An AR technology-based garden landscape travel guide system

Abstract

The application relates to the technical field of wisdom tourism, and provides a garden landscape tourism guide system based on AR technology, which comprises a reservation module, a high-precision positioning and navigation module and an intelligent voice explanation module. According to the weather, the number of people and the performance time in the garden, the reservation module is used for facilitating tourists to reserve the performance activities, characteristic experience projects and the like in the garden in advance. The high-precision positioning and navigation module is used for fusing GPS, a Bluetooth beacon and a visual SLAM technology to position the position of the tourists in real time and provide accurate navigation services. The intelligent voice explanation module is used for automatically triggering or manually triggering by scanning a code based on the position of the tourists, and is used for associating multilingual explanation contents of garden historical culture, plant knowledge and architectural anecdotes. Through the innovative space anchor point technology, virtual object skeleton nodes are accurately bound with physical marker points, the visual SLAM technology is combined to position and calibrate in real time, a 16-waveband multispectral environment perception array simulating the compound eye of mantis shrimps is matched, light field data are captured in real time to generate a three-dimensional environment model.

Description

Technical Field

[0001] This invention relates to the field of smart cultural tourism technology, specifically to a garden landscape cultural tourism guide system based on AR technology. Background Technology

[0002] Garden landscapes are a exquisite fusion of human wisdom and natural beauty, carrying profound historical and cultural connotations. They often feature well-arranged pavilions and towers, winding paths and corridors, flowers and trees that change with the seasons, and ingeniously designed rockeries and ponds. Strolling through them is like walking through a series of poetic paintings, where every scene seems to tell a story of the past, immersing visitors in a unique artistic atmosphere. However, without a garden landscape cultural tourism guide system, the visitor experience is often greatly diminished. Gardens typically cover a large area with a complex and intricate layout. Without clear guidance, visitors can easily lose their way, wandering through winding paths and potentially missing many key attractions, failing to appreciate their full beauty. For example, those hidden corners with unique cultural connotations deep within the garden are difficult to discover without guidance. Moreover, many attractions in gardens have rich historical allusions and cultural connotations; relying solely on observation and speculation makes it difficult for visitors to truly understand their underlying meaning and value. Without a guided tour system providing professional explanations, the ingenious architectural designs and unique plant arrangements become merely superficial visual presentations. Visitors struggle to appreciate the ingenuity of the ancient garden designers and the cultural heritage they embody. Their tours remain at a cursory, superficial level, failing to deliver a deeper aesthetic experience or cultural understanding. Therefore, a garden landscape cultural tourism guided tour system plays a crucial role in enhancing the quality of visitor experiences.

[0003] Garden landscapes carry rich cultural heritage, but traditional guided tour methods have many limitations, such as superficial cultural interpretation, insufficient navigation accuracy, and weak interactive immersion. Although existing AR guided tour systems attempt to improve this, they also have shortcomings; commercial apps cannot dynamically generate cultural interpretations and have high maintenance costs. Currently, there is an urgent need for a system that can dynamically adapt cultural content to the real scene based on high-precision spatial positioning, generate in-depth cultural interpretations in real time in weak network environments, and enhance immersion through multimodal interaction. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a garden landscape cultural tourism guide system based on AR technology, which solves the problems of superficial interpretation, insufficient navigation accuracy, and weak interactive immersion.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a garden landscape and cultural tourism guide system based on AR technology, comprising:

[0006] The reservation module calculates based on weather, visitor flow, and park performance times, making it convenient for tourists to book park performances, special experiences, and other activities in advance.

[0007] The high-precision positioning and navigation module integrates GPS, Bluetooth beacons, and visual SLAM technologies to locate tourists in real time and provide accurate navigation services.

[0008] The intelligent voice explanation module can be automatically triggered based on the visitor's location or manually triggered by scanning a QR code, and provides multilingual explanations related to garden history and culture, plant knowledge and architectural anecdotes.

