A method and system for generating a vehicle three-dimensional model with the aid of a large language model

Abstract

This invention discloses a method for generating a 3D vehicle model assisted by a large language model, comprising: acquiring sample vehicle structural 3D model information; inputting the image to be parsed, a sample image, and a first prompt word into the large language model, and outputting the target parsing text description information of the image to be parsed; inputting a second prompt word into the large language model, and outputting the text description information of the target sample vehicle structural 3D model that matches the target parsing text description information; determining the corresponding size data of the target sample vehicle structural 3D model from the sample vehicle structural 3D model information, and generating a third prompt word; inputting the third prompt word into the large language model, and outputting the vehicle structural coordinates; inputting the vehicle structural coordinates into a 3D modeling tool to generate the model of the target vehicle. This invention, based on a large language model, can save modeling time and improve modeling efficiency. This invention also discloses a system, electronic device, and computer-readable storage medium for implementing the above method.

Description

Technical Field

[0001] The present invention relates to the technical field of large language models, and in particular to a method, system, electronic device and storage medium for generating a three-dimensional vehicle model with the assistance of a large language model. Background Art

[0002] In recent years, the field of artificial intelligence, especially natural language processing (NLP) and machine learning (ML), has made remarkable progress. Among these advances, language models, especially large pre-trained language models such as ChatMindAi (Generative Pretrained Transformer) series and BERT (Bidirectional Encoder Representations from Transformers), have become the core technology for processing and understanding natural language.

[0003] In the modeling work for automobiles, modelers make models based on object pictures given by users. They need to find common model modifications or rebuild the model from white film. Modelers have to waste a lot of time looking for common model modifications or rebuilding the model from white film each time, which leads to low modeling efficiency. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a method, system, electronic device and storage medium for generating a three-dimensional vehicle model with the assistance of a large language model. The technical problem to be solved by the present invention is achieved through the following technical solutions:

[0005] A first aspect of an embodiment of the present invention provides a method for generating a three-dimensional vehicle model with the assistance of a large language model, comprising the following steps:

[0006] A method for generating a three-dimensional vehicle model with the aid of a large language model, characterized by comprising the following steps:

[0007] Acquire sample vehicle structure three-dimensional model information; wherein the sample vehicle structure three-dimensional model information includes: text description information of each sample vehicle structure three-dimensional model and size data of each sample vehicle structure three-dimensional model;

[0008] Inputting the image to be parsed, the example image and the first prompt word into the large language model, the large language model outputting the target parsed text description information of the image to be parsed; wherein the image to be parsed is obtained by splicing multiple images of the target vehicle, and the first prompt word includes a text description of the vehicle in the image to be parsed and the parsed text description information of the example image;

[0009] Inputting a second prompt word into the large language model, the large language model outputting text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information; wherein the second prompt word includes task instruction information, target parsed text description information and text description information of the sample vehicle structure three-dimensional model;

[0010] Determining the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generating a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information;

[0011] Inputting the third prompt word into the large language model, and the large language model outputs vehicle structure coordinates;

[0012] The vehicle structure coordinates are input into a three-dimensional modeling tool to generate a model of the target vehicle.

[0013] In one embodiment of the present invention, in the dimension data of the sample vehicle structure three-dimensional models, the sample vehicle structure three-dimensional models of the same type have the same proportions.

[0014] In one embodiment of the present invention, the specific content of the text description of the vehicle in the image to be analyzed is: respectively describing the wheel structure, vehicle design, exterior finish, interior layout, vehicle logo, window features and rearview mirror features in the image to be analyzed; wherein the image to be analyzed is a composite image of multiple angles of the target vehicle, and the text description of the vehicle in the image to be analyzed includes the appearance description of each structure of the vehicle;

[0015] The parsed text description information of the example image includes:

[0016] Descriptive information about the wheel structure’s hub material, hub style, hub structure, hub performance, and tire performance;

[0017] Descriptive information on the vehicle model lines, body contour shape, front end structure, side shape, and rear end lines and structure of the vehicle model design;

[0018] Description of the process, color and corresponding effects of the exterior finish;

[0019] Descriptive information about the style of the interior layout, the location and type of the instrument panel, the material and performance of the seats, and the structure and performance of the center console area;

[0020] Descriptive information about the shape, meaning, style, location and craftsmanship of the logo that identifies the vehicle;

[0021] Descriptive information on window features, such as the material of the window, the performance of the window lift system, and the sealing structure and performance around the window;

[0022] The features of the rearview mirror include the mirror's function, mirror structure, corresponding control system, shell color and lines.

[0023] In one embodiment of the present invention, it also includes:

[0024] generating a second prompt word according to the task instruction information, the target parsed text description information and the text description information of the sample vehicle structure three-dimensional model;

[0025] The second prompt word includes:

[0026] The text description information of the sample vehicle structure three-dimensional model is searched for the text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information.

[0027] In one embodiment of the present invention, the third prompt word includes:

[0028] The target sample vehicle structure three-dimensional model includes dimensional data, assembly task information, precautions and example information; wherein the precautions include: the origin is the center point of the vehicle, the X-axis is horizontal to the right, indicating the front and rear direction of the vehicle, the Z-axis is vertically upward, indicating the up and down direction of the vehicle, and the Y-axis is the out-of-screen direction, indicating the left and right direction of the vehicle.

