Automatic automobile modeling method and equipment based on machine learning and medium

Abstract

The invention discloses an automatic automobile modeling method and device based on machine learning and a medium, and relates to the field of electric digital data processing, and the method comprises the steps: obtaining part description information of an automobile modeling preliminary manuscript to obtain to-be-assembled accessories; arranging and combining the to-be-assembled accessories according to the types of the accessories and the size proportional relation among the parts to obtain a plurality of accessory combinations of a complete automobile model, and generating a plurality of to-be-modeled automobiles according to the curved surface model; detecting each to-be-modeled automobile, determining the to-be-modeled automobile with a to-be-optimized part, and identifying the to-be-optimized part and an abnormal type; determining a target to-be-optimized point location and actual coordinate information of the to-be-optimized part, and calculating a deviation degree of the to-be-optimized part in combination with the ideal coordinate information, so as to determine a compensation value of the to-be-optimized part according to the deviation degree; and adjusting the to-be-optimized part according to the compensation value to obtain an optimized to-be-modeled automobile, and rendering the optimized to-be-modeled automobile to generate a three-dimensional model of the optimized to-be-modeled automobile so as to realize automatic modeling of the automobile.

Description

Technical Field

[0001] The present application relates to the field of image recognition technology, and in particular to a method, device and medium for automated automobile modeling based on machine learning. Background Art

[0002] As consumers' demand for personalized and differentiated car appearance grows, design teams need to quickly generate and evaluate multiple design options to ensure that the final model can attract the target consumer group. However, the traditional car styling design process is cumbersome and time-consuming, usually requiring designers to gradually build the car styling through hand-drawing, modeling, rendering and other steps. This not only requires designers to have high professional skills, but also limits the diversity and innovation of design options, and greatly prolongs the time cycle from product design to market. Summary of the invention

[0003] The embodiments of the present application provide a method, device and medium for automatic automobile modeling based on machine learning to solve the above-mentioned technical problems.

[0004] On the one hand, the embodiment of the present application provides a method for automatic vehicle modeling based on machine learning, comprising:

[0005] Obtaining component description information in the preliminary draft of the automobile modeling, and obtaining the to-be-assembled accessories of the automobile modeling according to the component description information; wherein the component description information includes the component belonging area corresponding to each automobile component, the component shape outline, the component geometric features, and the size ratio relationship between the components;

[0006] According to the types of parts corresponding to the parts to be assembled and the size ratio relationship between the parts, the parts to be assembled are arranged and combined to obtain several combinations of parts corresponding to the complete car shape, and several cars to be shaped corresponding to the several combinations of parts are generated according to the surface shape;

[0007] Detecting each car to be shaped, determining the car to be shaped with a part to be optimized, and identifying the abnormality type corresponding to the part to be optimized; wherein the abnormality type includes unsmooth styling and incompatible specifications;

[0008] Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, so as to determine the compensation value corresponding to the part to be optimized according to the deviation degree;

[0009] The part to be optimized is adjusted according to the compensation value to obtain an optimized car to be shaped, and the optimized car to be shaped is rendered to dynamically display the three-dimensional model corresponding to the optimized car to be shaped, thereby realizing automatic styling of the car.

[0010] In one implementation of the present application, obtaining component description information in a preliminary draft of a car model includes:

[0011] Obtaining a predetermined preliminary draft of a car shape, and inputting the preliminary draft into an image recognition model for analysis, to determine a plurality of shapes in the preliminary draft;

[0012] For each of the multiple shapes, identifying the corresponding shape edge in the preliminary manuscript by an edge detection algorithm, and segmenting the corresponding shape contour in the preliminary manuscript according to the shape edge and by a graph segmentation algorithm;

[0013] According to the matching degree between the segmented shape contour and the preset automobile component, the component region to which the shape contour corresponds to the automobile component belongs is determined; wherein the component region includes: the front, rear, roof and body of the vehicle;

[0014] According to the preliminary manuscript, the size ratio relationship between the components corresponding to the design intention is determined, and the component geometric features corresponding to the automobile component are calculated in combination with the shape contour and the area to which the component belongs.

[0015] In one implementation of the present application, obtaining the to-be-assembled accessories of the automobile model according to the component description information specifically includes:

[0016] According to the preset mapping relationship, multiple accessories to be selected corresponding to the area to which the component belongs are determined in the preset accessory style library;

[0017] According to the shape outline in the component description information, determining whether there is a target to-be-assembled component corresponding to the region to which the component belongs from the components to be selected;

[0018] If yes, obtaining at least one target to-be-assembled accessory corresponding to the shape outline;

[0019] If not, construct the corresponding accessories to be assembled according to the area to which the components belong in the component description information and the characteristic points in the shape contour, and adjust the corresponding size of the constructed accessories to be assembled according to the component geometric features in the component description information to generate the target accessories to be assembled.

[0020] In one implementation of the present application, the parts to be assembled are arranged and combined according to the types of parts corresponding to the parts to be assembled and the size ratio relationship between the parts to be assembled, so as to obtain several combinations of parts corresponding to the complete automobile shape, and several automobiles to be shaped corresponding to the several combinations of parts are generated according to the curved surface shape, specifically including:

[0021] Classify all the parts to be assembled, determine the type of parts corresponding to each part to be assembled, and determine the detailed information of the parts to be assembled in the preset parts style library; wherein the types of parts at least include: vehicle body, door, window, rearview mirror and tire, and the detailed information of parts includes size, shape, curvature of curved surface and curvature direction of curved surface;

[0022] According to the characteristics of the automobile shape, the types of accessories required for each automobile to be shaped and the number of accessories required for each type of accessories are determined, and in combination with the size ratio relationship between the components in the component description information and the detailed information of each component to be assembled, all the components to be assembled are arranged and combined to generate a number of combinations of accessories corresponding to a number of complete automobile shapes;

[0023] For each of the plurality of accessory combinations, according to the accessory detailed information, the accessory parameters corresponding to each to-be-assembled accessory in the accessory combination are determined, and according to the connection mode between adjacent accessories, the accessory parameters of adjacent accessories in the same accessory combination are adjusted; wherein the accessory parameters at least include the length, width, height and curvature of the curve of the vehicle body;

[0024] Through the surface modeling algorithm, according to the adjusted accessory parameters, a plurality of automobile surfaces corresponding to the accessory combination are generated, and the plurality of automobile surfaces are spliced ​​to obtain the corresponding automobile to be modeled.

