Groove cutting method, device and system

Abstract

The invention provides a groove cutting method, device and system. The groove cutting method comprises the steps that S1, laser triangulation measurement data and flight time data of a target workpiece at multiple visual angles are obtained; s2, acquiring mixed resolution point cloud data of the target workpiece based on the laser triangulation data and the flight time data; s3, acquiring visual information data of the target workpiece, and performing real-time alignment on the target workpiece based on the mixed resolution point cloud data and the visual information data; s4, after the target workpiece is aligned in real time, a three-dimensional model of the target workpiece is reconstructed based on the mixed resolution point cloud data; and S5, cutting demand information is obtained, a cutting track is planned based on the cutting demand information and the three-dimensional model, and groove cutting is conducted on the target workpiece based on the cutting track and the three-dimensional model. According to the groove cutting method, the groove cutting efficiency can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of metal processing, and in particular to a groove cutting method, device and system. Background Art

[0002] Traditional bevel cutting technology mainly relies on manual operation or automated equipment based on two-dimensional vision, and has problems such as low positioning accuracy and poor adaptability. Especially when processing complex curved surfaces and special-shaped workpieces, two-dimensional vision is difficult to accurately capture three-dimensional spatial features, resulting in large errors in cutting path planning and the need for repeated manual calibration, which seriously affects production efficiency and processing quality. Although the existing cutting technology based on 3D vision can obtain three-dimensional information of the workpiece, the efficiency of processing three-dimensional point cloud data is insufficient in actual applications, and it is difficult to achieve real-time modeling under complex working conditions; at the same time, the visual positioning stability during the dynamic cutting process is easily affected by factors such as ambient light interference and metal reflection; in addition, the existing system has weak adaptive adjustment capabilities for multiple types of bevel parameters (such as processing angles and blunt edge sizes), and lacks a collaborative optimization mechanism with cutting equipment. Therefore, there is an urgent need for a bevel cutting method that can meet the requirements of high precision, high efficiency and high adaptability. Summary of the Invention

[0003] In view of this, the present invention provides a bevel cutting method, device and system, which can improve the accuracy, efficiency and adaptability of bevel cutting.

[0004] To solve at least one of the above technical problems, the present invention adopts the following technical solutions:

[0005] A first aspect of the present invention provides a bevel cutting method, comprising:

[0006] Step S1, obtaining laser triangulation data and time-of-flight data of a target workpiece at multiple viewing angles;

[0007] Step S2: acquiring mixed-resolution point cloud data of the target workpiece based on the laser triangulation data and the time-of-flight data, wherein the mixed-resolution point cloud data includes high-precision local point cloud data of the target workpiece from multiple perspectives and large-scale coarse-grained point cloud data;

[0008] Step S3: acquiring visual information data of the target workpiece, the visual information data including surface image texture information of the target workpiece, and performing real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data;

[0009] Step S4: after real-time alignment of the target workpiece, reconstructing a three-dimensional model of the target workpiece based on the mixed-resolution point cloud data;

[0010] Step S5: obtaining cutting requirement information, planning a cutting trajectory based on the cutting requirement information and the three-dimensional model, and performing bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

[0011] In one embodiment of the present invention, the method further comprises:

[0012] Preprocessing the laser triangulation data and the time-of-flight data respectively to obtain preprocessed laser triangulation data and time-of-flight data, wherein the preprocessing includes one or more of composite filtering, noise reduction, image enhancement, and edge detection;

[0013] In step S2, mixed-resolution point cloud data of the target workpiece is acquired based on the pre-processed laser triangulation data and time-of-flight data.

[0014] In one embodiment of the present invention, step S2 includes:

[0015] Step S21: obtaining high-precision local point cloud data based on laser triangulation data, and obtaining large-scale coarse-grained point cloud data based on time-of-flight data;

[0016] Step S22 : performing a primary registration on the high-precision local point cloud data from multiple perspectives and the large-scale coarse-grained point cloud data from multiple perspectives to form mixed-resolution point cloud data.

[0017] In one embodiment of the present invention, step S22 includes:

[0018] Identify missing areas in mixed-resolution point cloud data based on point cloud topology;

[0019] When it is determined that there is a missing area, the missing area is completed based on a completion algorithm, where the completion algorithm includes one or more of a surface fitting algorithm, a weighted interpolation algorithm, or a curvature-guided completion algorithm.

[0020] In one embodiment of the present invention, step S3 includes:

[0021] Step S31: extracting multimodal features based on the mixed-resolution point cloud data and the visual information data, where the multimodal features include image features and point cloud features;

[0022] Step S32: Based on the multimodal features, the target workpiece is aligned in real time using a hierarchical matching strategy.

