Method for measuring the pose of a machining-slm hybrid manufacturing joint based on point features

Abstract

The present application belongs to the technical field of composite manufacturing joint surface pose measurement, solves the problem that the existing measurement method can only satisfy the matching forming of parts of specific size and shape, and does not have universality. A machining-SLM composite manufacturing joint surface pose measurement method based on point features is provided: intermediate mark points and peripheral mark points are arranged at the base plate; images are collected, edge extraction of the intermediate mark points and the peripheral mark points and calculation of the centroid pixel coordinates are performed, the camera external parameter and the world coordinates of the peripheral mark points are solved by using the PNP method of pose solution; the part to be measured is placed on the replaced base plate, images are collected, edge extraction of the peripheral mark points and the joint surface is performed, and the camera external parameter at this time is calibrated; the centroid pixel coordinates of the joint surface are calculated, and the pose of the joint surface of the part to be measured is solved by using the PNP method of pose solution combined with the camera external parameter. The present application can solve the pose of the machining part base placed on the base plate by machine vision.

Description

Technical Field

[0001] This invention belongs to the field of composite manufacturing interface pose measurement technology, specifically relating to a method for measuring the pose of machining-SLM composite manufacturing interface based on point features. Background Technology

[0002] Selective laser melting (SLM) is an advanced laser additive manufacturing technology developed based on prototyping technology. It is often used to form complex thin-walled structures, internal flow channels, ordered porous structures, and other fine structures. It is increasingly being used to manufacture metal parts in the aerospace, automotive and other fields.

[0003] For turbine blade parts, the upper blade structure cannot be manufactured using conventional machining. For the base without a complex structure at the bottom of the part, the high cost of additive manufacturing makes the overall printing cost of the part even higher. Before SLM forming, a 3D model of the part to be formed needs to be input into the software, and the coordinates of the 3D model are also determined on the substrate. However, when the composite forming process is carried out, the pose of the machined part placed on the substrate is unknown. This requires measuring the pose of the machined part. This pose measurement on the plane can usually be simplified to the translation coordinates (X,Y) of the object and the rotation angle γ around the Z-axis.

[0004] Currently, most methods for measuring the pose of composite forming mating surfaces on the market involve designing specific positioning base plates and fixture systems to fix the position of the parts. However, this method has many requirements and limitations on the size and shape of the parts, and can only meet the matching and forming of parts of specific sizes and shapes, lacking versatility. When the demand for this specific part is small, the specific positioning base plate and fixture system will increase the cost. At the same time, this positioning method will also result in low overall forming accuracy due to the installation error of the base plate. Summary of the Invention

[0005] In order to solve at least one of the above-mentioned technical problems in the prior art, the present invention provides a method for measuring the pose of the machining-SLM composite manufacturing interface based on point features.

[0006] This invention employs the following technical solution: a method for measuring the pose of a machining-SLM composite manufacturing interface based on point features, comprising the following steps: S1: Arranging several intermediate mark points on the SLM substrate and several surrounding mark points around the substrate; S2: Calibrating the monocular camera to obtain its intrinsic parameters, and acquiring images of the substrate with the intermediate and surrounding mark points arranged; S3: Extracting the edges of the intermediate mark points and calculating the centroid pixel coordinates, and extracting the edges of the surrounding mark points and calculating the centroid pixel coordinates; S4: Using the known world coordinates, pixel coordinates, and camera intrinsic parameters of the intermediate mark points, obtaining the pose measurement method for the current position of the substrate. Camera extrinsic parameters; Based on the camera intrinsic parameters, the current camera extrinsic parameters, and the pixel coordinates of the surrounding mark points, calculate the world coordinates of the surrounding mark points to complete the positioning of the machine's working coordinate system; S5: Replace the substrate without the middle mark point, place the part to be tested on the substrate, acquire images of the mating surface to be tested and the surrounding mark points before processing, extract the edges of the surrounding mark points and the mating surface, calculate the centroid pixel coordinates of the surrounding mark points, and then calibrate the current camera extrinsic parameters by combining the known world coordinates of the surrounding mark points and the camera intrinsic parameters; S6: Calculate the centroid pixel coordinates of the mating surface, and then solve the pose of the mating surface of the part to be tested by combining the camera intrinsic parameters and the camera extrinsic parameters in step S5.

[0007] Preferably, in step S1, the groove formed by sintering the middle mark point on the substrate by a laser beam is milled out on the outer periphery of the substrate to accommodate the surrounding mark points. The surrounding mark points are thin sheet structures and are embedded in the grooves and are in the same plane as the end face of the substrate.

