A laser powder feeding connection positioning device and positioning method based on feature extraction
Through the laser powder feeding connection positioning device based on feature extraction, the reconstruction analysis and image recognition system are used to analyze the shape and position difference values in real time, and the telescopic positioning parts are controlled for precise positioning, which solves the problem of difficult adjustment of the positioning accuracy of split parts in the existing technology and realizes high-precision connection of split parts.
Patent Information
- Application Number
- CN202310748958.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-25
- Publication Date
- 2025-09-09
- Estimated Expiration
- 2043-06-25
AI Technical Summary
In the existing laser powder feeding connection technology, the positioning accuracy and clamping accuracy of the split parts are difficult to adjust flexibly, resulting in difficulty in ensuring the shape and position errors. Especially in the production of small batches of multiple complex components, the fixed tooling and fixtures have poor versatility and flexibility, and cannot meet the connection shape and position accuracy requirements of complex split parts.
A laser powder feeding connection and positioning device based on feature extraction is adopted. The overhead image of the split part is obtained through the shooting device. The shape and position difference values are analyzed in real time using the reconstruction analysis system and the image recognition processing system. The telescopic positioning parts are controlled for precise positioning to realize the shape and position control of the split part.
It realizes the precise positioning of the split parts, ensures the shape and position accuracy of the connection, adapts to split parts of different shapes, and improves the shape and position accuracy and positioning accuracy of complex split parts after connection.
Smart Images

Figure CN116809971B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of laser powder feeding connection of split parts, and particularly relates to a laser powder feeding connection positioning device and positioning method based on feature extraction. Background Art
[0002] At present, split parts are usually connected using laser powder feeding connection technology. Laser powder feeding connection technology is a derivative of laser melting deposition forming technology. It has the advantages of controllable microstructure of the connection area and minimized heat-affected zone, which solves the problem of thickness limitation of the traditional split part connection method.
[0003] For the laser powder feeding connection technology of split parts, the positioning accuracy and clamping accuracy of the connected split parts are the key factors affecting the final precision of the split parts connection shape and position. In the existing laser powder feeding connection technology, the split parts are usually connected by a "V"-shaped groove, which usually requires the bottom surface of the connected split parts to be flat, and the split parts need to be clamped and positioned as a whole on the horizontal plane. In the existing technology, for split connectors that can be mass-produced, fixed fixtures are usually used to clamp and position the split parts. This has the defects of complex fixed fixtures, high clamping and fixing costs, and long cycles. In addition, the versatility and flexibility of fixed fixtures are poor, and they cannot adapt to the frequency and requirements of small-batch development of multiple complex components. Moreover, once the split parts have shape and position errors before clamping, it is difficult to adjust the position of the split parts after clamping, and it is difficult to flexibly adjust the shape and position errors of the split parts; and using a simple pressure plate to clamp the split parts, especially for split parts with curved surfaces, although its shape and position adjustment is flexible, it is difficult to ensure the connection shape and position accuracy of the split parts, and it is also impossible to ensure the size and shape and position requirements of complex split parts.
[0004] Therefore, in order to address the defects in that the shape and position errors in the above-mentioned existing split parts connection process are difficult to flexibly adjust or the shape and position accuracy is difficult to ensure, the present invention discloses a laser powder feeding connection positioning device and positioning method based on feature extraction. Summary of the Invention
[0005] The purpose of the present invention is to provide a laser powder feeding connection positioning device and positioning method based on feature extraction, which can accurately detect the difference between the actual shape and position of the split part and the standard shape and position, and flexibly adjust the position of the split part according to the difference value, thereby realizing precise control of the shape and position of the split part, and ensuring the shape and position accuracy when the split part is finally connected.