[0025] In one implementation of the present application, each car to be styled is detected, the car to be styled having a part to be optimized is determined, and the abnormal type corresponding to the part to be optimized is identified, specifically including:

[0026] For each car to be styled, extract the geometric features of the car to be styled from the three-dimensional model of the car to be styled, and generate a curvature map corresponding to the three-dimensional model; wherein the geometric features of the car to be styled include: curvature, radius, angle and slope;

[0027] According to the curvature change in the curvature map, a transition area in the three-dimensional model is identified, and in combination with the transition area, whether the continuity of the same accessory surface in the three-dimensional model is abnormal is determined;

[0028] If yes, it is determined that the vehicle to be styled has a part to be optimized, and the abnormality type corresponding to the part to be optimized is unsmooth styling; if no, it is determined whether the connection between adjacent accessories matches according to the accessory parameters in the vehicle to be styled and the size ratio relationship between the components;

[0029] In the case of connection mismatch, it is determined that the vehicle to be styled has a part to be optimized, and the abnormality type corresponding to the part to be optimized is specification mismatch; wherein the connection matching is used to determine whether the accessory gap meets the preset gap and whether adjacent accessories are aligned.

[0030] In one implementation of the present application, the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized are determined according to the abnormality type, and the degree of deviation of the part to be optimized is calculated in combination with the ideal coordinate information of the target point to be optimized, specifically including:

[0031] For the part to be optimized whose abnormal type is unsmooth modeling, according to the curvature change in the curvature map, the target abnormal area of ​​the part to be optimized is identified in the transition area and the surface inflection point corresponding to the part to be optimized, and the target point to be optimized is determined in the target abnormal area;

[0032] Determine the ideal coordinate information of the target point to be optimized according to the component description information, and measure the actual coordinate information of the target point to be optimized in the three-dimensional model of the vehicle to be modeled by a three-dimensional coordinate measuring machine;

[0033] Determine the difference between the actual coordinate information and the ideal coordinate information in each dimension of the three-dimensional model, calculate the sum of the squares of the differences in each dimension, and take the square root of the sum of the squares in each dimension to obtain the degree of deviation of the part to be optimized;

[0034] According to the actual coordinate information, the ideal coordinate information and the preset compensation factor of each dimension, the deviation component corresponding to the deviation degree in each dimension is calculated.

[0035] In one implementation of the present application, the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized are determined according to the abnormality type, and the degree of deviation of the part to be optimized is calculated in combination with the ideal coordinate information of the target point to be optimized, specifically including:

[0036] For the part to be optimized whose abnormal type is incompatible specifications, the actual size of the part to be optimized and the part connected to the part to be optimized is obtained through the three-dimensional model of the car to be shaped;

[0037] Determine the size ratio between the part to be optimized and the connected part according to the component description information, and determine whether the part to be optimized needs to be resized in combination with the actual sizes of the part to be optimized and the connected part;

[0038] If yes, then the degree of deviation between the actual size of the part to be optimized and the ideal size is calculated; if no, then the actual coordinate information of the connection point between the part to be optimized and the connected part is measured in the three-dimensional model of the automobile to be shaped by a three-dimensional coordinate measuring machine;

[0039] Determine the ideal coordinate information of the connection point between the part to be optimized and the connected part, and calculate the degree of deviation of the connection point according to the actual coordinate information and the ideal coordinate information.

[0040] In one implementation of the present application, rendering the optimized car to be shaped to dynamically display the three-dimensional model corresponding to the optimized car to be shaped includes:

[0041] Exporting the optimized car to be styled to a rendering engine, and selecting material textures corresponding to the various accessories of the car to be styled in the rendering engine in combination with the car styling display intent corresponding to the component description information; wherein the material textures are used to simulate the color of car paint and metallic luster;

[0042] Selecting a lighting environment for the car to be styled in the rendering engine, simulating a real scene through the lighting environment, and rendering the car to be styled according to preset rendering parameters;

[0043] Animation is produced for the car to be modeled, a corresponding animation sequence is rendered by the rendering engine, and the rendered animation is encoded so as to dynamically display the animation in a video format; wherein the animation is used to simulate the movement of the car.

[0044] On the other hand, the embodiment of the present application also provides an automobile automated styling device based on machine learning, the device comprising:

[0045] at least one processor;

[0046] and, a memory communicatively coupled to the at least one processor;

[0047] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned automatic vehicle styling method based on machine learning.

[0048] On the other hand, an embodiment of the present application further provides a non-volatile computer storage medium storing computer executable instructions, which, when executed, implements the above-mentioned method for automated automobile styling based on machine learning.