[0023] In one embodiment of the present invention, step S32 includes:

[0024] Step S321: Based on multimodal features, the target workpiece is roughly registered by using curvature feature fast matching;

[0025] Step S322: After the rough registration, the target workpiece is finely registered using a dynamic weight factor based on the multimodal features;

[0026] Step S323: Acquire the real-time pose information of the target workpiece, and perform real-time alignment on the target workpiece based on the real-time pose information and the precise registration result.

[0027] In one embodiment of the present invention, step S4 includes:

[0028] Based on the mixed-resolution point cloud data, a reconstruction algorithm is used to perform three-dimensional reconstruction on the aligned target workpiece to obtain a three-dimensional model of the target workpiece. The reconstruction algorithm is a mesh reconstruction algorithm or a surface reconstruction algorithm.

[0029] In one embodiment of the present invention, step S5 includes:

[0030] Step S51: acquiring a processing area on the target workpiece and geometric information of the processing area based on the cutting requirement information and the three-dimensional model of the target workpiece;

[0031] Step S52: planning a path for the processing area based on the cutting requirement information and the geometric information of the processing area to obtain a cutting trajectory;

[0032] Step S53: performing bevel cutting on the processing area of the target workpiece based on the cutting trajectory and the geometric information of the processing area.

[0033] In one embodiment of the present invention, step S53 includes:

[0034] Step S531: setting the position parameters and process parameters of the cutting equipment based on the cutting trajectory, the geometric information of the processing area and the cutting requirement information;

[0035] Step S532: Obtain cutting feedback information during the bevel cutting process, and dynamically adjust the posture parameters and process parameters of the cutting equipment based on the cutting feedback information.

[0036] In a second aspect, the present invention provides a bevel cutting device, the bevel cutting device comprising:

[0037] A data acquisition module is used to acquire laser triangulation data, flight time data, and visual information data of a target workpiece from multiple viewing angles, wherein the visual information data includes surface image texture information of the target workpiece;

[0038] A point cloud data generation module is used to generate mixed-resolution point cloud data of the target workpiece based on the laser triangulation data and the time-of-flight data. The mixed-resolution point cloud data includes high-precision local point cloud data of the target workpiece from multiple perspectives and large-scale coarse-grained point cloud data;

[0039] Alignment module, which is used to align the target workpiece in real time based on mixed-resolution point cloud data and visual information data;

[0040] A 3D model reconstruction module is used to reconstruct a 3D model of a target workpiece based on mixed-resolution point cloud data after real-time alignment of the target workpiece;

[0041] The cutting module is used to plan the cutting trajectory based on the cutting requirement information and the three-dimensional model, and to perform bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

[0042] In a third aspect, the present invention provides a bevel cutting system, comprising:

[0043] 3D scanner, which is used to obtain laser triangulation data and time-of-flight data of the target workpiece from multiple perspectives;

[0044] A three-dimensional camera is used to obtain visual information data of a target workpiece, wherein the visual information data includes surface image texture information of the target workpiece;

[0045] a processor, the processor being electrically connected to the 3D scanner and the 3D camera, respectively, and being configured to acquire mixed-resolution point cloud data of the target workpiece based on the laser triangulation data and the time-of-flight data, the mixed-resolution point cloud data including high-precision local point cloud data and large-scale coarse-grained point cloud data from multiple perspectives of the target workpiece, the processor being further configured to perform real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data, and after the real-time alignment of the target workpiece, reconstruct a 3D model of the target workpiece based on the mixed-resolution point cloud data, and the processor being further configured to acquire cutting requirement information and plan a cutting trajectory based on the cutting requirement information and the 3D model;

[0046] The cutting device is electrically connected to the processor and is used to perform bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

[0047] In one embodiment of the present invention, the bevel cutting system further includes: a filtering device, the filtering device being electrically connected to the 3D scanner and the processor, respectively, the filtering device being configured to preprocess the laser triangulation data and the time-of-flight data to obtain preprocessed laser triangulation data and the time-of-flight data, the preprocessing including one or more of composite filtering, noise reduction, image enhancement, and edge detection;

[0048] The processor is used to obtain mixed-resolution point cloud data of the target workpiece based on the pre-processed laser triangulation data and the time-of-flight data.

[0049] The above technical solution of the present invention has at least one of the following beneficial effects:

[0050] The bevel cutting method of the present invention generates mixed-resolution point cloud data by acquiring laser triangulation data and time-of-flight data from multiple viewpoints of a target workpiece, and also acquires visual information data of the target workpiece. This allows for real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data, improving registration accuracy. A three-dimensional model of the target workpiece can then be reconstructed based on the mixed-resolution point cloud data. A cutting trajectory can then be planned based on the cutting requirement information and the three-dimensional model, and the target workpiece can be beveled based on the cutting trajectory and the three-dimensional model. This improves the accuracy, efficiency, and adaptability of bevel cutting. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic structural diagram of a bevel cutting system according to an embodiment of the present invention;

[0052] Figure 2 Flowchart of a groove cutting method according to one embodiment of the present invention;

