Digital cultural tourism management system based on multi-source data analysis

Abstract

The invention discloses a digital cultural tourism management system based on multi-source data analysis, and the system comprises a multi-source heterogeneous data collection module which is used for collecting tourist behavior data, environment data, cultural resource data and third-party platform data in real time; the user demand intelligent analysis module is used for constructing a dynamic user portrait through the multi-modal interaction data and identifying dominant and implicit demands; the culture knowledge graph construction module is used for generating a reasonable multi-dimensional knowledge network based on culture resource attributes and historical data; the travel route dynamic generation module is used for generating a personalized touring route in combination with the user portrait, the real-time environment and the resource state; according to the digital cultural tourism management system based on multi-source data analysis disclosed by the invention, the utilization rate of cultural resources is greatly improved, and the decision response speed is increased; the cultural cognition depth of tourists is improved, and the residence time of the tourists is prolonged; the damage rate of high-sensitivity cultural relics is reduced; and the special group culture acquisition efficiency is improved in a breakthrough manner.

Description

Technical Field

[0001] This invention relates to the field of tourism management technology, and in particular to a digital cultural tourism management system based on multi-source data analysis. Background Technology

[0002] Cultural tourism refers to a form of tourism that uses culture as the primary destination and tourist experience. It emphasizes tourists' understanding, appreciation, and experience of the culture, history, art, customs, and traditions of a destination. The purpose of cultural tourism is to promote cultural exchange, cultural heritage protection, cultural inheritance, and sustainable tourism development through tourism activities.

[0003] Shortcomings of existing technology:

[0004] 1. Data silos: Sensor data, visitor behavior data, and cultural resource data within scenic areas are scattered and isolated; 2. Homogeneous experiences: Traditional guided tours lack personalization and cultural depth; 3. Conflict between protection and utilization: It is impossible to quantify the carrying capacity of cultural relics and balance the needs of visitors; 4. Lack of accessibility services: Visually and hearing impaired groups have difficulty accessing cultural content. Summary of the Invention

[0005] This invention discloses a digital cultural tourism management system based on multi-source data analysis. It studies and improves the existing structure and its shortcomings, and provides a digital cultural tourism management system based on multi-source data analysis to achieve better practical value.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A digital cultural tourism management system based on multi-source data analysis includes:

[0008] A multi-source heterogeneous data acquisition module is used to collect tourist behavior data, environmental data, cultural resource data, and third-party platform data in real time;

[0009] The user demand intelligent analysis module is used to build dynamic user profiles through multimodal interaction data and identify explicit and implicit needs;

[0010] The cultural knowledge graph construction module is used to generate a reasonable multidimensional knowledge network based on cultural resource attributes and historical data.

[0011] The travel route dynamic generation module is used to generate personalized tour routes by combining user profiles, real-time environment and resource status.

[0012] The intelligent recommendation and feedback optimization module is used to achieve closed-loop optimization of service recommendation and user feedback through reinforcement learning algorithms;

[0013] The terminal interaction service module is used to support XR navigation, multilingual adaptive services, and accessibility interaction.

[0014] In some embodiments, the user demand intelligent analysis module includes:

[0015] A multi-module input parsing unit is used to integrate voice commands, text search, historical trajectory, and physiological sensor data;

[0016] The requirements analysis module is used to categorize user needs into multiple dimensions such as cultural exploration, leisure and entertainment, and teaching and research.

[0017] The real-time intent prediction unit is used to analyze tourists' dwell time, gaze focus, and interaction actions in a scene through an LSTM network, and dynamically adjust the demand weights.

[0018] In some embodiments, the travel route ecosystem generation module includes:

[0019] The demand matching unit is used to parse user cultural preference tags and real-time status data;

[0020] Ecological constraint modeling unit, used to quantify the carrying capacity of cultural heritage and environmental impact indicators;

[0021] A multi-objective optimization engine is used to generate a set of route options that balance experience quality and ecological protection;

[0022] Real-time adaptive units are used to dynamically adjust the route in response to sudden environmental events.

[0023] In some embodiments, the intelligent recommendation and feedback optimization module includes:

[0024] A context-aware recommendation engine is used to generate contextualized recommendations by integrating user real-time location, time period preferences, and relationships with fellow users.

[0025] An incremental learning feedback mechanism is used to dynamically update the recommendation model based on user ratings of recommended content, skipping behavior, and repeat visit rate.

[0026] The cross-platform data synchronization unit is used to connect to social media APIs, collect user sentiment and tag usage habits in shared content, and optimize cultural dissemination strategies.

[0027] In some embodiments, the travel route ecology generation module integrates cultural heritage protection strategies, including:

[0028] Vulnerable cultural avoidance units are used to automatically avoid sensitive areas with high visitor flow and set dynamic visitor capacity thresholds for vulnerable cultural relics.

[0029] The cultural heritage enhancement path unit is used to insert interactive nodes of intangible cultural heritage inheritors and AR historical event trigger points in the route;

[0030] The Sustainable Development Assessment Unit is used to calculate route carbon emissions and provide carbon alternatives.

[0031] In some embodiments, the terminal interaction service module includes:

[0032] Immersive pre-experience units are used to preview key cultural scenes of the route via VR settings, and support users to virtually mark points of interest;

[0033] A multi-terminal seamless connection unit allows AR tours that are interrupted by users on their mobile devices to be played continuously on VR devices in the venue.

[0034] The accessibility feedback channel provides a quick language evaluation interface for visually impaired users and a vibration feedback scoring mechanism for hearing-impaired users.

[0035] In some embodiments, a cultural consumption guidance submodule is also included, the cultural consumption guidance submodule comprising:

[0036] A consumer demand forecasting model is used to recommend purchases of cultural products, reservations for specialty restaurants, and performance ticket packages based on tourist profiles.

