Hand model processing precision cooperative control method and system based on vision intelligence

Abstract

The embodiment of the invention provides a hand model processing precision cooperative control method and system based on visual intelligence. The method comprises the following steps: dividing a large hand model into a plurality of processing sections according to a topological structure of the large hand model, and creating a corresponding digital twinborn model for each processing section; when any processing section is used as the current processing section for processing, acquiring surface point cloud data of the current processing section; comparing the surface point cloud data with a digital twin model corresponding to the current processing section to obtain actual processing error distribution of the current processing section; and based on the actual machining error distribution, performing feedback control on machining parameters of machining equipment. According to the method, through segmentation of multiple processing sections and feedback control based on actual processing error distribution, the processing error of a large-scale hand plate model can be reduced, so that the problems in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hand plate model processing, and particularly relates to a hand plate model processing precision collaborative control method and system based on visual intelligence. BACKGROUND

[0002] A hand plate model is an initial prototype made according to an appearance image or a structural drawing of a product (usually a sample) before a formal mold is opened during a product research and development stage. As a key link connecting product design and batch production, the hand plate model is mainly used for verifying appearance design, evaluating structural rationality, performing function testing and market research. Therefore, the processing precision of the hand plate model is directly related to the reliability of the verification result and has an important influence on subsequent mold development and product mass production.

[0003] At present, the hand plate model is usually manufactured as a whole by using numerical control processing or 3D printing technology. However, when facing a large-sized hand plate model (such as an automobile part, a large equipment shell, a sculpture artwork, etc.), this kind of processing method is easily limited by machine tool travel, tool wear, thermal deformation and material internal stress release and other factors, resulting in continuous accumulation of errors in the processing process, and finally affecting the processing precision of the hand plate model.

[0004] Therefore, the industry urgently needs a solution that can effectively improve the processing precision of a large-sized hand plate model. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a hand plate model processing precision collaborative control method and system based on visual intelligence, so as to solve the problem of insufficient processing precision of a large-sized hand plate model in the prior art.

[0006] In order to solve the above technical problems, the embodiments of the present application provide a hand plate model processing precision collaborative control method based on visual intelligence, comprising: According to the topological structure of a large-sized hand plate model, the large-sized hand plate model is divided into a plurality of processing sections, and a corresponding digital twin model is created for each processing section; When any one processing section is processed as a current processing section, surface point cloud data of the current processing section is acquired; The surface point cloud data is compared with the digital twin model corresponding to the current processing section to obtain an actual processing error distribution of the current processing section; Based on the actual processing error distribution, feedback control is performed on the processing parameters of the processing equipment.

[0007] Preferably, the method further comprises: The geometric boundary of the large-sized hand plate model is extracted by analyzing a structural design drawing of the large-sized hand plate model; For the region within the extracted geometric boundary, a topological connectivity analysis algorithm is used to identify each structural unit and the connection mode between each structural unit; By the geometric boundary of the large hand plate model, the identified each structural unit and the connection relationship between each structural unit, a topological feature map reflecting the topological structure of the large hand plate model is constructed.

[0008] Preferably, according to the topological structure of the large hand plate model, the large hand plate model is segmented into multiple processing sections, specifically including: Based on each structural unit in the topological feature map and the connection mode between each structural unit, each weak connection region in the large hand plate model is identified; wherein the weak connection region specifically includes a curved surface transition section and a non-key stress area; According to the identified each weak connection region, the large hand plate model is segmented into multiple processing sections.

[0009] Preferably, according to the identified each weak connection region, the large hand plate model is segmented into multiple processing sections, specifically including: Starting from the starting end of the large hand plate model, the first weak connection region is set as the current starting boundary; Starting from the current starting boundary, sequentially find the subsequent weak connection region along the model extension direction, and calculate the cumulative distance between the current starting boundary and the subsequent weak connection region; Judge whether the cumulative distance is greater than or equal to the preset threshold value for the first time; If yes, the weak connection region corresponding to the cumulative distance greater than or equal to the preset threshold value for the first time is determined as the current end boundary, and a processing section is segmented by the current starting boundary and the current end boundary; The current end boundary is taken as the new current starting boundary, and the subsequent weak connection region is sequentially found again along the model extension direction and the new cumulative distance is calculated, until there is no subsequent weak connection region, then the last weak connection region is forcibly determined as the new end boundary for segmentation.

[0010] Preferably, a corresponding digital twin model is created for each processing section, specifically including creating a corresponding digital twin model for each processing section by the following way: Based on the design parameters of the processing section, a section three-dimensional model is generated, wherein the design parameters include geometric size, form and position tolerance, and surface quality requirement; Embedding processing process information in the section three-dimensional model, the processing process information includes tool path planning, cutting parameters and recommended processing strategy; Integrate the processing section three-dimensional model and the processing information to form a digital twin model corresponding to the processing section.