[0006] The present invention is achieved through the following technical solutions:
[0007] A feature extraction-based laser powder feeding connection positioning device is used to position the connection position of laser powder feeding split parts, including a positioning substrate, the top end surface of the positioning substrate is a positioning surface, and a plurality of groups of telescopic positioning members are provided on both sides of the positioning surface, and a positioning space for tightening the split parts is formed between the positioning ends of the telescopic positioning members on both sides; a shooting device for shooting a top-view image of the connection position between the split parts is provided on the top of the positioning space, the shooting device is connected to a reconstruction analysis system, and the reconstruction analysis system is connected to an image recognition processing system; the reconstruction analysis system is used to reversely reconstruct the split part model and the positioning substrate model, and simulate the connection positioning of the split part model on the positioning substrate model to obtain the standard shape and position feature value of the split part model; the shooting device is used to shoot a top-view image of the split part that is actually connected and positioned on the positioning substrate, and the image recognition processing system extracts the actual shape and position feature value of the top-view image of the split part that is actually connected and positioned, and analyzes the shape and position difference value between the actual shape and position feature value and the standard shape and position feature value, and then controls the telescopic positioning member to stretch and contract to compensate for the shape and position difference value to achieve precise positioning of the split part.
[0008] In order to better implement the present invention, further, the telescopic positioning member includes at least two telescopic cylinders arranged in parallel, and the ends of the telescopic rods of the telescopic cylinders are provided with flexible positioning contacts.
[0009] In order to better implement the present invention, it further includes a downward pressing positioning member arranged at the top of the positioning space, and the downward pressing positioning member is used to press down and fix the split member with the positioning accuracy meeting the standard.
[0010] In order to better implement the present invention, further, the downward positioning member includes waist-shaped slide grooves arranged on both sides of the top of the positioning base plate, and the inner sliding of the waist-shaped slide grooves on both sides is provided with clamping bolts, and a pressure plate is arranged between the clamping bolts on both sides, and a nut is threadedly sleeved on the clamping bolt for pressing the pressure plate downward.
[0011] A laser powder feeding connection and positioning method based on feature extraction is implemented based on the above-mentioned laser powder feeding connection and positioning device. A reconstruction analysis system is used to reversely reconstruct and establish a split component model and a positioning substrate model, and the split component model is simulated and connected and positioned on the positioning substrate model to obtain standard shape and position feature values; actual connection and positioning of the split component are performed on the positioning substrate, and a top view image of the actually connected and positioned split component is photographed by a photographing device, and the top view image of the actually connected and positioned split component is analyzed by an image recognition processing system to obtain actual shape and position feature values, and a shape and position difference value between the actual shape and position feature value and the standard shape and position feature value is calculated. The shape and position difference value is used to control the telescopic positioning component to telescope and compensate for the shape and position difference value to achieve precise positioning of the split component.
[0012] In order to better implement the present invention, the following steps are further specifically included:
[0013] Step 1: Use a scanner to collect point cloud data of the split component and the positioning substrate, and the reconstruction analysis system reversely reconstructs the split component model and the positioning substrate model through the point cloud data;
[0014] Step 2: Import the split component model and the positioning base plate model into the modeling software, establish a positioning baseline on the positioning base plate model, and then simulate the connection and positioning of the split component model on the positioning base plate model until the connection line between the split component models coincides with the positioning baseline. Measure the standard shape and position feature values of the split component model based on the contour features of the positioning base plate model;
[0015] Step 3: actually connect and position the split components on the positioning substrate so that the connection lines between the split components coincide with the positioning baselines on the positioning substrate, and then use a camera to capture a top view of the actually connected and positioned split components;
[0016] Step 4. Import the top view image of the actually connected and positioned split part into the image recognition processing system, extract the actual shape and position feature values of the top view image of the actually connected and positioned split part through the image recognition processing system, and compare the actual shape and position feature values with the standard shape and position feature values to obtain shape and position difference values. According to the shape and position difference values, calculate the displacement required for the telescopic positioning part to push the split part to the standard positioning position, and control the telescopic positioning part to expand and contract according to the displacement to compensate for the shape and position difference value to achieve precise positioning of the split part.
[0017] In order to better implement the present invention, further, the step 2 specifically includes:
[0018] Step 2.1, establishing a positioning baseline on the positioning substrate model, wherein the positioning baseline includes a Y-axis positioning line parallel to the telescopic direction of the telescopic positioning member and an X-axis positioning line perpendicular to the Y-axis positioning line;
[0019] Step 2.2: Simulate the connection and positioning of the split component models on the positioning substrate model so that the connection line between the split component models coincides with the Y-axis positioning line, and the midpoint of the connection line between the split component models is located on the X-axis positioning line;
[0020] Step 2.3, collecting top-view images of the split component model and the positioning substrate model, and importing the top-view images into the image recognition processing system;
[0021] Step 2.4: Using the top view image outline of the positioning substrate model as a reference, mark the locations of several telescopic positioning parts on the top view image of the positioning substrate model. Then, use the image recognition processing system to calculate the distance between the locations of the telescopic positioning parts and the top view image outline of the split part model as the standard shape and position feature value.