[0049] The embodiments of the present application provide a method, device and medium for automatic modeling of an automobile based on machine learning, which at least have the following beneficial effects:

[0050] By obtaining the component description information in the preliminary car styling manuscript, covering the area to which the components belong, shape outline, geometric features and size ratio relationship between components, the parts to be assembled can be accurately determined, which greatly improves the conversion efficiency from creative manuscripts to actual styling construction and reduces the tedious process and errors of manual screening and matching of accessories; according to the types of accessories and the precise size ratio of components, the parts to be assembled are arranged and combined to generate multiple combinations of accessories and corresponding cars to be styled, fully exploring the possibilities of different combinations, which helps designers find the most innovative and practical car styling among many options; each car to be styled is automatically detected to identify abnormal ones The optimized parts and their corresponding abnormal types can timely discover the problems of unsmooth styling affecting the appearance or assembly problems caused by incompatible specifications, thereby avoiding rework due to design defects in subsequent production links and saving a lot of time and cost; the target points to be optimized are determined according to the abnormality types, and the degree of deviation is calculated based on the actual and ideal coordinate information, and then the compensation value is determined, so that the parts to be optimized can be fine-tuned in time to ensure that the adjusted car styling meets the design expectations and improve the overall quality and refinement of the car styling; by rendering the optimized car to be styled into a three-dimensional model, the design results are more intuitive, which makes it easier to observe the car styling from all angles and discover potential problems in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0052] Figure 1 A schematic diagram of a process flow of an automobile automated modeling method based on machine learning provided in an embodiment of the present application;

[0053] Figure 2 A schematic diagram of the internal structure of an automotive automated styling device based on machine learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0055] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0056] Figure 1 A schematic flow chart of a method for automated automobile modeling based on machine learning provided in an embodiment of the present application.

[0057] The analysis method involved in the embodiments of the present application can be implemented by a terminal device or a server, and the present application does not impose any special restrictions on this. For the convenience of understanding and description, the following embodiments are described in detail by taking a server as an example.

[0058] It should be noted that the server may be a single device or a system consisting of multiple devices, that is, a distributed server, and this application does not make any specific limitation on this.

[0059] like Figure 1 As shown, an automobile automatic modeling method based on machine learning provided in an embodiment of the present application includes:

[0060] 101. Obtain component description information in a preliminary draft of a car model, and obtain the to-be-assembled accessories of the car model according to the component description information.

[0061] The component description information includes the region to which each automobile component belongs, the shape outline of the component, the geometric features of the component, and the size ratio relationship between the components.

[0062] In one embodiment of the present application, obtaining component description information in a preliminary draft of a car model specifically includes:

[0063] Obtain a pre-determined preliminary draft of the automobile shape, and input the preliminary draft into the image recognition model for analysis to determine multiple shapes in the preliminary draft;

[0064] For each of the multiple shapes, identifying a corresponding shape edge in the preliminary manuscript by using an edge detection algorithm, and segmenting the corresponding shape contour in the preliminary manuscript according to the shape edge and using a graph segmentation algorithm;

[0065] According to the matching degree between the segmented shape contour and the preset automobile component, the component region to which the shape contour corresponds to the automobile component is determined; wherein the component region includes: the front, rear, roof and body of the vehicle;

[0066] According to the preliminary manuscript, determine the size ratio relationship between the components corresponding to the design intention, and calculate the component geometric features corresponding to the automobile parts in combination with the shape contour and the area to which the components belong.

[0067] In one embodiment, the preliminary manuscript is used as an input image and imported into a pre-trained image recognition model. The model uses deep learning technology to recognize basic shapes and features in the image. The model pre-processes the input image, including denoising, contrast enhancement, etc., to improve the accuracy of subsequent shape recognition.

[0068] Then, the preprocessed image is processed using an edge detection algorithm, such as Canny edge detection, to identify the edges of the shapes in the manuscript. The edge detection algorithm determines the edge position of the shape by calculating the gradient change of the pixel values ​​in the image. The identified edges are further processed using a graph segmentation algorithm, such as the GrabCut algorithm, to segment the shape outlines in the manuscript. The graph segmentation algorithm achieves accurate segmentation of the shape by analyzing the differences in color, texture and other features between the edge and the background.

[0069] After that, the segmented shape contour is matched with a preset automobile parts library. The parts library contains shape contour templates of various automobile parts, as well as their corresponding part names and regions, such as the front, rear, roof, body, etc. By calculating the similarity between the shape contour and the parts library template, such as using a shape context matching algorithm, the automobile parts and regions to which each shape contour corresponds are determined.

[0070] The geometric features of each automobile component are calculated by combining the shape outline and the region information to which the component belongs. It should be noted that the geometric features in the embodiment of the present application include the size (length, width, height, etc.) and shape (circular, rectangular, curved, etc.) of the component and the relative position relationship with other components. Finally, the parsed component description information (including component name, region, geometric features, etc.) is output in a structured format.

[0071] In one embodiment of the present application, obtaining the to-be-assembled accessories of the automobile model according to the component description information specifically includes:

[0072] According to the preset mapping relationship, multiple accessories to be selected corresponding to the area to which the component belongs are determined in the preset accessory style library;

[0073] According to the shape outline in the component description information, determine whether there is a target to-be-assembled component corresponding to the area to which the component belongs from the components to be selected;

[0074] If yes, then obtaining at least one target to-be-assembled accessory corresponding to the shape outline;

[0075] If not, the corresponding parts to be assembled are constructed according to the area to which the parts belong in the part description information and the characteristic points in the shape contour, and the corresponding sizes of the constructed parts to be assembled are adjusted according to the geometric features of the parts in the part description information to generate the target parts to be assembled.

[0076] In one embodiment, in the automobile design project, a set of preset mapping relationships has been established, which defines the correspondence between the regions to which the components belong (such as the front, rear, roof, body, etc.) and the components in the preset component style library. The preset component style library contains 3D models of various automobile components, which are classified and stored according to the regions to which the components belong.

[0077] According to the preset mapping relationship, multiple accessories to be selected corresponding to the area to which the component belongs are searched in the preset accessory style library. For example, if the area to which the component belongs is the front of the vehicle, all accessories related to the front of the vehicle are selected from the accessory style library as accessories to be selected. The shape outline in the component description information is matched with the accessory to be selected. The matching process can be achieved by calculating the similarity between the shape outline and the accessory outline, and the similarity can be calculated using a shape context matching algorithm or other shape matching algorithms. If there is an accessory to be selected that is completely matched or highly similar to the shape outline, it is determined as the target accessory to be assembled.