[0053] Figure 3 A flowchart of obtaining mixed-resolution point cloud data of a target workpiece in one embodiment of the present invention;

[0054] Figure 4 A flowchart of real-time alignment of a target workpiece according to an embodiment of the present invention;

[0055] Figure 5 A flowchart of performing bevel cutting on a target workpiece based on a cutting trajectory and a three-dimensional model in one embodiment of the present invention;

[0056] Figure 6 Schematic diagram of the structure of a bevel cutting device in one embodiment of the present invention. DETAILED DESCRIPTION

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0058] Figure 1 FIG. 1 shows a schematic structural diagram of a groove cutting system provided by an embodiment of the present invention. Figure 1As shown, the bevel cutting system 100 of an embodiment of the present invention may include: a three-dimensional scanner 110, a three-dimensional camera 120, a processor 130 and a cutting device 140. The processor 130 may be directly or indirectly electrically connected to the three-dimensional scanner 110, the three-dimensional camera 120 and the cutting device 140 through wired or wireless communication, and the embodiment of the present invention does not impose any restrictions on this.

[0059] The 3D scanner 110 can acquire laser triangulation data and time-of-flight data from multiple perspectives of the target workpiece. The 3D scanner 110 is also equipped with a narrowband filter and a polarization modulator to suppress and eliminate metallic reflections from the target workpiece's surface, thereby improving the data quality of the laser triangulation and time-of-flight data. The processor 130 can acquire mixed-resolution point cloud data of the target workpiece based on the laser triangulation and time-of-flight data. The mixed-resolution point cloud data includes high-precision local point cloud data from multiple perspectives and large-scale coarse-grained point cloud data. The 3D camera 120 can acquire visual information data of the target workpiece, including surface image and texture information. The processor 130 can perform real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data to improve registration accuracy. After real-time alignment of the target workpiece, the processor 130 can reconstruct a 3D model of the target workpiece based on the mixed-resolution point cloud data and plan a cutting trajectory based on the cutting requirement information and the 3D model. After the processor 130 has planned the cutting trajectory, the cutting device 140 can perform bevel cutting on the target workpiece based on the cutting trajectory and the 3D model. As a result, the accuracy, efficiency and adaptability of bevel cutting are improved, wherein the cutting device 140 may include a laser cutting device and a knife cutting device.

[0060] In one embodiment of the present invention, the bevel cutting system further includes a filtering device 150 electrically connected to the 3D scanner 110 and the processor 130. The filtering device 150 can preprocess the laser triangulation data and the time-of-flight data, thereby enabling the processor 130 to obtain the preprocessed laser triangulation data and the time-of-flight data and to acquire mixed-resolution point cloud data of the target workpiece based on the preprocessed laser triangulation data and the time-of-flight data. The preprocessing includes one or more of composite filtering, noise reduction, image enhancement, and edge detection. This provides the processor 130 with accurate workpiece information of the target workpiece, thereby improving registration speed and accuracy.

[0061] Reference Manual Figure 2 , which shows the process of the groove cutting method provided by an embodiment of the present invention, which can be applied to Figure 1 The bevel cutting system 100 is specifically as follows Figure 2 As shown, the method may include the following steps:

[0062] Step S1: Acquire laser triangulation data and time-of-flight data of a target workpiece at multiple viewing angles.

[0063] In this embodiment, the 3D scanner can simultaneously perform laser triangulation and time-of-flight measurements on a target workpiece from multiple viewpoints to obtain laser triangulation and time-of-flight data for the target workpiece. By synchronously collecting laser triangulation and time-of-flight data from multiple viewpoints of the target workpiece, the integrity and robustness of the point cloud data can be improved.

[0064] In other embodiments of the present invention, the 3D scanner is also equipped with a narrowband filter and a polarization modulator to suppress and eliminate metallic reflections from the target workpiece surface, thereby further improving the data quality of the laser triangulation and time-of-flight data. Furthermore, filtering devices can be used to preprocess the laser triangulation and time-of-flight data, respectively, to obtain preprocessed laser triangulation and time-of-flight data. The preprocessing can include one or more of composite filtering, noise reduction, image enhancement, and edge detection. This can remove environmental noise such as dust and light reflections, further improving the data quality of the laser triangulation and time-of-flight data.

[0065] Step S2: Acquire mixed-resolution point cloud data of the target workpiece based on the laser triangulation data and the time-of-flight data.

[0066] In this embodiment, a processor can be used to obtain mixed-resolution point cloud data of the target workpiece based on the pre-processed laser triangulation data and time-of-flight data. The laser triangulation data is the high-precision local point cloud data of the target workpiece, and the time-of-flight data is the large-scale coarse-grained point cloud data of the target workpiece. In other words, the mixed-resolution point cloud data can include high-precision local point cloud data and large-scale coarse-grained point cloud data from multiple perspectives of the target workpiece. In this way, an overall point cloud model of the target workpiece can be formed. Specifically, Figure 3 As shown, step S2 may include:

[0067] Step S21 : obtaining high-precision local point cloud data based on the laser triangulation data, and obtaining large-scale coarse-grained point cloud data based on the time-of-flight data.