[0009] The dynamic route recommendation module generates personalized tour routes based on tourists' interest tags and attraction popularity.

[0010] The AR culture generation unit analyzes real-world garden images based on a large language model, generates poems and allusions that match the cultural background of the scene in real time, and overlays them onto the AR view.

[0011] The virtual-real fusion interactive unit uses AR devices to accurately overlay virtual images of historical figures, scene restoration animations, and real-world scenes, supporting gesture / voice interaction.

[0012] Preferably, the reservation module includes:

[0013] The prediction unit uses LSTM neural network analysis of historical passenger flow data, real-time weather and holiday factors to predict the congestion index of each area in the next 2 hours.

[0014] The scheduling unit performs weighted scheduling based on user level and reservation time limit when multiple reservation requests conflict.

[0015] Preferably, the positioning and navigation module includes:

[0016] In areas with weak signals, switch to offline inertial navigation mode to maintain positioning accuracy through gyroscope and gait analysis;

[0017] When planning routes, routes with high cultural relevance are prioritized, and navigation instructions are generated in natural language.

[0018] A cluster of base stations is deployed at the highest point in the garden to build a local precision enhancement network. Tourist terminals receive differential signals from the base stations and correct their positioning coordinates in real time.

[0019] Preferably, the intelligent voice explanation module includes:

[0020] The voice unit analyzes tourists' voiceprint characteristics such as age, gender, and dialect, and dynamically adjusts speech rate, vocabulary complexity, and cultural depth; when a tourist's stay time is detected to be >90 seconds, it automatically associates historical events.

[0021] Preferably, the dynamic route recommendation module includes:

[0022] A reinforcement learning algorithm is introduced to dynamically optimize subsequent routes based on real-time feedback from tourists.

[0023] When groups of tourists have conflicting routes, priorities are dynamically allocated based on game theory Nash equilibrium.

[0024] The route is dynamically adjusted based on the flowering status of the plants.

[0025] Preferably, the AR culture generation unit includes:

[0026] The image recognition subunit extracts architectural style, plant species, and landscape layout features based on a convolutional neural network.

[0027] The cultural knowledge graph sub-unit links garden history documents, poetry databases, and cultural entity databases.

[0028] Dynamically generate sub-units and combine feature extraction results with knowledge graphs to generate scene-adapted cultural content in real time.

[0029] Preferably, the AR culture generation unit includes:

[0030] It supports multimodal input, allowing visitors to upload custom photos to trigger virtual restoration of specific historical scenes.

[0031] Preferably, the key areas of the garden are processed locally for image recognition and AR rendering tasks.

[0032] Preferably, the system is equipped with a layered triggering mechanism: automatic triggering starts AR explanation when tourists enter the preset geofence; manual triggering accesses in-depth cultural content by scanning the landscape QR code; and gesture triggering activates hidden interactive scenes by recognizing specific gestures.

[0033] Preferably, the virtual-real fusion interaction unit includes:

[0034] By employing spatial anchor point technology, virtual object skeletal nodes are bound to physical marker points, and virtual content is bound to physical coordinates, thus reducing the error range;

[0035] A multispectral environmental sensing array was constructed by simulating the compound eye structure of mantis shrimp to capture 16-band light field data in real time and generate a three-dimensional environmental model.

[0036] This invention provides a garden landscape and cultural tourism guide system based on AR technology. It has the following beneficial effects:

[0037] 1. This invention uses innovative spatial anchoring technology to precisely bind virtual object skeletal nodes to physical marker points. Combined with visual SLAM technology for real-time positioning and calibration, and a 16-band multispectral environmental perception array simulating the compound eyes of a mantis shrimp, it captures light field data in real time to generate a three-dimensional environment model. The fusion of multiple technologies reduces the positioning error between virtual content and physical landscape, improving upon traditional AR systems and enabling real-time synchronization of virtual light and shadow with real-world lighting, significantly enhancing the immersive experience for visitors.