[0008] Preferably, there are 9 central mark points arranged in a 3x3 matrix, with a standard circular shape, including one circle of a different radius to determine the direction; there are 4-9 surrounding mark points distributed around the perimeter of the substrate, and the material of the surrounding mark points is aluminum oxide.

[0009] Preferably, the image acquired in step S2 is subjected to smoothing filtering; in step S3, the sub-pixel edges are obtained by coarsely locating the edges of the middle mark points using the Canny algorithm, and a certain curvature threshold is set for filtering the coarsely located arc edges. A ray is drawn through the center of the middle mark point and intersects with the filtered arc edges to obtain multiple sub-pixel edge points. The centroid pixel coordinates of each middle mark point are obtained by fitting a circle using the least squares method; the complete edges of the surrounding mark points are directly extracted using the Canny algorithm.

[0010] Preferably, in step S5, the substrate is adjusted to ensure that the part to be tested and the surrounding mark points are on the same plane, the sub-pixel edges of the surrounding mark points and the plane to be tested are extracted using the Canny algorithm, and the pixel coordinates of the surrounding mark points are extracted using the least squares centroid extraction algorithm.

[0011] Preferably, in step S6, the shape of the bonding surface is a regular pattern of a single connected domain, and the pixel coordinates of the centroid of the bonding surface are extracted using a centroid extraction algorithm based on least squares contour.

[0012] Preferably, in step S6, the shape of the bonding surface is an irregular pattern of a single connected domain, and the pixel coordinates of the centroid of the bonding surface are extracted using a centroid extraction algorithm based on feature point detection and descriptor matching or based on edge detection and shape approximation.

[0013] Preferably, in step S6, the shape of the mating surface is a multi-connected mating surface. Depending on the actual processing situation, it is necessary to calculate the centroid of the multiple connected domains of the mating surface of the part as a whole or to calculate the centroid coordinates on each individual connected domain.

[0014] Compared with the prior art, the beneficial effects of the present invention are:

[0015] This invention proposes a point feature-based method for measuring the pose of a machining-SLM composite manufacturing interface. For a machined part base arbitrarily placed on a substrate, its pose is solved using the point feature (PNP) method in machine vision. The solved parameters are then input into an SLM device for composite molding. This invention does not impose requirements or limitations on the size and shape of the formed part, provided the substrate size allows. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the present invention;

[0018] Figure 2 This is a diagram showing the positional relationship between the substrate, the central mark point, and the surrounding mark points of the present invention.

[0019] Figure 3 This is a schematic diagram of the edge extraction of the middle mark point in this invention.

[0020] Figure 4 This is a diagram showing the positional relationship between the substrate, the component under test, and the surrounding mark points of the present invention.

[0021] Figure 5 This is a flowchart of the coordinate system positioning error evaluation process of the present invention;

[0022] Figure 6 This is a shape diagram of the mating surface to be tested (a regular pattern of a single connected domain) according to the present invention;

[0023] Figure 7 This is a shape diagram of the mating surface to be tested (an irregular pattern of a single connected domain) according to the present invention;

[0024] Figure 8 This is a shape diagram of the mating surface to be tested (multi-passage mating surface) of the present invention;

[0025] Figure 9 This is a diagram of the measurement system of the present invention.

[0026] In the figure: 1-substrate; 2-center mark point; 3-surround mark points; 4-part under test; 5-regular pattern of single connected domain; 6-irregular pattern of single connected domain; 7-multiple connected domain; 8-camera tripod; 9-notebook tray; 10-light source bracket; 11-camera; 12-quadrilateral light source. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described 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 implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] It should be noted that the structures, proportions, sizes, etc., shown in the accompanying drawings of this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportional relationships, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should fall within the scope of the technical content disclosed in the present invention. It should be noted that in this specification, relational terms such as "first" and "second" are only used to distinguish one entity from several other entities, and do not necessarily require or imply any actual relationship or order between these entities.

[0029] This invention provides an embodiment:

[0030] like Figures 1 to 8 As shown, a method for measuring the pose of a machining-SLM composite manufacturing interface based on point features includes the following steps:

[0031] S1: Arrange several central mark points on the SLM substrate and several surrounding mark points around the substrate;

[0032] S2: The monocular camera is calibrated using Zhang Zhengyou's calibration method to obtain the intrinsic parameters and distortion parameters of the monocular camera, and images of the central and surrounding mark points are acquired.