[0011] Preferably, the surface point cloud data of the current processing section is acquired, specifically including: Trigger the optical measurement device installed on the processing equipment to scan the processed surface to collect the surface point cloud data of the current processing section while the processing equipment is processing the current processing section or during the pause gap after completing a processing sub-step, wherein the optical measurement device includes a laser scanner or a laser radar.

[0012] Preferably, the surface point cloud data is compared with the digital twin model corresponding to the current processing section to obtain an actual processing error distribution, specifically including: Identify a plurality of key feature points common to the surface point cloud data and the digital twin model, wherein the key feature points include any one or more of the following: reference mark points, hole center points, hole edge points, and curved surface inflection points; Use the plurality of key feature points common to the surface point cloud data and the digital twin model to register the surface point cloud data with the theoretical surface of the digital twin model; Calculate the normal distance of each key feature point in the surface point cloud data to the theoretical surface of the digital twin model as the actual processing error of the key feature point to obtain the actual processing error distribution of each key feature point in the current processing section.

[0013] Preferably, based on the actual processing error distribution, the processing parameters of the processing equipment are feedback controlled, specifically including: Identify a local out-of-tolerance area in the actual processing error distribution; Generate a compensation processing path for the local out-of-tolerance area and send it to the processing equipment to drive the cutter to locally finish machining the local out-of-tolerance area.

[0014] Preferably, the surface point cloud data is registered with the theoretical surface of the digital twin model using the plurality of key feature points common to the surface point cloud data and the digital twin model, specifically including: Based on the plurality of key feature points common to the surface point cloud data and the digital twin model, a coordinate transformation matrix is calculated through a feature matching algorithm; The surface point cloud data is registered with the theoretical surface of the digital twin model using the coordinate transformation matrix.

[0015] To solve the above technical problems, the embodiment of the present application also provides a hand plate model machining precision collaborative control system based on visual intelligence, comprising: A segmentation creation unit is configured to segment the large hand plate model into multiple machining sections according to the topological structure of the large hand plate model, and create a corresponding digital twin model for each machining section. An acquisition unit is configured to acquire surface point cloud data of a current machining section when any one machining section is machined as the current machining section. A comparison unit is configured to compare the surface point cloud data with the digital twin model corresponding to the current machining section to obtain an actual machining error distribution of the current machining section. A feedback control unit is configured to perform feedback control on machining parameters of a machining device based on the actual machining error distribution.

[0016] The hand plate model machining precision collaborative control method based on visual intelligence provided by the embodiment of the present application comprises the following steps: first, segmenting a large hand plate model into multiple machining sections according to the topological structure of the large hand plate model, and creating a corresponding digital twin model for each machining section; then, acquiring surface point cloud data of a current machining section when any one machining section is machined as the current machining section; then, comparing the surface point cloud data with the digital twin model corresponding to the current machining section to obtain an actual machining error distribution of the current machining section; and then, performing feedback control on machining parameters of a machining device based on the actual machining error distribution. When machining the large hand plate model, on the one hand, the cumulative error can be reduced because the machining section is segmented into multiple machining sections for machining, and the size of the machining section is smaller than that of the large hand plate model; on the other hand, the actual machining error distribution of the current machining section is identified by comparing the surface point cloud data of the current machining section with the digital twin model, and then feedback control is performed on the machining parameters. Obviously, the machining error can be further reduced by identifying the actual machining error distribution and performing feedback control, thereby solving the problems in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0017] To more clearly illustrate the solutions in the present application, the drawings needed in the description of the embodiments of the present application will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0018] Figure 1 is the implementation flowchart of the hand plate model machining precision collaborative control method based on visual intelligence provided by the embodiment of the present application; Figure 2An implementation flowchart of creating a digital twin model in the hand plate model machining precision collaborative control method based on visual intelligence is provided in the embodiments of the present application. Figure 3 A structural schematic diagram of the hand plate model machining precision collaborative control system based on visual intelligence is provided in the embodiments of the present application. Figure 4 A structural schematic diagram of an embodiment of the computer device according to the present application. DETAILED DESCRIPTION

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs; the terminology used in the specification herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application; the description and the drawings of the specification herein should be considered in conjunction with the appended claims and the entire disclosure provided herein; the terms "comprise", "comprising", "include", "including", "have" and "having" as used herein, are meant to be interpreted inclusively rather than exclusively; the terms "first", "second", "third", etc. as used herein are meant to identify different objects, not to imply a particular order.

[0020] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. It is expressly understood that the embodiments described herein are merely examples from a whole class of comparable embodiments which those skilled in the art will readily identify.

[0021] As described previously, the hand plate model is currently generally manufactured by using numerical control machining or 3D printing technology, etc. When machining a large hand plate model, the machining process is easily limited by various factors such as machine tool stroke, tool wear, thermal deformation, and internal stress release of the material, etc., which causes errors to be accumulated during the machining process, and finally affects the machining precision of the hand plate model.