[0038] 2. This invention relies on a dynamic knowledge graph to construct a network linking garden historical documents, poems and allusions with cultural entities, accurately extracting scene features such as architectural style and plant species; when a visitor stays for more than 90 seconds, the system automatically triggers an in-depth explanation mechanism, dynamically calling up historical events, anecdotes of literati and extended cultural content related to the landscape through knowledge graph node links, and overlaying them onto the AR view in a multimodal form, thereby enhancing the depth of the visitor's cultural understanding.

[0039] 3. This invention effectively solves the latency problem in weak network environments through the localized deployment of edge computing nodes, resulting in a qualitative leap in system response speed and ensuring the stability and smoothness of the tour guide service. Based on a dynamic scheduling algorithm using intelligent behavior analysis, it achieves intelligent management of park visitor flow, significantly alleviating congestion during peak hours and improving the overall visitor experience by predicting hotspot areas in real time and optimizing route recommendations. This comprehensive intelligent tour guide model provides a new paradigm of immersive and personalized cultural tourism services for traditional garden scenic areas. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the system unit modules of the present invention;

[0041] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Example:

[0044] As one aspect of the present invention, please refer to the appendix. Figure 1 This invention provides a garden landscape cultural tourism guide system based on AR technology, comprising:

[0045] The reservation module accurately calculates the opening times of different attractions based on real-time weather conditions. This module utilizes an advanced deep learning model, the LSTM neural network, to comprehensively analyze historical visitor flow data, real-time weather conditions, and holiday factors. By mining time-series patterns in historical data, combining the impact of real-time weather changes on visitor travel intentions, and the peak visitor flow characteristics brought by holidays, it accurately predicts the congestion index of various areas of the park within the next two hours. Combined with the park's performance schedule, it provides visitors with clear information on performance times, locations, and content, allowing them to plan and book in advance according to their interests. Furthermore, the module uses clustering algorithms to deeply identify group visitor patterns, thereby optimizing the park's visitor flow management strategy, avoiding overcrowding in popular areas, and improving the overall visitor comfort.

[0046] When multiple reservation requests conflict, the reservation module will weight and schedule them according to user level (VIP / regular) and reservation time. This scientific weighted scheduling method ensures service fairness while maximizing the satisfaction of different visitors' needs, improving the efficiency and rationality of reservation processing, and ensuring the stable operation of the park's reservation system.

[0047] The module employs a fusion positioning solution combining GPS, Bluetooth beacons, and visual SLAM technologies. By fusing data from these three technologies, the module can generate precise location coordinates for visitors in real time, meeting the positioning needs in complex garden scenarios. It integrates a 3D model of the garden with real-time traffic data (such as pedestrian flow in various areas and temporarily closed paths), dynamically presenting the scenic layout, landmark locations, and optimal route planning through an AR view. Visitors can intuitively view the overlay effect of 3D reality and virtual navigation information, avoiding getting lost or missing attractions. The route planning algorithm incorporates a "cultural relevance" weight, prioritizing routes highly related to cultural elements such as garden history, architectural features, and plant landscapes. Traditional navigation coordinates or arrow indicators are transformed into natural language prompts, such as "Go forward 50 meters, turn right into the 'Secluded Path' corridor, where you will see a century-old ginkgo tree," improving navigation readability and immersive experience. In areas with weak GPS signals, it automatically switches to inertial navigation based on gyroscope and gait analysis. By analyzing data such as visitor stride length and direction changes, combined with a pre-constructed path topology on the park map, it maintains positioning accuracy within 1-2 meters. Differential positioning base station clusters are deployed at high points in the garden (such as mountaintops and towers) to construct local high-precision positioning areas. Tourist terminals receive differential signals sent by the base stations and correct positioning coordinates in real time, further reducing the error to sub-meter level, ensuring positioning accuracy at key scenic spots or complex paths;

