A laser welding positioning system and method for curved surface materials based on machine vision
The machine vision-based curved surface material laser welding positioning system solves the misjudgment and inaccuracy problems of curved surface material laser welding positioning, achieving high-precision and economical welding effects.
Patent Information
- Application Number
- CN202211694247.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-12-28
- Publication Date
- 2025-09-30
- Estimated Expiration
- 2042-12-28
AI Technical Summary
Existing technologies make it difficult to achieve high-precision laser welding positioning on curved materials, especially in the welding of platinum resistor pins and armored cables. There are problems of misjudgment and inaccuracy, and existing high-precision solutions are costly and complex.
A machine vision-based curved surface material laser welding positioning system is used, which includes a conveying device, a vertical moving mechanism, an image acquisition device and a laser processing device. Through image acquisition, preprocessing, blob analysis and calibration conversion, the processing point set of the weld is determined, and the laser processing device is controlled to perform welding.
It achieves high-precision and stable welding positioning, reduces costs, adapts to complex process requirements, and improves positioning accuracy and stability.
Smart Images
Figure CN116197559B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of laser processing, and in particular relates to a laser welding positioning system and method for curved surface materials based on machine vision. Background Art
[0002] With the rapid development of industrial equipment automation, the demand for the combination of machine vision technology and laser welding technology is becoming stronger and stronger. On the one hand, the development of machine vision technology is becoming more and more mature. On the other hand, the application field of laser welding has expanded to fine objects, and the requirements for precision are also more stringent.
[0003] Thin-film platinum resistors, a new generation of temperature measurement and control sensors, are commonly used in various temperature probes. In some high-temperature measurement applications, such as engine exhaust temperature monitoring, the probes consist of platinum resistor pins welded to armored cable. Both the platinum resistor pins and the armored cable core are cylindrical or needle-shaped, with the armored cable core diameter typically approximately four times that of the platinum resistor pins.
[0004] Traditional machine vision positioning methods for electronic components include template matching and edge extraction. While these methods offer high positioning accuracy, they also require high imaging quality. Without high-contrast, stable edge information in the target area, ideal results cannot be achieved. The stability of curved surface imaging is limited by the smoothness of the material and the lighting angle, resulting in unclear welds between curved surfaces and fluctuations in the size and edge shape of the arc-top spot. The pins of the platinum resistor are connected to the armored cable, and the grayscale information in the imaging area corresponding to the junction of the two cylindrical surfaces is unstable. Furthermore, because the pins of the platinum resistor are short and fine, they generally require external mechanical constraints to keep them tightly against the armored cable core, reducing the light source's degrees of freedom and creating significant difficulties with visual lighting. Consequently, positioning results are often misjudged and inaccurate, leading to defects such as processing termination, cold solder joints, and pin burnout.
[0005] Although new solutions with depth information acquisition technology, such as binocular cameras or 3D profile sensors, can achieve precise positioning of non-flat targets, they are economically expensive and have more complex control methods. Summary of the Invention
[0006] The purpose of the present invention is to address the deficiencies in the above-mentioned background technology and to provide a laser welding positioning system and method for curved surface materials based on machine vision.
[0007] The technical solution adopted by the present invention is: a machine vision-based positioning system for laser welding of curved surface materials, comprising a conveyor for placing equipment to be processed, a vertical movement mechanism, and a controller. The vertical movement mechanism is mounted with an image acquisition device and a laser processing device. The conveyor is used to carry the equipment to be processed below the vertical movement mechanism to perform XY plane motion, and the image acquisition device is used to capture images of the equipment to be processed.
[0008] The controller is used to determine the processing point set coordinates of all welding points on the equipment to be processed based on the image of the equipment to be processed, to control the conveying device to transport the equipment to be processed to the corresponding coordinate position below the laser processing device based on the processing point set coordinates, and to control the laser processing device to perform welding actions.
[0009] A method for laser welding positioning of curved surface materials based on machine vision, comprising the following steps:
[0010] Step 1: calibrate the point parameters of the image acquisition device and the laser processing device;
[0011] Step 2: Control the equipment to be processed to move below the image acquisition device, and acquire an image of the equipment to be processed based on the point parameters;
[0012] Step three, preprocessing the collected images;
[0013] Step 4: Analyze the pre-processed image to extract white patches and obtain physical information of the patches;
[0014] Step 5: Determine the processing point set of all welds based on the point parameters and the physical information of the plaque;
[0015] Step six: According to the processing point set, control the conveying device to transport the equipment to be processed to the corresponding coordinate position below the laser processing device, and control the laser processing device to perform welding.