[0078] If no accessories matching the shape profile are found in the accessories to be selected, it is necessary to construct a new accessory to be assembled based on the component description information. The construction process can be achieved through 3D modeling software, and the basic accessory shape is constructed based on the area to which the component belongs and the feature points in the shape profile. After the construction is completed, the size of the constructed accessory to be assembled is adjusted according to the component geometric features in the component description information. The adjustment process can be achieved through operations such as scaling and stretching to ensure that the generated accessory to be assembled meets the design requirements. After the matching or construction and adjustment steps, the target accessory to be assembled that meets the design requirements is finally generated. These accessories can be stored in the form of 3D models and used in subsequent automotive assembly simulation, rendering display and other links.

[0079] 102. According to the types of parts corresponding to the parts to be assembled and the size ratio relationship between the parts, the parts to be assembled are arranged and combined to obtain several combinations of parts corresponding to the complete car shape, and several cars to be shaped corresponding to the several combinations of parts are generated according to the surface shape.

[0080] In one embodiment of the present application, according to the types of accessories corresponding to the accessories to be assembled and the size ratio relationship between the components, the accessories to be assembled are arranged and combined to obtain several accessory combinations corresponding to the complete car shape, and several cars to be shaped corresponding to the several accessory combinations are generated according to the curved surface shape, specifically including:

[0081] Classify all the parts to be assembled, determine the type of parts corresponding to each part to be assembled, and determine the detailed information of the parts to be assembled in the preset parts style library; wherein the types of parts at least include: vehicle body, door, window, rearview mirror and tire, and the detailed information of parts includes size, shape, curvature of curved surface and curvature direction of curved surface;

[0082] According to the characteristics of automobile styling, the types of accessories required for each automobile to be styled and the number of accessories required for each type of accessories are determined, and all the accessories to be assembled are arranged and combined according to the size ratio relationship between the components in the component description information and the detailed information of each component to be assembled, so as to generate several combinations of accessories corresponding to several complete automobile styling;

[0083] For each of the plurality of accessory combinations, according to the accessory detailed information, the accessory parameters corresponding to each to-be-assembled accessory in the accessory combination are determined, and according to the connection mode between the adjacent accessories, the accessory parameters of the adjacent accessories in the same accessory combination are adjusted; wherein the accessory parameters at least include the length, width, height and curvature of the curve of the vehicle body;

[0084] Through the surface modeling algorithm, multiple automobile surfaces corresponding to the combination of accessories are generated according to the adjusted accessory parameters, and the multiple automobile surfaces are spliced ​​to obtain the corresponding automobile to be modeled.

[0085] In one embodiment, all the parts to be assembled are classified according to their functions and positions, which is helpful for the orderly combination and parameter adjustment of subsequent parts, including but not limited to the types of parts such as car body, door, window, rearview mirror and tire. In the preset parts style library, for each part to be assembled, the detailed information of the part is searched and determined, including key parameters such as size (length, width, height, etc.), shape, curvature of the surface and curvature direction of the surface.

[0086] According to the characteristics of the car's styling, determine the types and quantities of accessories required for each car to be styled. Combine the size ratios between the parts in the part description information and the detailed information of each part to be assembled, and arrange and combine all the parts to be assembled. Through algorithm optimization, generate several combinations of accessories corresponding to the complete car styling that meet the design requirements.

[0087] For each accessory combination, the initial accessory parameters of each accessory to be assembled are determined based on the accessory detailed information. These parameters include but are not limited to the length, width, height of the vehicle body, and the curvature of the curve. According to the connection method between adjacent accessories (such as hinge connection, bolt connection, etc.), the accessory parameters of adjacent accessories in the same accessory combination are adjusted to ensure that the connection between accessories is smooth and meets the actual assembly requirements.

[0088] Apply surface modeling algorithms (such as NURBS surface algorithms, subdivision surface algorithms, etc.) to generate multiple car surfaces corresponding to each accessory combination according to the adjusted accessory parameters. These surfaces should have a high degree of smoothness and continuity to ensure that the final generated car model has a realistic appearance and texture. The generated multiple car surfaces are spliced ​​to obtain the corresponding three-dimensional model of the car to be modeled. The splicing process may involve smooth transition processing between surfaces to ensure the coherence and consistency of the entire car model.

[0089] 103. Detect each car to be shaped, determine the car to be shaped with parts to be optimized, and identify the abnormality type corresponding to the parts to be optimized; wherein the abnormality type includes unsmooth styling and incompatible specifications.

[0090] In one embodiment of the present application, each car to be styled is detected, the car to be styled having a part to be optimized is determined, and the abnormal type corresponding to the part to be optimized is identified, specifically including:

[0091] For each car to be styled, the geometric features of the car to be styled are extracted from the three-dimensional model of the car to be styled, and a curvature map corresponding to the three-dimensional model is generated; wherein the geometric features of the car to be styled include: curvature, radius, angle and slope;

[0092] According to the curvature change in the curvature map, the transition area in the three-dimensional model is identified, and combined with the transition area, it is determined whether the continuity of the same accessory surface in the three-dimensional model is abnormal;

[0093] If yes, it is determined that the to-be-modeled vehicle has a part to be optimized, and the abnormality type corresponding to the to-be-optimized part is that the modeling is not smooth; if no, it is determined whether the connection between adjacent parts matches according to the parameters of the parts in the to-be-modeled vehicle and the size ratio relationship between the parts;

[0094] In the case of connection mismatch, it is determined that the vehicle to be styled has a part to be optimized, and the abnormality type corresponding to the part to be optimized is specification mismatch; wherein, the connection matching is used to determine whether the gap between the accessories meets the preset gap and whether adjacent accessories are aligned.