[0022] In view of this, the embodiments of the present application provide a hand plate model machining precision collaborative control method and system based on visual intelligence, which can be used to solve the problems in the prior art. In order for the personnel in the technical field to better understand the present application scheme, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings.

[0023] In the present application, an optical measuring device can be arranged on a machining equipment (including a numerical control machining machine tool, a 3D printer, etc.), which can include a laser scanner or a laser radar. The optical measuring device can be used to collect surface point cloud data of a relevant machining section during the machining process or during the intermittent period of the machining equipment.

[0024] As Figure 1 The application provides a specific process flow diagram of a hand plate model machining precision collaborative control method based on visual intelligence. The method can be applied to local or cloud controllers in machining equipment. The method comprises the following steps: Step S11: According to the topological structure of the large hand plate model, the large hand plate model is divided into multiple machining sections.

[0025] The large hand plate model refers to a hand plate model with large size. In actual application, the large hand plate model can be a car part, a large equipment shell, or a sculpture artwork. The size of these hand plate models is large, and the error accumulation in the machining process can easily affect the machining precision.

[0026] The topological structure of the large hand plate model reflects the structural units of the large hand plate model and the connection relationship between the units. The structural unit refers to a subpart in the large hand plate model that has relative structural independence and functional integrity. From the structure, the structural unit is relatively closed or continuous in geometric form and can be regarded as an independent mechanical component. For example, the structural unit can be a rigid connection area (such as an integrally formed side plate, a panel, or a reinforcing rib in a shell). From the function, the structural unit has a clear and specific function in the large hand plate model, such as a mounting hole group area, a heat dissipation window array, or an interface boss for connecting other components. In addition, the structural unit usually has basic consistency in process, that is, the same structural unit can have the same or similar machining strategy, tool, or precision standard. For example, a flat area on a large curved surface and a high-curvature transition area can be identified as different structural units.

[0027] The connection relationship between the structural units refers to the geometric correlation and mechanical interaction mode between different structural units. For example, the connection relationship can include weak connection (which means that there is an explicit and easy-to-separate boundary between two structural units) and rigid connection (which means that two structural units seamlessly transition in geometry, jointly bear load, and are difficult to separate without damaging the material).

[0028] In the present application, the topology of the large hand plate model can be obtained in advance in the following manner. Specifically, a structural design drawing of the large hand plate model can be obtained first, which can include CAD data, digital model data, sample images of the large hand plate model, etc. Then, the structural design drawing of the large hand plate model can be analyzed to extract the geometric boundaries of the large hand plate model, such as by using an image recognition algorithm to recognize the outer contour lines in the structural design drawing as the extracted geometric boundaries of the large hand plate model.

[0029] After obtaining the geometric boundaries of the large hand plate model, the regions within the geometric boundaries can be analyzed using a topological connectivity analysis algorithm or the like to identify the individual structural units and the connection relationships between the structural units. Alternatively, an active calibration method can be used to identify the individual structural units and the connection relationships between the structural units. Then, the topology feature map reflecting the topology of the large hand plate model can be constructed based on the geometric boundaries of the large hand plate model, the identified individual structural units, and the connection relationships between the structural units, such as a topological feature map represented by a graph structure, in which the nodes represent the corresponding structural units and the edges (connection lines between nodes) represent the connection relationships.

[0030] After constructing the topology feature map reflecting the topology of the large hand plate model, in step S11, the individual weakly connected regions in the large hand plate model can be identified based on the individual structural units in the topology feature map and the connection relationships between the structural units, wherein the weakly connected regions specifically include curved surface transition sections and non-critical load-bearing regions. In the present application, considering that in a complex component, the principal stress flow generally passes along the main direction of the structure and the region with the greatest rigidity, and the curved surface transition section (which has a relatively large curvature change, especially in regions with a more drastic change in curvature) is not usually the distribution path of the principal stress flow, the curved surface transition section therefore bears less load compared to the main load-bearing region or mounting point, so that the damage to the overall function is relatively small when the curved surface transition section is segmented as a weakly connected region in the subsequent process. Similarly, the non-critical load-bearing region refers to a region in the large hand plate model that does not participate in the main load transfer, does not provide critical mounting and positioning, and is not involved in the core sealing function, so it is also considered as a weakly connected region.

[0031] After identifying each weak connection region in the large hand plate model, the large hand plate model can be segmented into multiple processing sections according to the identified weak connection regions. For example, one segmentation method can be to directly segment the weak connection regions as the segmentation boundary, thereby obtaining multiple processing sections. However, in actual application, this method can easily result in too many processing sections.

[0032] Therefore, according to the identified weak connection regions, the large hand plate model can be segmented into multiple processing sections. The specific implementation manner can also be to set the first weak connection region as the current starting boundary from the starting end of the large hand plate model. The starting end of the large hand plate model can be any end, and the extension direction of the model is from the other end. Thus, the first weak connection region is identified, and the first weak connection region is set as the current starting boundary.

