Virtual tourism experience method based on large model

Abstract

The invention provides a virtual tourism experience method based on a large model, and aims to provide personalized and immersive tourism experience by integrating advanced artificial intelligence and virtual / augmented reality technologies. The method comprises the following steps: firstly, collecting interest and preference information of a user, such as historical travel records, scenic spot preferences, travel time preferences, social media activities and the like, so as to capture detailed demands of the user; thirdly, generating personalized tourist routes and suggestions according to the information by utilizing a large-scale artificial intelligence model and combining deep learning and natural language processing technologies; in addition, the system renders scenic spots in a tour route in real time through a virtual reality or augmented reality technology, and the reality sense of vision and hearing is improved by applying advanced graphic and physical simulation technologies. And finally, an interactive interface is designed, so that the user can interact with the virtual tour guide through voice, text or gestures, more scenic spot information is obtained, a tour route is adjusted in real time, and the individuation and participation sense of user experience are greatly enhanced.

Description

Technical Field

[0001] The invention relates to the technical field of large-scale model scenic spot guide, in particular to a large-scale model-based virtual tourism experience method. Background Art

[0002] With the rapid development of information technology and artificial intelligence, virtual reality (VR), augmented reality (AR) and mixed reality (MR) technologies have begun to shine in the tourism field, bringing users unprecedented immersive tourism experience. Through these technologies, users can experience the natural scenery and cultural landscapes around the world through virtual environments at home. This experience not only breaks through the limitations of geographical location, but also saves travel costs and time to some extent. However, although these technologies have made certain progress in providing visual immersion, there are still many challenges in terms of personalized experience, interactivity and realism.

[0003] First, current virtual tourism experiences often lack personalization. Most virtual tourism applications rely on pre-recorded videos or 3D models, which results in a monotonous user experience and fails to meet the personalized needs of different users. It is difficult for users to customize travel routes according to their interests and needs, and it is also difficult to obtain travel content recommendations that truly meet their personal preferences. To solve this problem, it is necessary to develop more intelligent personalized recommendation systems that can learn user behavior and preferences to provide a more personalized travel experience.

[0004] Secondly, many virtual tourism applications have poor interactivity. In these applications, users often play the role of a passive observer rather than an active participant who can interact with elements in the virtual environment in real time. This passive observation limits the user's participation and the depth of experience. In order to improve interactivity, more interactive elements can be introduced, such as allowing users to interact with objects in the virtual environment through voice, text or gestures, and even participate in changes and decisions in the virtual environment.

[0005] Furthermore, the existing VR and AR technologies still need to be improved in terms of realism. Although they can provide a certain immersive experience visually, the simulation of sound, touch and other senses often cannot fully replicate the complexity and details of the real world. In addition, the simulation of light and shadow effects and physical behaviors in virtual environments usually cannot achieve completely realistic effects. This requires further improvement through more advanced graphics rendering technology and more sophisticated sensory feedback systems.

[0006] Finally, virtual tourism usually lacks social experience. Tourism is not only a personal experience, but also a form of social activity. Most current virtual tourism solutions ignore the interaction and social needs between users and lack the function of supporting multiple users to interact online and share experiences at the same time. Introducing social functions, such as virtual group tours and interactive chat rooms between users, can enhance the social experience of users and make virtual tourism a shared activity among friends and family members.

[0007] In summary, although existing VR, AR and MR technologies have shown great potential in the tourism industry, to fully tap the advantages of these technologies, more innovation and improvement are needed in terms of personalized experience, interactivity, realism and social experience.

[0008] To this end, we designed a virtual tourism experience method based on a large model to solve the above problems. Summary of the invention

[0009] The purpose of the present invention is to provide a virtual tourism experience method based on a large model to solve the problems raised in the above background technology.