[0068] Step S22 : performing a primary registration on the high-precision local point cloud data from multiple perspectives and the large-scale coarse-grained point cloud data from multiple perspectives to form mixed-resolution point cloud data.

[0069] In this embodiment, methods such as feature point matching or edge correspondence can be used to input high-precision local point cloud data from multiple perspectives and large-scale coarse-grained point cloud data from multiple perspectives into the same coordinate system. By calculating the rigid transformation relationship between multiple perspectives, initial alignment is performed to form mixed-resolution point cloud data, thereby obtaining the overall structural characteristics of the workpiece while retaining high-precision detail information.

[0070] Furthermore, missing regions can be identified in mixed-resolution point cloud data based on the point cloud topology. If missing regions are identified, they are completed using a completion algorithm, including one or more of a surface fitting algorithm, a weighted interpolation algorithm, or a curvature-guided completion algorithm. This completion algorithm improves the integrity and continuity of the point cloud, further optimizing the spatial continuity and geometric integrity of the mixed-resolution point cloud data.

[0071] Step S3: Acquire visual information data of the target workpiece, and perform real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data.

[0072] In this embodiment, while using a 3D scanner to acquire laser triangulation data and time-of-flight data, a 3D camera can also be used to acquire visual information data of the target workpiece. The visual information data may include surface image texture information of the target workpiece. The target workpiece can be aligned in real time based on the mixed-resolution point cloud data and the visual information data. Specifically, Figure 4 As shown, step S3 includes:

[0073] Step S31: extracting multimodal features based on the mixed-resolution point cloud data and the visual information data, where the multimodal features include image features and point cloud features.

[0074] In this embodiment, a visual-point cloud fusion network model including a cascaded feature extraction module can be first constructed, and the geometric structure information in the mixed-resolution point cloud data of the target workpiece and the surface image texture information in the visual information data can be processed by the visual-point cloud fusion network model to obtain the multimodal features of the target workpiece including image features and point cloud features. Among them, the image features can be used to extract two-dimensional information such as color, texture, and edge of the target workpiece through the deep convolutional neural network (such as ResNet, HRNet, etc.) of the visual-point cloud fusion network model. Point cloud features can be extracted through a three-dimensional point cloud network (such as PointNet++, KPConv, SparseConvNet, etc.). By extracting multimodal features, the recognition robustness and accuracy of complex grooves, small structures or occluded areas can be improved in the subsequent alignment process.

[0075] Step S32: Based on the multimodal features, the target workpiece is aligned in real time using a hierarchical matching strategy.

[0076] Specifically, step S32 may include:

[0077] Step S321 : Based on the multimodal features, the target workpiece is roughly registered by using curvature feature fast matching.

[0078] In this embodiment, a coarse registration can be performed based on the curvature features of the point cloud features in the multimodal features to determine the initial pose of the target workpiece. This can reduce computational complexity, achieve rapid screening and coarse positioning, and narrow the search space for subsequent fine registration.

[0079] Step S322: After the rough registration is performed, the target workpiece is finely registered using dynamic weight factors based on the multimodal features.

[0080] In this embodiment, a dynamic weight factor can be introduced into the Iterative Closest Point (ICP) algorithm based on multimodal features. During the point cloud registration process, weights are applied to different point pairs, and adaptive adjustments are made based on regional feature significance, geometric stability, or matching credibility. That is, lower weights are assigned to point pairs with larger errors, and higher weights are assigned to point pairs with clear edges, obvious features, and high matching credibility. This allows for precise registration of the target workpiece, which can accelerate convergence speed while improving registration accuracy. Specifically, the optimization goal of the Iterative Closest Point algorithm is:

[0081]

[0082] Among them, P i is the source point cloud after coarse registration, Q i is the point in the desired target point cloud, R is the rotation matrix, and t is the translation vector.

[0083] Step S323: Acquire the real-time pose information of the target workpiece, and perform real-time alignment on the target workpiece based on the real-time pose information and the precise registration result.

[0084] In this embodiment, a calibration target can be installed at a feature point on the target workpiece, and a three-dimensional scanner or a three-dimensional camera can be used to observe the calibration target in real time to obtain real-time posture information of the target workpiece. Then, a posture compensation matrix can be constructed based on the deviation between the real-time posture information and the previously obtained precise alignment result to perform error correction on the posture of the target workpiece, thereby achieving real-time alignment of the target workpiece and completing closed-loop self-correction of the error.

[0085] Step S4: After the target workpiece is aligned in real time, a three-dimensional model of the target workpiece is reconstructed based on the mixed-resolution point cloud data.