[0037] A virtual-physical integrated consumption system is used to support the cross-scenario circulation of digital currency between physical stores and virtual exhibits and collections;

[0038] The consumer feedback knowledge feedback unit is used to inject purchase behavior data back into the knowledge graph, enhancing the connection between products and cultural IPs.

[0039] In some embodiments, the multi-source heterogeneous data acquisition module includes:

[0040] The sensory computing unit is used to capture tourists' real-time emotional fluctuations through facial expression recognition and voice emotion analysis.

[0041] The environmental perception enhancement unit is used to scan the building space structure using LiDAR to provide high-precision three-dimensional topology data for route generation.

[0042] The user-generated content mining unit is used to extract unstructured preference data from short videos and travelogues uploaded by tourists.

[0043] In some embodiments, the system further includes a mixed reality navigation optimization unit, the mixed reality navigation optimization unit comprising:

[0044] The lighting adaptive rendering unit is used to adjust the display parameters of the AR cultural relic restoration model according to the real-time ambient lighting.

[0045] Multi-user collaborative view unit, used to support group visitors to view each other's annotations and markers in a shared AR scene;

[0046] The physical-virtual interaction mapping unit is used to define the force feedback rule base when visitors touch virtual exhibits.

[0047] In some embodiments, a cultural value assessment system is also provided, the cultural value assessment system comprising:

[0048] The knowledge transfer effectiveness indicator unit is used to quantify the educational effect by comparing cultural cognition tests before and after the guided tour.

[0049] The cultural identity analysis unit is used to calculate the breadth of regional cultural dissemination based on sentiment analysis of content shared on social media.

[0050] The resource utilization health unit is used to generate a protection early warning heat map by combining cultural relic monitoring data with tourist density.

[0051] The digital cultural tourism management system based on multi-source data analysis provided by this invention has the following advantages:

[0052] (1) By utilizing the intelligent analysis module for user needs to connect multiple data sources inside and outside the scenic area, the utilization rate of cultural resources is greatly improved and the decision-making response speed is accelerated; (2) Through the cultural knowledge graph construction module and the lighting adaptive rendering unit, personalized cultural exploration paths are provided for tourists, enhancing the depth of tourists' cultural cognition and increasing the length of tourists' stay; (3) The ecological constraint modeling unit of the travel route ecological generation module accurately quantifies the balance point between protection and utilization, reducing the damage rate of highly sensitive cultural relics; (4) Through the barrier-free feedback channel of the terminal interactive service module, visually impaired users can perceive cultural relics through tactile encoding, and hearing impaired users can participate in cultural discussions through real-time sign language translation, significantly improving the cultural acquisition efficiency of special groups. This system has significant industrial value in improving management efficiency, strengthening cultural inheritance, and ensuring sustainable development when applied to multiple cultural venues. Attached Figure Description

[0053] Figure 1 This is a framework diagram of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0054] Figure 2 This is a framework diagram of a multi-source heterogeneous data acquisition module for a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0055] Figure 3 This is a framework diagram of the intelligent user demand analysis module of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0056] Figure 4This is a framework diagram of the travel route dynamic generation module of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0057] Figure 5 This is a flowchart of the travel route dynamic generation module of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0058] Figure 6 This is a flowchart of a multi-objective optimization engine for a travel route ecosystem generation module in a digital cultural tourism management system based on multi-source data analysis, as proposed in this invention.

[0059] Figure 7 This is a framework diagram of the intelligent recommendation and feedback optimization module of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0060] Figure 8 This is a framework diagram of the terminal interactive service module of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0061] Figure 9 This is a technical framework diagram of the terminal interactive service module of a digital cultural tourism management system based on multi-source data analysis proposed in this invention.

[0062] Figure 10 This is a framework diagram of the cultural consumption guidance submodule of a digital cultural tourism management system based on multi-source data analysis proposed in this invention. Detailed Implementation

[0063] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0064] Reference Figures 1 to 10 In a preferred embodiment, a digital cultural tourism management system based on multi-source data analysis includes:

[0065] A multi-source heterogeneous data acquisition module is used to collect tourist behavior data, environmental data, cultural resource data, and third-party platform data in real time;

[0066] The user demand intelligent analysis module is used to build dynamic user profiles through multimodal interaction data and identify explicit and implicit needs;

[0067] The cultural knowledge graph construction module is used to generate a reasonable multidimensional knowledge network based on cultural resource attributes and historical data.

[0068] The travel route dynamic generation module is used to generate personalized tour routes by combining user profiles, real-time environment and resource status.

[0069] The intelligent recommendation and feedback optimization module is used to achieve closed-loop optimization of service recommendation and user feedback through reinforcement learning algorithms;

[0070] The terminal interaction service module is used to support XR navigation, multilingual adaptive services, and accessibility interaction.

[0071] Specifically, such as Figure 2 As shown, the multi-source heterogeneous data acquisition module includes an IoT sensing unit, a visitor behavior capture unit, a cultural resource digitization unit, and a third-party data integration unit. The IoT sensing unit is deployed in a sensor network cluster within the cultural venue to collect real-time physical environment data and equipment status; the visitor behavior capture unit uses computer vision and mobile terminal positioning technology to acquire visitor activity trajectories and interactive behaviors; the cultural resource digitization unit performs high-precision 3D modeling and attribute extraction of culture and cultural carriers; and the third-party data integration unit connects to the open data interface of external platforms to obtain relevant auxiliary information.