[0016] Furthermore, the calibrated point parameters include:
[0017] According to the parameters of the equipment to be processed and the diameter of the laser spot emitted by the laser processing device, set the offset distance of the laser focus center relative to the center of the equipment to be processed that meets the welding process requirements;
[0018] Calibrate the relative position relationship between the two central focal points of the image acquisition device and the laser processing device in the first coordinate system;
[0019] The conversion matrix from the camera image coordinate system to the physical coordinate system of the image acquisition device is obtained through the nine-point calibration method.
[0020] Furthermore, an image of the equipment to be processed is acquired based on the first coordinate in the first coordinate system, and a coordinate system formed with the planar movement direction of the conveying device and the vertical movement direction of the vertical moving mechanism as a reference is used as the first coordinate system; the incoming axis coordinate when the image acquisition device takes the image is used as the first coordinate.
[0021] Furthermore, the preprocessing of the collected image includes sequentially performing image denoising, filtering and smoothing, adjusting contrast, and binarization.
[0022] Furthermore, the process of obtaining the physical information of the patches is as follows: performing Blob analysis on the preprocessed image, extracting a number of patches formed by the minimum circumscribed rectangle of the equipment to be processed, and determining the physical information of the patches based on the transformation matrix.
[0023] Furthermore, based on the point parameters and the physical information of several patches, the uniformly distributed coordinate point set P2{(x1,y1),(x2,y2),…,(xn,yn)} of all the welding points relative to the center of the patch is determined.
[0024] xn=((L-ΦA)*n / N)-(N+1) / 2N*(L-ΦA)=((2n-N-1)*(L-ΦA)) / 2N;
[0025] y1=y2=…yn=±D1;
[0026] x1, x2, ..., xn are the X-axial distances of the first, second, ..., and nth welding points relative to the center of the plaque, respectively;
[0027] y1, y2, …, yn are the Y-axis distances of the first, second, …, and nth solder joints relative to the center of the patch, respectively;
[0028] L is the physical length of the long axis of the plaque; ΦA is the laser spot diameter, D1 is the center offset distance; N is the number of welds, and n is the weld serial number.
[0029] Furthermore, the set of all welding points P3 in the physical coordinate system is determined based on the uniformly distributed coordinate point set P2: For each patch, the rotation transformation formula is used:
[0030]
[0031] The uniformly distributed coordinate point set P2 is rotated around the origin by an angle θ, and then added to the coordinates of the center point of the plaque circumscribed rectangle to obtain the total weld point set P3, where x is the X value of the coordinate of each point in the uniformly distributed coordinate point set P2, y is the Y value of the coordinate of each point in the uniformly distributed coordinate point set P2, x' is the X value of the coordinate of each point in the total weld point set P3, and y' is the Y value of the coordinate of each point in the total weld point set P3.
[0032] Furthermore, a processing point set P4 of all welding points in the first coordinate system is determined based on the point set P3 of all welding points: the processing point set P4 is obtained by adding the point set P3 of all welding points to the relative position relationship and the first coordinate P1.
[0033] The beneficial effects of the present invention are:
[0034] The present invention realizes nine-point calibration from the camera coordinate system to the transmission device coordinate system based on the nine-point position information preset by the camera and the transmission device. The coordinate value of the visual target in the transmission coordinate system only needs to be simply added to the current motion coordinate. The method is simple and the calculation is convenient.
[0035] The present invention performs Blob analysis on the image, extracts the outline of the target object, and obtains the minimum bounding rectangle. For the minimum bounding rectangle, the physical coordinates of the center point, the physical length of the major axis, and the physical angle of the major axis are obtained using calibration information. This can stably obtain the target center and posture, and can resist interference from some unclean conditions.
[0036] The present invention calculates the physical coordinate value of the target weld point based on the preset theoretical distance value between the weld point and the center of the curved surface using the center coordinates, length, and angle. Then, the actual processing physical coordinate value of the target weld point is calculated based on the preset relative hardware offset between the laser processing center and the camera center. The processing is guided by the controller, which improves the positioning accuracy and stability, adapts to complex process requirements, and is highly operational and economical. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] Figure 1 Schematic diagram of the system structure of the present invention.