[0010] To achieve the above object, the present invention provides the following technical solution: a virtual tourism experience method based on a large model, comprising the following steps:

[0011] Collect information about users’ interests and preferences, including but not limited to users’ historical travel records, preferred attraction types, travel time preferences, interactive feedback and social media activities, and direct questionnaire survey results provided by users;

[0012] Generate personalized travel itineraries based on the interests and preferences information using a large-scale artificial intelligence model that analyzes user data through deep learning, natural language processing technology, and contextual understanding capabilities to generate travel recommendations that best suit the user;

[0013] Combine virtual reality or augmented reality technology to render the attractions in the tourist route in real time, use advanced graphics rendering technology and physical simulation to provide a highly realistic visual and auditory experience, including but not limited to 3D model generation, environmental lighting simulation and virtual sound field creation;

[0014] Design an interactive interface that enables users to interact with the virtual tour guide through voice, text, or gestures to obtain more attraction information and travel suggestions, and support real-time feedback and adjustment of travel routes.

[0015] As a preferred technical solution of the present invention: the large-scale artificial intelligence model is Wenxin Yiyan or other models with natural language processing and machine learning capabilities, which can understand the user's needs and preferences based on the natural language input by the user and dynamically generate response content.

[0016] As a preferred technical solution of the present invention: the user's interest and preference information includes the user's historical travel records, attraction preference types, and travel time preferences, and the user model is continuously learned and updated through a machine learning algorithm to improve the accuracy of recommendations, wherein the algorithm includes but is not limited to support vector machines, random forests, and deep neural networks.

[0017] As a preferred technical solution of the present invention: the method further comprises: realizing the interaction between the user and the virtual tour guide through natural language processing (NLP) technology, wherein the virtual tour guide can provide services in multiple languages, adapt to the needs of international users, and can understand and respond to complex user queries;

[0018] The machine learning algorithm uses an incremental learning algorithm based on user feedback:

[0019] P new =P old +η(FP old )

[0020] Among them, P new and P old They represent the user preference parameters before and after the update, F is the preference information extracted from the latest user feedback, and η is the learning rate.

[0021] As a preferred technical solution of the present invention: the interaction includes users asking questions through voice or text, and the system responds according to the responses generated by the large artificial intelligence model, and supports multiple rounds of dialogue to deeply understand the needs of users, including providing historical and cultural background information, local food recommendations and guidance on dealing with emergencies.

[0022] As a preferred technical solution of the present invention: the generation of the personalized travel route includes determining the recommended weight of the scenic spot using the following calculation formula

[0023] W i =α×S i +β×T i +γ×D i +δ×F i +∈×C i

[0024] Among them, W i represents the recommendation weight of the i-th scenic spot, S i represents the user's interest score for the i-th scenic spot, which is calculated by the user's evaluation and interaction frequency for similar scenic spots; T i represents the theme matching score of the i-th attraction, which is calculated based on the correlation between the attraction type and the user's preference; D i represents the matching degree between the i-th scenic spot and the user's preferred travel time, taking into account the user's preferred season, time period and actual opening hours; Fi represents the user’s circle of friends feedback and social media interaction related to the i-th attraction, based on social network analysis; C i It represents the crowding level prediction of the scenic spot, which is calculated based on historical data and current trends; α, β, γ, δ,∈ are weight coefficients adjusted by machine learning algorithm.

[0025] As a preferred technical solution of the present invention: the interest score is determined based on the user's past evaluation and interaction frequency of similar attractions, and is adjusted through a user feedback loop to adapt to changes in user interests, wherein the feedback loop includes the user's evaluation and suggestions for the travel experience.

[0026] As a preferred technical solution of the present invention: according to user feedback and behavior data, the recommendation algorithm for subsequent travel routes is adjusted, and optimization is performed using reinforcement learning or other machine learning algorithms, including a combination of online learning and batch learning, to reflect changes in user preferences in real time;

[0027] Reinforcement learning is used to dynamically adjust the recommendation strategy based on real-time feedback from users. The reward update is expressed using the following formula:

[0028]

[0029] Where Q(s,a) is the quality function of taking action a in state s, R is the immediate reward received, γ is the discount factor, and α is the learning rate.

[0030] As a preferred technical solution of the present invention: the adjustment includes optimizing the α, β, γ, δ,∈ weight coefficients using a machine learning algorithm to improve the accuracy of personalized recommendations, and dynamically adjusting by real-time analysis of user behavior data, including the user's movement trajectory, residence time, and interactive behavior in the virtual environment;

[0031] The optimization of weight coefficients can be achieved by gradient descent method, the formula is as follows:

[0032]

[0033] Among them, θ represents the weight coefficient, J(θ) is the loss function, α is the learning rate, is the gradient of the loss function with respect to the weights.