[0022] In order to better implement the present invention, further, the step 4 specifically includes:
[0023] Step 4.1. Using the top view image outline of the positioning substrate as a reference, mark the locations of several telescopic positioning members on the top view image of the positioning substrate. Then, use the image recognition processing system to extract the actual top view image outline of the actually connected and positioned split components, and calculate the distance between the locations of the telescopic positioning members and the actual top view image outline as the actual shape and position feature value.
[0024] Step 4.2, calculating the difference between the actual shape and position feature value and the standard shape and position feature value as the shape and position difference value;
[0025] Step 4.3: Control the telescopic positioning members to perform a first-stage extension so that the distances between the positioning ends of all the telescopic positioning members and the actual top-view image outline are equal; then control the telescopic positioning members to perform a second-stage synchronous extension so that the positioning ends of all the telescopic positioning members are synchronously aligned with the actual top-view image outline;
[0026] Step 4.4: Control the telescopic positioning member to perform three-stage telescopic expansion and contraction to compensate for the shape and position difference.
[0027] In order to better implement the present invention, further, steps 4.1-4.4 are repeated to compensate for the shape and position difference values multiple times.
[0028] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0029] (1) The present invention reversely reconstructs the split component model and the positioning substrate model, and simulates the connection and positioning of the split component model to the standard position on the positioning substrate model, thereby obtaining the standard shape and position characteristic value of the split component model; then the split component is actually connected and positioned on the positioning substrate, and a top view image of the split component is captured by a shooting device, and the top view image is analyzed in real time by an image recognition processing system to obtain the actual shape and position characteristic value of the split component when it is actually connected and positioned; and the shape and position difference value between the standard shape and position characteristic value and the actual shape and position characteristic value is used to control the expansion and contraction of a plurality of telescopic positioning components in real time to compensate for the shape and position difference value, thereby achieving real-time and flexible regulation of the shape and position of the split component, and effectively ensuring the shape and position accuracy of the connection between the split components after connection;
[0030] (2) The present invention sets a plurality of groups of telescopic positioning members on both sides of the positioning substrate according to the outline of the split parts, and then adjusts the position of the split parts in real time at multiple points according to the shape and position difference values between the positioning ends of the telescopic positioning members at each point and the outline of the split parts. This not only satisfies the shape and position adjustment requirements for split parts of different shapes, but also greatly improves the shape and position accuracy of the connection between complex split parts. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 This is a schematic diagram of the structure of the laser powder feeding connection positioning device;
[0032] Figure 2 It is a schematic diagram of the simulated connection positioning of the split component model on the positioning substrate model;
[0033] Figure 3 A top view image for simulating the connection and positioning of the split component model;
[0034] Figure 4 A top view image for actual connection positioning of split components;
[0035] Figure 5 A schematic diagram of a telescopic positioning member performing a certain extension;
[0036] Figure 6 A schematic diagram of the telescopic positioning member performing two-stage telescopic extension;
[0037] Figure 7 Schematic diagram of the three-stage telescopic positioning component to compensate for the shape and position difference.
[0038] Among them: 1- positioning substrate; 2- telescopic positioning member; 3- pressing positioning member; 4- shooting device; 5- reconstruction analysis system; 6- image recognition and processing system. DETAILED DESCRIPTION
[0039] Example 1:
[0040] A laser powder feeding connection positioning device based on feature extraction, such as Figure 1 As shown, it is used to locate the connection position of the laser powder feeding split parts, including a positioning substrate 1, the top end surface of the positioning substrate 1 is a positioning surface, and a plurality of groups of telescopic positioning members 2 are provided on both sides of the positioning surface, and the positioning ends of the telescopic positioning members 2 on both sides form a positioning space for tightening the split parts; a shooting device 4 for shooting a top view image of the connection position between the split parts is provided on the top of the positioning space, the shooting device 4 is connected to the reconstruction analysis system 5, and the reconstruction analysis system 5 is connected to the image recognition processing system 6; the reconstruction analysis system 5 is used for reverse reconstruction A split component model and a positioning substrate model are prepared, and the split component model is simulated and connected and positioned on the positioning substrate model to obtain the standard shape and position feature values of the split component model; the shooting device 4 is used to shoot the top view image of the split component that is actually connected and positioned on the positioning substrate 1, and the image recognition processing system 6 extracts the actual shape and position feature values of the top view image of the split component that is actually connected and positioned, and analyzes the shape and position difference values between the actual shape and position feature values and the standard shape and position feature values, and then controls the telescopic positioning component 2 to telescope to compensate for the shape and position difference values to achieve precise positioning of the split component.