[0072] Furthermore, the IoT sensing unit includes temperature and humidity sensors, air quality detectors, and noise decibel meters, all networked and transmitted via the LoRa protocol. It also includes smart meters for monitoring display case lighting energy consumption, vibration sensors for detecting minute displacements of artifacts (accuracy ±0.1mm), and access control counters for counting visitor flow. The IoT sensing unit's positioning beacon network is based on indoor positioning base stations using UWB technology (positioning accuracy 10cm), interacting with visitor mobile apps or rental terminals. The visitor behavior capture unit includes a video analytics subsystem, a mobile data acquisition unit, and a wearable device interface. The video analytics subsystem uses wide-angle cameras at key nodes combined with edge computing devices to run a YOLOv5 model to identify visitor posture, gaze direction, and hotspots. The mobile data acquisition unit obtains GPS tracks, screen operation logs, and Bluetooth beacon trigger records via the visitor mobile app. The wearable device interface supports devices such as Apple Watch for collecting physiological data such as heart rate and steps, which is transmitted after AES-256 encryption. The cultural resource digitization unit includes a 3D scanning device, an attribute extraction engine, and a degradation monitoring module. The 3D scanning device comprises a handheld laser scanner and a texture camera. The handheld laser scanner generates point cloud models of the cultural relics, while the texture camera captures surface color information. The attribute extraction engine uses OCR to recognize inscribed characters and speech-to-text technology to extract textual metadata from the audio explanations. The degradation monitoring module includes a miniature spectrometer used to periodically detect changes in the molecules of mural pigments. A third-party data integration unit integrates: real-time traffic data (specifically, obtaining bus arrival times and parking space availability information around the scenic area via municipal transportation API); social media sentiment analysis (specifically, crawling Twitter / Weibo hashtags and using the BERT model to analyze sentiment and topic popularity); and a weather warning interface (accessing short-term rainfall forecast data from the National Meteorological Administration with an accuracy of 1km grid). The IoT sensing unit also includes a microenvironment control link for cultural relics. This link automatically triggers the PID controller of the temperature-controlled display case to adjust parameters when temperature and humidity sensor data exceeds the cultural relic preservation threshold; and triggers the monitoring camera to turn to the alarm area and start recording evidence after the vibration sensor detects abnormal vibration. The tourist behavior capture unit includes a privacy protection mechanism. This mechanism involves real-time extraction of skeletal joint data using edge computing in video analysis, automatic local erasure of the original video stream, and uploading of mobile phone GPS trajectory data to the server after k-anonymization (k≥5). Physiological data is stored in a local secure enclave, with only the fatigue index output interface being open. The multi-source heterogeneous data acquisition module also includes a data preprocessing pipeline, comprising a spatiotemporal alignment engine, an outlier filter, and a metadata annotation unit. The spatiotemporal alignment engine is used to uniformly convert UWB positioning coordinates and GPS data to the scenic area's local coordinate system.The outlier filter uses the isolated forest algorithm to remove abnormal sensor readings. The metadata annotation unit adds structured tags such as artifact ID, acquisition time, and spatial location to the scanning model. The multi-source heterogeneous data acquisition module interacts with other modules of the system through a unified data bus. Specifically, it outputs to the user demand analysis module: real-time visitor location, physiological fatigue index, and distribution of interaction hotspots; inputs to the cultural knowledge graph construction module: 3D model of artifacts, degradation monitoring data, and semantic vectors of inscription text; and receives feedback from the intelligent recommendation and feedback optimization module: when visitors skip recommended exhibit IDs, it triggers adjustments to the behavioral data acquisition weights. The multi-source heterogeneous data acquisition module adopts a hierarchical energy management strategy: key artifact monitoring sensors use energy harvesting technology (solar power + vibration power generation) to ensure 24 / 7 operation; low-priority sensors (such as noise monitoring) enter sleep mode when visitor flow is less than a threshold; and edge computing nodes dynamically adjust their CPU frequency (1.2GHz-2.5GHz) based on task load.

[0073] like Figure 3 As shown, in some embodiments, the intelligent user demand analysis module includes: a multi-module input parsing unit for integrating voice commands, text search, historical trajectories, and physiological sensor data; a demand analysis module for classifying user demands into multi-dimensional tags such as cultural exploration, leisure and entertainment, and educational research; and a real-time intent prediction unit for analyzing tourists' dwell time, gaze focus, and interactive actions in the scene using an LSTM network, and dynamically adjusting demand weights. This intelligent user demand analysis module also includes a demand verification and optimization unit: calibrating model parameters through interactive feedback.

[0074] Specifically, the multi-module input parsing unit includes:

[0075] The voice command processing subunit is used to employ an end-to-end speech recognition model (such as Conformer-Transducer) to support dialect recognition and a dictionary of cultural proper nouns (such as "bronze taotie pattern").

[0076] The text semantic analysis subunit is used to combine the BERT model in the cultural domain to parse search keywords and travelogue reviews, and extract entity-sentiment pairs (such as <Terracotta Warriors, amazement>).

[0077] A behavior trajectory decoder is used to convert UWB positioning data into a dwelling heat map and identify high-frequency clustering areas through a clustering algorithm (accuracy ±0.5 meters).

[0078] The physiological signal adapter is used to analyze the PPG signal of the smart bracelet to calculate the real-time arousal index. The formula is as follows: Where E(t) is the index value at time t, HRV(t) is the heart rate variability at time t, and HRV baseIt is the baseline value for heart rate variability, and GSR(t) is the skin conductance response value at time t. max The maximum value of the skin conductance response is α = 0.6, and β = 0.4 is the cultural context calibration coefficient.