[0034] As a preferred technical solution of the present invention: the real-time rendering of the virtual reality or augmented reality technology includes 360-degree panoramic display, high-definition texture mapping, physical lighting simulation and environmental sound fusion to provide an immersive experience, and also includes the fusion of environmental sound and local cultural elements to enhance the user's sense of immersion, and supports multi-user synchronous participation and interaction, including virtual group photos and social functions.

[0035] Beneficial effects:

[0036] Provide personalized travel experience: Utilize large-scale artificial intelligence models to generate personalized travel routes based on multi-dimensional data such as users' historical behaviors, preferences, and social media activities, which can accurately meet the personalized needs of different users.

[0037] Enhance the immersion and realism of virtual tourism: By combining virtual reality (VR) and augmented reality (AR) technologies, real-time rendering of scenic spots and provision of 360-degree panoramic displays, high-definition texture mapping and physical lighting simulation, users can be provided with an immersive tourism experience.

[0038] Support for multi-language and multi-mode interaction: The system supports multiple language services and multiple interaction methods (including voice, text, and gestures) to meet the needs of users in different countries and regions, improving the system's universality and ease of use.

[0039] Promote social interaction and multi-user experience: Support multiple users to participate in the same virtual tour online at the same time, provide social interaction functions such as virtual selfies, shared photo walls, etc., which increase the fun and social nature of the travel experience.

[0040] Dynamically adjust and optimize the travel experience: The system collects user behavior data and feedback in real time in the virtual environment, dynamically adjusts and optimizes the travel experience, and ensures that users always get the best experience.

[0041] Continuous learning and service optimization: Use a variety of machine learning technologies to continuously learn and optimize recommendation algorithms and user experience, ensuring that the system continues to improve its recommendation accuracy and experience quality as technology advances and user data accumulates.

[0042] In general, by combining advanced artificial intelligence and virtual reality technologies, it can not only provide personalized and highly realistic virtual tourism experiences, but also enhance user participation and satisfaction through multi-user interaction and continuous optimization, and has broad application prospects and market potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a flowchart of the large model-based virtual tourism experience method proposed in the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The following is a specific implementation method for the above claims. This section will elaborate on how to implement each key step of the method:

[0046] See also Figure 1 , the embodiment of the present invention provides a virtual tourism experience method based on a large artificial intelligence model. This method uses the latest artificial intelligence, virtual reality (VR), augmented reality (AR) technology, combined with user personalized data, to provide a highly customized and immersive virtual tourism experience. The following is a detailed description through specific embodiments:

[0047] Step 1: Collect information about users’ interests and preferences

[0048] In the first step of the present invention, the system first needs to collect users' personal information extensively, including but not limited to historical travel records, attraction preference types, travel time preferences, interactive feedback, and social media activities. In addition, the system also collects direct input from users through online questionnaires and direct user interviews, such as the user's interest in specific attractions and specific expectations for travel experience.

[0049] For example, user A has visited coastal cities and cultural sites many times, indicating that he has a high preference for such attractions. The content and interaction data shared by users on social media will also be mined by the system's big data analysis tools to further determine the user's interests. At this stage, the system will use machine learning algorithms such as decision trees and cluster analysis to extract useful information from a large amount of user behavior data and form a user interest model.

[0050] In addition, the system will further improve the user portrait by analyzing the user's interaction records with travel-related apps, including the scenic spot details pages browsed by the user, the travel guides and comments collected. For example, if a user leaves a detailed positive review of a specific scenic spot, the system will use this data as strong positive feedback to increase the user's preference weight for this type of scenic spot.

[0051] Step 2: Generate personalized travel itineraries using large-scale AI models

[0052] In this step, the large-scale artificial intelligence model in the system (such as GPT-4o) will generate recommendations for travel routes and attractions that are most suitable for users based on the data collected in step 1 through deep learning, natural language processing technology, and contextual understanding capabilities.

[0053] Taking user A as an example, based on his preferences and historical data, the AI ​​model may recommend a travel route that includes multiple coastal cities and historical sites. The model will analyze the user's travel time preferences and the seasonal characteristics of each attraction, and dynamically adjust the recommended route based on the theme matching of the attractions and the user's past evaluations.