[0029] In one embodiment of the present invention, the assembly task information includes: assembling a complete vehicle according to the size data of the three-dimensional model of the target sample vehicle structure in accordance with the arrangement of x, y, and z coordinates, with the center point of the vehicle as the origin of the coordinate system, and outputting the coordinates of each structure.

[0030] In one embodiment of the present invention, the example information includes: example dimension data, thought process description information, and output results of the example dimension data;

[0031] The thought process description information includes:

[0032] The center point of the vehicle is taken as the origin, and the front-to-back and left-to-right symmetry axes of the vehicle are along the Z axis and the X axis respectively;

[0033] The vehicle design and exterior finish are identical in size, with the body covering the wheels and extending beyond them, with the bottom in contact with the wheels;

[0034] The center point of the car logo is at the same height as the center point of the vehicle;

[0035] The rearview mirrors are installed on both sides of the vehicle, where the rearview mirror is located near the A-pillar;

[0036] Determine boundary values ​​of the length, width and height of the vehicle based on the example dimension data;

[0037] The center point coordinates of the wheels are set according to the width and length of the vehicle.

[0038] The coordinates of the vehicle model design are the coordinates of the center point of the vehicle;

[0039] The exterior finishes are designed to the same coordinates as the model;

[0040] The coordinates of the car logo are the coordinates of the center point of the car logo;

[0041] The coordinates of the rearview mirror are the coordinates of the center point of each rearview mirror.

[0042] A second aspect of an embodiment of the present invention provides a system for generating a three-dimensional vehicle model with the assistance of a large language model, comprising:

[0043] An acquisition module is used to acquire sample vehicle structure three-dimensional model information; the sample vehicle structure three-dimensional model information includes: text description information of each sample vehicle structure three-dimensional model and size data of each sample vehicle structure three-dimensional model;

[0044] A first input module is used to input the image to be analyzed, the example image and the first prompt word into the large language model; wherein the image to be analyzed is obtained by splicing multiple images of the target vehicle, and the first prompt word includes a text description of the vehicle in the image to be analyzed and the analysis text description information of the example image;

[0045] A large language model, used to output target parsed text description information of the image to be parsed;

[0046] The first input module is further used to input a second prompt word into the large language model; wherein the second prompt word includes task instruction information, target parsed text description information and text description information of the sample vehicle structure three-dimensional model;

[0047] The large language model is further used for the large language model to output text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information;

[0048] A determination module, configured to determine the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generate a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information;

[0049] The first input module is further used to input the third prompt word into the large language model;

[0050] The large language model is also used for the large language model to output vehicle structure coordinates;

[0051] A second input module, used to input the vehicle structure coordinates into a three-dimensional modeling tool;

[0052] 3D modeling tool, used to generate a model of the target vehicle.

[0053] A third aspect of an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for generating a three-dimensional vehicle model with the assistance of a large language model provided in the first aspect of an embodiment of the present invention is implemented.

[0054] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for generating a vehicle three-dimensional model with the assistance of a large language model provided by the first aspect of an embodiment of the present invention is implemented.

[0055] Beneficial effects of the present invention:

[0056] The present invention is based on a large language model and can quickly find the data of the model to be modeled in a large number of known models, and generate the coordinate data required for modeling from these data through the large language model, and then input the coordinate data into a three-dimensional modeling application to complete the modeling, thereby saving modeling time and improving modeling efficiency.

[0057] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0058] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0060] Figure 1 A schematic diagram of a flow chart of a method for generating a vehicle three-dimensional model with the assistance of a large language model provided by an embodiment of the present invention;

[0061] Figure 2 A block diagram of a system for generating a three-dimensional vehicle model using a large language model assisted by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The present invention is further described in detail below with reference to specific embodiments, but the embodiments of the present invention are not limited thereto.

[0063] like Figure 1 As shown, a first aspect of an embodiment of the present invention provides a method for generating a vehicle three-dimensional model with the assistance of a large language model, comprising the following steps:

[0064] Step 11, obtaining the three-dimensional model information of the sample vehicle structure.

[0065] The sample vehicle structure three-dimensional model information includes: text description information of each sample vehicle structure three-dimensional model and size data of each sample vehicle structure three-dimensional model.

[0066] Step 12: input the image to be parsed, the example image and the first prompt word into the large language model, and the large language model outputs target parsed text description information of the image to be parsed.

[0067] The image to be analyzed is obtained by splicing multiple images of the target vehicle, and the first prompt word includes a text description of the vehicle in the image to be analyzed and analyzed text description information of the example image.

[0068] Step 13: input the second prompt word into the large language model, and the large language model outputs text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information.

[0069] The second prompt word includes: task instruction information, target parsing text description information and text description information of the sample vehicle structure three-dimensional model.

[0070] Step 14, determining the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generating a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information.

[0071] Step 15: input the third prompt word into the large language model, and the large language model outputs the vehicle structure coordinates.

[0072] Step 16, input the vehicle structure coordinates into a three-dimensional modeling tool to generate a model of the target vehicle.

[0073] In this embodiment, based on the large language model, the data of the model to be modeled can be quickly found in a large number of known models, and the coordinate data required for modeling can be generated from these data through the large language model. The coordinate data can then be input into a three-dimensional modeling application to complete the modeling, thereby saving modeling time and improving modeling efficiency.