[0038] Figure 2 It is a schematic diagram of the equipment to be processed according to the present invention.
[0039] Figure 3 The camera of the present invention collects the original image.
[0040] Figure 4 This is the visual processing diagram of the present invention.
[0041] Figure 5 This is the actual solder joint distribution diagram of the present invention.
[0042] In the figure: 1-camera and lens; 2-light source; 3-vertical movement mechanism; 4-laser processing head; 5-jig; 6-transmission device; 7-platinum resistor pin; 8-armored battery cell; 9-soldering point location. DETAILED DESCRIPTION
[0043] The following is a further description of specific embodiments of the present invention in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is intended to facilitate understanding of the present invention and does not constitute a limitation of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
[0044] like Figure 1 As shown, the present invention provides a laser welding positioning system for curved surface materials based on machine vision, comprising a conveying device 6 for placing the equipment to be processed, a vertical moving mechanism 3 and a controller (not shown in the figure), the vertical moving mechanism is located above the area where the conveying device is located, and an image acquisition device and a laser processing device are installed on the vertical moving mechanism. The conveying device is used to carry the equipment to be processed to perform XY plane movement below the vertical moving mechanism, and the image acquisition device is used to collect images of the equipment to be processed and output image information to the controller; the equipment to be processed is a platinum resistor 7 and an armored battery cell 8, which are placed on a jig 5, as shown in FIG. Figure 2 Position relationship shown.
[0045] The controller has high-speed computing and motion control capabilities. The controller is used to determine the processing point set coordinates of all welding point positions 9 on the equipment to be processed based on the image of the equipment to be processed, and is used to control the conveying device to transport the equipment to be processed to the corresponding coordinate position below the laser processing device based on the processing point set coordinates, and is used to control the laser processing device to perform welding actions.
[0046] In the above scheme, the specific equipment to be processed is mounted on a jig, which is then mounted on a conveyor mechanism. The image acquisition device includes a camera and lens 1, and a light source 2. The camera is preferably an industrial camera, the lens is preferably a telecentric lens, and the light source is preferably a dome light source. The vertical movement mechanism is a mechanical support structure mounted above the conveyor mechanism and capable of moving up and down.
[0047] Based on the above-mentioned curved surface material laser welding positioning system, the present invention also provides a curved surface material laser welding positioning method based on machine vision, comprising the following steps:
[0048] The first step is to calibrate the point parameters of the image acquisition device and the laser processing device, including
[0049] According to the diameter of the platinum resistor pin 7 and the laser spot diameter ΦA emitted by the laser processing device, set the offset distance D1 of the laser focus center relative to the center of the platinum resistor pin to meet the welding process requirements;
[0050] The relative position relationship H(X) between the camera and the lens and the two central focal points of the laser processing head 4 in the first coordinate system is obtained by theoretical calculation or other testing means. h ,Y h); a coordinate system formed based on the planar motion direction of the conveying device and the vertical motion direction of the vertical moving mechanism is used as the first coordinate system C1; after the camera calibration is completed, the center point of the camera is the origin. When shooting, the axis coordinates are input, and the camera center coordinates are P1, which is used as the first coordinate P1 (x1, y1).
[0051] The transformation matrix from the camera image coordinate system Cp to the physical coordinate system C2 is obtained through the nine-point calibration method, which is marked as the calibration matrix TM1. The nine-point calibration method is based on the nine point information preset by the camera and the transmission device to realize the nine-point calibration of the camera coordinate system (i.e., Cp) to the transmission device coordinate system (i.e., C2). Finally, the C2 coordinate system has the same unit as the C1 coordinate system, but the axial direction is opposite.
[0052] Assuming the image pixel size is X*Y and the field of view size is WX*WY, the nine-point calibration process is as follows: 1. Place a feature object on the motion platform, such as a circular feature figure; 2. Move the platform so that the center of the feature object is located at the center of the image, and note that point 5: PT5 (0.5*X, 0.5*Y) has a physical coordinate of (0,0); 3. The physical coordinates of the remaining 8 points are centered on PT5, and 0.25*WX and 0.25*WY are used as two axial steps to generate 8 orientation points: (-0.25*WX, 0.25*WY), (0, 0.25*WY), (0.25*WX 0.25*WY), ...; 4. Relative to point 5, the motion platform moves 8 orientation points, and uses the visual algorithm to obtain the image pixel coordinates of the center of the feature object at the 8 points respectively; 5. Use the pixel coordinates of the nine points and their corresponding physical coordinates to solve the affine transformation matrix.
