Method for assisting in generating forest fire-fighting emergency special automobile model by artificial intelligence

Abstract

The invention discloses a method for assisting in generating a special fire emergency car model through artificial intelligence, and belongs to the crossing field of artificial intelligence and car design. The method aims to overcome the defects of traditional forest fire fighting vehicle design, and the forest fire rescue efficiency and the firefighter safety are improved. Data such as fire fighting truck 2D drawings, forest fire cases, fire fighting demand feedback and the like are widely collected through multiple channels and are cleaned, labeled and structured, and effective data are divided into a training set, a verification set and a test set according to a specific proportion. A deep learning model (such as GAN or VAE) is selected and optimized according to the design characteristics of the forest fire-fighting vehicle, and the model is stabilized through a large amount of iterative training. A design keyword is input to generate a 2D sketch, big data analysis is utilized to screen, evaluate and optimize for multiple times, a 3D model framework is generated by means of a 3D generation AI model, fine optimization is performed after multi-dimensional evaluation, and finally a 3D model which is vivid in color and meets actual combat requirements is generated. By utilizing the artificial intelligence technology, the design efficiency is greatly improved, the adaptability of the automobile to the forest environment is enhanced, the safety of firefighters is guaranteed, the industrial technology progress is promoted, and remarkable economic and social benefits are achieved.

Description

Technical Field

[0001] The present invention relates to the cross - field of artificial intelligence and automotive design, and particularly to a method for using artificial intelligence to assist in generating a special forest fire fighting emergency vehicle model. Background Art

[0002] Forest fires pose a serious threat to the ecological environment and the safety of human life and property. In recent years, forest fires have occurred frequently. For example, in 2019, a forest fire occurred in Muli County, Liangshan Prefecture, Sichuan Province, with a burned area of about 20 hectares; in 2020, a forest fire occurred in Shan Shenba, Shuangmei Village, Qinglong Sub - district, Anning, Yunnan, with a burned area of 170.1 hectares.

[0003] Traditional forest fire fighting methods have many drawbacks when facing complex forest environments. Existing fire trucks are difficult to quickly reach the core area of the fire, and the information obtained is lagging and inaccurate, resulting in a lack of reliable basis for rescue decisions. Firefighters operate in dangerous environments and their lives are at great risk. At the same time, in the field of automotive design, traditional design methods are less efficient and it is difficult to fully consider the complex requirements and diverse scenarios of forest fire fighting. With the rapid development of big data and artificial intelligence technologies, applying them to the design of special forest fire fighting emergency vehicle models has important practical significance. Currently, the application of artificial intelligence in the field of automotive design is still in the exploratory stage, lacking a mature and systematic method to assist in generating vehicle models that are highly adaptable to forest fire fighting scenarios. Summary of the Invention

[0004] The present invention aims to provide a method for using artificial intelligence to assist in generating a special fire fighting emergency vehicle model, using artificial intelligence technology to solve the problems existing in the design of traditional forest fire fighting vehicles, improve the adaptability of fire trucks to complex forest environments, enhance the efficiency of forest fire rescue, ensure the safety of firefighters, and at the same time improve the efficiency and quality of vehicle model design.

[0005] Extensively collect data: Collect data from multiple channels such as the Internet, professional databases, academic literature, and actual cases of the fire department. Include obtaining 2D drawings and technical parameters of different types of fire trucks (traditional fire trucks, special fire trucks, etc.), with the number of collections being no less than 8; collect at least 5 forest fire cases covering different regions of the world, and obtain topographic data (such as slope, altitude, etc.) of the fire occurrence locations, the spread of the fire, information on the fire fighting and rescue process, etc.; collect feedback on the functional requirements of forest fire fighting vehicles from the fire department, firefighters, and relevant experts.

[0006] Data cleaning and annotation: Clean the collected data to remove duplicate, incorrect, and incomplete data. Standardize the format of the 2D drawings of fire trucks and correct dimension deviations; perform structured processing on the forest fire case data and annotate key information such as fire type, ignition point, and fire development stage; convert the fire demand feedback into quantifiable indicators such as vehicle passability indicators and the number of equipment carried.

[0007] Data division: Use artificial intelligence to generate content (AIGC) technology to divide the preprocessed data into a training set, a validation set, and a test set, with a division ratio of 8:1:1.

[0008] Model selection and optimization: Select a deep learning model suitable for processing images and data, such as a generative adversarial network (GAN) or a variational autoencoder (VAE), etc., and optimize it according to the design characteristics of forest fire trucks. Adjust the model parameters to enable it to focus on learning the design features of forest fire trucks, such as special body structures (high ground clearance, narrow body, etc. suitable for complex terrains), fire equipment layouts (rational placement of efficient fire extinguishing equipment), and adaptability to forest fire scenarios (ability to handle different terrains and fire intensities).