[0074] A second aspect of an embodiment of the present invention provides a method for generating a three-dimensional vehicle model with the assistance of a large language model, comprising the following steps:

[0075] Step 21, obtaining the three-dimensional model information of the sample vehicle structure.

[0076] The sample vehicle structure 3D model information includes: text description information of each sample vehicle structure 3D model and size data of each sample vehicle structure 3D model. In this embodiment, in the size data of the sample vehicle structure 3D model, the sample vehicle structure 3D models of the same type have the same proportion.

[0077] Prior to this step, the sample model is preprocessed to adjust the size ratios of the same type of models to be consistent, such as the ratios of all wheel models, the ratios of all car logo models, etc., and generate text description information of the model after adjusting the ratios, obtain the sample vehicle model structure information, and store the sample vehicle model structure information in the local database. Here, the sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information includes a wheel structure three-dimensional model, a vehicle design model (overall appearance model), an exterior finish model, an interior layout model, a car logo model, a window model and a rearview mirror model, all of which are three-dimensional models. The content of the text description information of the sample vehicle structure three-dimensional model includes, but is not limited to, the structural features of each model, its position relative to the overall vehicle, craftsmanship, performance, effects, cultural meanings, and the like.

[0078] In a feasible implementation, the text description information of the sample vehicle structure three-dimensional model can also be generated through a large prediction model or other methods.

[0079] Step 22, obtain multiple images of the target vehicle, and stitch the multiple images in sequence to obtain the image to be analyzed. In this step, the target vehicle is the vehicle that the user needs to model, and at least one front view, at least one rear view, at least one left view, at least one right view, at least one top view, at least one bottom view, at least one stereoscopic view, and at least one interior layout view of the target vehicle are obtained, and these images are stitched one by one to obtain the image to be analyzed.

[0080] Step 23: input the image to be parsed, the example image and the first prompt word into the large language model, and the large language model outputs target parsed text description information of the image to be parsed.

[0081] The first prompt word includes a text description of the vehicle in the image to be analyzed by taking the analyzed text description information of the example image as an example and the analyzed text description information of the example image.

[0082] In this step, the specific contents of the text description of the vehicle in the image to be analyzed are: taking the analyzed text description information of the sample image as an example, a text description is performed on the wheel structure, vehicle design, exterior finish, interior layout, vehicle logo, window features and rearview mirror features in the image to be analyzed.

[0083] The image to be analyzed is a composite image of the target vehicle at different angles, and the text description of the vehicle in the image to be analyzed includes a description of the appearance of each structure of the vehicle. The example image can be a composite image of a vehicle at multiple angles.

[0084] The parsed text description information of the example image includes:

[0085] 1. Descriptive information about the wheel structure’s hub material, hub style, hub structure, hub performance, and tire performance.

[0086] 2. Descriptive information about the vehicle model’s lines, body contour shape, front end structure, side shape, and rear end lines and structure.

[0087] 3. Description of the craftsmanship, color and corresponding effects of the exterior finish.

[0088] 4. Descriptive information about the style of the interior layout, the location and type of the instrument panel, the material and performance of the seats, and the structure and performance of the central control area.

[0089] 5. Descriptive information on the shape, meaning, style, location and craftsmanship of the vehicle logo for vehicle identification.

[0090] 6. Descriptive information on the window characteristics, including the material of the window, the performance of the window lifting system, and the sealing structure and performance around the window.

[0091] 7. Descriptive information on the functions, mirror structure, corresponding control system, shell color and lines of the rearview mirror features.

[0092] The parsed text description information of the example image includes at least the above seven aspects, and the content in each aspect also includes but is not limited to the above description, which can be expanded as needed.

[0093] In a feasible implementation, an example of the first prompt word is as follows:

[0094] Please perform in-depth analysis on the vehicle in the image to be analyzed (please analyze according to the case in Example 1) and output the analysis text description information: subdivided into wheel structure, vehicle design, exterior finish, interior layout, vehicle logo, window characteristics and rearview mirror features. Note: 1. The image to be analyzed is a combination of different angles of a vehicle. 2. Please subdivide the vehicle into a detailed description of the appearance of each part in the vehicle in-depth analysis.

[0095] Example 1:

[0096] According to the example picture, I can deeply analyze:

[0097] 1. Wheel structure

[0098] The wheels of cars are usually made of high-quality alloys, and the design style combines dynamism and elegance. Common designs include five-, ten-, or multi-spoke wheels, among which the "turbine blade" shape is particularly typical, which not only enhances the visual sense of sportiness, but also optimizes aerodynamic performance. In terms of tires, ×× brands often choose low rolling resistance tires, which are designed to improve fuel efficiency and driving comfort, while ensuring good grip and handling.

[0099] 2. Model design

[0100] The model is known for its classic streamlined design, which combines simple lines with a modern sense of technology. The body contour is smooth, and the front end adopts the iconic single-frame hexagonal air intake grille, which is matched with LED or matrix LED headlights, showing a strong family recognition. The side waistline design of the body is dynamic, emphasizing the sense of power and speed, and the rear of the car creates a high-end and dynamic visual effect through the layered lines and integrated taillight design.

