration describes the direction of the laser beam output from the laser processing head, i.e., the path of the laser beam in space. It is one of the parameters of the laser processing head tool coordinate system relative to the robot coordinate system. Ignoring the tilt angle will lead to large calibration errors, resulting in inaccurate focal length during processing and affecting processing quality. Accurate tilt angle calibration ensures that the focal length remains consistent at different turning angles when the laser head is tilted, resulting in consistent processing quality. Figure 2 As shown: the solid line represents the robot's travel range when the laser head is perpendicular to the processing surface, and the dashed line represents the area to be processed. At this time, tilting the laser head in all directions can expand the robot's processing range.

[0034] V. Laser processing head focal length calibration: The laser processing head moves along the Z-axis of the tool coordinate system. The industrial camera captures the image features of the red light spot indicating the laser processing head. The image features are compared with the pre-stored image features of the red light spot indicating the laser head. The robot position is adjusted so that the focus of the laser head is located at the predetermined focal length on the surface to be processed of the part.

[0035] like Figure 3 As shown: the center position is the spot size at the standard focal length. As the light head moves away from the calibration plate, the focal length increases, the spot area becomes larger, and the feedback after reading the image shows it approaching; as the light head moves closer to the calibration plate, the focal length decreases, the spot area becomes larger, and the feedback after reading the image shows it moving away; during operation, it always moves from far to near, and the spot gradually becomes smaller.

[0036] The tilt angle calibration and focal length calibration of the laser processing head determine the information of the processing head installed on the robot's execution end, such as the size and specifications of the processing head itself, the tool center of the processing head, and the spatial positional relationship between the tool coordinate system of the processing head and the robot's world coordinate system.

[0037] VI. Generate processing procedures.

[0038] The advantages of the method of the present invention are: First, during machine vision calibration, the image features of the indicator light spot serve as the focal length reference, eliminating the need for manual determination of the focal length reference and ensuring accurate detection; Second, the tilt angle calibration of the laser processing head further guarantees accurate focal length during processing, thereby improving processing quality.

[0039] Example 2

[0040] When vision is used as an aid, the laser head focal length is used as a reference, and manual point-to-point programming is used, the method of this invention is Method 1, which specifically includes: calibration of the intrinsic parameters of the industrial camera, calibration of the focal length and spot image features of the laser processing laser head—work site preparation—affixing the calibration plate to the surface to be processed—calibration of the tilt angle of the laser processing laser head—visual inspection and calibration of the corner points of the surface to be processed to determine the focal length of the laser processing laser head—generating the processing program.

[0041] The specific steps for visually inspecting and calibrating the focal length of the laser processing head at each corner of the surface to be processed are as follows: A specific corner of the surface to be processed is manually selected for focal length calibration, which is then automatically completed by the equipment. After the focal length calibration of multiple corners is completed, these points are connected to form a closed shape, which is the outline of the surface to be processed. The calibrated corner robot coordinate values ​​are then input into a pre-written program for data replacement, generating the processing program.

[0042] The advantages of this method are: visual inspection is fast and accurate, while manual point positioning is flexible and convenient.

[0043] Example 3

[0044] When vision is the primary factor, machine vision inspection results are the benchmark, and the program is automatically generated, the method of this invention is Method Two, which specifically includes: calibration of the intrinsic parameters of the industrial camera, calibration of the focal length and spot image features of the laser processing head—work site preparation—affixing the calibration plate to the surface to be processed—calibration of the extrinsic parameters of the industrial camera, calibration of the coordinate system of the part—calibration of the tilt angle of the laser processing head—calibration of the focal length of the laser processing head—contour extraction, motion trajectory node extraction, motion trajectory generation—generation of the processing program.

[0045] Compared to Method 1, Method 2 adds "industrial camera extrinsic parameter calibration, part coordinate system calibration" and "part machining contour extraction, motion trajectory node extraction, and motion trajectory generation". It can detect the size and specifications of the surface to be processed and the spatial position information of the processing equipment relative to the measured part. Through data calculation, it automatically generates a processing program. Specifically, the industrial camera captures an image of the part surface and transmits it to an industrial computer. The industrial computer obtains the contour information of the part to be processed through digital image processing algorithms, calculates the spatial relationship between the coordinates of the part contour nodes and the coordinates of the robot processing head tool through coordinate transformation algorithms, automatically plans the cladding processing path trajectory and generates trajectory node coordinate values, generates a motion control program, and controls the laser processing equipment to complete the cladding processing.