[0033] Then, the subsequent weak connection regions are sequentially searched along the extension direction of the model from the current starting boundary, and the cumulative distance between the current starting boundary and the subsequent weak connection region is calculated. Then, it is determined whether the cumulative distance is greater than or equal to the preset threshold value for the first time. If so, the weak connection region corresponding to the cumulative distance greater than or equal to the preset threshold value for the first time is determined as the current ending boundary, and the current starting boundary and the current ending boundary are segmented to obtain a processing section.

[0034] For example, the cumulative distance between the first weak connection region and the second weak connection region (which is sequentially searched along the extension direction of the model) can be calculated, and it is determined whether the cumulative distance is greater than or equal to the preset threshold value for the first time. If so, it indicates that the distance between the two is not too short. At this time, the weak connection region corresponding to the cumulative distance greater than or equal to the preset threshold value for the first time, that is, the second weak connection region, is determined as the current ending boundary, and the current starting boundary and the current ending boundary are segmented to obtain a processing section. The preset threshold value can be set according to the related parameters of the processing equipment to avoid too many processing sections due to too short processing sections, thereby affecting the final processing precision and overall processing efficiency.

[0035] Then, the current ending boundary (that is, the second weak connection region) can be used as a new current starting boundary, and the same method can be used to sequentially search for subsequent weak connection regions along the extension direction of the model and calculate the new cumulative distance. When there is no subsequent weak connection region, the last weak connection region is forcibly determined as a new ending boundary for segmentation, thereby finally segmenting the large hand plate model into multiple processing sections.

[0036] Of course, in the above example, the cumulative distance between the first weak connection region and the second weak connection region is less than the preset threshold, which means that the cumulative distance is not greater than or equal to the preset threshold for the first time, so the third weak connection region can be sequentially searched along the model extension direction, the cumulative distance between the first weak connection region and the third weak connection region is calculated, and it is judged whether the cumulative distance is greater than or equal to the preset threshold for the first time. If so, the third weak connection region is determined as the current end boundary, and a processing section is obtained by dividing the current starting boundary (i.e. the first weak connection region) and the current end boundary (i.e. the third weak connection region). Of course, at this time, if the cumulative distance is still less than the preset threshold, it means that it is not greater than or equal to the preset threshold for the first time, and the fourth weak connection region can be sequentially searched along the model extension direction, and based on the same idea, it is determined whether the fourth weak connection region can be used as the current end boundary.

[0037] Step S12: Create a corresponding digital twin model for each processing section.

[0038] After the large hand plate model is divided into multiple processing sections through the above step S11, in this step S12, a corresponding digital twin model can be further created for each processing section. In the present application, the digital twin model created for each processing section is the core and basis for subsequent high-precision comparison and intelligent control. Therefore, the specific implementation of this step S12 of creating a digital twin model can be further described as follows: Figure 2 As shown in the figure, this step S12 can create a corresponding digital twin model for each processing section by the following way: Step S121: Generate a section three-dimensional model based on the design parameters of the processing section.

[0039] Among them, the processing section can be any one of the processing sections obtained by step S11.

[0040] In practical applications, the design parameters of the processing section can be obtained by combining the structural design drawing of the large hand plate model and the topological structure of the large hand plate model. For example, the processing section in the structural design drawing can be located by the topological structure of the large hand plate model, and then the design parameters of the processing section can be obtained from the structural design drawing, wherein the design parameters include geometric dimensions, shape and position tolerances, and surface quality requirements (such as surface roughness specified for different regions).

[0041] The specific implementation of the step S121 can be that, after obtaining the design parameters of the machining section, the three-dimensional model of the section can be generated in a manner known in the art, such as a direct extraction reconstruction method. Here, the manner of generating the three-dimensional model of the section is not limited.

[0042] Step S122: embedding machining process information in the three-dimensional model of the section, the machining process information including tool path planning, cutting parameters and recommended machining strategies.

[0043] The step S122 is to convert the static three-dimensional model of the section into a process model that can guide production by embedding machining process information.

[0044] In the step S122, the tool path planning, cutting parameters and recommended machining strategies in the machining process information can be embedded in sequence. For example, the embedding of the tool path planning can be embedding the tool type (such as an end mill, a ball nose tool, a center drill), the machining sequence (such as the sequence of rough machining, semi-finishing and finishing), and the tool trajectory geometry data (the motion path of the tool center point) into the three-dimensional model of the section. Specifically, in a computer-aided manufacturing (CAM) software environment, based on the geometric features of the three-dimensional model of the section, numerical control programming can be performed, and then based on the embedding program obtained by the numerical control programming, the tool path planning can be embedded into the three-dimensional model of the section.