[0101] 3. Appearance and finish

[0102] The exterior finish of the car is finely processed, usually using high-quality metallic paint or pearlescent paint, with a variety of color options. The body surface is smooth and flat, and advanced spraying technology is used to ensure uniform color and long-term gloss. Chrome-plated decorative parts, such as window frames, door handles and other details, add a touch of luxury. In addition, some models are also equipped with S-Line kits, which further enhance the sporty atmosphere.

[0103] 4. Interior layout

[0104] The interior is known for its simple yet luxurious style, with attention to detail and material selection. The dashboard layout is clear and leans towards the driver's side for easy operation. The Virtual Cockpit system is a highlight, which replaces traditional instruments with a high-resolution fully digital instrument panel that can be customized to display information, including navigation maps, vehicle information, etc. The seats are usually wrapped in high-grade leather and provide a variety of adjustment functions to ensure comfort during long-distance driving. The center console integrates the latest MMI multimedia interactive system, which supports touch operation and voice control, enhancing the technological experience.

[0105] 5. Car logo

[0106] The car logo is composed of four interlocking rings, symbolizing the alliance of the original four independent car manufacturers, reflecting the brand's heritage and spirit of cooperation. The car logo is usually located in the center of the front grille and the rear of the car, as well as in the center of the steering wheel. The design is simple and recognizable, and it is chrome-plated or brightly finished to highlight the sense of quality.

[0107] 6. Window features

[0108] The windows are made of double-layer soundproof glass, which effectively isolates external noise and improves the quietness of the car. The window lifting system is smooth and quiet, and most models are equipped with one-button lifting and anti-pinch functions to ensure safe use. The sealing strip design around the windows is optimized, which is not only waterproof and dustproof, but also improves the vehicle's sealing and overall NVH (noise, vibration and harshness) performance.

[0109] 7. Rearview mirror features

[0110] The design of the rearview mirror of the car focuses on practicality and aesthetics. It usually has electric adjustment, heating and defrosting, memory function, and automatic folding function on some high-end models. The mirror adopts a wide-angle design to reduce blind spots and improve driving safety. Some models are also equipped with a blind spot monitoring system, which is integrated in the rearview mirror and warns the driver to pay attention to vehicles in adjacent lanes through lights. In terms of appearance, the rearview mirror housing is coordinated with the body color, and some models also adopt a streamlined design to reduce wind resistance.

[0111] }

[0112] Step 24: Generate a second prompt word according to the task instruction information, the target parsed text description information and the text description information of the sample vehicle structure three-dimensional model.

[0113] In this step, the specific content of the second prompt word is: search the text description information of the target sample vehicle structure 3D model that matches the target parsed text description information from the text description information of the sample vehicle structure 3D model. Here, searching for ... that matches ... from ... is also the task instruction information.

[0114] Specifically, since the large language data model cannot directly call the local database, in an embodiment of the present application, in order to realize the writing of text description information of all sample vehicle structure three-dimensional models in the local database in the prompt words of the large model, it is limited by the maximum number of words in the prompt words of the large model. When there is no way to list all the text description information in the local database through one prompt word, a new prompt word is written. The new prompt word only has the text description information about the sample vehicle structure three-dimensional model in the local database and is different from the previous prompt word, and so on, until the text description information of all sample vehicle structure three-dimensional models in the local database is input into the large language model in the form of prompt words.

[0115] In a feasible implementation, an example of the second prompt word is as follows (a wheel is used as an example below):

[0116] Please find the text description information of the 3D model that is most similar to the one I want, where the hub body is composed of five wide straight bars or slightly curved spokes (the text description information of the wheel in the target parsed text description information) from all known wheel descriptions (i.e. the text description information of all sample wheel 3D models in the local database), and output it to me.

[0117] The following is the text description information of all sample wheel 3D models in the local database:

[0118] [(Note: obtained in the local database)

[0119] 1. Five-spoke sports wheel: This design is commonly seen on vehicles that pursue a sense of speed and sports performance. The main body of the wheel is composed of five wide straight or slightly curved spokes, which are simple and powerful. It can not only effectively reduce weight, but also facilitate heat dissipation, giving people a strong and dynamic impression visually.

[0120] 2. Multi-spoke luxury wheels: Compared with the simple five-spoke wheels, multi-spoke wheels have more spokes, and ten, twelve or even more spokes are common. Each spoke is relatively slender and closely arranged, presenting a more complex and delicate visual effect. They are commonly seen on luxury cars or business cars, emphasizing refinement and stability, while also improving the overall grade of the vehicle.

[0121] 3. Blade-shaped wheels: This design emphasizes the sharpness and dynamics of the lines. The spoke shape imitates the shape of a sharp knife or dart, and the edges may have sharp angles or streamlined curves, giving people a feeling of cutting through the air and extraordinary speed. Blade-shaped wheels are commonly seen on sports cars or sports utility vehicles, emphasizing the speed and aggressiveness of the vehicle. They are not only beautiful, but also can improve the aerodynamic performance of the vehicle.

[0122] … ]

[0124] Step 25, input the second prompt word into the large language model, and the large language model outputs the text description information of the target sample vehicle structure 3D model that matches the target parsed text description information. In this step, the large language model outputs the text description information of the sample vehicle structure 3D model that is closest to the target parsed text description information as the text description information of the target sample vehicle structure 3D model. The closest can be understood as, for example, the closest semantics of the two.