[0079] The implementation of the demand analysis module includes demand dimension segmentation, dynamic weight allocation, and demand conflict arbitration. Demand dimension segmentation includes: cultural exploration dimension: quantifying the depth of cultural relic knowledge demand (based on dwell time / AR call count); leisure and social dimension: assessing the frequency of interaction among companions and the distribution of photo hotspots; and educational study dimension: detecting the completion rate of guided tour playback and participation in test questions. Dynamic weight allocation includes: initial weights based on user profiles (age / nationality / occupation); and a real-time adjustment factor: when the rate of change in physiological excitement ΔE / Δt > a threshold, the leisure weight is increased by 15%. Demand conflict arbitration is specifically manifested in the activation of a compromise solution generation mechanism when the difference between the weight of cultural exploration for adults and the weight of leisure for children among family tourists is >40%.

[0080] The real-time intent prediction unit specifically includes:

[0081] The spatiotemporal context engine automatically increases the weight of educational research and study tours during the morning (8:00-10:00) as a time factor; and predicts the demand for handicraft experiences when near intangible cultural heritage workshops as a spatial factor.

[0082] The multi-source fusion prediction model consists of an input layer including current trajectory coordinates, gaze direction vector, and historical preference vector; a processing layer including a dual-channel LSTM network to process temporal behavior and spatial relationships respectively; and an output layer including Softmax to generate demand probability distributions (P_cultural exploration, P_leisure, P_education).

[0083] Emergency intent recognition specifically involves marking "departure request" as the highest priority when a tourist is detected repeatedly looking at the exit sign and their heart rate increases by more than 20%.

[0084] The requirement verification and optimization unit is implemented in the following ways:

[0085] The active verification interface is specifically implemented by popping up a request confirmation pop-up window on the AR interface (such as "Do you want to learn more about ceramic firing techniques?"), and the voice robot initiates multiple rounds of dialogue to verify implicit requests.

[0086] A negative feedback learning mechanism reduces the weight of the corresponding requirement tag when a user skips recommended content.

[0087] W new =W old ·e -λ·skip_count

[0088] Among them, W new W represents the updated weights or values.old This represents the current weight or initial value, λ = 0.2 is the decay coefficient, and skip_count represents the number of times or intervals to "skip" something.

[0089] Cross-scene transfer learning specifically involves transferring the requirement model for museum scene verification to ancient town scenic areas, and reducing distribution differences through Domain Adaptive Network (DANN).

[0090] The user demand intelligent analysis module also includes a cultural cognition assessment sub-unit, which includes:

[0091] The knowledge absorption detection unit is used to compare the similarity of tourists' descriptive texts of key cultural concepts before and after the guided tour (based on Sentence-BERT encoding).

[0092] The cultural identity calculation unit is used to analyze the frequency of regional cultural symbols (such as Qinqiang opera masks and Hui-style architecture) in social media shared content;

[0093] Generate personalized report units to generate a visual curve of user cultural literacy improvement and weaknesses (such as "insufficient knowledge of Tang Dynasty clothing").

[0094] The interaction between the user demand intelligent analysis module and other modules of the system includes:

[0095] The input end receives the raw sensor data stream from the multi-source heterogeneous data acquisition module and obtains the cultural entity popularity ranking from the cultural knowledge graph construction module.

[0096] The output sends a demand vector to the travel route dynamic generation module: {coordinates: (x, y), demand tags: [in-depth explanation: 0.8, rest: 0.3], effective duration: 120s}, and transmits the intent prediction result to the intelligent recommendation and feedback optimization module: {probability of visiting the porcelain exhibition area in the next 5 minutes: 82%}.

[0097] The feedback loop receives purchase records from the consumption guidance module and adjusts the consumption demand weights; when it receives evacuation instructions from the emergency module, it freezes the analysis of non-urgent demands.

[0098] The user demand intelligent analysis module also implements privacy protection strategies, including deleting audio files immediately after converting raw voice data to text locally; desensitizing behavioral trajectory data on the terminal and uploading only gridded area numbers (such as the A3 exhibition area); and establishing a user data sandbox, where demand models between different scenic spots are shared through federated learning.

[0099] In some embodiments, the cultural knowledge graph construction module includes: a multi-source cultural entity extraction unit, a spatiotemporal relationship reasoning unit, a graph dynamic evolution unit, and a cross-modal fusion unit, wherein: the multi-source cultural entity extraction unit extracts cultural concepts and their attributes from structured and unstructured data sources; the spatiotemporal relationship reasoning unit establishes historical evolution and geographical association rules between cultural entities; the graph dynamic evolution unit updates node weights based on real-time tourist behavior data; and the cross-modal fusion unit realizes knowledge association between text, image, and audio data.

[0100] Specifically, the multi-source cultural entity extraction unit includes:

[0101] The ancient text analysis subunit is used to extract historical figures, events, and technical terms using a named entity recognition model, which is pre-trained on a cultural classics corpus.

[0102] The cultural relic attribute extraction subunit is used to automatically label the age, material, and craftsmanship of cultural relics through geometric feature analysis of 3D scanning models;

[0103] Intangible cultural heritage linkers are used to extract the steps of a technique from oral recordings of inheritors and establish causal dependency chains between these steps.

[0104] The spatiotemporal relationship reasoning unit includes:

[0105] The historical timeline mapping unit is used to convert dynastic dates into a unified time coordinate system and mark the start / end timestamps of cultural entities;

[0106] Geospatial correlation is used to construct a topological map of cultural transmission paths based on the geographical coordinates of the places where cultural relics are unearthed and the locations recorded in historical documents. Furthermore, geospatial correlation specifically includes: assigning a spatiotemporal attenuation factor to cultural transmission paths, with the path weight decreasing as the transmission distance increases; establishing a model of the blocking / enhancing effects of historical events such as war / trade on transmission paths; and outputting a cultural influence radiation map to visualize the core cultural areas and their radiation range.