[0045] The cutting parameters can include spindle speed (RPM), cutting feed rate (mm / min), cutting depth / step distance (mm), coolant switch state, etc. In a computer-aided manufacturing software environment, based on the geometric features of the three-dimensional model of the section, numerical control programming can be performed, and then the cutting parameters can be embedded into the three-dimensional model of the section.

[0046] The recommended machining strategy can include the name of the machining strategy and the key parameter setting logic, etc. In this application, the recommended machining strategy can also be embedded into the three-dimensional model of the section by means of numerical control programming.

[0047] In this way, in the step S122, the tool path planning, cutting parameters and recommended machining strategies can be obtained first, and then in a computer-aided manufacturing (CAM) software environment, the tool path planning, cutting parameters and recommended machining strategies can be embedded into the three-dimensional model of the section by means of the embedding program obtained by numerical control programming.

[0048] Step S123: associating and integrating the three-dimensional model of the section with the machining process information to form a digital twin model corresponding to the machining section.

[0049] The specific implementation of this step S123 can be to first associate and map the data, and then generate an interactive digital twin model. For example, a central data model (for example, an engineering database entry based on XML or JSON format) can be first created, which takes the process section three-dimensional model as the geometric core, and then all the machining process information embedded in step S122 is accurately mapped to the features in the central data model through a pointer or a unique identifier (ID), and then the whole is packaged in a software object that can be called by the upper layer application (such as machine tool control system, machining equipment, monitoring system), thereby serving as the digital twin model corresponding to the machining section.

[0050] Step S13: When any one machining section is processed as a current machining section, the surface point cloud data of the current machining section is acquired.

[0051] The current machining section refers to the machining section that is currently being processed or is ready to be processed. Obviously, the current machining section can be any one of the plurality of machining sections segmented in step S11.

[0052] In this step S13, the surface point cloud data of the current machining section can be acquired when the current machining section is processed. For example, the optical measuring device installed on the machining equipment can be triggered to scan the machined surface to collect the surface point cloud data of the current machining section at the same time when the machining equipment processes the current machining section or during the pause gap after completing a machining sub-step (referring to a machining sub-step in the current machining section). Obviously, the surface point cloud data of the current machining section can reflect the processing condition of the surface of the current machining section.

[0053] In an embodiment of the present application, the optical measuring device can be arranged on the spindle or the tool turret of the machining equipment, so that it can move with the spindle and maintain a relative position relationship with the tool. In this way, the optical measuring device can collect surface point cloud data in real time during the processing of the tool of the machining equipment. Real-time collection in this way can enable more timely feedback control.

[0054] Of course, the optical measuring device can also be arranged on a certain fixed mounting point of the machining equipment, so that the surface point cloud data can be collected during the pause gap after completing a machining sub-step. This method collects data during the pause gap, which can avoid the blocking of the line of sight caused by the movement of the tool at this time (the tool is usually reset at this time and does not block the collection line of sight of the optical measuring device), thereby more comprehensively scanning the machined surface.

[0055] Step S14: comparing the surface point cloud data with the digital twin model corresponding to the current machining section to obtain the actual machining error distribution of the current machining section.

[0056] In this step S14, a plurality of key feature points common to the surface point cloud data and the digital twin model can be identified first, the key feature points including any one or more of the following: a reference mark point, a hole center, a hole edge point, and a curved surface inflection point. That is, in this step S14, the surface point cloud data and the digital twin model are compared to obtain a plurality of key feature points common to both.

[0057] After obtaining the plurality of key feature points common to both, the surface point cloud data and the theoretical surface of the digital twin model can be registered by using the plurality of key feature points common to both, so as to improve the accuracy of subsequent processing results by registering the surface point cloud data and the theoretical surface of the digital twin model. In this application, the specific implementation of registering the surface point cloud data and the theoretical surface of the digital twin model can be that, first, based on the plurality of key feature points common to both, a coordinate transformation matrix for aligning the surface point cloud data from a measurement coordinate system to a digital twin model coordinate system is calculated by a feature matching algorithm, wherein the feature matching algorithm can be a 4PCS algorithm, a Phase Correlation algorithm, etc., and any one of the feature matching algorithms can be selected by a person skilled in the art according to the quality of the point cloud and the type of the feature to calculate the coordinate transformation matrix.

[0058] Then, the surface point cloud data and the theoretical surface of the digital twin model are registered by using the coordinate transformation matrix, and in this process, each point (each data point) in the surface point cloud data can be transformed to the digital twin model coordinate system by a transformation formula P_aligned = R • P_measured + t. In this formula, P_aligned is the coordinate in the new coordinate system (i.e. the digital twin model coordinate system) after transformation; P_measured is the data point in the surface point cloud data; R is the coordinate transformation matrix; and t is the translation vector. In this way, each point in the surface point cloud data can be transformed to the digital twin model coordinate system by the formula, so as to register the surface point cloud data and the theoretical surface of the digital twin model.