[0048] The intelligent voice guide module utilizes location awareness, voiceprint analysis, and behavior recognition technologies to provide tourists with personalized, scenario-based multilingual guided tours. Based on the tourist's real-time location, the system automatically activates corresponding guided tour content when the tourist enters a preset geofence, covering multi-dimensional information such as garden history and culture, plant knowledge, and architectural anecdotes. Simultaneously, tourists can actively access in-depth cultural content by scanning QR codes on scenic spots, satisfying their personalized exploration needs. For example, scanning the QR code of an ancient tree provides access to its varietal characteristics, cultivation history, and other professional knowledge. The guided tour content supports multilingual switching to adapt to the needs of tourists of different nationalities and is synchronously linked to virtual annotations in the AR view, achieving an immersive experience of "voice narration + visual presentation."

[0049] By analyzing tourists' speech patterns to identify age, gender, and dialect characteristics, the system adjusts its narration strategies in real time. For older tourists or non-native speakers, the speaking speed is automatically reduced and technical vocabulary simplified; for younger tourists or culture enthusiasts, the speaking speed is increased, and more historical research or technical terms are incorporated. The system also dynamically adjusts the depth of the narration based on the tourists' cultural background, as determined by voiceprint analysis. For example, local historical anecdotes can be incorporated for tourists speaking a particular dialect; for international tourists, the focus is on comparing Chinese and Western garden cultures to enhance the content's relatability.

[0050] When the system detects that a visitor's stay at a scenic spot exceeds 90 seconds, it automatically activates the "Deep Exploration Mode." Based on the scenic spot's geographical coordinates and historical timeline, it plays related historical events or stories of famous figures. Combined with the visitor's behavior, it simultaneously pushes extended content such as seasonal changes in vegetation and architectural details, enriching the visitor's knowledge acquisition.

[0051] The dynamic route recommendation module integrates tourist behavior data, attraction popularity, and dynamic environmental factors to construct an intelligent system for generating and optimizing personalized routes.

[0052] By integrating tourists' historical travel preferences and the length of stay at each attraction, a personalized interest map is formed; real-time data such as visitor flow, visitor reviews, and check-in frequency of each attraction in the park are captured, and a dynamic popularity list is generated by combining seasonal factors; based on the weight ratio of interest tags and popularity data, a "interest-first + popularity-balanced" tour route is automatically generated and avoids peak hours.

[0053] The module incorporates reinforcement learning algorithms to improve route recommendation accuracy through a closed loop of "feedback-learning-optimization." Tourists can rate the current route using AR devices or actively mark "want to add / skip certain attractions," and the system collects this feedback data in real time. Based on the feedback data, the system continuously adjusts the route recommendation strategy. For example, if multiple tourists report that a certain route has "insufficient cultural relevance," the model will automatically increase the cultural weight of attractions in that area and prioritize routes associated with historical anecdotes in subsequent recommendations.

[0054] By monitoring the overlap of routes and time conflicts among groups of tourists in real time, if 10 groups of tourists are found to be planning to visit the "Lion Grove Rockery" from 11:00 to 11:30, a conflict scheduling mechanism is triggered. Priority weights are assigned to each group of tourists, taking into account factors such as user level, booking time, and group size. For example, VIP family groups have a higher weight than ordinary individual tourists. The optimal allocation scheme is solved using game theory models to achieve balanced resource scheduling in conflict scenarios while ensuring the core needs of each group of tourists are met, thus avoiding large-scale route congestion.

[0055] By connecting with the park's plant maintenance system, the system obtains real-time forecasts of flowering periods and flowering status for various plants, and automatically optimizes route priorities based on phenological data. For example, when cherry blossoms first bloom in mid-March, the "Cherry Blossom Garden" is included in the priority recommended routes, and the best viewing time is marked. If a sudden drop in temperature delays the flowering period of a certain plant, the system automatically adjusts the recommended order of related routes to prevent visitors from missing the best viewing conditions. While adjusting routes, the system also links the cultural content corresponding to the plants (such as pairing cherry blossom routes with explanations of Japanese garden culture, and lotus routes incorporating the poetic allusion of "rising from the mud unsullied"), achieving a dual value enhancement of "landscape viewing + cultural experience".