[0095] In one embodiment, for each 3D model of the car to be modeled, professional 3D modeling software or algorithm is used to extract its geometric features, including key parameters such as curvature, radius, angle and slope. These features can reflect the shape, contour and surface changes of the car model. Based on the extracted geometric features, a curvature map corresponding to the 3D model is generated. The curvature map is a visualization tool that can intuitively display the curvature changes of the model surface, which is helpful for subsequent analysis and optimization.

[0096] The curvature change information in the curvature map is used to identify the transition area in the 3D model. The transition area is usually located at the junction of different surfaces or accessories, and is the part of the car that is prone to unsmoothness or breakage in the car styling. Combined with the information of the transition area, the continuity of the same accessory surface in the 3D model is checked. By comparing parameters such as the curvature change and radius difference between adjacent surfaces, it is determined whether the surface is continuous and smooth. If the same accessory surface is found to have continuity anomalies (such as sudden changes in curvature, inconsistent radii, etc.), it is determined that there are parts to be optimized in the car to be styled, and the anomaly type is unsmooth styling.

[0097] After confirming that the surface continuity is normal, the connection between adjacent accessories is verified for matching according to the accessory parameters in the car to be modeled and the size ratio relationship between the components. The connection matching verification includes two aspects: one is to check whether the gap between the accessories is consistent with the preset gap to ensure that there will be no gap that is too large or too small during the assembly process; the other is to check whether the adjacent accessories are aligned to ensure that the assembled car model has a neat appearance and a stable structure. If the connection between adjacent accessories is found to be mismatched (such as too large gaps, misaligned accessories, etc.), it is determined that there are parts to be optimized in the car to be modeled, and the abnormality type is specification mismatch.

[0098] Develop corresponding optimization plans for the identified parts to be optimized and their abnormal types. For parts with unsmooth shapes, they can be improved by adjusting surface parameters, adding transition surfaces, etc. For parts with inappropriate specifications, it is necessary to redesign the size of accessories, adjust the assembly method, etc. The optimization plan is applied to the 3D model, and the steps of geometric feature extraction, curvature map generation, transition area identification, continuity check, and connection matching verification are re-performed to ensure that the optimized car model meets the design requirements and has a high degree of manufacturing feasibility.

[0099] 104. Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, so as to determine the compensation value corresponding to the part to be optimized according to the deviation degree.

[0100] In one embodiment of the present application, the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized are determined according to the abnormality type, and the deviation degree of the part to be optimized is calculated in combination with the ideal coordinate information of the target point to be optimized, specifically including:

[0101] For the part to be optimized whose abnormal type is unsmooth modeling, according to the curvature change in the curvature map, the target abnormal area of ​​the part to be optimized is identified in the transition area and the inflection point of the surface corresponding to the part to be optimized, and the target point to be optimized is determined in the target abnormal area;

[0102] Determine the ideal coordinate information of the target point to be optimized according to the component description information, and measure the actual coordinate information of the target point to be optimized in the three-dimensional model of the vehicle to be modeled by a three-dimensional coordinate measuring machine;

[0103] Determine the difference between the actual coordinate information and the ideal coordinate information in each dimension of the three-dimensional model, calculate the sum of the squares of the differences in each dimension, and take the square root of the sum of the squares in each dimension to obtain the degree of deviation of the part to be optimized;

[0104] According to the actual coordinate information, the ideal coordinate information and the preset compensation factor of each dimension, the deviation component corresponding to the deviation degree in each dimension is calculated.

[0105] In one embodiment, for the part to be optimized with an abnormal type of unsmooth modeling, firstly, according to the curvature change information in the curvature map, the target abnormal area is carefully identified in the transition area and the inflection point of the surface. These areas usually show characteristics such as sudden change of curvature, discontinuous surface or too many inflection points. In the target abnormal area, the target points to be optimized are further determined. These points are usually the key positions that cause unsmooth modeling, such as inflection points of the surface, extreme points of curvature, etc.

[0106] Based on the component description information, determine the ideal coordinate information of the target point to be optimized. Use a high-precision coordinate measuring machine (CMM) to accurately measure the actual coordinate information of the target point to be optimized in the three-dimensional model of the car to be styled. The three-dimensional coordinate measuring machine can provide high-precision spatial coordinate measurement to ensure the accuracy of the data.

[0107] Determine the difference between the actual coordinate information and the ideal coordinate information in each dimension (X, Y, and Z axis) of the 3D model. These differences reflect the deviation between the actual position of the part to be optimized in space and the designed position. Calculate the sum of the squares of the differences in each dimension and take the square root of the sum of squares to obtain the overall deviation of the part to be optimized. This step uses the concept of Euclidean distance, which can intuitively reflect the deviation of the part to be optimized in 3D space.

[0108] According to the actual coordinate information, ideal coordinate information and preset compensation factor of each dimension, the deviation component corresponding to the degree of deviation in each dimension is calculated. The preset compensation factor is an adjustment coefficient preset based on design experience and manufacturing tolerances, which is used to adjust the calculation results of the deviation component to make it more consistent with the actual situation. The calculation of the deviation component helps designers to have a deeper understanding of the deviation of the part to be optimized in each dimension.

[0109] Based on the calculation results of the degree of deviation and the deviation component, a detailed optimization plan is formulated. The optimization plan may include adjusting surface parameters, adding transition surfaces, modifying the size of accessories, and other measures. The optimization plan is applied to the 3D model, and the steps of curvature map analysis, three-coordinate measurement, and degree of deviation calculation are re-performed to ensure that the optimized car model meets the design requirements and has a high degree of manufacturing feasibility. This can accurately identify and quantify the target abnormal area and its degree of deviation of the part to be optimized, which not only improves the accuracy and quality of the car design, but also ensures that the final generated car model has a high degree of manufacturing feasibility and market competitiveness.