[0059] After registering the two data points, the normal distance from each key feature point in the surface point cloud data to the theoretical surface of the digital twin model can be calculated. This distance is taken as the actual machining error of the key feature point, thus obtaining the actual machining error distribution of each key feature point in the current machining section. In this step, the normal distance from each key feature point in the surface point cloud data to the theoretical surface of the digital twin model can be calculated, which is the distance perpendicular to the surface where the key feature point is located. This normal distance obviously reflects the deviation between the actual machined point and the theoretical point, so it can be used as the actual machining error of the key feature point. In this way, the actual machining error of each key feature point can be obtained, thus obtaining the actual machining error distribution of each key feature point in the current machining section.

[0060] Step S15: Based on the actual processing error distribution, perform feedback control on the processing parameters of the processing equipment.

[0061] After obtaining the actual processing error distribution through step S14, since this actual processing error distribution represents the actual processing errors of each key feature point, step S15 can identify local out-of-tolerance regions in the actual processing error distribution. These local out-of-tolerance regions refer to areas where the actual processing errors of key feature points exceed a preset maximum error. For example, it can be determined whether the actual processing error of each key feature point is greater than the preset maximum error, thereby filtering out key feature points whose actual processing errors exceed the preset maximum error, and further reducing the actual processing error... The area surrounding key feature points with a difference greater than the preset maximum error is designated as the local out-of-tolerance region. Obviously, the machining error in this local out-of-tolerance region is relatively large, exceeding the tolerable limit. After identifying the local out-of-tolerance region, a compensation machining path can be generated for this region and sent to the machining equipment to drive the tool to perform local finishing on the local out-of-tolerance region. For example, a lower machining step and a finer machining path (as a compensation machining path) can be set for this local out-of-tolerance region to perform local finishing, thereby further reducing the machining error in the local out-of-tolerance region.

[0062] The method for collaborative control of machining accuracy of a large prototype model based on visual intelligence, as provided in this application embodiment, includes first dividing the large prototype model into multiple machining segments according to its topology, and creating a corresponding digital twin model for each machining segment. Then, when machining any machining segment as the current machining segment, the surface point cloud data of that current machining segment is acquired. This surface point cloud data is then compared with the corresponding digital twin model to obtain the actual machining error distribution of the current machining segment. Finally, based on this actual machining error distribution, the machining parameters of the machining equipment are controlled by feedback. When machining a large prototype model, this method, firstly, reduces accumulated errors because the segment is divided into multiple machining segments, and since each machining segment is smaller than the large prototype model. Secondly, it combines the surface point cloud data of the current machining segment with the digital twin model to identify the actual machining error distribution of the current machining segment, thereby enabling feedback control of the machining parameters. Clearly, by identifying and controlling the actual machining error distribution, machining errors can be further reduced, thus solving the problems in the prior art.

[0063] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0064] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0065] Based on the inventive concept of the visual intelligence-based prototyping model processing accuracy collaborative control method provided in the embodiments of this application, the embodiments of this application also provide a visual intelligence-based prototyping model processing accuracy collaborative control system. For any unclear points regarding the content of this system embodiment, please refer to the corresponding content in the method embodiment. Figure 3 The diagram shows the specific structure of the visual intelligence-based prototyping model processing accuracy collaborative control system (hereinafter referred to as system 30). System 30 includes: a segmentation creation unit 301, an acquisition unit 302, a comparison unit 303, and a feedback control unit 304, wherein: The segmentation creation unit 301 is used to segment the large prototype model into multiple processing segments according to the topology of the large prototype model, and create a corresponding digital twin model for each processing segment. The acquisition unit 302 is used to acquire the surface point cloud data of the current processing segment when any processing segment is used as the current processing segment for processing. The comparison unit 303 is used to compare the surface point cloud data with the digital twin model corresponding to the current processing section to obtain the actual processing error distribution of the current processing section. The feedback control unit 304 is used to perform feedback control on the processing parameters of the processing equipment based on the actual processing error distribution.

[0066] Since the system 30 adopts the same inventive concept as the visual intelligence-based prototyping model processing accuracy collaborative control method provided in the embodiments of this application, it can also solve the problems in the prior art, which will not be elaborated here.

[0067] The system 30 may further include a topology construction unit, used to extract the geometric boundaries of the large prototype model by parsing the structural design drawing of the large prototype model; for the region within the extracted geometric boundaries, a topology connectivity analysis algorithm is used to identify each structural unit and the connection methods between each structural unit; and a topology feature map reflecting the topology of the large prototype model is constructed based on the geometric boundaries of the large prototype model, the identified structural units, and the connection relationships between each structural unit.

[0068] Specifically, based on the topological structure of the large prototype model, the large prototype model is divided into multiple processing sections, which may include: Based on the structural units in the topological feature map and the connection methods between the structural units, weak connection regions in the large prototype model are identified; wherein, the weak connection regions specifically include curved transition sections and non-critical stress areas. Based on the identified weak connection regions, the large prototype model is divided into multiple processing sections.