[0107] The spatiotemporal conflict detection unit triggers an expert review process when the deviation between newly acquired artifact dating data and existing maps exceeds a threshold.

[0108] The dynamic evolutionary unit of the map includes:

[0109] The node popularity calculation unit updates the node importance score based on visitor dwell time, AR interaction count, and social media mentions according to a preset weighting formula;

[0110] Knowledge gap identification is used to detect concepts not covered by the knowledge graph in high-frequency user queries and generate a list of knowledge nodes to be expanded; further, knowledge gap identification includes: analyzing tourists' unmatched search terms in the AR interface;

[0111] Detect unanswerable questions in the guide robot's dialogue; automatically generate data collection tasks when the frequency of similar knowledge gaps exceeds a threshold;

[0112] The automatic relationship completion unit uses a knowledge embedding model (such as TransE) to predict missing entity relationships, and automatically adds them if the confidence level is greater than a set value.

[0113] The cross-modal fusion unit includes:

[0114] Image semantic anchoring involves extracting feature vectors from artifact images using a convolutional neural network and aligning them with text description vectors in a shared space.

[0115] Audio knowledge association: Identify the names of cultural entities from the audio explanations and dynamically bind them to the corresponding graph nodes;

[0116] The multimodal search interface allows users to search for relevant cultural relic nodes using hand-drawn sketches.

[0117] The cultural knowledge graph construction module also includes an authenticity verification mechanism, which specifically includes: multi-source cross-validation, requiring at least two independent data sources (such as archaeological reports + local chronicles) to confirm the entry of the same entity's attributes; a version traceability system, where all graph update operations are recorded on the blockchain, allowing for the tracing of any historical version; and an expert collaboration platform, which grants certified scholars the right to mark disputed nodes and uses a majority voting mechanism to confirm modifications.

[0118] The cultural knowledge graph construction module also includes a copyright management sub-unit, which includes: digital copyright anchoring, which associates intellectual property registration numbers with each intangible cultural heritage skill node; derivative product authorization tracking, which records the reference relationship between cultural derivative product design drawings and original nodes of the graph; and a revenue distribution engine, which automatically calculates the copyright sharing ratio based on the contribution data of the knowledge graph.

[0119] Interaction between the cultural knowledge graph construction module and other modules of the system:

[0120] Input: Cultural relic scanning data and UGC content from the multi-source heterogeneous data acquisition module; real-time query intent vector from the user demand module;

[0121] Output: Sends a heat map of the spatial distribution of cultural nodes to the travel route dynamic generation module; provides an entity association matrix (e.g., the association strength of "Terracotta Warriors → Qin Terracotta Warriors Manufacturing Techniques" is 0.92) to the intelligent recommendation and feedback optimization module;

[0122] Feedback loop: Adjust the accessibility score of cultural nodes based on the deviation between the actual tour route and the recommended route.

[0123] like Figures 4 to 6As shown, in some embodiments, the travel route ecosystem generation module includes:

[0124] The demand matching unit is used to parse user cultural preference tags and real-time status data;

[0125] Ecological constraint modeling unit, used to quantify the carrying capacity of cultural heritage and environmental impact indicators;

[0126] A multi-objective optimization engine is used to generate a set of route options that balance experience quality and ecological protection;

[0127] Real-time adaptive units are used to dynamically adjust the route in response to sudden environmental events.

[0128] Specifically, the demand matching unit includes

[0129] The cultural appeal parser maps preference tags (such as "Tang Dynasty ceramics" and "religious architecture") from user profiles to knowledge graph nodes. The formula for calculating the cultural appeal matching degree is as follows:

[0130]

[0131] in, A matrix of cultural resource characteristics. Let ω be the user preference vector. i The real-time heat weight of node i.

[0132] A physical fitness assessment tool that calculates the physical exertion index of a route based on slope sensor data and the user's age.

[0133] The group demand arbitrator automatically inserts parent-child interaction nodes (such as the rubbing experience area) when it detects a conflict in preferences between children and adults among family visitors.

[0134] The demand matching unit also includes a demand-resource matching model, which includes:

[0135] Group demand arbitration algorithm:

[0136]

[0137] Where, α p Let p be the decision weight. To determine the preference similarity (based on cosine distance), Δ interact Enhanced parent-child interaction

[0138] The ecological constraint modeling unit includes: a dynamic carrying capacity model for cultural relics and a carbon footprint tracker. The calculation formula for the dynamic carrying capacity model for cultural relics is as follows:

[0139]

[0140] Where C0 is the initial load capacity (unit: person / hour), and α is the material attenuation coefficient (pottery α = 0.02, mural α = 0.05).

[0141] The formula for calculating the path carbon footprint of a carbon footprint tracker is:

[0142]

[0143] Where, d e ρ is the length of the road segment. e For road surface material coefficient (stone pavement ρ) e =1.0, lawn ρ e =1.3), δ e Carbon factor for tourists' gait.

[0144] The multi-objective optimization engine's solution formula is:

[0145]

[0146] Constrained by:

[0147] g1:

[0148] g2:

[0149] g3:

[0150] in, For the set of path access nodes, I i H represents the cultural value density of nodes. i For real-time popularity, A collection of sensitive cultural relics, This is a set of intangible cultural heritage activity periods. The multi-objective optimization engine adopts an improved NSGA-II algorithm: the crossover operator adapts to the discrete characteristics of cultural nodes, preserving continuous cultural theme fragments; the constraint processing mechanism transforms cultural relic overload into a penalty term.

[0151] The improved NSGA-II algorithm includes a cultural theme-preserving crossover operator:

[0152]

[0153] in, It is a random mask vector. As a theme enhancement factor, Δ theme This is a theme continuity correction item. The constraint violation penalty mechanism is as follows:

[0154]

[0155] Where λ1 and λ2 are adaptive penalty coefficients.