[0110] In one embodiment of the present application, the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized are determined according to the abnormality type, and the deviation degree of the part to be optimized is calculated in combination with the ideal coordinate information of the target point to be optimized, specifically including:

[0111] For the to-be-optimized parts whose abnormal type is incompatible specifications, the actual sizes of the to-be-optimized parts and parts connected to the to-be-optimized parts are obtained through the three-dimensional model of the to-be-optimized car;

[0112] Determine the size ratio between the part to be optimized and the connected part according to the component description information, and determine whether the part to be optimized needs to be resized in combination with the actual sizes of the part to be optimized and the connected part;

[0113] If yes, the deviation between the actual size of the part to be optimized and the ideal size is calculated; if no, the actual coordinate information of the connection points between the part to be optimized and the connected parts is measured by a three-dimensional coordinate measuring machine in the three-dimensional model of the car to be shaped;

[0114] Determine the ideal coordinate information of the connection point between the part to be optimized and the connected part, and calculate the degree of deviation of the connection point based on the actual coordinate information and the ideal coordinate information.

[0115] In one embodiment, first, the actual size of the part to be optimized and the part connected to the part to be optimized is accurately obtained through the three-dimensional model of the car to be shaped. This includes key dimensional parameters such as length, width, and height. Then, based on the component description information, the size ratio relationship between the part to be optimized and the connected part is determined. These proportional relationships are usually determined in the early stages of design and serve as a benchmark for subsequent design and manufacturing. Combined with the actual size and size ratio relationship between the part to be optimized and the connected part, a comprehensive analysis is performed to determine whether the part to be optimized needs to be resized. If the actual size does not match the preset ratio relationship, or does not match the size of the connected part, size adjustment is required.

[0116] If it is determined that the part to be optimized needs to be sized, the degree of deviation between the actual size of the part to be optimized and the ideal size is further calculated. This can be achieved by comparing the actual size with the preset ideal size (usually derived from the CAD model at the beginning of the design or verified design data). The degree of deviation can be calculated by a variety of methods, such as directly calculating the difference, calculating the relative error, etc. The specific method to be selected depends on factors such as design requirements and manufacturing tolerances.

[0117] If it is determined that the part to be optimized does not need to be resized, or after resizing, the accuracy of the connection point still needs to be further verified, the actual coordinate information of the connection point between the part to be optimized and the connected part is measured in the three-dimensional model of the car to be styled by a high-precision three-dimensional coordinate measuring machine (CMM). At the same time, the ideal coordinate information of the connection point is determined based on the component description information or design data. These ideal coordinate information usually represent the expected position at the beginning of the design. Based on the actual coordinate information and the ideal coordinate information, the degree of deviation of the connection point is calculated. This can be achieved by comparing the difference between the two coordinate points in each dimension (X, Y, Z axis).

[0118] 105. The parts to be optimized are adjusted according to the compensation values ​​to obtain an optimized car to be shaped, and the optimized car to be shaped is rendered to generate a three-dimensional model corresponding to the optimized car to be shaped, thereby realizing automatic styling of the car.

[0119] Specifically, in one embodiment of the present application, rendering the optimized car to be styled to dynamically display the three-dimensional model corresponding to the optimized car to be styled includes:

[0120] Export the optimized car to be styled to the rendering engine, and select the material texture corresponding to each accessory of the car to be styled in the rendering engine in combination with the car styling display intent corresponding to the component description information; the material texture is used to simulate the car paint color and metallic luster;

[0121] Select a lighting environment for the car to be modeled in the rendering engine, simulate the real scene through the lighting environment, and render the car to be modeled according to preset rendering parameters;

[0122] Create an animation for the car to be modeled, render the corresponding animation sequence through a rendering engine, and encode the rendered animation to dynamically display the animation in a video format; wherein the animation is used to simulate the movement of the car.

[0123] In one embodiment, after the optimization and adjustment of the car to be styled is completed, it is exported to a format recognizable by the rendering engine (such as OBJ, FBX, etc.) and imported into the rendering engine. In combination with the car styling display intent corresponding to the component description information, appropriate material textures are selected for each accessory of the car to be styled in the rendering engine. These material textures should be able to accurately simulate the appearance characteristics of the paint color, metallic luster, and other materials such as plastic and glass. When selecting material textures, the reflection, refraction, and shadow effects between different materials must also be considered to ensure the authenticity and three-dimensionality of the rendering results.

[0124] Select a suitable lighting environment for the car to be modeled in the rendering engine. The lighting environment should be able to simulate the lighting conditions in real scenes, including natural light and artificial light. By adjusting the position, intensity, color and other parameters of the light source, different visual effects and atmospheres can be created. At the same time, the impact of lighting on material textures must also be considered to ensure the accuracy and authenticity of the rendering results.

[0125] Render the modeled car according to the preset rendering parameters (such as resolution, sampling quality, anti-aliasing, etc.). The rendering process will generate high-quality images or video frames to show the appearance and details of the car. During the rendering process, you also need to pay close attention to the rendering progress and results, and adjust the rendering parameters and material textures in time to ensure rendering quality and efficiency.

[0126] Create animations for the car to be modeled to simulate the car's motion. Animations can include simple rotations, translations, and other actions, or they can include complex driving scenes and interactive effects. Render the corresponding animation sequence in the rendering engine. The rendering process will generate a series of continuous images or video frames to show the dynamic effects of the car.

[0127] Encode the rendered animation and convert it into common video formats (such as MP4, AVI, etc.). The encoding process should ensure the quality and smoothness of the video. Dynamically display the encoded animation through video playback software or online platforms. The display process can include functions such as loop playback and slow motion playback, so that customers can have a more comprehensive understanding of the appearance and performance of the car.