[0046] The advantages of Method 2 over Method 1 are: 1. It achieves vision-driven, machine vision-based inspection results, with the machine determining the calibration point instead of manually determining it, thus overcoming the subjective problems inherent in human work and making the obtained motion trajectory more accurate and comprehensive, and the cladding quality more reliable; 2. It automatically completes the generation of the processing program without relying on any human experience, greatly improving work efficiency and achieving automation; 3. The processing program can be obtained directly from the industrial camera installed on the outer wall of the processing head, avoiding the sensitivity of manual inspection and calibration to heights, narrow areas, and dangerous and harsh environments, and has good accessibility.

[0047] Example 4

[0048] In Method 2: the calibration of industrial camera extrinsic parameters and part coordinate system is specifically as follows: the industrial robot takes pictures of the surface to be processed at different distances and angles, extracts the pixel data of each feature point of the calibration board, and calculates the spatial positional relationship between each feature point of the calibration board; based on the spatial relationship between the camera coordinate system and the robot world coordinate system, and based on the pixel standard size specification and the ratio of image pixel size, a part coordinate system relative to the robot world coordinate system is constructed, and the extrinsic parameters of the camera are calculated.

[0049] The calibration of the industrial camera's extrinsic parameters and the calibration of the part's coordinate system determine the coordinate system of the machining head tool and the coordinate system of the part, respectively. The combination of the two can determine the size and specifications of the part and its spatial position and angle relative to the robot (machining head). By establishing a part coordinate system based on the robot's world coordinate system, a spatial relationship reference is provided for subsequent machining. That is, the contour of the surface to be machined can be constructed by extracting the pixel coordinates of the motion trajectory nodes in the image.

[0050] The positions of "Industrial camera extrinsic parameter calibration, part coordinate system calibration" and "laser processing head tilt angle calibration, laser processing head focal length calibration" can be interchanged.

[0051] The processing area outline is a transformation between pixel coordinates and robot processing tool coordinates (through pixel coordinates → camera coordinates → robot world coordinates → robot flange coordinates → robot processing tool coordinates).

[0052] Example 5

[0053] In Method 2, the extraction of part machining contours, extraction of motion trajectory nodes, and generation of motion trajectories are specifically as follows: a calibration board composed of two-dimensional squares is used for macroscopic dimension calibration; distant images of the robot are collected; a rough vertical contour is extracted from the distant view; the shooting area is divided according to the shooting accuracy and shooting size range of the industrial camera; clear images are taken sequentially according to the shooting area; the position of the camera coordinate system relative to the robot world coordinate system is recorded simultaneously; the images are processed to obtain the contour and corner coordinates of the area to be processed; after obtaining the edge of the image contour, the pixel coordinates of the motion trajectory nodes (contour corner points) are converted into tool coordinate system coordinate points under the robot world coordinate system (i.e., motion trajectory node extraction); the area enclosed by multiple corner points is the machining area, i.e., the machining contour of the part, and thus the motion trajectory is generated.

[0054] 1. When the macroscopic dimensions of the part are small, a single image is sufficient to complete data extraction, calculation and meet measurement accuracy, so only a single image is taken; when the macroscopic dimensions of the part are large, and the accuracy of a single image cannot meet the requirements, multiple images need to be taken to extract pixel contour information to form the overall contour information of the part.

[0055] 2. The specific steps of image processing are: input - grayscale conversion - noise reduction - binarization - morphological processing - threshold segmentation - edge extraction - contour feature extraction.