[0156] The dynamic carrying capacity model for cultural relics sets up a mechanism to strengthen cultural heritage preservation, which includes identifying the location p of the intangible cultural heritage inheritor in the knowledge graph. When the distance between the user's route and the location p of the intangible cultural heritage inheritor is less than 100m and the time period matches, a 15-minute interactive session is inserted. A virtual battlefield is automatically loaded at the coordinates (x, y) of the ancient battlefield site. The triggering conditions are that the user stays for more than 2 minutes and the light intensity is greater than 1000 lux.

[0157] The ecological constraint modeling unit also includes an ecological assessment model, which includes:

[0158] Cultural Communication Effectiveness Index:

[0159]

[0160] Among them, R i Let η be the knowledge test accuracy of node i. i It is a factor of cultural identity.

[0161] Ecological protection score:

[0162]

[0163] Where β = 0.5 and ω = 1.2 are calibration parameters.

[0164] The real-time adaptive unit includes:

[0165] In response to sudden environmental changes, after receiving a weather warning, indoor route libraries are activated to replace open-air routes; when the temperature and humidity of cultural relics exceed the standards, routes are replanned to avoid the area.

[0166] For group behavior coordination, when the population density in a local area is detected to be greater than 5 people / ㎡, time-sharing reservation and diversion will be initiated; alternative routes will be pushed to tourists' mobile phones, and navigation instructions will be updated after confirmation.

[0167] The real-time adaptive unit also includes a dynamic regulator, which executes a function to respond to sudden environmental changes.

[0168]

[0169] Where ΔE is the real-time environmental change matrix (such as rainfall intensity, temperature and humidity of cultural relics), and γ is the emergency adjustment weight.

[0170] like Figure 7 As shown, in some embodiments, the intelligent recommendation and feedback optimization module includes:

[0171] A context-aware recommendation engine is used to generate contextualized recommendations by integrating user real-time location, time period preferences, and relationships with fellow users.

[0172] An incremental learning feedback mechanism is used to dynamically update the recommendation model based on user ratings of recommended content, skipping behavior, and repeat visit rate.

[0173] The cross-platform data synchronization unit is used to connect to social media APIs, collect user sentiment and tag usage habits in shared content, and optimize cultural dissemination strategies.

[0174] Specifically, the context-aware recommendation engine generates recommendation strategies:

[0175]

[0176] in, For (location, fatigue level, cultural preference), C t For (weather, crowd density, time window, activity status), This represents the vector of cultural nodes along the current path. These elements may be integrated into a decision-making system or path planning algorithm. For example, the system first evaluates the entity's current state (location, fatigue level, cultural preferences). Next, the system considers current contextual factors (weather, crowd density, time window, activity status). Based on this information, the system may recommend one or more next cultural nodes to visit and update the path. Over time, this process will be repeated, thus forming a dynamic and adaptive path planning scheme.

[0177] The incremental learning feedback mechanism adopts a two-stage update mechanism:

[0178] Short-term online updates (every 5 minutes): in, It is a hybrid reward function, where 0.7 and 0.3 are the weighting coefficients for two reward signals. The choice of these weights depends on the specific application scenario and optimization objective. A higher immediate reward weight (e.g., 0.7) may be suitable for tasks that require rapid response and clear feedback. A higher expected reward weight (e.g., 0.3) focuses more on long-term planning and strategic decision-making.

[0179] Long-term offline updates (daily):

[0180] in, A value network is a neural network used to estimate the expected long-term reward for a given state or state-action pair. In reinforcement learning, it is often used to predict the sum of future rewards, thereby guiding the agent in making decisions.

[0181] Cross-platform data synchronization unit implementation:

[0182] Federal cultural preference migration:

[0183]

[0184] Where K is the number of participating scenic spots, i∈D k This is local data for the kth scenic area.

[0185] Social media knowledge distillation:

[0186]

[0187] Where T is the temperature parameter, z local , z social These are the eigenvectors.

[0188] The intelligent recommendation and feedback optimization module also includes a multimodal feedback collector, which includes:

[0189] Explicit feedback quantification:

[0190]

[0191] Multimodal implicit feedback modeling:

[0192]

[0193] in, For AR gesture operation trajectory.

[0194] The multimodal feedback collector also includes a cultural dissemination effectiveness assessment, which includes:

[0195] Knowledge transfer measurement:

[0196]

[0197] Cultural identity diffusion:

[0198]

[0199] Final reward function:

[0200] The intelligent recommendation and feedback optimization module also includes a privacy protection mechanism:

[0201] Homomorphic recommendation calculation:

[0202]

[0203] Differential privacy feedback:

[0204] Δf=maxr t -r′ t

[0205] In some embodiments, the travel route ecology generation module integrates cultural heritage protection strategies, including:

[0206] Vulnerable cultural avoidance units are used to automatically avoid sensitive areas with high visitor flow and set dynamic visitor capacity thresholds for vulnerable cultural relics.

[0207] The cultural heritage enhancement path unit is used to insert interactive nodes of intangible cultural heritage inheritors and AR historical event trigger points in the route;

[0208] The Sustainable Development Assessment Unit is used to calculate route carbon emissions and provide carbon alternatives.

[0209] Such as 8 and Figure 9 As shown, in some embodiments, the terminal interaction service module includes:

[0210] Immersive pre-experience units are used to preview key cultural scenes of the route via VR settings, and support users to virtually mark points of interest;

[0211] A multi-terminal seamless connection unit allows AR tours that are interrupted by users on their mobile devices to be played continuously on VR devices in the venue.