[0069] Specifically, dividing the large prototype model into multiple processing segments based on the identified weak connection regions may include: Starting from the beginning of the large prototype model, the first weakly connected region is set as the current starting boundary; Starting from the current starting boundary, sequentially search for subsequent weakly connected regions along the model extension direction, and calculate the cumulative distance between the current starting boundary and the subsequent weakly connected regions. Determine whether the cumulative distance is greater than or equal to the preset threshold for the first time; If so, the weak connection region corresponding to the first cumulative distance greater than or equal to the preset threshold is determined as the current end boundary, and a processing segment is obtained by dividing the current start boundary and the current end boundary. The current ending boundary is used as the new current starting boundary, and the subsequent weak connection regions are searched sequentially along the model extension direction and the new cumulative distance is calculated until there are no subsequent weak connection regions. Then, the last weak connection region is forcibly determined as the new ending boundary for segmentation.

[0070] Specifically, creating a corresponding digital twin model for each processing stage can include creating a corresponding digital twin model for each processing stage in the following ways: Based on the design parameters of the processing section, a three-dimensional model of the section is generated, wherein the design parameters include geometric dimensions, geometric tolerances and surface quality requirements; Machining process information is embedded in the three-dimensional model of the work section. The machining process information includes tool path planning, cutting parameters, and recommended machining strategies. The 3D model of the processing section is associated and integrated with the processing technology information to form a digital twin model corresponding to the processing section.

[0071] Specifically, acquiring the surface point cloud data of the current processing section may include: While the processing equipment is processing the current processing section, or during a pause after completing a processing sub-step, an optical measuring device installed on the processing equipment is triggered to scan the processed surface to collect surface point cloud data of the current processing section. The optical measuring device includes a laser scanner or a lidar.

[0072] Specifically, comparing the surface point cloud data with the digital twin model corresponding to the current processing segment to obtain the actual processing error distribution may include: Identify multiple key feature points common to the surface point cloud data and the digital twin model, wherein the key feature points include any one or more of the following: reference marker points, hole center, hole edge points, and surface inflection points; Using multiple key feature points shared by the surface point cloud data and the digital twin model, the surface point cloud data is registered with the theoretical surface of the digital twin model; The normal distance from each key feature point in the surface point cloud data to the theoretical surface of the digital twin model is calculated respectively, and used as the actual processing error of the key feature point, so as to obtain the actual processing error distribution of each key feature point in the current processing section.

[0073] Specifically, feedback control of the processing parameters of the processing equipment based on the actual processing error distribution can include: Identify local out-of-tolerance regions in the actual processing error distribution; A compensation machining path is generated for the local out-of-tolerance area and sent to the machining equipment to drive the tool to perform local finishing on the local out-of-tolerance area.

[0074] Specifically, registering the surface point cloud data with the theoretical surface of the digital twin model using multiple key feature points shared by the surface point cloud data and the digital twin model can include: Based on multiple key feature points shared by the surface point cloud data and the digital twin model, a coordinate transformation matrix is ​​calculated using a feature matching algorithm; The surface point cloud data is registered with the theoretical surface of the digital twin model using the coordinate transformation matrix.

[0075] To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed] for details. Figure 4 , Figure 4 This is a basic structural block diagram of a computer device according to an embodiment of this application.

[0076] The computer device 400 includes a memory 410, a processor 420, and a network interface 430 that are interconnected via a system bus. It should be noted that only the computer device 400 with components 410-430 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0077] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0078] The memory 410 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 410 may be an internal storage unit of the computer device 400, such as the hard disk or memory of the computer device 400. In other embodiments, the memory 410 may also be an external storage device of the computer device 400, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. Of course, the memory 410 may also include both internal storage units and external storage devices of the computer device 400. In this embodiment, the memory 410 is typically used to store the operating system and various application software installed on the computer device 400, such as computer-readable instructions of the method provided in this embodiment. Furthermore, the memory 410 can also be used to temporarily store various types of data that have been output or will be output.

[0079] In some embodiments, the processor 420 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 420 is typically used to control the overall operation of the computer device 400. In this embodiment, the processor 420 is used to execute computer-readable instructions stored in the memory 410 or to process data, for example, to execute computer-readable instructions of the methods provided in this embodiment.

[0080] The network interface 430 may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 400 and other electronic devices.

[0081] This application also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the method described above.

[0082] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0083] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A method for collaborative control of the processing accuracy of a prototype model based on visual intelligence, characterized in that, include: Based on the topology of the large prototype model, the large prototype model is divided into multiple processing sections, and a corresponding digital twin model is created for each processing section. When any processing segment is used as the current processing segment, the surface point cloud data of the current processing segment is acquired; The surface point cloud data is compared with the digital twin model corresponding to the current processing section to obtain the actual processing error distribution of the current processing section; Based on the actual processing error distribution, feedback control is performed on the processing parameters of the processing equipment.