[0033] S3: Perform edge extraction and centroid pixel coordinate calculation for the middle mark point, and perform edge extraction and centroid pixel coordinate calculation for the surrounding mark points;

[0034] S4: Calculate the camera extrinsic parameters at this time using the known world coordinates (instrument input), pixel coordinates, and camera intrinsic parameters of the intermediate mark point; calculate the world coordinates of the surrounding mark points based on the camera intrinsic parameters, the current camera extrinsic parameters, and the pixel coordinates of the surrounding mark points to complete the positioning of the machine's working coordinate system.

[0035] S5: Replace the substrate without the middle mark point, place the part to be tested on the substrate, and before processing, collect images of the mating surface and the four marks. Extract the edges of the four marks and the mating surface, calculate the centroid pixel coordinates of the four marks, and then combine the known world coordinates of the four marks and the camera intrinsic parameters to calibrate the camera extrinsic parameters at this time.

[0036] S6: Calculate the centroid pixel coordinates of the mating surface, and then combine the camera intrinsic parameters and the camera extrinsic parameters in step S5 to solve the pose of the mating surface of the part under test, including the translation coordinates (X,Y) and the rotation angle γ around the Z axis.

[0037] In this embodiment, the imaging system includes a camera tripod, a CCD camera, and four strip light sources. The four strip light sources are fixed on a specially designed light source suspension. The angle of each individual light source is adjusted to achieve the final illumination conditions. The CCD camera takes pictures to complete the image acquisition process. The camera interface is connected to a computer, where the acquired image information can be displayed and processed. Due to the limited internal space of the SLM equipment's molding chamber, taking the EasyAdd 3D EP-150 equipment as an example, the camera's working space is approximately 200mm vertically from the substrate, with a substrate size D = 150mm. Including the surrounding mark points, the camera's field of view also needs to be controlled to approximately 200mm × 200mm. Furthermore, while SLM equipment often has its own light source, this light source only provides simple illumination for observation and is not suitable for precise mark point positioning. The four strip light sources, composed of four freely combined strips of light, allow for free adjustment of the angle of each strip to achieve different effects. They are highly versatile and can be adjusted according to the requirements of the object being measured, making them particularly effective for mark point positioning. Based on this research, a wide-angle imaging principle and professional lighting image acquisition scheme were developed, providing effective raw image data for the accurate extraction of composite manufacturing interface based on machine vision.

[0038] The method for accurately acquiring the working coordinate system of an SLM machine based on point features involves the design and layout of preset marker points. The model of the marker matrix is ​​input into the SLM device, and relevant process parameters are set. The high-energy laser beam of the SLM device is used to directly ablate the substrate to form the central marker matrix. The central marker point is a standard circle, with one circle of a different radius used to determine the orientation, ensuring a one-to-one correspondence between the input world coordinates and pixel coordinates of the central marker point. This provides a reliable basis for subsequent camera extrinsic parameter calibration and the calculation of the world coordinates of the surrounding marker points. Four to nine grooves are milled around the substrate, and appropriately sized alumina ceramic sheets are embedded in these grooves to complete the layout of the surrounding marker points. The surrounding marker points are thin sheet structures embedded in the grooves and are coplanar with the end faces of the substrate. Each SLM device only needs to calibrate the world coordinates of the surrounding marker points once using the central marker point. When laying out the surrounding marker points, it is necessary to ensure that they are coplanar with the substrate plane and that the connection with the device is reliable and interference-free. Alumina ceramic material is chosen because of its wear resistance, insensitivity to changes in lighting, and good contrast.

[0039] When acquiring images, ensure the machine is stopped. The image acquired in step S2 is smoothed using a Gaussian filter. In step S3, since the edge formation quality of the middle mark point is poor, the Canny algorithm is used to coarsely locate the edge of the middle mark point to obtain sub-pixel edges. A certain curvature threshold is set for the coarsely located arc edges for filtering. A ray is drawn through the center of the middle mark point and intersects with the filtered arc edges to obtain multiple sub-pixel edge points. The centroid pixel coordinates of each middle mark point are obtained by fitting a circle using the least squares method. The edge quality of the surrounding mark points is better, and the Canny algorithm is used to directly extract their complete edges.

[0040] Point feature-based pose estimation methods are essentially PNP problems in pose measurement. This problem refers to calculating the camera pose (i.e., the camera's position and orientation in 3D space) given the 3D point coordinates of the object model and the 2D point coordinates captured by the camera, thereby achieving the measurement of the object's pose. Generally, more than three points are needed to ensure the accuracy of the pose solution of the relevant algorithms.