[0041] The shooting device 4 uses a high-definition industrial camera with a shooting speed of 3s / time. At the same time, an alarm module is provided in the reconstruction analysis system 5. If the reconstruction analysis system 5 detects that the simulation connection positioning accuracy between the split-part model and the positioning substrate model does not meet the standard, an alarm will be issued through the alarm module. At this time, the staff should re-simulate the connection positioning of the split-part model until the simulation connection positioning accuracy of the split-part model meets the standard.
[0042] The split parts to be connected and positioned are placed in the positioning space formed by the positioning ends of the telescopic positioning parts 2 on both sides, and the shape and position difference values between the actual shape and position feature values and the standard shape and position feature values are analyzed by the image recognition processing system 6. Then the image recognition processing system 6 controls the telescopic positioning part 2 to expand and contract according to the shape and position difference values existing at the current point position, and then pushes the split parts located in the positioning space to the standard position. The above-mentioned standard position refers to the position of the split part model after it is connected and positioned on the positioning substrate model with the standard shape and position feature values.
[0043] Example 2:
[0044] This embodiment is improved on the basis of the above embodiment 1. Figure 1 As shown, the telescopic positioning member 2 includes at least two telescopic cylinders arranged in parallel, and the ends of the telescopic rods of the telescopic cylinders are provided with flexible positioning contacts, which are made of rubber material.
[0045] When positioning the plane of the split component, the telescopic positioning component 2 generally includes two telescopic cylinders arranged in parallel. Two points are determined on the plane of the split component by the two telescopic cylinders, and then a plane is determined by the two positionings.
[0046] When positioning a curved surface of a split component, the telescopic positioning member 2 should include at least two parallel telescopic cylinders. These cylinders locate at least two points on the curved surface of the split component, thereby ensuring curvature accuracy during surface positioning. Preferably, for curved surfaces with long spans, as many telescopic cylinders as possible should be installed based on the spatial layout to increase the number of points on the curved surface, thereby improving curvature accuracy during surface positioning.
[0047] The rest of this embodiment is the same as that of embodiment 1, so it will not be described again.
[0048] Example 3:
[0049] This embodiment is improved on the basis of the above embodiment 1 or 2. Figure 1 As shown, the laser powder feeding connection positioning device further includes a downward pressing positioning member 3 arranged at the top of the positioning space, and the downward pressing positioning member 3 is used to press down and fix the split parts that meet the positioning accuracy standards.
[0050] Furthermore, the downward positioning member 3 includes waist-shaped grooves arranged on both sides of the top of the positioning base plate 1, and the inner sliding parts of the waist-shaped grooves on both sides are provided with clamping bolts, and a pressure plate is provided between the clamping bolts on both sides. The threaded sleeve on the clamping bolt is provided with a nut for pressing the pressure plate downward.
[0051] When the split part is positioned, in order to prevent the split part from moving during subsequent processing, the nut on the top of the clamping bolt can be turned to press the pressure plate downward, so that the pressure plate presses the split part downward and fixes the split part. The bottom end of the clamping bolt is connected to the T-slot of the machine tool to achieve positioning and fixation.