[0053] The second step is to control the equipment to be processed to move to the bottom of the image acquisition device, and obtain the camera image of the equipment to be processed through the preset coordinates P1 (x1, y1) in the C1 coordinate system, such as Figure 3 shown.
[0054] Step 3: Preprocess the captured image, including smoothing it using Gaussian and median filters, and adjusting contrast to enhance image information. This improves the contrast between the seven cylindrical surfaces of the platinum resistor leads and the eight cylindrical surfaces of the armored battery cells. If the target image has cluttered grayscale information, overexposure can be used to enhance the integrity of the target. The preprocessed image is then binarized, and morphological opening and closing operations are used to further filter out interfering points inside and outside the target surface (i.e., remove spots inside and outside the seven cylindrical surfaces of the platinum resistor leads and the eight cylindrical surfaces of the armored battery cells).
[0055] Step 4: Extract white patches from the above image and perform Blob analysis. After extracting the outline of the target object, obtain the minimum bounding rectangle and set the combined screening conditions [min, max]: patch area, rectangularity, patch short axis width, patch long axis length, and quantity until all extracted patches are all cylindrical bright spots of the platinum resistor pin 7. Through the image calibration matrix TM1, obtain all the physical information of the patches (such as the physical coordinates of the center point, the physical length of the long axis, and the physical angle of the long axis) and sort them in ascending order in the Y direction, such as Figure 4 shown.
[0056] By setting the tolerance value range of the center point physical coordinates, you can filter out abnormal pins in assembly; by setting the tolerance value range of the major axis physical length, you can filter out incomplete pins; by setting the tolerance value range of the major axis physical angle, you can filter out bent pins.
[0057] Step 5: Using the extracted physical information of the patch: the physical length L of the patch's long axis and the center offset distance D1 required by the process, for the target pin patch, calculate the uniformly distributed coordinate point set P2{(x1,y1),(x2,y2),…,(xn,yn)} of the n solder joints relative to the center of the patch required by the process, where
[0058] y1=y2=…yn=±D1;
[0059] xn=((L-ΦA)*n / N)-(N+1) / 2N*(L-ΦA)=((2n-N-1)*(L-ΦA)) / 2N;
[0060] x1, x2, ..., xn are the X-axial distances of the first, second, ..., and nth welding points relative to the center of the plaque, respectively;
[0061] y1, y2, …, yn are the Y-axis distances of the first, second, …, and nth solder joints relative to the center of the patch, respectively;
[0062] L is the physical length of the long axis of the plaque; ΦA is the laser spot diameter, D1 is the center offset distance; N is the number of welds, and n is the weld serial number.
[0063] Step 6: Using the extracted physical information of the patch: the physical coordinates of the center point of the patch's circumscribed rectangle and the angle θ of the patch's long axis. For the target pin patch, use the rotation transformation formula:
[0064]
[0065] First, rotate the point set P2 around the origin by an angle θ, and then add it to the coordinates of the center point of the plaque circumscribed rectangle to obtain the new point set P3. The point set P3 is the point set of all weld points in the C2 coordinate system. x is the X value of the coordinate of each point in the uniformly distributed coordinate point set P2, y is the Y value of the coordinate of each point in the uniformly distributed coordinate point set P2, x' is the X value of the coordinate of each point in the all weld point set P3, and y' is the Y value of the coordinate of each point in the all weld point set P3.
[0066] Step 7: Given the relative position relationship H(Xh, Yh) and coordinates P1(x1, y1) of the two focal points of the camera lens 1 and the laser processing head 4 in the C1 coordinate system, add the point set P3 to H(Xh, Yh) and P1(x1, y1) to obtain the processing point set P4.
[0067] P4=P3+H(Xh,Yh)+P1(x1,y1);
[0068] That is the actual processing point set of all solder joints of the target pin in the C1 coordinate system, such as Figure 5 shown.
[0069] Step 8: Repeat steps 5 to 7 for the remaining pins, and calculate the actual processing point set of all solder joints in sequence.
[0070] Step 9: Based on the processing point set, the controller guides the laser processing head to perform welding action.
[0071] The above are only specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered by the scope of protection of the present invention. Matters not described in detail in this specification belong to the prior art known to those skilled in the art.