[0054] In addition, the model will also consider the convenience and cost efficiency of actual travel, and automatically plan the most reasonable travel route for users. For example, in order to avoid unnecessary round trips and traffic jams, the AI ​​model will prioritize recommending attractions near the user's location, and then recommend attractions in remote areas, while providing users with a variety of transportation options and estimated costs.

[0055] To calculate the recommendation weight W for each scenic spot, the system uses the following formula:

[0056] W i =α×S i +β×T i +γ×D i +δ×F i +∈×C i

[0057] Among them, W i represents the recommendation weight of the i-th scenic spot, S i represents the user's interest score for the i-th scenic spot, which is calculated by the user's evaluation and interaction frequency for similar scenic spots; T i represents the theme matching score of the i-th attraction, which is calculated based on the correlation between the attraction type and the user's preference; D i represents the matching degree between the i-th scenic spot and the user's preferred travel time, taking into account the user's preferred season, time period and actual opening hours; F i represents the user’s circle of friends feedback and social media interaction related to the i-th attraction, based on social network analysis; C i It represents the crowding level prediction of the scenic spot, which is calculated based on historical data and current trends; α, β, γ, δ,∈ are weight coefficients adjusted by machine learning algorithm.

[0058] For example, if a coastal city is a peak tourist season in summer, but users prefer to avoid crowds, the system will reduce the recommendation weight of this attraction during the peak season and increase its recommendation weight during the off-season. Such dynamic adjustments ensure the flexibility and personalization of the recommendation system.

[0059] Step 3: Combine VR / AR technology to render scenic spots in real time

[0060] After the user selects the travel route and attractions, the system will use virtual reality (VR) or augmented reality (AR) technology to render these attractions in real time. This process includes the use of advanced graphics rendering technology, physical simulation and virtual sound field creation to provide a highly realistic visual and auditory experience.

[0061] For example, users can use VR helmets to watch a 360-degree panoramic view of a coastal city and feel as if they were there. The system also simulates local environmental sounds and cultural elements, such as the sound of waves and street music, to enhance the sense of immersion. Users can interact with the virtual environment through gestures or controllers, such as "approaching" a scenic spot to view details, or talking to virtual characters to learn more about the scenic spot.

[0062] In addition, the system can also adjust the perspective and content display based on the user's body movements and eye tracking data, so that users can experience a more personalized interactive experience. For example, if the user shows a strong interest in a certain exhibit, the system can automatically zoom in on the details of the exhibit and even provide audio explanations.

[0063] To enhance the sense of reality, the system also uses a physical engine to simulate the light and shadow effects in the natural environment, such as the changes in light at sunrise and sunset, the shimmering of water, etc. These highly realistic visual effects, combined with customized sound effects and music, make the virtual tourism experience no less than real tourism.

[0064] Step 4: Design the interface for user interaction

[0065] The system provides an interface for users to interact with the virtual tour guide through voice, text or gestures. The virtual tour guide is able to provide services in multiple languages ​​and understand complex user queries.

[0066] For example, users can ask the virtual tour guide through voice commands: "What special food is there in this place?" The virtual tour guide will provide detailed food recommendations and related information based on the analysis of large-scale artificial intelligence models. Users can also request to adjust the tour route or get help and guidance in emergency situations.

[0067] In addition, the interface design takes into account multiple user needs, including barrier-free access for people with visual impairments. For example, for users with limited vision, the system provides voice recognition and feedback functions, allowing them to interact with the system through hearing. For users with limited hearing, the system provides a text interface with high contrast and large fonts, as well as sign language recognition.

[0068] Users can also adjust various parameters of the virtual experience through the interface, such as the brightness and contrast of visual effects, the volume and type of sound effects, and the sensitivity of the interaction method. This high degree of customization not only improves user comfort, but also ensures that users with different backgrounds and needs can get a satisfactory experience.

[0069] Step 5: Dynamically adjust and optimize the travel experience

[0070] The system not only plans and recommends for users before traveling, but also collects feedback in real time during the user experience, dynamically adjusts and optimizes the travel experience. The system determines the user's satisfaction and interests by analyzing the user's behavior in the virtual environment, such as the length of stay, objects of frequent interaction, facial expressions, and physiological reactions (such as heart rate and eye movements).