[0052] Example 4:
[0053] A laser powder feeding connection and positioning method based on feature extraction is implemented based on the laser powder feeding connection and positioning device described in any one of Examples 1-3, and utilizes a reconstruction analysis system 5 to reversely reconstruct and establish a split component model and a positioning substrate model, and simulates the connection and positioning of the split component model on the positioning substrate model to obtain standard shape and position feature values; the actual connection and positioning of the split component is performed on the positioning substrate 1, and a top view image of the actually connected and positioned split component is captured by a shooting device 4, and the top view image of the actually connected and positioned split component is analyzed by an image recognition processing system 6 to obtain the actual shape and position feature values, and the shape and position difference value between the actual shape and position feature value and the standard shape and position feature value is calculated, and the telescopic positioning component 2 is controlled by the shape and position difference value to perform telescopic compensation for the shape and position difference value to achieve precise positioning of the split component.
[0054] The standard form and position feature values include data such as the contour features of the split component model, the form and position connection relationship between the split component models, and the form and position relationship between the split component model and the positioning substrate model; the actual form and position feature values include data such as the contour features of the split component, the form and position connection relationship between the split component models, and the form and position relationship between the split component and the positioning substrate 1.
[0055] By comparing the shape and position difference values between the standard shape and position characteristic values and the actual shape and position characteristic values, and then compensating the shape and position difference values of the corresponding points on the split part by the telescopic positioning part 2, the error between the actual position of the actual connection positioning of the split part under actual circumstances and the standard position of the simulated connection positioning is effectively reduced, thereby improving the shape and position requirement accuracy and positioning accuracy of the connection positioning of the split part.
[0056] The rest of this embodiment is the same as that of the above-mentioned embodiment 1, and therefore will not be described in detail.
[0057] Example 5:
[0058] This embodiment is improved on the basis of the above-mentioned embodiment 4, and specifically includes the following steps:
[0059] Step 1: Figure 2 As shown, a scanner is used to collect point cloud data of the split component and the positioning substrate 1, and the reconstruction analysis system 5 obtains the split component model and the positioning substrate model through reverse reconstruction of the point cloud data;
[0060] While ensuring reconstruction accuracy and efficiency, the chord height difference method is used to filter noise points in the point cloud data that cannot be visually deleted. The uniform grid method is used to simplify the scattered points in the point cloud data, and the equal distribution density method is used to simplify the gridded points in the point cloud data. When reverse reconstructing split parts with curved surfaces, an approximate fitting algorithm is used to reconstruct the surfaces. The accuracy error between the reverse reconstructed split part model and the actual split part should be within ±0.2mm. The accuracy error between the reverse reconstructed positioning substrate model and the actual positioning substrate 1 should be within ±0.2mm.
[0061] Step 2: Import the split component model and the positioning base plate model into the modeling software, establish a positioning baseline on the positioning base plate model, and then simulate the connection and positioning of the split component model on the positioning base plate model until the connection line between the split component models coincides with the positioning baseline. Measure the standard shape and position feature values of the split component model based on the contour features of the positioning base plate model;
[0062] Use modeling software to import the reconstructed split component model and positioning substrate model. The positioning substrate model is a flat plate model with a thickness of 1 mm and the same positioning baseline as the actual positioning substrate 1. After the model is successfully imported, the simulated connection positioning of the split component model can be performed on the positioning substrate model.
[0063] Specifically:
[0064] The connection line between the split component models is used as the positioning control line, so that the positioning control line coincides with the positioning baseline on the positioning substrate model to ensure that the entire split component model is within the controllable range of connection, positioning, and clamping, and then determine the final placement position of the split component model, that is, the standard position.
[0065] Since the contour features of the positioning substrate model are fixed and have obvious characteristic rectangles, the contour features of the positioning substrate model are used as a benchmark, and measuring tools are used to measure the important shape and position feature values of the split component model, such as curvature, shape and position relationship between the split component models, shape and position relationship between the split component model and the positioning substrate model, and other parameters. The top view image of the split component model performing simulated connection and positioning on the positioning substrate model is extracted, and then the top view data is imported into the image recognition processing system 6 to analyze the standard shape and position feature values of the split component model.
[0066] Step 3: The split components are actually connected and positioned on the positioning substrate 1 so that the connection lines between the split components coincide with the positioning baseline on the positioning substrate 1, and then a top view image of the actually connected and positioned split components is captured using the camera 4;
[0067] Step 4. Import the top view image of the actually connected and positioned split part into the image recognition processing system 6, extract the actual shape and position feature values of the top view image of the actually connected and positioned split part through the image recognition processing system 6, and compare the actual shape and position feature values with the standard shape and position feature values to obtain shape and position difference values, calculate the displacement required for the telescopic positioning part 2 to push the split part to the standard positioning position according to the shape and position difference values, and control the telescopic positioning part 2 to expand and contract according to the displacement to compensate for the shape and position difference value to achieve precise positioning of the split part.