[0056] The AR culture generation unit, relying on large language models and multimodal recognition technology, performs semantic analysis on real-world garden images, generating scene-appropriate poetry, allusions, historical explanations, and other content in real time, which is then overlaid onto the AR view. Based on convolutional neural networks (CNNs), it extracts features from the architectural style, plant species, and landscape layout of the real-world garden. A knowledge graph is constructed, encompassing historical garden documents, a poetry database, and a cultural entity database. Through graph association, the architectural features of "Xiangzhou" are linked to knowledge nodes such as Ming Dynasty garden design concepts and literati boat dwelling culture; combining image feature extraction results with knowledge graph data, scene-appropriate cultural content is generated in real time. For example, by combining the boat-shaped structural features of "Xiangzhou" with the concept of "literati expressing their feelings through mountains and rivers" in the knowledge graph, an explanatory text is generated stating, "This building, shaped like a boat, metaphorically represents the literati's 'detachment from worldly affairs' sentiments," which is then transformed into dynamic annotations in the AR view.

[0057] Visitors can take and upload photos of a corner of the garden. The system uses image recognition to match corresponding scenes in historical documents, triggering a virtual restoration. In addition to photos, it can also combine GPS location, time information, and other dimensions to accurately match historical scenes.

[0058] Edge computing nodes are deployed in key areas of the park to directly process image data uploaded by visitors' terminals, avoiding cloud transmission delays and ensuring real-time overlay of AR content; high-frequency image recognition tasks are completed locally, reducing the overall network bandwidth consumption of the park and ensuring system stability; images taken by visitors are processed locally, avoiding sensitive data from being uploaded to the cloud and complying with data security standards.

[0059] The system employs a triple triggering method to achieve full-scene coverage of "passive perception, active exploration, and deep interaction." When tourists enter a preset geofence, the AR device automatically starts providing explanations; in-depth cultural content can be accessed by scanning the landscape's QR code; and hidden interactive scenes can be activated by recognizing specific gestures, achieving an immersive interaction of "gestures, historical scenes, and sensory experiences." Through the integration of large language models and multimodal technologies, the AR cultural generation unit transforms garden landscapes from "visual objects" into "interactive cultural carriers," enabling tourists to obtain in-depth cultural interpretations in real time during their virtual tours, achieving an immersive cultural tourism experience of "one scene, one story at every step."

[0060] Leveraging the visual computing capabilities of AR devices, virtual images of historical figures and scene reconstruction animations are precisely integrated with the real-world environment, creating an immersive experience where "ancient people and tourists stroll together." A skeletal animation system is constructed for virtual objects, and key motion nodes are matched with physical markers in the real garden scenery using algorithms to prevent virtual images from "floating" or "clipping." By continuously acquiring position sensor data from the AR devices, the world coordinates of the virtual content are updated in real time. When tourists move their viewpoint, the relative position and light and shadow projection angle of the virtual objects change synchronously with the real scene, with an error range controllable within 2 centimeters, ensuring a realistic fusion of virtual and real elements.

[0061] Simulating the multi-band photosensitive characteristics of the compound eyes of mantis shrimp, the sensing array covers 16 bands including ultraviolet, visible light, and near-infrared, capturing light field information in the garden environment in real time; based on multispectral data, the system can quickly reconstruct a high-precision 3D model of the environment, including the material properties and spatial structure of objects; this model provides an accurate physical spatial reference for the overlay of virtual content, ensuring that the shadows and reflections of virtual objects are consistent with the real scene;

[0062] The combination of spatial anchoring technology and multispectral perception enables the positioning accuracy of virtual content to reach the sub-centimeter level, which is 5-10 times better than traditional AR navigation systems, solving the industry pain point of "virtual object drifting"; the 3D environment modeling supported by 16-band light field data can realistically restore the material texture of garden landscapes, and the lighting effects of virtual content are synchronized with the real-world lighting conditions in real time, enhancing the sense of immersion.