[0212] The accessibility feedback channel provides a quick language evaluation interface for visually impaired users and a vibration feedback scoring mechanism for hearing-impaired users.

[0213] Specifically, the immersive pre-experience units include:

[0214] AR scene dynamic generation algorithm:

[0215] ARConrnt = f render (K v (E, L)

[0216] Among them, K v The knowledge graph represents the multimedia resources associated with a specific node v. These multimedia resources can be 3D models, historical images, or other types of visual data, which provide rich information and context for the node. E is the ambient lighting matrix E = [Lux, color temperature, contrast]. L represents the user's position and posture in three-dimensional space.

[0217] Illumination adaptive compensation model: Automatically increase the brightness of virtual objects when the ambient lux is less than 1000.

[0218] The accessibility feedback channel includes a visual impairment assistive subsystem and a hearing impairment assistive subsystem. The impairment assistive subsystem includes spatial audio navigation: Gain L=Left channel gain decreases exponentially with the deviation from the target direction θ; the hearing-impaired assistive subsystem includes sign language motion capture: generating 3D skeletal point sequences via a TOF camera; real-time sign language translation:

[0219] Classification based on spatiotemporal graph convolutional networks.

[0220] The terminal interaction service module also includes a cultural behavior analyzer, which executes the following:

[0221] Attention hotspot detection:

[0222] Where λt = 0.1 is the time decay factor;

[0223] Cultural cognition depth assessment: Calculation of the matching degree between AR interaction duration and knowledge graph node depth:

[0224]

[0225] The terminal interaction service module also includes a real-time collaboration system, which includes:

[0226] Multi-user shared AR space:

[0227] Virtual annotation persistence: The arrows drawn by user A must meet the following requirements to be displayed in user B's field of view:

[0228]

[0229] The terminal interaction service module also includes a privacy protection mechanism, which includes:

[0230] Local processing of biometrics:

[0231] Eye-tracking data retains only the gaze coordinates (x, y), and the original image is discarded immediately; voice commands are converted to text on the terminal and then the audio is deleted; data transmission is minimized: behavioral analysis results are uploaded in aggregate form: $\text{Report}={\text{Hotspot Area ID},\text{Average Dwell Time}}$

[0232] like Figure 10 As shown, in some embodiments, a cultural consumption guidance submodule is also included, which includes:

[0233] A consumer demand forecasting model is used to recommend purchases of cultural products, reservations for specialty restaurants, and performance ticket packages based on tourist profiles.

[0234] A virtual-physical integrated consumption system is used to support the cross-scenario circulation of digital currency between physical stores and virtual exhibits and collections;

[0235] The consumer feedback knowledge feedback unit is used to inject purchase behavior data back into the knowledge graph, enhancing the connection between products and cultural IPs.

[0236] In some embodiments, the multi-source heterogeneous data acquisition module includes:

[0237] The sensory computing unit is used to capture tourists' real-time emotional fluctuations through facial expression recognition and voice emotion analysis.

[0238] The environmental perception enhancement unit is used to scan the building space structure using LiDAR to provide high-precision three-dimensional topology data for route generation.

[0239] The user-generated content mining unit is used to extract unstructured preference data from short videos and travelogues uploaded by tourists.

[0240] In some embodiments, the digital cultural tourism management system based on multi-source data analysis further includes a mixed reality tour guide optimization unit, which comprises:

[0241] The lighting adaptive rendering unit is used to adjust the display parameters of the AR cultural relic restoration model according to the real-time ambient lighting.

[0242] Multi-user collaborative view unit, used to support group visitors to view each other's annotations and markers in a shared AR scene;

[0243] A physical-virtual interaction mapping unit is used to define a rule base for force feedback when visitors touch virtual exhibits. In some embodiments, a cultural value assessment system is also provided, which includes:

[0244] The knowledge transfer effectiveness indicator unit is used to quantify the educational effect by comparing cultural cognition tests before and after the guided tour.

[0245] The cultural identity analysis unit is used to calculate the breadth of regional cultural dissemination based on sentiment analysis of content shared on social media.

[0246] The resource utilization health unit is used to generate a protection early warning heat map by combining cultural relic monitoring data with tourist density.

[0247] In some embodiments, the digital cultural tourism management system based on multi-source data analysis further includes a dynamic pricing strategy module, which operates in the following manner:

[0248] Demand-responsive pricing: Adjusting prices for unique experiences based on route popularity and real-time capacity.

[0249] Cultural value added calculation: Set a premium coefficient for route nodes that include explanations by inheritors of intangible cultural heritage;

[0250] Public welfare compensation mechanism: A portion of the proceeds will be automatically allocated to the cultural relics digital protection fund.

[0251] In some embodiments, the digital cultural tourism management system based on multi-source data analysis also includes cross-scenic area collaborative management, which is achieved through the following methods:

[0252] Cultural theme connection: Recommend routes that extend across regional cultural contexts based on users' historical travel data;

[0253] Resource scheduling cloud platform: Unified and intelligent allocation of tour guide vehicles and tour guide resources across multiple scenic areas;

[0254] Emergency response network: Tourist evacuation plans for emergencies are automatically synchronized with surrounding related scenic spots.

[0255] Example 1 illustrates the operational plan of a digital cultural tourism management system based on multi-source data analysis in distinctive scenarios such as stone culture, Zen culture, and karst landforms. Details are as follows:

[0256] Multi-source heterogeneous data acquisition module startup:

[0257] Internet of Things (IoT) sensing units: Temperature and humidity sensors (to monitor the microenvironment of marble carving) and laser scanners (to collect 3D point clouds of intangible cultural heritage stone carvings) are deployed at the Yunfu Stone Craft Expo Center;

[0258] Tourist behavior capture unit: Tracks the movement of tourists in Guoen Temple (the hometown of the Sixth Patriarch of Zen Buddhism) through the UWB positioning system, and analyzes the gaze duration of the Platform Sutra stone carving by combining AR glasses gaze capture.