[0071] For example, if the system detects that a user has spent a long time in front of a virtual exhibition at a certain attraction, the system can infer that the user is interested in this type of exhibition and recommend more similar exhibitions in the future. On the contrary, if the user quickly skips certain content, the system will reduce the frequency of recommending such content.

[0072] In addition, the system will collect direct feedback from users after their experience, such as satisfaction surveys and open comments. This data will be used to train and optimize the AI ​​model to improve its accuracy and personalization. For example, through sentiment analysis of user comments, the system can more accurately understand the user's specific needs and preferences, so as to make more appropriate adjustments in future recommendations.

[0073] Step 6: Provide multi-user interaction and social functions

[0074] In order to enhance the user's social experience, the system supports multiple users to participate in the same virtual tour online at the same time, and provides rich social interaction functions. Users can invite family and friends to enter the virtual environment and tour and explore together.

[0075] In this multi-user mode, each user can see the virtual avatars of other users and communicate with them through voice or text. The system also provides a series of interactive tools, such as virtual selfies, shared photo walls, real-time location sharing, etc., so that users can enjoy the fun of real travel in virtual travel.

[0076] For example, users can take a photo in front of a virtual attraction, and the system will automatically generate a virtual photo with the background of the attraction, allowing users to edit, share on social media or send to friends. In addition, users can also create and participate in various mini-games and activities in the virtual environment to increase the fun and participation of the interaction.

[0077] Step 7: Continuous learning and service optimization

[0078] The system's continuous learning and service optimization are key to ensuring long-term user satisfaction and system stability. By collecting and analyzing large amounts of user data, the system is able to continuously improve its algorithms and functions.

[0079] The system uses a variety of machine learning techniques, including but not limited to supervised learning, unsupervised learning, reinforcement learning, and transfer learning, to achieve full process optimization from user behavior to system feedback. For example, through reinforcement learning, the system can autonomously adjust the weight parameters α, β, γ, δ,∈ in the recommendation algorithm to adapt to changes in user preferences and emerging trends. The optimization of weight coefficients can be achieved through the gradient descent method, and the formula is as follows:

[0080]

[0081] Among them, θ represents the weight coefficient, J(θ) is the loss function, α is the learning rate, is the gradient of the loss function with respect to the weights;

[0082] Reinforcement learning is used to dynamically adjust the recommendation strategy based on real-time feedback from users. The reward update is expressed using the following formula:

[0083]

[0084] Where Q(s,a) is the quality function of taking action a in state s, R is the immediate reward received, γ is the discount factor, and α is the learning rate.

[0085] The system will also be regularly updated and maintained to ensure the advancement of technology and consistency of user experience, including updating the content of virtual attractions, adding new interactive functions and optimizing interface design.

[0086] The present invention provides a new virtual tourism experience method by combining a large-scale artificial intelligence model and VR / AR technology. This method can not only provide personalized travel recommendations, but also greatly improve user satisfaction and participation through real-time interaction and immersive experience. With the continuous advancement of technology and the accumulation of user data, the recommendation accuracy and experience quality of this method will continue to improve, providing users with a richer, more realistic and personalized virtual tourism experience.

[0087] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0088] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A virtual tourism experience method based on a large model, characterized in that: The steps include: Collect information about users’ interests and preferences, including but not limited to users’ historical travel records, preferred attraction types, travel time preferences, interactive feedback and social media activities, and direct questionnaire survey results provided by users; Generate personalized travel itineraries based on the interests and preferences information using a large-scale artificial intelligence model that analyzes user data through deep learning, natural language processing technology, and contextual understanding capabilities to generate travel recommendations that best suit the user; Combining virtual reality or augmented reality technology to render the attractions in the tourist route in real time, using graphics rendering technology and physical simulation to provide a highly realistic visual and auditory experience, including but not limited to 3D model generation, environmental lighting simulation and virtual sound field creation; Design an interactive interface that enables users to interact with the virtual tour guide through voice, text, or gestures to obtain more attraction information and travel suggestions, and support real-time feedback and adjustment of travel routes.