[0068] The components are placed on the positioning substrate 1, with the connecting lines between the components aligned with the positioning baseline. A top-down image of the components and the positioning substrate 1 is captured using the camera 4 positioned directly above the components. Using the contour features of the positioning substrate 1 as a reference, the image recognition and processing system 6 obtains the actual shape and position feature values of the components.
[0069] Furthermore, the step 2 specifically includes:
[0070] Step 2.1, establishing a positioning baseline on the positioning substrate model, wherein the positioning baseline includes a Y-axis positioning line parallel to the telescopic direction of the telescopic positioning member 2 and an X-axis positioning line perpendicular to the Y-axis positioning line;
[0071] Step 2.2: Simulate the connection and positioning of the split component models on the positioning substrate model so that the connection line between the split component models coincides with the Y-axis positioning line, and the midpoint of the connection line between the split component models is located on the X-axis positioning line;
[0072] Specifically:
[0073] The image recognition processing system 6 extracts the position of the connecting line between the separate components in the overhead image and determines the overlap between the connecting line and the Y-axis positioning line. If the connecting line and the Y-axis positioning line do not overlap, an alarm is issued to remind the user to reposition the separate components. During this process, the camera 4 captures images at a rate of 3 seconds per image. It should be noted that the overlap between the connecting line and the Y-axis positioning line means that the overlap error between the connecting line and the Y-axis positioning line is within ±0.3 mm.
[0074] Step 2.3, collecting the top view images of the split component model and the positioning substrate model, and importing the top view images into the image recognition processing system 6;
[0075] Step 2.4: Based on the top-view image contour of the positioning substrate model, mark several points where the telescopic positioning members 2 are located on the top-view image of the positioning substrate model. Then, through the image recognition and processing system 6, calculate the distance value between the points where the telescopic positioning members 2 are located and the top-view image contour of the split part model as the standard geometric feature value.
[0076] As Figure 3 shown, eight points A1, B1, D1, E1, A2, B2, D2, and E2 are marked on the positioning substrate model as the positions of the eight telescopic cylinders. At the same time, points C1 and C2 are marked on the Y-axis central positioning line as reference points to improve the calibration accuracy of the positions of the eight telescopic cylinders. Then, through the image recognition and processing system 6, calculate the distance value between the positions of the telescopic cylinders and the top-view image contour of the split part model, and obtain S A2’ -S E1 and S A2 -S E2 ten values as the standard geometric features. For example, S A1 represents the distance between the positioning end of the telescopic cylinder at point A1 and the contour of the split part model under the simulated connection and positioning condition.
[0077] Further, step 4 specifically includes:
[0078] Step 4.1: Based on the top-view image contour of the positioning substrate 1, mark several points where the telescopic positioning members 2 are located on the top-view image of the positioning substrate 1. Then, through the image recognition and processing system 6, extract the actual top-view image contour of the split part under actual connection and positioning, and calculate the distance value between the points where the telescopic positioning members 2 are located and the actual top-view image contour as the actual geometric feature value;
[0079] As Figure 4 shown, after actually placing the split part on the positioning substrate 1, eight points A1, B1, D1, E1, A2, B2, D2, and E2 are marked on the positioning substrate 1 as the actual positions of the eight telescopic cylinders. At the same time, points C1 and C2 are marked on the Y-axis central positioning line as reference points to improve the calibration accuracy of the positions of the eight telescopic cylinders, and then S A1’ -S E1’ and S A2’ -S E2’ ten values are obtained as the actual geometric feature values. For example, S A1’ represents the distance between the positioning end of the telescopic cylinder at point A1 and the contour of the split part under actual connection and positioning.