[0063] As another aspect of this invention application, please refer to the appendix. Figure 2 To better illustrate the data acquisition and feedback implementation scheme of this invention, an AR-based garden landscape and cultural tourism guide system is described below:

[0064] I. System Deployment Environment

[0065]

[0066] II. Data Collection in the Operation Example of the Reservation Module

[0067] Data categories Collected values Data source Real-time temperature 28℃ API (American Weather Bureau) Precipitation probability 40％ Miniature weather station in the park Current number of visitors 1240 people Turnstile counter performance venue "XXX Show" 11:00 (200 seats) Park ERP System

[0068] LSTM congestion index prediction

[0069] area 10:00 Prediction 11:00 Prediction 12:00 Prediction Action Recommendations Point A 68％ 92％ 89％ 20% of reservations open Point B 42％ 55％ 63％ 50% of reservations open

[0070] Conflict scheduling

[0071] User ID User type Appointment time limit Demand Conflict Priority result U1001 VIP 7 days in advance 0.3 1.58 success U2034 ordinary That day 0.7 0.71 wait

[0072] III. Recommended Cultural Pathways

[0073] Path type Navigation instructions Cultural relevance Shortest path Walk straight for 200 meters to reach the finish line. none Cultural Path Walk 50 meters along the left-hand corridor to enjoy XXXX. History of Ming Dynasty Calligraphy and Garden Design

[0074] IV. Intelligent Voice Explanation Example: Voiceprint Adaptive Strategy

[0075] Voiceprint characteristics Explanation Strategy Example instructions Elderly + Dialect Explanation Strategy The Jianshan Tower before you was rebuilt during the Guangxu era. Youth + Mandarin Speech speed increased by 20%, incorporating technical terms. This window features a 'cracked ice' lattice pattern, reflecting the Song Dynasty style.

[0076] In-depth explanation of trigger logs

[0077] time Location Duration of stay Trigger content 10:23:45 Point C 92 seconds Play the corresponding audio

[0078] V. Examples of Dynamic Route Recommendation

[0079] Plant status adjustment

[0080] Plant Points Location Original route weight Adjusted weights action Lotus Breeze Pavilion Peak bloom 6.2 8.2 Prioritize and link with explanations of classical poems

[0081] VI. Comparison of Local Rendering Performance of AR Culture Generation and Interactive Examples

[0082] Task type Cloud latency Edge node latency promote Image recognition 1,200ms 85ms 93％ AR scene rendering 1,800ms 110ms 94％

[0083] Layered Trigger Statistics

[0084] Triggering method percentage Typical scenarios Automatic geofencing 62％ Enter "Point D" to begin the explanation of the construction history of the covered bridge. Manual triggering via QR code scanning 28％ Scan ancient ginkgo trees to obtain tree age testing reports Gesture Interaction 10％ Heart-shaped hand gesture activates AR love-themed poems

[0085] VII. Verification of Virtual-Real Fusion Effect: Spatial Anchor Point Accuracy Test (Unit: mm)

[0086] Scene Traditional AR systems Edge node latency promote Taihu stone surface 5.2 0.3 94％ wooden columns 3.8 0.2 95％

[0087] Multispectral reconstruction quality

[0088]

[0089]

[0090] After applying this system, tourists' stay time has been significantly extended, their cultural understanding has been greatly enhanced, AR positioning accuracy has been optimized to a breakthrough, and the system's response speed has been accelerated dramatically.