[0125] Step 26, determining the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generating a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information.

[0126] In this step, first, according to the text description information of the target sample vehicle structure three-dimensional model, the dimensional data of the corresponding sample vehicle structure three-dimensional model is searched in the local database. The dimensional data of the sample vehicle structure three-dimensional model found is the dimensional data of the target sample vehicle structure three-dimensional model. After obtaining the dimensional data of the target sample vehicle structure three-dimensional model, the third prompt word is generated.

[0127] In this step, the content of the third prompt word includes: size data of the target sample vehicle structure three-dimensional model, assembly task information, precautions and example information.

[0128] Among them, the points of attention include: the origin is the center point of the vehicle, the X-axis is horizontal to the right, indicating the front and rear direction of the vehicle, the Z-axis is vertically upward, indicating the up and down direction of the vehicle, and the Y-axis is the out-of-screen direction (the direction from the screen to the user), indicating the left and right direction of the vehicle.

[0129] The dimension data of the target sample vehicle structure three-dimensional model are concatenated into character strings as a part of the third prompt word.

[0130] The assembly task information includes: according to the dimension data of the target sample vehicle structure 3D model, the coordinates of a complete vehicle are assembled according to the arrangement of x, y, and z coordinates, the center point of the vehicle is the origin of the coordinate system, and the coordinates of each structure are calculated and output. Here, in order to improve the calculation efficiency, specifically, the coordinates of each structure output are the coordinates of the center point of each structure. The coordinates of the center point of each structure can be calculated based on the center point of the vehicle as the origin of the coordinate system and the dimension data, and the modeling can be completed in the subsequent 3D modeling tool based on the coordinates and dimension data of the center point.

[0131] The example information includes: example dimension data, thinking process description information and output results of the example dimension data.

[0132] The thought process description includes:

[0133] 1. The center point of the vehicle is taken as the origin, and the front and rear and left and right symmetry axes of the vehicle are along the Z axis and Y axis respectively.

[0134] 2. The vehicle design and exterior finish are the same size, the vehicle body covers the wheels and extends beyond the wheels, with the bottom in contact with the wheels.

[0135] 3. The center point of the car logo is at the same height as the center point of the vehicle.

[0136] 4. The rearview mirrors are installed on both sides of the vehicle, with the rearview mirror located close to the A-pillar.

[0137] 5. Determine the boundary values ​​of the length, width and height of the vehicle based on the example dimension data.

[0138] 6. The center point coordinates of the wheels are set according to the width and length of the vehicle.

[0139] 7. The coordinates of vehicle model design are the coordinates of the center point of the vehicle.

[0140] 8. The coordinates of the exterior finish are the same as those of the vehicle model design.

[0141] 9. The coordinates of the car logo are the coordinates of the center point of the car logo.

[0142] 10. The coordinates of the rearview mirror are the coordinates of the center point of each rearview mirror.

[0143] In the content of the third prompt word, the content in each aspect also includes but is not limited to the above description and can be expanded as needed.

[0144] For example, the content of the third prompt word is as follows:

[0145] Target size data:

[0146] 1. Wheel size: length 65 width 20 height 65

[0147] 2. The design size of the vehicle is 480cm long, 200cm wide and 160cm high.

[0148] 3. The exterior dimensions are 480cm long, 200cm wide and 160cm high.

[0149] 4. The size of the logo is 12mm long, 12mm wide and 12mm high.

[0150] 5. The size of the rearview mirror is 21cm long, 21cm wide and 12cm high.

[0151] Now please help me assemble the known model size mentioned above into the center point coordinates of a complete vehicle structure according to the arrangement of x, y, and z coordinates. Note that the center point of the vehicle should be the origin of the coordinate system (you can refer to Example 1). Finally, please output the center point coordinates of each structure.

[0152] Here are some things to note:

[0153] 1. The origin is (0,0,0): the center point of the vehicle.

[0154] 2. Each structure is placed according to the coordinates of the center point of each structure.

[0155] 3. X-axis: Horizontally to the right, representing the front and rear direction of the vehicle, with the front of the vehicle facing the positive direction of the x-axis.

[0156] 4. Z axis: vertically upward, representing the up and down direction of the vehicle, with the roof facing the positive direction of the z axis.

[0157] 5.Y axis: screen direction, representing the left and right direction of the vehicle.

[0158] 6. The vehicle logo is at the same height as the center point of the vehicle by default.

[0159] 7. The default height of the left and right rearview mirrors is placed above the center point of the vehicle.

[0160] Example 1:

[0161] Known:

[0162] 1. Wheels, size: length 60, width 20, height 60

[0163] 2. The vehicle model is designed to be 450cm long, 180cm wide and 150cm high.

[0164] 3. Appearance and finish: 450cm long, 180cm wide, 150cm high

[0165] 4. Vehicle logo recognition, size is 10 in length, 10 in width, and 10 in height

[0166] 5. Rearview mirror, size is 20mm long, 20mm wide, 10mm high

[0167] Thought Process:

[0168] 1. Planning and layout

[0169] The center point of the vehicle is taken as the origin (0,0,0), which means that the front and rear and left and right symmetry axes of the wheel are along the Z axis and X axis respectively. The size of the wheel is 60 length, 20 width, and 60 height, which means the diameter is 60cm and the thickness is 20cm.