[0086] In this embodiment, a reconstruction algorithm can be used to perform three-dimensional reconstruction on the aligned target workpiece based on mixed-resolution point cloud data to obtain a three-dimensional model of the target workpiece with spatial topological continuity and surface smoothness, providing an input basis for subsequent path planning and cutting control. Specifically, the reconstruction algorithm can be a mesh reconstruction algorithm or a surface reconstruction algorithm, wherein the mesh reconstruction algorithm can include algorithms such as Poisson Reconstruction and Triangulated Surface Reconstruction. By using the mesh reconstruction algorithm, discrete points in the point cloud can be connected to construct a closed triangular facet mesh so that the three-dimensional model has spatial topological continuity. The surface reconstruction algorithm can include algorithms such as surface fitting algorithms based on ordered point clouds, implicit function methods (such as reconstruction based on Signed Distance Function), etc., so that the three-dimensional model can restore the surface of the smooth or complex curved surface of the target workpiece. Therefore, reconstructing the three-dimensional model of the target workpiece based on mixed-resolution point cloud data can facilitate processing area extraction and cutting trajectory planning, thereby improving the accuracy and efficiency of groove cutting.

[0087] Step S5: obtaining cutting requirement information, planning a cutting trajectory based on the cutting requirement information and the three-dimensional model, and performing bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

[0088] In this embodiment, the cutting path of the target workpiece can be automatically planned based on the actual bevel cutting requirement information and the three-dimensional model of the target workpiece, and the cutting device can be controlled to perform the bevel cutting operation. Figure 5 As shown, step S5 may include:

[0089] Step S51 : acquiring a processing area on the target workpiece and geometric information of the processing area based on the cutting requirement information and the three-dimensional model of the target workpiece.

[0090] In this embodiment, geometric features of the 3D model surface can be identified in conjunction with cutting requirement information. For example, methods such as edge extraction, plane fitting, and weld location can be used to identify the processing area. Furthermore, geometric information of the processing area can be extracted from the 3D model, such as its surface normal vector, spatial contour, local curvature, and normal continuity, providing basic data for subsequent path planning.

[0091] Step S52: performing path planning on the processing area based on the cutting requirement information and the geometric information of the processing area to obtain a cutting trajectory.

[0092] In this embodiment, when the cutting device is a tool cutting device, in order to reduce the cutting time and wear on the cutting tool of the cutting device, path planning can be performed based on the shortest path. The key cutting points extracted from the processing area are constructed as nodes in the graph structure. The Euclidean distance between the paths is used as the edge weight. The shortest path is obtained by solving the objective function, which is:

[0093]

[0094] Among them, P i and P i+1 Of course, path planning can also be performed using sampling path generation or multi-objective optimization strategies based on actual cutting requirements, such as processing continuity, path smoothness, equipment motion constraints, and other factors, which are not limited by the present invention.

[0095] Furthermore, in order to improve the overall cutting efficiency and path quality, multiple optimization weight factors and execution constraints can be introduced during path planning to generate the optimal cutting trajectory. Specifically, multiple cutting efficiency weight coefficients can be introduced during path planning, and optimization can be performed with the goal of shortest processing time and strongest path continuity. During the optimization process, the posture change between the path segments in the cutting trajectory (such as the pitch and yaw angle changes of the cutting equipment tool) and the path curvature are used as weight factors. By adjusting the various cutting efficiency weight coefficients in the cutting efficiency weight function, the paths are preferentially sorted, and the cutting paths with shorter processing time and higher path continuity are given priority, where the cutting efficiency weight function J eff for:

[0096] J eff =α*T cut +β*∑|△θ|+γ*∑|k|

[0097] Among them, T cut is the theoretical cutting time, α is the cutting efficiency weighting coefficient corresponding to the theoretical cutting time, ∑|△θ| is the posture change, β is the cutting efficiency weighting coefficient corresponding to the posture change, ∑|k| is the path curvature change, and γ is the cutting efficiency weighting coefficient corresponding to the posture change. For two paths covering the same processing area, the cutting efficiency weighting function can be calculated to prioritize the path with smaller posture changes and fewer path turns. This reduces switching time and non-cutting motions, thereby increasing the effective processing volume per unit time. This reduces non-cutting motion time, improving overall operation time and cutting efficiency.

[0098] Furthermore, a stress distribution map of the target workpiece surface can be pre-constructed based on the geometric information of the target workpiece processing area and the material information of the target workpiece. High-stress areas can be marked on the stress distribution map, and areas with sharp attitude changes can be marked on the stress distribution map based on the cutting requirements. The processor can assign lower weights to high-stress areas or areas with sharp attitude changes, thereby reducing their priority in path planning. Then, based on the pre-acquired tool wear coefficient, the processor can avoid frequent cutting into high-stress areas or areas with sharp attitude changes in the target workpiece when the current cutting equipment tool wear is high, thereby reducing additional tool wear and extending the tool life cycle.