[0009] Iterative training: Perform iterative training on the model using the training set, enabling the generator to learn the design characteristics of fire trucks, the requirements for forest fire response, and fire demands. By continuously adjusting the model parameters, the drawings or models generated by the generator gradually approach the ideal design, while also enhancing the discriminator's judgment ability. The training process is carried out on a server equipped with high-performance computing devices (such as multiple NVIDIA RTX series GPUs). After no less than [M] iterations, the model shows stable performance on the validation set, achieving a good generation effect.

[0010] Initial drawing generation: Input design keywords such as "accurate fire source positioning", "efficient fire extinguishing system", "remote monitoring function", etc. into the trained model, and the model generates multiple groups of 2D sketches of forest fire trucks in a short time. These sketches present different body layouts, equipment configurations, and appearance designs.

[0011] Screening and evaluation: Select 4 groups of the most promising sketches from the generated ones and use big data analysis tools to compare and evaluate them with existing forest fire trucks on the market and actual fire rescue needs. Analyze the rationality of the body structure (such as whether it is convenient to drive on forest roads and whether the turning radius is appropriate), the scientific nature of the equipment layout (such as whether the operation of fire equipment is convenient and whether there is interference between various equipment), etc.

[0012] Optimization and improvement: According to the evaluation results, the selected sketches are input into the AI model again for optimization, adjusting details such as lines, dimensions, and component positions. After no less than 3 rounds of optimization, a more perfect 2D drawing is obtained. Finally, use professional image editing software (such as Adobe Photoshop) to modify the details of the drawing to ensure the drawing quality.

[0013] Initial generation of 3D model: Input the processed 2D drawing of the forest fire fighting vehicle into the trained 3D generation AI model, and the model generates a preliminary framework of the 3D model based on the learned knowledge. This framework includes the main body of the vehicle body, the basic forms and approximate positions of the main fire fighting equipment, but there are deficiencies in details, such as the lack of texture on the vehicle body surface and the non-refinement of fire fighting equipment components.

[0014] Multi-dimensional evaluation: Use big data analysis tools to compare the initially generated 3D model with the actual requirements and existing mature 3D models of fire fighting vehicles. Evaluate the model from multiple dimensions such as the rationality of the spatial structure (such as whether the interior space meets the requirements for carrying personnel and equipment), the convenience of equipment operation (such as the flexibility of the operation angle of the fire fighting water gun), and the feasibility of actual combat application (such as the passability and stability in different forest terrains).

[0015] Fine optimization: According to the evaluation results, input the model into AI again for optimization, supplement details, such as adding vehicle body material texture (such as the texture of fireproof materials), refining fire fighting equipment components (such as nozzles, valves, etc.), and adjusting the structure (such as optimizing the center of gravity distribution of the vehicle body). After multiple rounds of optimization, the 3D model becomes more refined and realistic, meeting the actual combat requirements of forest fire fighting. Finally, use AIGC technology to add colors to the model to generate a high-quality 3D model.

[0016] 1. Improve design efficiency: With the help of artificial intelligence and big data technologies, a large number of design schemes can be quickly generated and screened for optimization. Compared with traditional design methods, the design cycle is significantly shortened, improving design efficiency.

[0017] 2. Enhance adaptability: By learning a large number of forest fire cases and fire fighting demand data, the generated vehicle model can better adapt to complex forest environments and diverse rescue needs, improving the actual effect of forest fire fighting.

[0018] 3. Ensure the safety of firefighters: The optimized vehicle model fully considers the convenience and safety of firefighters in design. For example, a reasonable equipment layout facilitates firefighters to operate quickly, improving the safety of firefighters during the rescue process.

[0019] 4. Promote industry development: The method of the present invention provides innovative ideas and practical experiences for the application of artificial intelligence in the field of vehicle design, helping to promote the technological progress and development of the entire industry. Brief Description of the Drawings

[0020] Figure 1 It is a flow chart of a special fire emergency vehicle model assisted by artificial intelligence, introducing the process framework from prompt words to the generation of 2D models and 3D models assisted by artificial intelligence.

[0021] Figure 2 It is an example of a 2D sketch of a forest fire truck, showing different body layouts and equipment configurations; Figure 3 It is a preliminary framework of a 3D model, reflecting the basic forms of the body main part and main fire-fighting equipment; Figure 4 It is an optimized 3D model, presenting fine details and realistic effects. Detailed Implementation Modes

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention. Embodiment

[0023] 1. Data collection and preprocessing stage: 150 2D drawings of fire trucks are collected from professional fire websites, automobile design forums, etc. through web crawler technology, 120 forest fire cases are obtained from the internal database of the fire department, and 80 pieces of feedback from the fire department are collected. Duplicate data is removed using data cleaning algorithms, the drawing formats are unified using image conversion tools, the fire case data is structurally organized, and key information is marked. After processing, the proportion of valid data reaches 75%.