[0080] Step 4.2: Calculate the difference between the actual geometric feature value and the standard geometric feature value as the geometric difference value. For example, the geometric difference value of the telescopic cylinder at point A1 is equal to |SA1’ –S A1 |;
[0081] Step 4.3. Control the telescopic positioning member 2 to perform a first-stage telescoping so that the distances between the positioning ends of all the telescopic positioning members 2 and the contour of the actual top-view image are equal. As shown in Figure 5 the figure, after the first-stage telescoping, the distances between the positioning ends of all the telescopic positioning members 2 and the contour of the split member are all equal to S1. As shown in Figure 6 the figure, then control the telescopic positioning member 2 to perform synchronous second-stage telescoping so that the positioning ends of all the telescopic positioning members 2 are synchronously fitted with the contour of the split member, ensuring that the positioning ends of the telescopic positioning members 2 simultaneously contact the contour of the split member and avoiding serious rotational misalignment of the split member caused by a travel difference between the telescopic positioning members 2.
[0082] Further, S1 is preferably 10 mm - 15 mm, and the positioning ends of the telescopic positioning members 2 are telescoped at a speed of 6 mm / s - 10 mm / s.
[0083] Step 4.4. As shown in Figure 7 the figure, control the telescopic positioning member 2 to perform third-stage telescoping to compensate for the form and position difference value. For example, the telescopic cylinder at point A1 extends by a distance of |S A1’ –S A1 |, thereby compensating for the form and position difference value of the telescopic cylinder at point A1. Further, repeat steps 4.1 - 4.4 to compensate for the form and position difference value multiple times.
[0084] The above is only a preferred embodiment of the present invention, and does not impose any form of limitation on the present invention. Any simple modification or equivalent change made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.
Claims
1. A laser powder feeding connection positioning device based on feature extraction, used for positioning the connection position of laser powder feeding split parts, comprising a positioning substrate (1), characterized in that: The top end surface of the positioning substrate (1) is a positioning surface, and a plurality of groups of telescopic positioning members (2) are provided on both sides of the positioning surface, and a positioning space for tightening the split parts is formed between the positioning ends of the telescopic positioning members (2) on both sides; a shooting device (4) for shooting a top view image of the connection position between the split parts is provided on the top of the positioning space, and the shooting device (4) is connected to a reconstruction analysis system (5), and the reconstruction analysis system (5) is connected to an image recognition processing system (6); the reconstruction analysis system (5) is used to reversely reconstruct the split part model and the positioning substrate model, and simulate the connection and positioning of the split part model on the positioning substrate model to obtain the standard shape and position characteristic value of the split part model; The photographing device (4) is used to photograph a top view image of the split component that is actually connected and positioned on the positioning substrate (1); the image recognition processing system (6) extracts the actual shape and position feature values of the top view image of the split component that is actually connected and positioned, analyzes the shape and position difference values between the actual shape and position feature values and the standard shape and position feature values, and then controls the telescopic positioning component (2) to telescope to compensate for the shape and position difference values to achieve accurate positioning of the split component.
2. The laser powder feeding connection positioning device based on feature extraction according to claim 1 is characterized in that: The telescopic positioning member (2) comprises at least two telescopic cylinders arranged in parallel, and the ends of the telescopic rods of the telescopic cylinders are provided with flexible positioning contacts.
3. A laser powder feeding connection positioning device based on feature extraction according to claim 1 or 2, characterized in that: It also includes a downward pressing positioning member (3) arranged at the top of the positioning space, and the downward pressing positioning member (3) is used to press downward and fix the split member with the positioning accuracy meeting the standard.
4. The laser powder feeding connection positioning device based on feature extraction according to claim 3 is characterized in that: The downward-pressing positioning member (3) comprises waist-shaped chutes arranged on both sides of the top of the positioning base plate (1), and clamping bolts are slidably arranged inside the waist-shaped chutes on both sides, and a pressure plate is arranged between the clamping bolts on both sides, and a nut for pressing the pressure plate downward is threadedly sleeved on the clamping bolt.