[0170] Vehicle design and exterior finish: The dimensions of the vehicle design and exterior finish are consistent, 450x180x150, indicating that the vehicle body is 450cm long, 180cm wide and 150cm high. The vehicle body usually covers the wheels and extends beyond the wheels, with the bottom touching the wheels.

[0171] The dimensions of the vehicle are 450 cm long, 180 cm wide and 150 cm high, so the boundaries of the vehicle body are: length direction x: -225 ~225 cm, width direction y: -90 ~ 90 cm, height direction z: -75 ~ 75 cm.

[0172] Vehicle logo: The center point of the vehicle logo is at the same height as the center point of the vehicle.

[0173] Rearview mirrors: need to be considered for installation on both sides of the vehicle, usually near the A-pillar.

[0174] 2. Set the coordinate system

[0175] The origin is set to the center point of the vehicle.

[0176] 3. Component coordinate setting

[0177] wheel:

[0178] The wheel size is 60mm long, 20mm wide and 60mm high.

[0179] The center points of the wheels need to be located at the four corners of the vehicle (front, back, left, and right).

[0180] Front left wheel center coordinates: (225 - 30, -90 + 10, -75 + 30) = (195, -80, -45)

[0181] Front right wheel center coordinates: (225 - 30, 90 - 10, -75 + 30) = (195, 80, -45)

[0182] The coordinates of the center point of the rear left wheel: (-225 + 30, -90 + 10, -75 + 30) = (-195, -80, -45)

[0183] Rear right wheel center coordinates: (-225 + 30, 90 - 10, -75 + 30) = (-195, 80, -45)

[0184] Vehicle design and exterior finish: The exterior finish is the same size as the vehicle, so its center point is consistent with the center point of the vehicle, located at (0, 0, 0). It does not need to be set separately, but the surface details of the vehicle body need to be considered when modeling.

[0185] Vehicle logo recognition: The dimensions of the vehicle logo recognition are 10 in length, 10 in width, and 10 in height, assuming that the vehicle logo is located above the front center of the vehicle.

[0186] Car logo center coordinates: (225 - 5, 0, 0) = (220, 0, 0)

[0187] Rearview mirror: The dimensions of the rearview mirror are 20 in length, 20 in width and 10 in height.

[0188] Left rearview mirror center coordinates: (225 / 2 - 10, -90 - 10, 10) = (102.5, -100, 10)

[0189] Coordinates of the center point of the right rearview mirror: (225 / 2 - 10, 90 + 10, 10) = (102.5, 100, 10)

[0190] Output:

[0191] Front left wheel center coordinates: (195, -80, -45)

[0192] Front right wheel center coordinates: (195, 80, -45)

[0193] Rear left wheel center coordinates: (-195, -80, -45)

[0194] Rear right wheel center coordinates: (-195, 80, -45)

[0195] The center point coordinates of the vehicle design: (0, 0, 0)

[0196] The center point coordinates of the exterior surface: (0, 0, 0)

[0197] The center coordinates of the car logo: (220, 0, 0)

[0198] Left rearview mirror center coordinates: (102.5, -100, 10)

[0199] The coordinates of the center point of the right rearview mirror are: (102.5, 100, 10).

[0200] In Example 1 of the third prompt word, the coordinates of the center point can be estimated based on the origin and dimension data and the structural position of the actual vehicle. For example, the coordinates of the center point of the front left wheel are: (225 – wheel length / 2, -90 + wheel width / 2, -75 + wheel height / 2) = (195, -80, -45). In subsequent steps, the large language model can perform corresponding calculations on the dimension data of the three-dimensional model of the target sample vehicle structure combined with the actual structure of the vehicle.

[0201] Step 27, input the third prompt word into the large language model, and the large language model outputs the vehicle structure coordinates.

[0202] Step 28, input the vehicle structure coordinates into a three-dimensional modeling tool to generate a model of the target vehicle.

[0203] like Figure 2 As shown, a third aspect of an embodiment of the present invention provides a system for generating a vehicle three-dimensional model with the assistance of a large language model, including:

[0204] The acquisition module 31 is used to acquire the sample vehicle structure three-dimensional model information; the sample vehicle structure three-dimensional model information includes: text description information of each sample vehicle structure three-dimensional model and size data of each sample vehicle structure three-dimensional model;

[0205] A first input module 32 is used to input the image to be analyzed, the example image and the first prompt word into the large language model; wherein the image to be analyzed is obtained by splicing multiple images of the target vehicle, and the first prompt word includes a text description of the vehicle in the image to be analyzed and the analysis text description information of the example image;

[0206] A large language model 33, used for outputting target parsed text description information of the image to be parsed;

[0207] The first input module 32 is further used to input the second prompt word into the large language model; wherein the second prompt word includes task instruction information, target parsed text description information and text description information of the sample vehicle structure three-dimensional model;

[0208] The large language model 33 is also used for the large language model to output text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information;

[0209] A determination module 34 is used to determine the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generate a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information;

[0210] The first input module 32 is further used to input the third prompt word into the large language model;

[0211] The large language model 33 is also used for the large language model to output the vehicle structure coordinates;

[0212] A second input module 35, for inputting the vehicle structure coordinates into the three-dimensional modeling tool;

[0213] The three-dimensional modeling tool 36 is used to generate a model of the target vehicle.