[0099] In addition, since in the actual scenario of bevel cutting, there are often fixtures, support structures, locators or other interferences around the target workpiece, in order to ensure the stability and safety of the cutting process, obstacle avoidance constraints can be further introduced during path planning. Specifically, sensors can be used to pre-collect environmental information of the target workpiece processing area, and a spatial obstacle voxel map can be constructed based on the environmental information. In the voxel map, the spatial voxels occupied by fixtures, support structures, adjacent workpieces, locators, the equipment itself or other interferences are marked as prohibited areas, thereby setting constraint boundaries for the cutting equipment. During the path planning process, the processor can dynamically plan the pitch angle, yaw angle and other posture parameters of the cutting head of the cutting equipment to avoid the prohibited areas where fixtures, support structures, adjacent workpieces, the cutting equipment itself or other interferences are located, ensuring that the cutting trajectory is executed in a reachable and safe space. Therefore, by introducing multiple optimization weight factors and execution constraints, the stability of the bevel cutting process and the availability of the equipment are effectively improved.

[0100] Step S53: performing bevel cutting on the processing area of the target workpiece based on the cutting trajectory and the geometric information of the processing area.

[0101] In one embodiment of the present invention, step S53 may include:

[0102] Step S531: setting the position parameters and process parameters of the cutting equipment based on the cutting trajectory, the geometric information of the processing area and the cutting requirement information.

[0103] In this embodiment, the parameters of the cutting equipment can first be configured based on the cutting trajectory generated in the early stage and the geometric information of the target workpiece processing area to ensure that the cutting process can be executed accurately. Specifically, the posture parameters of the cutting equipment can be first obtained based on the spatial distribution of the cutting path points on the cutting trajectory and the changes in the geometric information of the processing area, such as the position, inclination, normal orientation and other information of the cutting head of the cutting tool of the cutting equipment in space, to ensure that the cutting angle of the equipment always maintains a reasonable angle with the groove surface, thereby obtaining a high-quality groove section. Then, the process parameters during cutting, such as the power, cutting speed, auxiliary gas pressure and other parameters of the cutting equipment, can be set based on the cutting requirement information. The process parameter configuration during cutting can be dynamically optimized according to the material thickness, groove width and precision requirements in the cutting requirement information, so as to achieve a reasonable match between energy density and melting efficiency, ensure that the cross-section cutting is uniform, burr-free and slag-free, and provide stable boundary conditions for subsequent welding or assembly processes.

[0104] Step S532: Obtain cutting feedback information during the bevel cutting process, and dynamically adjust the posture parameters and process parameters of the cutting equipment based on the cutting feedback information.

[0105] In this embodiment, a variety of sensors can be set on the cutting equipment, such as vibration sensors or temperature sensors. By setting up a variety of sensors, cutting feedback information of the cutting equipment or the target workpiece during the bevel cutting process can be obtained, and the posture parameters and process parameters of the cutting equipment can be dynamically adjusted based on the cutting feedback information.

[0106] Specifically, the vibration sensor on the cutting equipment can capture the vibration signal of the interaction area between the cutting equipment and the target workpiece in real time. The processor can perform Fourier transform on the vibration signal to extract the main frequency component and amplitude of the vibration signal, and then obtain the vibration offset data of the target workpiece based on the main frequency component and amplitude. At the same time, the temperature sensor can detect the temperature of the target workpiece processing area in real time. When the temperature sensor detects that the temperature of the target workpiece processing area has risen, the processor can calculate the thermal deformation trend displacement of the target workpiece processing area based on the thermal deformation coefficient of the target workpiece material that has been pre-entered. The sensor can then use the pre-established dynamic path compensation model to vector superimpose the obtained vibration offset data and thermal deformation trend displacement in the same spatial coordinate system to obtain the correction value of the posture parameter, and then fine-tune the original planned path according to the correction value, while correcting the error and maintaining the geometric consistency of the cutting trajectory and the processing area.

[0107] The cutting equipment is also equipped with a photoelectric sensor and an edge camera, which can assess factors such as penetration of the target workpiece cut point, edge cleanliness, and smoothness. Specifically, the photoelectric sensor can detect the depth of the cutting kerf in real time. If the kerf depth is less than a predetermined depth within a predetermined time, the processor can determine that the cutting point penetration is insufficient and, in turn, increase the power of the cutting equipment or slow down the cutting speed. If the kerf depth is about to exceed the predetermined depth within a predetermined time, the processor can determine that overpenetration or overburning is imminent and, in turn, reduce the power of the cutting equipment or speed up the cutting speed. This balance between energy density and cutting rate is achieved during the processing process, avoiding substandard cutting results or problems such as overburning, overpenetration, and blackened cut edges caused by heat accumulation or excessive power. Furthermore, the edge camera on the cutting equipment can scan the edge of the cutting kerf. Based on the scanned image of the cutting kerf edge, the processor can use an edge detection operator to determine the contour fluctuation amplitude of the cutting kerf edge and the reflective consistency amplitude of the cutting kerf edge through pixel brightness variance. Specifically, tiny burrs or irregular jitters on the edge of the cutting seam will cause the contour fluctuation amplitude of the cutting seam edge to fluctuate, and when the edge roughness exceeds the set threshold, it will cause the reflection consistency amplitude of the cutting seam edge to fluctuate. At this time, the processor can adjust the laser focal length or tool cutting angle of the cutting equipment to improve the quality of the cutting seam. Therefore, by dynamically adjusting the posture parameters and process parameters of the cutting equipment through a closed loop, not only the cutting quality and process stability of the bevel cutting are improved, but also the service life of key components on the cutting equipment such as lasers and nozzles is effectively extended.