2. The method for virtual tourism experience based on a large model according to claim 1, characterized in that: The large-scale artificial intelligence model is Wenxin Yiyan or other models with natural language processing and machine learning capabilities, which can understand the user's needs and preferences based on the natural language input by the user and dynamically generate response content.

3. The method for virtual tourism experience based on a large model according to claim 1, characterized in that: The user's interest and preference information includes the user's historical travel records, preferred attraction types, and travel time preferences, and the user model is continuously learned and updated through machine learning algorithms to improve the accuracy of recommendations, wherein the algorithms include but are not limited to support vector machines, random forests, and deep neural networks; The machine learning algorithm uses an incremental learning algorithm based on user feedback: P new =P old +η(FP old ) Among them, P new and P old They represent the user preference parameters before and after the update, F is the preference information extracted from the latest user feedback, and η is the learning rate.

4. The method for virtual tourism experience based on a large model according to claim 1, characterized in that: The method further includes: enabling interaction between the user and the virtual tour guide through natural language processing technology, wherein the virtual tour guide can provide services in multiple languages, adapt to the needs of international users, and can understand and respond to complex user queries.

5. The method for virtual tourism experience based on a large model according to claim 4, characterized in that: The interaction includes users asking questions via voice or text, and the system responding based on responses generated by large artificial intelligence models, supporting multiple rounds of conversation to gain a deeper understanding of the user's needs, including providing historical and cultural background information, local food recommendations, and guidance on how to respond to emergencies.

6. The method for virtual tourism experience based on a large model according to claim 1, characterized in that: The generation of the personalized travel route includes determining the recommended weight W of the scenic spot using the following calculation formula: IN i =α×S i +β×T i +γ×D i +δ×F i +∈×C i Among them, W i represents the recommendation weight of the i-th scenic spot, S i represents the user's interest score for the i-th scenic spot, which is calculated by the user's evaluation and interaction frequency for similar scenic spots; T i represents the theme matching score of the i-th attraction, which is calculated based on the relevance between the attraction type and the user's preference; D i represents the matching degree between the i-th scenic spot and the user's preferred travel time, taking into account the user's preferred season, time period and actual opening hours; F i represents the user’s circle of friends feedback and social media interactions related to the i-th attraction, based on social network analysis; C i Represents the crowding level forecast of the attraction, calculated based on historical data and current trends; α, β, γ, δ,∈ are weight coefficients adjusted by the machine learning algorithm.

7. The method for virtual tourism experience based on a large model according to claim 6, characterized in that: The interest score S i It is determined based on users' past evaluations and interaction frequency of similar attractions, and adjusted through a user feedback loop to adapt to changes in user interests, where the feedback loop includes users' evaluations and suggestions for travel experiences.

8. The method for virtual tourism experience based on a large model according to claim 1, characterized in that: Further including: Adjust the recommendation algorithm for subsequent travel routes based on user feedback and behavior data, using reinforcement learning or other machine learning algorithms for optimization, including a combination of online learning and batch learning to reflect changes in user preferences in real time; Reinforcement learning is used to dynamically adjust the recommendation strategy based on real-time feedback from users. The reward update is expressed using the following formula: Where Q(s,a) is the quality function of taking action a in state s, R is the immediate reward received, γ is the discount factor, and α is the learning rate.

9. The method for virtual tourism experience based on a large model according to claim 8, characterized in that: The adjustment includes optimizing the α, β, γ, δ,∈ weight coefficients using a machine learning algorithm to improve the accuracy of personalized recommendations, and making dynamic adjustments by analyzing user behavior data in real time, including the user's movement trajectory, residence time, and interactive behavior in the virtual environment; The optimization of weight coefficients can be achieved by gradient descent method, the formula is as follows: Among them, θ represents the weight coefficient, J(θ) is the loss function, α is the learning rate, is the gradient of the loss function with respect to the weights.

10. The method for virtual tourism experience based on a large model according to claim 1, characterized in that: The method described herein, wherein the real-time rendering of the virtual reality or augmented reality technology includes 360-degree panoramic display, high-definition texture mapping, physical lighting simulation and environmental sound fusion to provide an immersive experience, and also includes the fusion of environmental sound and local cultural elements to enhance the user's sense of immersion, and supports multi-user synchronous participation and interaction, including virtual group photo and social functions.

Images