2. The method according to claim 1, characterized in that, The method further includes: The geometric boundaries of the large prototype model are extracted by analyzing its structural design drawings. For the region within the extracted geometric boundary, a topological connectivity analysis algorithm is used to identify each structural unit and the connection methods between each structural unit; By using the geometric boundaries of the large prototype model, the identified structural units, and the connection relationships between the structural units, a topological feature map reflecting the topological structure of the large prototype model is constructed.

3. The method according to claim 2, characterized in that, Based on the topology of the large prototype model, the large prototype model is divided into multiple processing sections, specifically including: Based on the structural units in the topological feature map and the connection methods between the structural units, weak connection regions in the large prototype model are identified; wherein, the weak connection regions specifically include curved transition sections and non-critical stress areas. Based on the identified weak connection regions, the large prototype model is divided into multiple processing sections.

4. The method according to claim 3, characterized in that, The step of dividing the large prototype model into multiple processing segments based on the identified weak connection regions specifically includes: Starting from the beginning of the large prototype model, the first weakly connected region is set as the current starting boundary; Starting from the current starting boundary, sequentially search for subsequent weakly connected regions along the model extension direction, and calculate the cumulative distance between the current starting boundary and the subsequent weakly connected regions. Determine whether the cumulative distance is greater than or equal to the preset threshold for the first time; If so, the weak connection region corresponding to the first cumulative distance greater than or equal to the preset threshold is determined as the current end boundary, and a processing segment is obtained by dividing the current start boundary and the current end boundary. The current ending boundary is used as the new current starting boundary, and the subsequent weak connection regions are searched sequentially along the model extension direction and the new cumulative distance is calculated until there are no subsequent weak connection regions. Then, the last weak connection region is forcibly determined as the new ending boundary for segmentation.

5. The method according to claim 1, characterized in that, Create corresponding digital twin models for each processing stage, specifically by creating corresponding digital twin models for each processing stage in the following ways: Based on the design parameters of the processing section, a three-dimensional model of the section is generated, wherein the design parameters include geometric dimensions, geometric tolerances and surface quality requirements; Machining process information is embedded in the three-dimensional model of the work section. The machining process information includes tool path planning, cutting parameters, and recommended machining strategies. The 3D model of the processing section is associated and integrated with the processing technology information to form a digital twin model corresponding to the processing section.

6. The method according to claim 1, characterized in that, Obtaining the surface point cloud data of the current processing section specifically includes: While the processing equipment is processing the current processing section, or during a pause after completing a processing sub-step, an optical measuring device installed on the processing equipment is triggered to scan the processed surface to collect surface point cloud data of the current processing section. The optical measuring device includes a laser scanner or a lidar.

7. The method according to claim 1, characterized in that, The surface point cloud data is compared with the digital twin model corresponding to the current processing section to obtain the actual processing error distribution, specifically including: Identify multiple key feature points common to the surface point cloud data and the digital twin model, wherein the key feature points include any one or more of the following: reference marker points, hole center, hole edge points, and surface inflection points; Using multiple key feature points shared by the surface point cloud data and the digital twin model, the surface point cloud data is registered with the theoretical surface of the digital twin model; The normal distance from each key feature point in the surface point cloud data to the theoretical surface of the digital twin model is calculated respectively, and used as the actual processing error of the key feature point, so as to obtain the actual processing error distribution of each key feature point in the current processing section.

8. The method according to claim 7, characterized in that, Based on the actual processing error distribution, feedback control is performed on the processing parameters of the processing equipment, specifically including: Identify local out-of-tolerance regions in the actual processing error distribution; A compensation machining path is generated for the local out-of-tolerance area and sent to the machining equipment to drive the tool to perform local finishing on the local out-of-tolerance area.

9. The method according to claim 7, characterized in that, Using multiple key feature points shared by the surface point cloud data and the digital twin model, the surface point cloud data is registered with the theoretical surface of the digital twin model, specifically including: Based on multiple key feature points shared by the surface point cloud data and the digital twin model, a coordinate transformation matrix is ​​calculated using a feature matching algorithm; The surface point cloud data is registered with the theoretical surface of the digital twin model using the coordinate transformation matrix.

10. A collaborative control system for the processing accuracy of hand-made prototype models based on visual intelligence, characterized in that, include: The segmentation creation unit is used to segment the large prototype model into multiple processing segments according to the topology of the large prototype model, and create a corresponding digital twin model for each processing segment. The acquisition unit is used to acquire the surface point cloud data of any processing segment when processing any processing segment as the current processing segment; The comparison unit is used to compare the surface point cloud data with the digital twin model corresponding to the current processing section to obtain the actual processing error distribution of the current processing section. The feedback control unit is used to perform feedback control on the processing parameters of the processing equipment based on the actual processing error distribution.

Images