[0024] 2. Model training stage: The generative adversarial network (GAN) model is selected, and the data is divided into a training set, a validation set, and a test set according to the ratio of 70%:15%:15%. Training is carried out on a server equipped with 4 NVIDIA RTX 3090 GPUs. After 5000 iterations, the generation effect of the model on the validation set is stable, and the discriminator accuracy reaches 85%.

[0025] 3. 2D drawing generation and optimization stage: Keywords such as "intelligent obstacle avoidance" and "rapid water replenishment" are input into the trained model, and the model generates 20 groups of 2D sketches. Eight groups are selected from them, and comparative evaluation is carried out using big data analysis tools. After 3 rounds of optimization, high-quality 2D drawings are obtained, and then Adobe Photoshop is used for detail processing.

[0026] 4. 3D Model Generation and Optimization Phase: Input the optimized 2D drawings into the 3D generation AI model to generate a preliminary framework of the 3D model. After evaluating from aspects such as spatial structure, operational convenience, and practical feasibility, and through 4 rounds of optimization, a fine and realistic 3D model is generated and colors are added.

[0027] In practical applications, according to the forest characteristics (such as terrain, vegetation type) and changes in fire-fighting requirements in different regions, the data can be updated and supplemented, and the model can be retrained to continuously optimize the generated vehicle model to better meet the actual needs of forest fire-fighting. At the same time, the method of the present invention can also be extended to the design of other types of fire-fighting emergency vehicles, having a wide range of application prospects.

[0028] It should be noted that the protection scope of the present invention is not limited to the above embodiments, and any improvements, equivalent replacements, etc. made on the basis of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for artificial intelligence-assisted generation of a special vehicle model for forest fire emergency, characterized in that: The method comprises the following steps: S1. Data collection and preprocessing: Collect 2D drawings and technical parameters of fire trucks, forest fire case information in different regions around the world, and feedback on the functional requirements of forest fire trucks from fire departments, firefighters and relevant experts from multiple channels such as the Internet, professional databases, academic literature, and actual cases of fire departments; The collected data were cleaned to remove duplicate, erroneous and incomplete data, standardize the 2D drawing format of fire trucks, and correct dimensional deviations. 2.S2, model training: The pre-processed data was divided into training set, validation set and test set in a ratio of 8:1:1 using artificial intelligence generated content (AIGC) technology; Select deep learning models suitable for processing images and data, such as generative adversarial networks (GANs) or variational autoencoders (VAEs), and optimize them for the design characteristics of forest fire fighting vehicles, adjusting model parameters to focus on learning the design characteristics of forest fire fighting vehicles; The model is iteratively trained on the training set on a server equipped with high-performance computing equipment. After multiple iterations, the model performs stably on the validation set. 3.S3, 2D drawing generation and optimization: Input design keywords into the trained model, and the model generates multiple sets of 2D sketches of forest fire trucks; Select the four most promising groups from the generated sketches and use big data analysis tools to compare and evaluate them with existing forest fire fighting vehicles on the market and actual fire rescue needs; Based on the evaluation results, the selected sketches are input into the AI ​​model again for optimization. After no less than 3 rounds of optimization, a more complete 2D drawing is obtained. Finally, professional image editing software is used to modify the details of the drawing. 4.S4, 3D model generation and optimization: Input the processed 2D drawings of forest fire trucks into the trained 3D generation AI model to generate the preliminary framework of the 3D model; Using big data analysis tools, the initially generated 3D model is compared and evaluated with actual needs and existing mature fire truck 3D models from multiple dimensions such as rationality of spatial structure, convenience of equipment operation, and feasibility of actual application; Based on the evaluation results, the model is input into AI again for optimization, details are added, and the structure is adjusted. After multiple rounds of optimization, AIGC technology is used to add color to the model to generate a high-quality 3D model.

5. The method for artificial intelligence-assisted generation of a forest fire emergency special vehicle model according to claim 1 is characterized in that: The number of 2D drawings of fire trucks collected is no less than 8, and the number of forest fire cases collected is at least 5.

6. The method for artificial intelligence-assisted generation of a forest fire emergency special vehicle model according to claim 1, characterized in that: The high-performance computing device is a plurality of NVIDIA RTX series GPUs.

7. The method for artificial intelligence-assisted generation of a forest fire emergency special vehicle model according to claim 1, characterized in that: The design keywords include "precise fire source positioning", "efficient fire extinguishing system", "remote monitoring function", "intelligent obstacle avoidance", "rapid water replenishment", etc.

8. The method for artificial intelligence-assisted generation of a forest fire emergency special vehicle model according to claim 1, characterized in that: The professional image editing software is Adobe Photoshop.

9. The method for artificial intelligence-assisted generation of a forest fire emergency special vehicle model according to claim 1, characterized in that: The supplementary details to the 3D model include adding body material textures and refining fire equipment parts, and the adjustment of the structure includes optimizing the body center of gravity distribution.

Images