[0214] In one embodiment of the present invention, in the dimension data of the sample vehicle structure three-dimensional models, the sample vehicle structure three-dimensional models of the same type have the same proportions.

[0215] In one embodiment of the present invention, the specific content of the text description of the vehicle in the image to be analyzed is: respectively describing the wheel structure, vehicle design, exterior finish, interior layout, vehicle logo, window features and rearview mirror features in the image to be analyzed; wherein the image to be analyzed is a composite image of multiple angles of the target vehicle, and the text description of the vehicle in the image to be analyzed includes the appearance description of each structure of the vehicle;

[0216] The parsed text description information of the example image includes:

[0217] Descriptive information about the wheel structure’s hub material, hub style, hub structure, hub performance, and tire performance;

[0218] Descriptive information on the vehicle model lines, body contour shape, front end structure, side shape, and rear end lines and structure of the vehicle model design;

[0219] Description of the process, color and corresponding effects of the exterior finish;

[0220] Descriptive information about the style of the interior layout, the location and type of the instrument panel, the material and performance of the seats, and the structure and performance of the center console area;

[0221] Descriptive information about the shape, meaning, style, location and craftsmanship of the logo that identifies the vehicle;

[0222] Descriptive information on window features, such as the material of the window, the performance of the window lift system, and the sealing structure and performance around the window;

[0223] The features of the rearview mirror include the mirror's function, mirror structure, corresponding control system, shell color and lines.

[0224] In one embodiment of the present invention, a second prompt word is generated according to the task instruction information, the target parsing text description information and the text description information of the sample vehicle structure three-dimensional model;

[0225] The second prompt words include:

[0226] The text description information of the target sample vehicle structure 3D model that matches the target parsed text description information is searched from the text description information of the sample vehicle structure 3D model.

[0227] In one embodiment of the present invention, the third prompt word includes:

[0228] The dimensional data, assembly task information, precautions and example information of the target sample vehicle structure three-dimensional model; among which, the precautions include: the origin is the center point of the vehicle, the X-axis is horizontal to the right, indicating the front and rear direction of the vehicle, the Z-axis is vertically upward, indicating the up and down direction of the vehicle, and the Y-axis is the out-of-screen direction, indicating the left and right direction of the vehicle.

[0229] In one embodiment of the present invention, the assembly task information includes: assembling a complete vehicle according to the size data of the three-dimensional model of the target sample vehicle structure according to the arrangement of x, y, and z coordinates, with the center point of the vehicle as the origin of the coordinate system, and outputting the coordinates of each structure.

[0230] In one embodiment of the present invention, the example information includes: example dimension data, thought process description information, and output results of the example dimension data;

[0231] The thought process description includes:

[0232] The center point of the vehicle is taken as the origin, and the front-to-back and left-to-right symmetry axes of the vehicle are along the Z axis and the X axis respectively;

[0233] The vehicle design and exterior finish are identical in size, with the body covering the wheels and extending beyond them, with the bottom in contact with the wheels;

[0234] The center point of the car logo is at the same height as the center point of the vehicle;

[0235] The rearview mirrors are installed on both sides of the vehicle, where the rearview mirror is located near the A-pillar;

[0236] Determine boundary values ​​of the length, width and height of the vehicle based on the example dimension data;

[0237] The center point coordinates of the wheels are set according to the width and length of the vehicle.

[0238] The coordinates of the vehicle model design are the coordinates of the center point of the vehicle;

[0239] The exterior finishes are designed to the same coordinates as the model;

[0240] The coordinates of the car logo are the coordinates of the center point of the car logo;

[0241] The coordinates of the rearview mirror are the coordinates of the center point of each rearview mirror.

[0242] A fourth aspect of an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for generating a three-dimensional vehicle model with the assistance of a large language model provided in the above-mentioned embodiment of the present invention is implemented.

[0243] The fifth aspect of an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of a method for generating a vehicle three-dimensional model with the assistance of a large language model provided in the above embodiment of the present invention are implemented.

[0244] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0245] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware systems.

[0246] The method provided in the embodiment of the present invention can be applied to electronic devices. Specifically, the electronic device can be: a desktop computer, a portable computer, an intelligent mobile terminal, a server, etc. This is not limited here, and any electronic device that can implement the present invention belongs to the protection scope of the present invention.

[0247] As for the system / electronic device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0248] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or multiple boxes.

[0249] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0250] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0251] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for generating a three-dimensional vehicle model using a large language model, characterized in that: The following steps are involved: Acquire sample vehicle structure three-dimensional model information; wherein the sample vehicle structure three-dimensional model information includes: text description information of each sample vehicle structure three-dimensional model and size data of each sample vehicle structure three-dimensional model; Inputting the image to be parsed, the example image and the first prompt word into the large language model, the large language model outputting the target parsed text description information of the image to be parsed; wherein the image to be parsed is obtained by splicing multiple images of the target vehicle, and the first prompt word includes a text description of the vehicle in the image to be parsed and the parsed text description information of the example image; Inputting a second prompt word into the large language model, the large language model outputting text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information; wherein the second prompt word includes: task instruction information, target parsed text description information and text description information of the sample vehicle structure three-dimensional model; Determining the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generating a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information; Inputting the third prompt word into the large language model, and the large language model outputs vehicle structure coordinates; The vehicle structure coordinates are input into a three-dimensional modeling tool to generate a model of the target vehicle.