5. A laser powder feeding connection positioning method based on feature extraction, implemented based on the laser powder feeding connection positioning device based on feature extraction according to any one of claims 1 to 4, characterized in that: Reconstructing the split component model and the positioning substrate model by reverse reconstruction using the reconstruction analysis system (5), and performing simulated connection and positioning of the split component model on the positioning substrate model to obtain standard shape and position characteristic values; The actual connection and positioning of the split part is performed on the positioning substrate (1), and a top view image of the actually connected and positioned split part is photographed by a photographing device (4). The top view image of the actually connected and positioned split part is analyzed by an image recognition processing system (6) to obtain an actual shape and position characteristic value, and a shape and position difference value between the actual shape and position characteristic value and the standard shape and position characteristic value is calculated. The shape and position difference value is used to control the telescopic positioning part (2) to telescope and compensate for the shape and position difference value to achieve accurate positioning of the split part.
6. The laser powder feeding connection positioning method based on feature extraction according to claim 5 is characterized in that: The specific steps include: Step 1: using a scanner to collect point cloud data of the split component and the positioning substrate (1); and a reconstruction analysis system (5) reversely reconstructs the point cloud data to obtain a split component model and a positioning substrate model; Step 2: Import the split component model and the positioning base plate model into the modeling software, establish a positioning baseline on the positioning base plate model, and then simulate the connection and positioning of the split component model on the positioning base plate model until the connection line between the split component models coincides with the positioning baseline. Measure the standard shape and position feature values of the split component model based on the contour features of the positioning base plate model; Step 3: actually connect and position the split parts on the positioning substrate (1) so that the connection lines between the split parts coincide with the positioning baseline on the positioning substrate (1), and then use a camera (4) to capture a top view of the actually connected and positioned split parts; Step 4: Import the top view image of the actually connected and positioned split component into the image recognition processing system (6), extract the actual shape and position feature value of the top view image of the actually connected and positioned split component through the image recognition processing system (6), and compare the actual shape and position feature value with the standard shape and position feature value to obtain the shape and position difference value, calculate the displacement required for the telescopic positioning component (2) to push the split component to the standard positioning position according to the shape and position difference value, and control the telescopic positioning component (2) to expand and contract according to the displacement value to compensate for the shape and position difference value to achieve accurate positioning of the split component.
7. The laser powder feeding connection positioning method based on feature extraction according to claim 6 is characterized in that: The step 2 specifically includes: Step 2.1, establishing a positioning baseline on the positioning substrate model, wherein the positioning baseline includes a Y-axis positioning line parallel to the telescopic direction of the telescopic positioning member (2) and an X-axis positioning line perpendicular to the Y-axis positioning line; Step 2.2: Simulate the connection and positioning of the split component models on the positioning substrate model so that the connection line between the split component models coincides with the Y-axis positioning line, and the midpoint of the connection line between the split component models is located on the X-axis positioning line; Step 2.3, collecting top-view images of the split component model and the positioning substrate model, and importing the top-view images into the image recognition processing system (6); Step 2.4: Using the top view image outline of the positioning substrate model as a reference, mark the positions of several telescopic positioning members (2) on the top view image of the positioning substrate model, and then calculate the distance value between the position of the telescopic positioning member (2) and the top view image outline of the split member model through the image recognition processing system (6) as the standard shape and position feature value.
8. The laser powder feeding connection positioning method based on feature extraction according to claim 6 is characterized in that: The step 4 specifically includes: Step 4.1, using the top view image outline of the positioning substrate (1) as a reference, mark the positions of several telescopic positioning members (2) on the top view image of the positioning substrate (1), then extract the actual top view image outline of the split parts that are actually connected and positioned through the image recognition processing system (6), and calculate the distance value between the position of the telescopic positioning member (2) and the actual top view image outline as the actual shape and position feature value; Step 4.2, calculating the difference between the actual shape and position feature value and the standard shape and position feature value as the shape and position difference value; Step 4.3, controlling the telescopic positioning member (2) to extend and retract for a certain period of time, so that the distances between the positioning ends of all the telescopic positioning members (2) and the actual top view image contour are equal; Step 4.4, controlling the telescopic positioning member (2) to perform two-stage telescopic expansion to compensate for the shape and position difference value.
9. The laser powder feeding connection positioning method based on feature extraction according to claim 8, characterized in that: Repeat steps 4.1-4.4 to compensate for the positional discrepancy multiple times.
Citation Information
Patent Citations
Complex structure blank processing and positioning method based on three-dimensional inverse solution
CN108917600A
Distributed-monocular-vision-based automatic positioning method for connecting intersecting hole system of aircraft parts
CN109373894A