[0108] In summary, the bevel cutting method of the present invention generates mixed-resolution point cloud data by acquiring laser triangulation data and time-of-flight data from multiple viewpoints of a target workpiece, and also acquires visual information data of the target workpiece. This allows for real-time alignment of the target workpiece based on the mixed-resolution point cloud data and visual information data, improving registration accuracy. A three-dimensional model of the target workpiece can then be reconstructed based on the mixed-resolution point cloud data, a cutting trajectory can be planned based on cutting requirement information and the three-dimensional model, and bevel cutting can be performed on the target workpiece based on the cutting trajectory and the three-dimensional model. This improves the accuracy, efficiency, and adaptability of bevel cutting.

[0109] Thirdly, as Figure 6 As shown, the present invention provides a bevel cutting device, which includes:

[0110] A data acquisition module 610 is used to acquire laser triangulation data, time-of-flight data, and visual information data of a target workpiece from multiple viewing angles, wherein the visual information data includes surface image texture information of the target workpiece;

[0111] A point cloud data generation module 620 is configured to generate mixed-resolution point cloud data of the target workpiece based on the laser triangulation data and the time-of-flight data. The mixed-resolution point cloud data includes high-precision local point cloud data from multiple perspectives of the target workpiece and large-scale coarse-grained point cloud data.

[0112] An alignment module 630 is used to align the target workpiece in real time based on the mixed-resolution point cloud data and the visual information data;

[0113] A three-dimensional model reconstruction module 640 is used to reconstruct a three-dimensional model of the target workpiece based on the mixed-resolution point cloud data after real-time alignment of the target workpiece;

[0114] The cutting module 650 is used to plan a cutting trajectory based on the cutting requirement information and the three-dimensional model, and perform bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

[0115] It should be noted that the devices provided in the above embodiments are only illustrated by the division of the above functional modules when implementing their functions. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the devices provided in the above embodiments and the corresponding method embodiments are based on the same concept. The specific implementation process is detailed in the corresponding method embodiments and will not be repeated here.

[0116] Unless otherwise defined, the technical or scientific terms used in the present invention shall have the usual meanings understood by persons of ordinary skill in the field to which the present invention belongs. The words "first", "second" and similar terms used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "a" do not indicate a quantity limitation, but rather indicate the existence of at least one. Words such as "connected" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0117] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A bevel cutting method, characterized in that: include: Step S1, obtaining laser triangulation data and time-of-flight data of a target workpiece at multiple viewing angles; Step S2: acquiring mixed-resolution point cloud data of the target workpiece based on the laser triangulation data and the time-of-flight data, wherein the mixed-resolution point cloud data includes high-precision local point cloud data of multiple viewing angles of the target workpiece and large-scale coarse-grained point cloud data; Step S3: acquiring visual information data of the target workpiece, the visual information data including surface image texture information of the target workpiece, and performing real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data; Step S4: after real-time alignment of the target workpiece, reconstructing a three-dimensional model of the target workpiece based on the mixed-resolution point cloud data; Step S5: obtaining cutting requirement information, planning a cutting trajectory based on the cutting requirement information and the three-dimensional model, and performing bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

2. The groove cutting method according to claim 1, characterized in that: The method further comprises: Preprocessing the laser triangulation data and the time-of-flight data respectively to obtain preprocessed laser triangulation data and the time-of-flight data, wherein the preprocessing includes one or more of composite filtering, noise reduction, image enhancement, and edge detection; Wherein, in the step S2, mixed-resolution point cloud data of the target workpiece is acquired based on the pre-processed laser triangulation data and the time-of-flight data.

3. The groove cutting method according to claim 1, characterized in that: The step S2 comprises: Step S21: acquiring the high-precision local point cloud data based on the laser triangulation data, and acquiring large-scale coarse-grained point cloud data based on the time-of-flight data; Step S22 : performing a primary registration on the high-precision local point cloud data of multiple perspectives and the large-scale coarse-grained point cloud data of multiple perspectives to form the mixed-resolution point cloud data.