2. The method according to claim 1, characterized in that In the dimension data of the sample vehicle structure three-dimensional models, the sample vehicle structure three-dimensional models of the same type have the same proportions.

3. The method according to claim 1, characterized in that The specific content of the text description of the vehicle in the image to be analyzed is: respectively describing the wheel structure, vehicle model design, exterior finish, interior layout, vehicle logo, window features and rearview mirror features in the image to be analyzed; wherein the image to be analyzed is a composite image of multiple angles of the target vehicle, and the text description of the vehicle in the image to be analyzed also includes a description of the appearance of each structure of the vehicle; The parsed text description information of the example image includes: Descriptive information about the wheel structure’s hub material, hub style, hub structure, hub performance, and tire performance; Descriptive information on the vehicle model lines, body contour shape, front end structure, side shape, and rear end lines and structure of the vehicle model design; Description of the process, color and corresponding effects of the exterior finish; Descriptive information about the style of the interior layout, the location and type of the instrument panel, the material and performance of the seats, and the structure and performance of the center console area; Descriptive information about the shape, meaning, style, location and craftsmanship of the logo that identifies the vehicle; Descriptive information on window features, such as the material of the window, the performance of the window lift system, and the sealing structure and performance around the window; The features of the rearview mirror include the mirror's function, mirror structure, corresponding control system, shell color and lines.

4. The method according to claim 1, characterized in that Also includes: generating a second prompt word according to the task instruction information, the target parsed text description information and the text description information of the sample vehicle structure three-dimensional model; The second prompt word includes: The text description information of the sample vehicle structure three-dimensional model is searched for the text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information.

5. The method according to claim 1, characterized in that The third prompt word includes: The target sample vehicle structure three-dimensional model includes dimensional data, assembly task information, precautions and example information; wherein the precautions include: the origin is the center point of the vehicle, the X-axis is horizontal to the right, indicating the front and rear direction of the vehicle, the Z-axis is vertically upward, indicating the up and down direction of the vehicle, and the Y-axis is the out-of-screen direction, indicating the left and right direction of the vehicle.

6. The method according to claim 5, characterized in that The assembly task information includes: assembling a complete vehicle according to the size data of the target sample vehicle structure three-dimensional model in accordance with the arrangement of x, y, and z coordinates, with the center point of the vehicle as the origin of the coordinate system, and outputting the coordinates of each structure.

7. The method according to claim 5, characterized in that The example information includes: example size data, thinking process description information and output results of the example size data; The thought process description information includes: The center point of the vehicle is taken as the origin, and the front-to-back and left-to-right symmetry axes of the vehicle are along the Z axis and the X axis respectively; The vehicle design and exterior finish are identical in size, with the body covering the wheels and extending beyond them, with the bottom in contact with the wheels; The center point of the car logo is at the same height as the center point of the vehicle; The rearview mirrors are installed on both sides of the vehicle, where the rearview mirror is located near the A-pillar; Determine boundary values ​​of the length, width and height of the vehicle based on the example dimension data; The center point coordinates of the wheels are set according to the width and length of the vehicle. The coordinates of the vehicle model design are the coordinates of the center point of the vehicle; The exterior finishes are designed to the same coordinates as the model; The coordinates of the car logo are the coordinates of the center point of the car logo; The coordinates of the rearview mirror are the coordinates of the center point of each rearview mirror.

8. A system for generating a three-dimensional vehicle model using a large language model, characterized in that: include: An acquisition module, used to acquire the three-dimensional model information of the sample vehicle structure; The sample vehicle structure three-dimensional model information includes: text description information of each sample vehicle structure three-dimensional model and size data of each sample vehicle structure three-dimensional model; A first input module is used to input the image to be analyzed, the example image and the first prompt word into the large language model; wherein the image to be analyzed is obtained by splicing multiple images of the target vehicle, and the first prompt word includes a text description of the vehicle in the image to be analyzed and the analysis text description information of the example image; A large language model, used to output target parsed text description information of the image to be parsed; The first input module is further used to input a second prompt word into the large language model; wherein the second prompt word includes task instruction information, target parsed text description information and text description information of the sample vehicle structure three-dimensional model; The large language model is further used for the large language model to output text description information of the target sample vehicle structure three-dimensional model that matches the target parsed text description information; a determination module, configured to determine the size data of the corresponding target sample vehicle structure three-dimensional model in the sample vehicle structure three-dimensional model information according to the text description information of the target sample vehicle structure three-dimensional model, and generate a third prompt word according to the size data of the target sample vehicle structure three-dimensional model and the assembly task information; The first input module is further used to input the third prompt word into the large language model; The large language model is also used for the large language model to output vehicle structure coordinates; A second input module, used to input the vehicle structure coordinates into a three-dimensional modeling tool; 3D modeling tool, used to generate a model of the target vehicle.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for generating a vehicle three-dimensional model with the assistance of a large language model as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for generating a three-dimensional vehicle model with the assistance of a large language model according to any one of claims 1 to 7 is implemented.

Images