[0041] In this patent, the pixel coordinates p of the intermediate mark point are used. i =[u i ,v i ,] T (i = 1, 2, ..., n), and their corresponding world coordinates (i.e., machine working coordinates) P i =[X i ,Y i Z i ,] T To solve the pose problem for (i = 1, 2, ..., n), we first normalize each world coordinate point to eliminate scale factors and better estimate the pose.

[0042]

[0043] For each normalized world coordinate point P′ i and the corresponding image coordinates p i Establish projection equations

[0044]

[0045] Where f is the focal length of the camera, c x and c y These are the principal point coordinates of the image plane, and the camera rotation matrix. Rearranging the projection equations into matrix form, we obtain the following equations:

[0046]

[0047] This simplifies to p = AP′, where p is a vector containing all image coordinates, A is a matrix containing all projection matrices, and P′ is a vector containing all normalized world coordinates. Singular value decomposition of matrix A yields the solution vector p′.

[0048]

[0049] The camera's rotation matrix R and translation vector t can be extracted from the camera's intrinsic parameter matrix and solution vector p′, thus obtaining the camera's pose (extrinsic parameters).

[0050] Furthermore, to further improve the accuracy of pose estimation, minimizing the reprojection error is introduced to optimize pose estimation. This involves using point P with known world coordinates... i Based on the current pose estimation, the image plane is projected to obtain the reprojected coordinate points. Calculate reprojection coordinates Compared with the actual observed image coordinates p i The error ε between i .

[0051]

[0052] The objective function is minimized using the numerical optimization method least squares optimization to update the pose estimate; the reprojection calculation, error calculation and optimization process are repeated until the error is minimized.

[0053] The middle mark point on the substrate is designated as the first mark point, and the surrounding mark points are designated as the second mark points. The camera extrinsic parameter A is calculated using the world coordinates A and pixel coordinates of the first mark point. The world coordinates of the second mark point are then solved using the camera extrinsic parameter A and the pixel coordinates of the second mark point, thus completing the positioning of the coordinate system.

[0054] To ensure positioning accuracy, the camera is recalibrated using the pixel coordinates of the second marker point and the world coordinates to obtain the camera's extrinsic parameter B. Combined with the pixel coordinates of the first marker point, the world coordinates B of the first marker point are calculated. The difference between the world coordinates A and the world coordinates B can be used to determine the pose calculation error.

[0055] In step S5, the substrate is adjusted to ensure that the part under test and the surrounding mark points are on the same plane. The Canny algorithm is used to extract the sub-pixel edges of the surrounding mark points and the plane under test, and the least squares centroid extraction algorithm is used to extract the pixel coordinates of the surrounding mark points.

[0056] For the phenomenon that the shape of the interface may have different connectivity characteristics, and may be composed of single or multiple connected regions, and the geometric features of the contour plane are diverse, and may be composed of regular or irregular shapes, a corresponding solution scheme is designed.

[0057] The shape of the bonding surface is a regular pattern of a simply connected domain. The pixel coordinates of the centroid of the bonding surface are extracted using a centroid extraction algorithm based on least squares contours.

[0058] The shape of the bonding surface is an irregular pattern of a simply connected domain. A centroid extraction algorithm based on feature point detection and descriptor matching or based on edge detection and shape approximation is adopted.

[0059] The shape of the mating surface is a multi-connected mating surface. Depending on the actual processing conditions, it is necessary to calculate the centroid of the multiple connected domains of the mating surface of the part as a whole or to calculate the centroid coordinates of each individual connected domain.

[0060] The above-listed cases allow for the selection of different centroid extraction methods based on the characteristics of the mating surfaces of the actual machined parts, providing a basis for their pose calculation. After completing these tasks, composite forming processing can be carried out.

[0061] This invention has no requirements or limitations on the size and shape of the formed parts, provided the substrate size allows, and thus has wide applicability. The method utilizes an imaging system with a tripod as a platform; during implementation, the imaging system simply needs to be inserted into the SLM forming chamber, and removed after pose measurement. For machined parts of different shapes placed on the same substrate, this method can quickly complete the pose measurement of the mating surfaces.

[0062] This patent requires only one imaging system. The structure of the imaging system does not need to be too complex; it only needs to support the camera and the imaging system. The field of view of the camera is determined by the size of the substrate. For a substrate of a fixed size, only one suitable camera is needed to complete the pose measurement of all composite parts formed on the substrate, resulting in low cost.