4. The groove cutting method according to claim 3, characterized in that: The step S22 includes: Identifying missing areas of the mixed-resolution point cloud data based on a point cloud topology structure; When it is determined that there is a missing region, the missing region is completed based on a completion algorithm, wherein the completion algorithm includes one or more of a surface fitting algorithm, a weighted interpolation algorithm, or a curvature-guided completion algorithm.

5. The groove cutting method according to claim 3, characterized in that: The step S3 comprises: Step S31: extracting multimodal features based on the mixed-resolution point cloud data and the visual information data, where the multimodal features include image features and point cloud features; Step S32: Based on the multimodal features, a hierarchical matching strategy is used to perform real-time alignment on the target workpiece.

6. The groove cutting method according to claim 5, characterized in that: The step S32 includes: Step S321: Based on the multimodal features, the target workpiece is roughly registered by using curvature feature fast matching; Step S322: After the rough registration, fine registration is performed on the target workpiece using a dynamic weight factor based on the multimodal features; Step S323 : Acquire real-time position and posture information of the target workpiece, and perform real-time alignment on the target workpiece based on the real-time position and posture information and the precise registration result.

7. The groove cutting method according to claim 5, characterized in that: The step S4 comprises: Based on the mixed-resolution point cloud data, a reconstruction algorithm is used to perform three-dimensional reconstruction on the aligned target workpiece to obtain the three-dimensional model of the target workpiece, and the reconstruction algorithm is a mesh reconstruction algorithm or a surface reconstruction algorithm.

8. The groove cutting method according to claim 1, characterized in that: The step S5 comprises: Step S51: acquiring a processing area on the target workpiece and geometric information of the processing area based on the cutting requirement information and the three-dimensional model of the target workpiece; Step S52: performing path planning on the processing area based on the cutting requirement information and the geometric information of the processing area to obtain the cutting trajectory; Step S53: performing bevel cutting on the processing area of the target workpiece based on the cutting trajectory and the geometric information of the processing area.

9. The groove cutting method according to claim 8, characterized in that: The step S53 includes: Step S531, setting the posture parameters and process parameters of the cutting equipment based on the cutting trajectory, the geometric information of the processing area and the cutting requirement information; Step S532: Obtain cutting feedback information during the groove cutting process, and dynamically adjust the posture parameters and process parameters of the cutting equipment based on the cutting feedback information.

10. A bevel cutting device, characterized in that: The bevel cutting device comprises: a data acquisition module, the data acquisition module being used to acquire laser triangulation data, time-of-flight data, and visual information data of a target workpiece at multiple viewing angles, the visual information data including surface image texture information of the target workpiece; a point cloud data generation module, the point cloud data generation module being configured to generate mixed-resolution point cloud data of a target workpiece based on the laser triangulation data and the time-of-flight data, the mixed-resolution point cloud data comprising high-precision local point cloud data of the target workpiece from multiple perspectives and large-scale coarse-grained point cloud data; an alignment module, configured to perform real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data; a three-dimensional model reconstruction module, configured to reconstruct a three-dimensional model of the target workpiece based on the mixed-resolution point cloud data after real-time alignment of the target workpiece; A cutting module is used to plan a cutting trajectory based on cutting requirement information and the three-dimensional model, and to perform bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

11. A bevel cutting system, characterized in that: include: A three-dimensional scanner for acquiring laser triangulation data and time-of-flight data from multiple viewing angles of a target workpiece; A three-dimensional camera, wherein the three-dimensional camera is used to obtain visual information data of the target workpiece, wherein the visual information data includes surface image texture information of the target workpiece; a processor, the processor being electrically connected to the 3D scanner and the 3D camera, respectively, the processor being configured to acquire mixed-resolution point cloud data of a target workpiece based on the laser triangulation data and the time-of-flight data, the mixed-resolution point cloud data comprising high-precision local point cloud data and large-scale coarse-grained point cloud data from multiple perspectives of the target workpiece, the processor being further configured to perform real-time alignment of the target workpiece based on the mixed-resolution point cloud data and the visual information data, and after real-time alignment of the target workpiece, reconstruct a 3D model of the target workpiece based on the mixed-resolution point cloud data, the processor being further configured to acquire cutting requirement information and plan a cutting trajectory based on the cutting requirement information and the 3D model; A cutting device is electrically connected to the processor, and is used to perform bevel cutting on the target workpiece based on the cutting trajectory and the three-dimensional model.

12. The bevel cutting system according to claim 11, wherein: Also includes: a filtering device, the filtering device being electrically connected to the 3D scanner and the processor, respectively, and configured to preprocess the laser triangulation data and the time-of-flight data to obtain preprocessed laser triangulation data and the time-of-flight data, wherein the preprocessing includes one or more of composite filtering, noise reduction, image enhancement, and edge detection; The processor is used to acquire mixed-resolution point cloud data of a target workpiece based on the preprocessed laser triangulation data and the time-of-flight data.

Images