[0128] The above is an embodiment of the method proposed in this application. Based on the same inventive concept, the embodiment of this application also provides an automobile automatic modeling device based on machine learning, and its structure is as follows: Figure 2 shown.

[0129] Figure 2 The internal structure diagram of an automobile automated styling device based on machine learning provided in an embodiment of the present application. Figure 2 As shown, the device includes:

[0130] at least one processor;

[0131] and, a memory communicatively coupled to the at least one processor;

[0132] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor to enable the at least one processor to:

[0133] Obtaining component description information in the preliminary draft of the automobile modeling, and obtaining the to-be-assembled accessories of the automobile modeling according to the component description information; wherein the component description information includes the component belonging area corresponding to each automobile component, the component shape outline, the component geometric features, and the size ratio relationship between the components;

[0134] According to the types of parts corresponding to the parts to be assembled and the size ratio relationship between the parts, the parts to be assembled are arranged and combined to obtain several combinations of parts corresponding to the complete car shape, and several cars to be shaped corresponding to the several combinations of parts are generated according to the surface shape;

[0135] Detect each car to be shaped, determine the car to be shaped with parts to be optimized, and identify the abnormality type corresponding to the parts to be optimized; wherein the abnormality type includes unsmooth styling and incompatible specifications;

[0136] Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, so as to determine the compensation value corresponding to the part to be optimized according to the deviation degree;

[0137] The part to be optimized is adjusted according to the compensation value to obtain an optimized car to be shaped, and the optimized car to be shaped is rendered to generate a three-dimensional model corresponding to the optimized car to be shaped, thereby realizing automatic styling of the car.

[0138] The present application also provides a non-volatile computer storage medium storing computer executable instructions. When the computer executable instructions are executed, they can:

[0139] Obtaining component description information in the preliminary draft of the automobile modeling, and obtaining the to-be-assembled accessories of the automobile modeling according to the component description information; wherein the component description information includes the component belonging area corresponding to each automobile component, the component shape outline, the component geometric features, and the size ratio relationship between the components;

[0140] According to the types of parts corresponding to the parts to be assembled and the size ratio relationship between the parts, the parts to be assembled are arranged and combined to obtain several combinations of parts corresponding to the complete car shape, and several cars to be shaped corresponding to the several combinations of parts are generated according to the surface shape;

[0141] Detect each car to be shaped, determine the car to be shaped with parts to be optimized, and identify the abnormality type corresponding to the parts to be optimized; wherein the abnormality type includes unsmooth styling and incompatible specifications;

[0142] Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, so as to determine the compensation value corresponding to the part to be optimized according to the deviation degree;

[0143] The part to be optimized is adjusted according to the compensation value to obtain an optimized car to be shaped, and the optimized car to be shaped is rendered to generate a three-dimensional model corresponding to the optimized car to be shaped, thereby realizing automatic styling of the car.

[0144] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0145] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for automatic modeling of an automobile based on machine learning, characterized in that: The method comprises: Obtaining component description information in the preliminary draft of the automobile modeling, and obtaining the to-be-assembled accessories of the automobile modeling according to the component description information; wherein the component description information includes the component belonging area corresponding to each automobile component, the component shape outline, the component geometric features, and the size ratio relationship between the components; According to the types of parts corresponding to the parts to be assembled and the size ratio relationship between the parts, the parts to be assembled are arranged and combined to obtain several combinations of parts corresponding to the complete car shape, and several cars to be shaped corresponding to the several combinations of parts are generated according to the surface shape; Detecting each car to be shaped, determining the car to be shaped with a part to be optimized, and identifying the abnormality type corresponding to the part to be optimized; wherein the abnormality type includes unsmooth styling and incompatible specifications; Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, so as to determine the compensation value corresponding to the part to be optimized according to the deviation degree; The part to be optimized is adjusted according to the compensation value to obtain an optimized car to be shaped, and the optimized car to be shaped is rendered to dynamically display the three-dimensional model corresponding to the optimized car to be shaped, thereby realizing automatic styling of the car.

2. The method for automatic modeling of an automobile based on machine learning according to claim 1, characterized in that: Obtain component description information from the preliminary draft of the car model, including: Obtaining a predetermined preliminary draft of a car shape, and inputting the preliminary draft into an image recognition model for analysis, to determine a plurality of shapes in the preliminary draft; For each of the multiple shapes, identifying the corresponding shape edge in the preliminary manuscript by an edge detection algorithm, and segmenting the corresponding shape contour in the preliminary manuscript according to the shape edge and by a graph segmentation algorithm; According to the matching degree between the segmented shape contour and the preset automobile component, the component region to which the shape contour corresponds to the automobile component belongs is determined; wherein the component region includes: the front, rear, roof and body of the vehicle; According to the preliminary manuscript, the size ratio relationship between the components corresponding to the design intention is determined, and the component geometric features corresponding to the automobile component are calculated in combination with the shape contour and the area to which the component belongs.

3. The method for automatic modeling of an automobile based on machine learning according to claim 1, characterized in that: Obtaining the to-be-assembled accessories of the automobile shape according to the component description information specifically includes: According to the preset mapping relationship, multiple accessories to be selected corresponding to the area to which the component belongs are determined in the preset accessory style library; According to the shape outline in the component description information, determining whether there is a target to-be-assembled component corresponding to the region to which the component belongs from the components to be selected; If yes, obtaining at least one target to-be-assembled accessory corresponding to the shape outline; If not, construct the corresponding accessories to be assembled according to the area to which the components belong in the component description information and the characteristic points in the shape contour, and adjust the corresponding size of the constructed accessories to be assembled according to the component geometric features in the component description information to generate the target accessories to be assembled.