Claims
1. A method for positioning laser welding of curved surface materials implemented by a laser welding positioning system for curved surface materials based on machine vision, wherein the laser welding positioning system for curved surface materials comprises a conveying device for placing a device to be processed, a vertical moving mechanism, and a controller, wherein an image acquisition device and a laser processing device are installed on the vertical moving mechanism, the conveying device is used to carry the device to be processed and perform XY plane movement under the vertical moving mechanism, and the image acquisition device is used to capture an image of the device to be processed; the controller is used to determine the processing point set coordinates of all welding points on the device to be processed based on the image of the device to be processed, and is used to control the conveying device to transport the device to be processed to the corresponding coordinate position under the laser processing device based on the processing point set coordinates, and is used to control the laser processing device to perform welding action, characterized in that: The method comprises the following steps: Step 1: calibrate the point parameters of the image acquisition device and the laser processing device; Step 2: Control the equipment to be processed to move below the image acquisition device, and acquire an image of the equipment to be processed based on the point parameters; Step three, preprocessing the collected images; Step 4: Analyze the pre-processed image to extract white patches and obtain physical information of the patches; Step 5: Determine the processing point set of all welds based on the point parameters and the physical information of the plaque; Step 6: According to the processing point set, control the conveying device to transport the device to be processed to the corresponding coordinate position below the laser processing device, and control the laser processing device to perform welding; Based on the point parameters and the physical information of several patches, the uniformly distributed coordinate point set P2{(x1,y1),(x2,y2),…,(xn,yn)} of all the weld points relative to the center of the patch is determined. Among them, xn=((L-ΦA)*n / N)-(N+1) / 2N*(L-ΦA)=((2n-N-1)*(L-ΦA)) / 2N; y1=y2=…yn=±D1; x1, x2, ..., xn are the X-axial distances of the first, second, ..., and nth welding points relative to the center of the plaque, respectively; y1, y2, …, yn are the Y-axis distances of the first, second, …, and nth solder joints relative to the center of the patch, respectively; L is the physical length of the long axis of the plaque; ΦA is the diameter of the laser spot, D1 is the center offset distance; N is the number of welds, and n is the weld serial number; Based on the uniformly distributed coordinate point set P2, determine the set of all welding points P3 in the physical coordinate system: for each patch, use the rotation transformation formula: ； The uniformly distributed coordinate point set P2 is rotated around the origin by an angle θ, and then added to the coordinates of the center point of the plaque circumscribed rectangle to obtain the total weld point set P3. x is the X value of the coordinate of each point in the uniformly distributed coordinate point set P2, y is the Y value of the coordinate of each point in the uniformly distributed coordinate point set P2, x' is the X value of the coordinate of each point in the total weld point set P3, and y' is the Y value of the coordinate of each point in the total weld point set P3. Determine the processing point set P4 of all welding points in the first coordinate system based on the all welding point set P3: add the all welding point set P3, the relative position relationship, and the first coordinate P1 to obtain the processing point set P4.
2. The method for laser welding and positioning of curved surface materials according to claim 1, characterized in that: The calibrated point parameters include: According to the parameters of the equipment to be processed and the diameter of the laser spot emitted by the laser processing device, set the offset distance of the laser focus center relative to the center of the equipment to be processed that meets the welding process requirements; Calibrate the relative position relationship between the two central focal points of the image acquisition device and the laser processing device in the first coordinate system; The conversion matrix from the camera image coordinate system to the physical coordinate system of the image acquisition device is obtained through the nine-point calibration method.
3. The method for laser welding and positioning of curved surface materials according to claim 2, characterized in that: An image of the equipment to be processed is acquired based on the first coordinate in the first coordinate system, and a coordinate system formed by taking the planar movement direction of the conveying device and the vertical movement direction of the vertical moving mechanism as a reference is used as the first coordinate system; the input axis coordinate when the image acquisition device takes the image is used as the first coordinate.
4. The method for laser welding and positioning of curved surface materials according to claim 1, characterized in that: The preprocessing of the collected images includes image denoising, filtering and smoothing, contrast adjustment and binarization processing in sequence.
5. The method for laser welding and positioning of curved surface materials according to claim 1, characterized in that: The process of obtaining the physical information of the patches is as follows: performing Blob analysis on the preprocessed image, extracting several patches formed by the minimum circumscribed rectangle of the equipment to be processed, and determining the physical information of several patches based on the transformation matrix.