[0063] This patented pose measurement method is essentially a visual measurement method. Visual measurement offers advantages such as simple system structure, ease of movement, rapid and convenient data acquisition, convenient operation, low measurement cost, and the potential for online, real-time 3D measurement. Furthermore, machine vision systems can simultaneously measure multiple dimensions, enabling rapid completion of measurement tasks and making them suitable for online measurement with higher efficiency. Measuring minute dimensions is also a strength of machine vision systems; they can utilize high-magnification lenses to magnify the measured object, achieving measurement accuracy below the micrometer level. With appropriate camera and lens selection, this measurement method can achieve a measurement accuracy below 0.1 mm.

[0064] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for measuring the pose of a machining-SLM composite manufacturing interface based on point features, characterized in that, Includes the following steps: S1: Arrange several central mark points on the SLM substrate and several surrounding mark points around the substrate; S2: Calibrate the monocular camera, obtain the intrinsic parameters of the monocular camera, and acquire images of the substrate with the center mark point and the surrounding mark points set up. S3: Perform edge extraction and centroid pixel coordinate calculation for the middle mark point, and perform edge extraction and centroid pixel coordinate calculation for the surrounding mark points; S4: Using the known world coordinates, pixel coordinates, and camera intrinsic parameters of the intermediate mark point, obtain the camera extrinsic parameters at this time; based on the camera intrinsic parameters, the current camera extrinsic parameters, and the pixel coordinates of the surrounding mark points, obtain the world coordinates of the surrounding mark points to complete the positioning of the machine's working coordinate system. S5: Replace the substrate without the middle mark point, place the part to be tested on the substrate, and before processing, collect images of the mating surface and the four marks. Extract the edges of the four marks and the mating surface, calculate the centroid pixel coordinates of the four marks, and then combine the known world coordinates of the four marks and the camera intrinsic parameters to calibrate the camera extrinsic parameters at this time. S6: Calculate the centroid pixel coordinates of the mating surface, and then combine the camera intrinsic parameters and the camera extrinsic parameters in step S5 to solve for the pose of the mating surface of the part to be tested.

2. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 1, characterized in that: In step S1, the groove formed by the laser beam sintering the middle mark point on the substrate is milled into a number of grooves on the outer periphery of the substrate to accommodate the surrounding mark points. The surrounding mark points are thin sheet structures and are embedded in the grooves and are in the same plane as the end face of the substrate.

3. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 2, characterized in that: There are 9 central mark points arranged in a 3x3 matrix, with a standard circular shape. One of the circles has a different radius to determine the direction. There are 4-9 mark points around the perimeter of the substrate, and the material of the perimeter mark points is aluminum oxide.

4. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 3, characterized in that: The image acquired in step S2 is smoothed and filtered. In step S3, the Canny algorithm is used to coarsely locate the edges of the middle mark points to obtain sub-pixel edges. A certain curvature threshold is set for the coarsely located arc edges for filtering. A ray is drawn through the center of the middle mark point and intersects with the filtered arc edges to obtain multiple sub-pixel edge points. The centroid pixel coordinates of each middle mark point are obtained by fitting a circle using the least squares method. The complete edges of the surrounding mark points are directly extracted using the Canny algorithm.

5. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 4, characterized in that: In step S5, the substrate is adjusted to ensure that the part under test and the surrounding mark points are on the same plane. The Canny algorithm is used to extract the sub-pixel edges of the surrounding mark points and the plane under test, and the least squares centroid extraction algorithm is used to extract the pixel coordinates of the surrounding mark points.

6. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 5, characterized in that: In step S6, the shape of the bonding surface is a regular pattern of a simply connected domain, and the pixel coordinates of the centroid of the bonding surface are extracted using a centroid extraction algorithm based on least squares contour.

7. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 5, characterized in that: In step S6, the shape of the bonding surface is an irregular pattern of a single connected domain. The pixel coordinates of the centroid of the bonding surface are extracted using a centroid extraction algorithm based on feature point detection and descriptor matching or based on edge detection and shape approximation.

8. The method for measuring the pose of the machining-SLM composite manufacturing interface based on point features according to claim 5, characterized in that: In step S6, the shape of the mating surface is a multi-connected mating surface. Depending on the actual processing situation, it is necessary to calculate the centroid of the multiple connected domains of the mating surface of the part as a whole or to calculate the centroid coordinates on each individual connected domain.

Images