4. The method for automatic modeling of an automobile based on machine learning according to claim 1, characterized in that: According to the types of parts to be assembled and the size ratio between the parts, the parts to be assembled are arranged and combined to obtain several combinations of parts corresponding to the complete car shape, and several cars to be shaped corresponding to several combinations of parts are generated according to the surface shape, including: Classify all the parts to be assembled, determine the type of parts corresponding to each part to be assembled, and determine the detailed information of the parts to be assembled in the preset parts style library; wherein the types of parts at least include: vehicle body, door, window, rearview mirror and tire, and the detailed information of parts includes size, shape, curvature of curved surface and curvature direction of curved surface; According to the characteristics of the automobile shape, the types of accessories required for each automobile to be shaped and the number of accessories required for each type of accessories are determined, and in combination with the size ratio relationship between the components in the component description information and the detailed information of each component to be assembled, all the components to be assembled are arranged and combined to generate a number of combinations of accessories corresponding to a number of complete automobile shapes; For each of the plurality of accessory combinations, according to the accessory detailed information, the accessory parameters corresponding to each to-be-assembled accessory in the accessory combination are determined, and according to the connection mode between adjacent accessories, the accessory parameters of adjacent accessories in the same accessory combination are adjusted; wherein the accessory parameters at least include the length, width, height and curvature of the curve of the vehicle body; Through the surface modeling algorithm, according to the adjusted accessory parameters, a plurality of automobile surfaces corresponding to the accessory combination are generated, and the plurality of automobile surfaces are spliced ​​to obtain the corresponding automobile to be modeled.

5. The method for automatic modeling of an automobile based on machine learning according to claim 1, characterized in that: Each car to be shaped is tested to determine the car to be shaped with a part to be optimized, and the abnormal type corresponding to the part to be optimized is identified, specifically including: For each car to be styled, extract the geometric features of the car to be styled from the three-dimensional model of the car to be styled, and generate a curvature map corresponding to the three-dimensional model; wherein the geometric features of the car to be styled include: curvature, radius, angle and slope; According to the curvature change in the curvature map, a transition area in the three-dimensional model is identified, and in combination with the transition area, whether the continuity of the same accessory surface in the three-dimensional model is abnormal is determined; If yes, it is determined that the vehicle to be styled has a part to be optimized, and the abnormality type corresponding to the part to be optimized is unsmooth styling; if no, it is determined whether the connection between adjacent accessories matches according to the accessory parameters in the vehicle to be styled and the size ratio relationship between the components; In the case of connection mismatch, it is determined that the vehicle to be styled has a part to be optimized, and the abnormality type corresponding to the part to be optimized is specification mismatch; wherein the connection matching is used to determine whether the accessory gap meets the preset gap and whether adjacent accessories are aligned.

6. The method for automatic modeling of an automobile based on machine learning according to claim 1, characterized in that: Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, specifically including: For the part to be optimized whose abnormal type is unsmooth modeling, according to the curvature change in the curvature map, the target abnormal area of ​​the part to be optimized is identified in the transition area and the surface inflection point corresponding to the part to be optimized, and the target point to be optimized is determined in the target abnormal area; Determine the ideal coordinate information of the target point to be optimized according to the component description information, and measure the actual coordinate information of the target point to be optimized in the three-dimensional model of the vehicle to be modeled by a three-dimensional coordinate measuring machine; Determine the difference between the actual coordinate information and the ideal coordinate information in each dimension of the three-dimensional model, calculate the sum of the squares of the differences in each dimension, and take the square root of the sum of the squares in each dimension to obtain the degree of deviation of the part to be optimized; According to the actual coordinate information, the ideal coordinate information and the preset compensation factor of each dimension, the deviation component corresponding to the deviation degree in each dimension is calculated.

7. The method for automatic modeling of an automobile based on machine learning according to claim 1, characterized in that: Determine the target point to be optimized in the part to be optimized and the actual coordinate information of the target point to be optimized according to the abnormality type, and calculate the deviation degree of the part to be optimized in combination with the ideal coordinate information of the target point to be optimized, specifically including: For the part to be optimized whose abnormal type is incompatible specifications, the actual size of the part to be optimized and the part connected to the part to be optimized is obtained through the three-dimensional model of the car to be shaped; Determine the size ratio between the part to be optimized and the connected part according to the component description information, and determine whether the part to be optimized needs to be resized in combination with the actual sizes of the part to be optimized and the connected part; If yes, then the degree of deviation between the actual size of the part to be optimized and the ideal size is calculated; if no, then the actual coordinate information of the connection point between the part to be optimized and the connected part is measured in the three-dimensional model of the automobile to be shaped by a three-dimensional coordinate measuring machine; Determine the ideal coordinate information of the connection point between the part to be optimized and the connected part, and calculate the degree of deviation of the connection point according to the actual coordinate information and the ideal coordinate information.

8. The method for automatic automobile modeling based on machine learning according to claim 1, characterized in that: Rendering the optimized car to be shaped to dynamically display the three-dimensional model corresponding to the optimized car to be shaped, specifically including: Exporting the optimized car to be styled to a rendering engine, and selecting material textures corresponding to the various accessories of the car to be styled in the rendering engine in combination with the car styling display intent corresponding to the component description information; wherein the material textures are used to simulate the color of car paint and metallic luster; Selecting a lighting environment for the car to be styled in the rendering engine, simulating a real scene through the lighting environment, and rendering the car to be styled according to preset rendering parameters; Animation is produced for the car to be modeled, a corresponding animation sequence is rendered by the rendering engine, and the rendered animation is encoded so as to dynamically display the animation in a video format; wherein the animation is used to simulate the movement of the car.

9. An automobile automated modeling device based on machine learning, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; Wherein, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the automatic automobile styling method based on machine learning as described in any one of claims 1-8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: When the computer executable instructions are executed, an automatic automobile modeling method based on machine learning as described in any one of claims 1 to 8 is implemented.

Images