[0056] The process involves: image grayscale conversion, transforming the captured image into a grayscale image set; Gaussian and median noise reduction processing of the grayscale image; and grayscale image binarization: ROI region extraction is performed on the first image, selecting the processing area's contour edge as the ROI region, obtaining the grayscale change threshold, and then statistically analyzing the threshold for each image (using grayscale histograms of the ROI regions, denoting the grayscale values ​​with the highest probability in each row as m0-mn, and those with low probability and grayscale values ​​close to 0 as p0-pn), and finally calculating the median of m and p. g, with it as the threshold; Image morphological processing: erosion / dilation (contraction / expansion) rounding edges, noise elimination, i.e., filling closure, removing burrs; Image thresholding segmentation; Use threshold g to segment the image, segmenting the processing area contour and neighboring areas (grayscale abrupt changes occur at the boundary between the target object and the background, the pixel coordinates of the grayscale abrupt change are determined by threshold comparison, and the edge coordinates are determined) The pixel coordinates of the grayscale abrupt change location (X, Y coordinates of the pixel coordinate system) are the image contour edges; Edge extraction: Subtract the eroded and shrunken image from the original image.

[0057] 3. The specific steps for generating the motion trajectory are as follows: After obtaining the coordinate values ​​of the corner points of the processing area contour through image processing, the contour boundary points are compared according to the program motion rules. If the point exceeds the robot's motion range, the robot pauses and outputs a prompt message, which includes the coordinates of the point that exceeds the range, the direction of the exceedance, and the value of the exceedance. If the point exceeds the range, the robot first calculates whether the tilt of the laser head within a 20° angle covers the exceedance range. If the result still exceeds the range, a prompt result is output, and the processing contour is re-planned and generated according to the robot's reachable range.

[0058] 4. When the area of ​​the part to be processed is large, splicing can be performed; at the same time, areas on the surface to be processed that do not need to be processed can be identified through image processing. In this way, the processing program can achieve the effect of partial processing and partial non-processing by using different motion commands or adding on / off light commands for different block contour edges.

[0059] Example 6

[0060] In Method 2: Generate the processing program: Substitute the pre-set process parameter data, such as the cladding layer spot width, cladding layer thickness, cladding layer overlap size, cladding speed, processing direction, line-by-line offset direction, start point, and end point; the control system automatically calculates and assigns values; logic control instruction information is inserted; the point verification program is executed to generate the cladding processing program.

[0061] The data obtained from each step is fed into a pre-programmed detection program. When the coordinates of the image contour corner points are obtained, a motion program is automatically generated based on pre-set parameters and programming rules, passing through each contour corner point to the end point. At this point, a low-speed trial run is required. During this trial run, the operator checks and judges whether the program has any errors and whether the accuracy meets the requirements, manually intervening to correct any problems found. Once confirmed to be accurate, the processing program is executed to perform the cladding process.

[0062] Example 7

[0063] The industrial robot laser processing device in the method of this invention is an existing device, which includes, but is not limited to: processing software operation and computing system (host computer): (hardware) processor, memory, input / output interface, communication control, display operation terminal + (software) database, computing module; detection and acquisition system (industrial camera): acquisition input device (camera, lens, data transmission), auxiliary marking accessories (calibration plate, magnetic back attraction, standard calibration pattern combining ring and checkerboard), brightness adjustment device (lighting); processing mechanism execution system (industrial robot, special machine tool, powder feeder, automatic regulating valve (automatically adjusts powder feeding and carrier gas according to the output signal of the host computer) and other actuators).

Claims

1. A method for detecting laser cladding points and programming automated equipment, comprising the following steps: Industrial camera intrinsic parameter calibration, laser processing head focal length and spot image feature calibration—work site preparation—affixing calibration board to the surface to be processed—laser processing head tilt angle calibration—laser processing head focal length calibration—generating processing program; wherein: Industrial camera intrinsic parameter calibration: a planar two-dimensional calibration board is placed on the front of the industrial camera. The calibration board consists of regular squares and marker circles. The size and positional relationship of the squares and marker circles are known; adjust the distance and angle between the industrial camera and the calibration board so that the field of view of the industrial camera covers the image of the calibration board and the image of the calibration board is clear and distinct; the industrial camera takes multiple images of the calibration board from different angles and integrates them into an image set a1; use image processing software to process and obtain the camera intrinsic parameter data, including focal length, pixel size, and the number of pixels between the center and origin of the image; laser processing head focal length and spot image feature calibration: the laser processing head along the tool coordinate system Z The laser head is moved along an axis to measure the distance between the laser processing head and the calibration plate, as well as the size of the laser spot on the calibration plate, thereby adjusting the position of the laser processing head to meet the focal length requirements of the cladding process. The laser processing head's red light spot image features are captured and stored using an industrial camera, and this process is repeated multiple times to build a dataset of laser processing head focal length spot image features. On-site preparation includes mounting the laser processing head onto the tool end of the industrial robot, attaching the industrial camera to the side outer wall of the laser processing head, and setting up low-angle side lighting to achieve diffuse dark-field illumination. Laser processing head tilt angle calibration involves the robot moving along the tool end... The laser head reciprocates along the Z-axis of the calibration system. An industrial camera captures the trajectory of the red light spot indicating the laser processing head relative to the calibration point. The trajectory is calculated to obtain the skew angle of the laser processing head relative to the surface to be processed. The robot is then fed back to correct the skew angle so that the laser light output direction is perpendicular to the surface to be processed. Laser head focal length calibration: The laser processing head moves along the Z-axis of the tool coordinate system. The industrial camera captures the image features of the red light spot indicating the laser processing head. This image is compared with the pre-stored image features of the red light spot indicating the laser processing head. The robot position is adjusted so that the focus of the laser head is located at the predetermined focal length on the surface to be processed.