[0091] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A garden landscape and cultural tourism guide system based on AR technology, characterized in that, include: The reservation module calculates based on weather, visitor flow, and park performance times, making it convenient for tourists to book park performances, special experiences, and other activities in advance. The high-precision positioning and navigation module integrates GPS, Bluetooth beacons, and visual SLAM technologies to locate tourists in real time and provide accurate navigation services. The intelligent voice explanation module can be automatically triggered based on the visitor's location or manually triggered by scanning a QR code, and provides multilingual explanations related to garden history and culture, plant knowledge and architectural anecdotes. The dynamic route recommendation module generates personalized tour routes based on tourists' interest tags and attraction popularity. The AR culture generation unit analyzes real-world garden images based on a large language model, generates poems and allusions that match the cultural background of the scene in real time, and overlays them onto the AR view. The virtual-real fusion interactive unit uses AR devices to accurately overlay virtual images of historical figures, scene restoration animations, and real-world scenes, supporting gesture / voice interaction.

2. The garden landscape and cultural tourism guide system based on AR technology according to claim 1, characterized in that, The reservation module includes: The prediction unit uses LSTM neural network analysis of historical passenger flow data, real-time weather and holiday factors to predict the congestion index of each area in the next 2 hours. The scheduling unit performs weighted scheduling based on user level and reservation time limit when multiple reservation requests conflict.

3. The garden landscape and cultural tourism guide system based on AR technology according to claim 1, characterized in that, The positioning and navigation module includes: In areas with weak signals, switch to offline inertial navigation mode to maintain positioning accuracy through gyroscope and gait analysis; When planning routes, routes with high cultural relevance are prioritized, and navigation instructions are generated in natural language. A cluster of base stations is deployed at the highest point in the garden to build a local precision enhancement network. Tourist terminals receive differential signals from the base stations and correct their positioning coordinates in real time.

4. A garden landscape and cultural tourism guide system based on AR technology according to claim 1, characterized in that, The intelligent voice explanation module includes: The voice unit analyzes tourists' voiceprint characteristics such as age, gender, and dialect, and dynamically adjusts speech rate, vocabulary complexity, and cultural depth; when it detects that a tourist's stay time is >90 seconds, it automatically associates historical events.

5. A garden landscape and cultural tourism guide system based on AR technology according to claim 1, characterized in that, The dynamic route recommendation module includes: A reinforcement learning algorithm is introduced to dynamically optimize subsequent routes based on real-time feedback from tourists. When groups of tourists have conflicting routes, priorities are dynamically allocated based on game theory Nash equilibrium. The route is dynamically adjusted based on the flowering status of the plants.

6. A garden landscape and cultural tourism guide system based on AR technology according to claim 1, characterized in that, The AR culture generation unit includes: The image recognition subunit extracts architectural style, plant species, and landscape layout features based on a convolutional neural network. The cultural knowledge graph sub-unit links garden history documents, poetry databases, and cultural entity databases. Dynamically generate sub-units and combine feature extraction results with knowledge graphs to generate scene-adapted cultural content in real time.

7. A garden landscape and cultural tourism guide system based on AR technology according to claim 6, characterized in that, The AR culture generation unit includes: It supports multimodal input, allowing visitors to upload custom photos to trigger virtual restoration of specific historical scenes.

8. A garden landscape and cultural tourism guide system based on AR technology according to claim 6, characterized in that, The key areas of the garden are processed locally for image recognition and AR rendering tasks.

9. A garden landscape and cultural tourism guide system based on AR technology according to claim 6, characterized in that, The system is set up with a hierarchical triggering mechanism. It can be automatically triggered when tourists enter a preset geographical fence to start AR explanation; or manually triggered when scanning the landscape QR code to access in-depth cultural content. Gesture trigger: Recognize specific gestures to activate hidden interactive scenes.

10. A garden landscape and cultural tourism guide system based on AR technology according to claim 9, characterized in that, The virtual-real fusion interaction unit includes: By employing spatial anchor point technology, virtual object skeletal nodes are bound to physical marker points, and virtual content is bound to physical coordinates, thus reducing the error range; A multispectral environmental sensing array was constructed by simulating the compound eye structure of mantis shrimp to capture 16-band light field data in real time and generate a three-dimensional environmental model.

Images