2. The method according to claim 1, characterized in that: When vision is used as an aid, the laser head focal length is used as a reference, and manual point-to-point programming is used, the specific method is as follows: calibration of the intrinsic parameters of the industrial camera, calibration of the focal length and spot image features of the laser processing laser head — preparation of the work site — affixing the calibration plate to the surface to be processed — calibration of the tilt angle of the laser processing laser head — visual inspection and calibration of the focal length of the laser processing laser head at each corner point of the surface to be processed — generation of the processing program.

3. The method according to claim 1, characterized in that: The specific steps for visually inspecting and calibrating the focal length of the laser processing head at each corner of the surface to be processed are as follows: the focal length is calibrated at a corner of the surface to be processed manually, and the calibration is completed automatically by the equipment. After the focal length calibration of multiple corners is completed, these points are connected to form a closed figure, which is the outline of the surface to be processed. The robot coordinate values ​​of the calibrated corners are then input into a pre-written program for data replacement to generate the processing program.

4. The method according to claim 1, characterized in that: When vision is the primary factor, machine vision inspection results are the benchmark, and the program is automatically generated, the specific method is as follows: industrial camera intrinsic parameter calibration, laser processing head focal length and spot image feature calibration—work site preparation—affixing calibration plate to the surface to be processed—industrial camera extrinsic parameter calibration, part coordinate system calibration—laser processing head tilt angle calibration—laser processing head focal length calibration—contour extraction, motion trajectory node extraction, motion trajectory generation—generating the processing program.

5. The method according to claim 3, characterized in that: The calibration of industrial camera extrinsic parameters and part coordinate system involves: an industrial robot taking pictures of the surface to be processed from different distances and angles, extracting pixel data of each feature point on the calibration board, and calculating the spatial positional relationship between each feature point on the calibration board; based on the spatial relationship between the camera coordinate system and the robot world coordinate system, and based on the pixel standard size specification and the ratio of image pixel size, constructing a part coordinate system relative to the robot world coordinate system, and calculating the camera's extrinsic parameters.

6. The method according to claim 3, characterized in that: The specific steps for part machining contour extraction, motion trajectory node extraction, and motion trajectory generation are as follows: A calibration board composed of two-dimensional squares is used for macroscopic dimensional calibration. A distant view image of the robot is acquired, and a rough vertical contour is extracted. The shooting area is divided according to the industrial camera's shooting accuracy and size range. Clear images are taken sequentially for each shooting area, and the camera coordinate system position relative to the robot's world coordinate system is recorded simultaneously. The image is processed to obtain the contour and corner coordinates of the area to be processed. After obtaining the image contour edge, the pixel coordinates of the contour corner points are converted to tool coordinate system coordinates under the robot's world coordinate system, thus extracting the motion trajectory nodes. The area enclosed by multiple corner points is the machining area, i.e., the machining contour of the part, which in turn generates the motion trajectory.

7. The method according to claim 5, characterized in that: The specific steps of image processing are: input – grayscale conversion – noise reduction – binarization – morphological processing – threshold segmentation – edge extraction – contour feature extraction.

Images