[0259] Third-party data integration unit: Access to the meteorological bureau's rainstorm warning (affecting Panlong Cave karst landform tours) and the social media topic popularity of "Yunfu Stone Art".

[0260] User demand intelligent analysis module operation:

[0261] Identify the following tags for Guangzhou study tour groups: [Stone craftsmanship: 0.9, Zen culture: 0.7, Cave geology: 0.6];

[0262] Real-time intent prediction unit: Detecting tourists repeatedly lingering in the stone carving master workshop → triggering the demand for "stone carving DIY experience" with a weight increase of 0.3;

[0263] Arbitrator of group needs: Balancing the conflict between elderly tourists (deep Zen culture) and children (interactive entertainment), by inserting Zen sand painting experience nodes.

[0264] The response of the cultural knowledge graph construction module:

[0265] Dynamic updates: Expanding the knowledge branch of "Huineng's verses" based on questions asked by visitors in the AR explanation of the "Platform Sutra of the Sixth Patriarch";

[0266] Spatiotemporal reasoning: Connect the ancient stone transportation route (Nanjiang Ancient Post Road) with the contemporary stone art industry chain to generate a path for the dissemination of "ancient and modern stone art inheritance".

[0267] Travel route ecosystem generation module decision:

[0268] Ecological constraint modeling: The carrying capacity of the karst cave has been reduced to 50 people / hour (originally 120 people) due to the rainstorm warning. Low-carbon path calculation: detour via cement road and priority to stone slab ancient road (carbon emission reduction of 22%).

[0269] Multi-objective optimization engine: Output route: Stone Museum, Stone Carving Workshop (during the time when intangible cultural heritage inheritors are on site), Zen Sand Painting Area (to avoid rain), Guoen Temple (after the rain stops), satisfying: maximizing cultural value + physical exertion < elderly threshold + avoiding karst cave risk areas.

[0270] The intelligent recommendation and feedback optimization module is executed as follows:

[0271] AR scene push: Trigger a "virtual chiseling tutorial" in the stone carving workshop, and adjust the teaching progress according to the strength of the tourist's operation;

[0272] Negative feedback learning: 3 tourists skipped the "stone identification explanation", the priority of similar recommendations was reduced, and the weight of practical content was increased;

[0273] Cultural consumption guidance: Based on visitor stay data, we recommend customized stone seal services (engraved with surname + Zen saying).

[0274] Terminal interaction service module implementation:

[0275] Accessibility: Visually impaired tourists touch the stone sculpture replica → tactile coding vibrations transmit the "marble pattern" texture (short-long pulse = flowing water pattern); hearing impaired tourists ask questions through gestures → sign language recognition system displays animated subtitles of the Platform Sutra.

[0276] Cross-terminal synchronization: AR stone tracing tours that are not completed on the mobile device can be played on the museum's VR equipment.

[0277] Cultural behavior analysis: It was detected that teenagers spent an average of 8 minutes in the Zen culture area → Optimize the knowledge depth of the study tour route.

[0278] The system reflects the characteristics of Yunfu in the following ways:

[0279] Stone Intangible Cultural Heritage Protection: Laser displacement sensors monitor micro-cracks in stone carvings in real time (accuracy 0.01mm), and the data is stored on the blockchain. When the vibration exceeds the standard, the travel route ecosystem generation module automatically closes the exhibition area and pushes an alternative solution.

[0280] Immersive Zen Culture Experience: The cultural knowledge graph construction module builds a transmission chain of "Sixth Patriarch - Southern School of Zen - Japanese Tea Ceremony"; the lighting adaptive rendering unit overlays a scene of Tang Dynasty debate under the Bodhi tree at Guoen Temple (the lighting adaptive algorithm takes into account the shade of the banyan tree).

[0281] Sustainable tourism in karst landscapes: The CO2 concentration sensor in the cave is linked to the ecological constraint modeling unit to dynamically adjust the number of visitors; a dual mode of "geological research route" and "family adventure route" is generated to meet different needs.

[0282] Visitor ratings of the Zen-inspired sand art (explicit feedback) and AR interaction duration (implicit feedback) are fed back to the incremental learning optimizer, resulting in a 38% improvement in the accuracy of recommendations for similar groups the following day.

[0283] Any content not described in detail in this specification is prior art known to those skilled in the art.

[0284] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The substitutions may be replacements of some structures, devices, or method steps, or they may be complete technical solutions. Equivalent substitutions or modifications made to the technical solutions and inventive concepts of the present invention should all be covered within the scope of protection of the present invention.

Claims

1. A digital cultural tourism management system based on multi-source data analysis, characterized in that, include: A multi-source heterogeneous data acquisition module is used to collect tourist behavior data, environmental data, cultural resource data, and third-party platform data in real time; The user demand intelligent analysis module is used to build dynamic user profiles through multimodal interaction data and identify explicit and implicit needs; The cultural knowledge graph construction module is used to generate a reasonable multidimensional knowledge network based on cultural resource attributes and historical data. The travel route dynamic generation module is used to generate personalized tour routes by combining user profiles, real-time environment and resource status. The intelligent recommendation and feedback optimization module is used to achieve closed-loop optimization of service recommendation and user feedback through reinforcement learning algorithms; The terminal interaction service module is used to support XR navigation, multilingual adaptive services, and accessibility interaction.

2. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, The intelligent user demand analysis module includes: A multi-module input parsing unit is used to integrate voice commands, text search, historical trajectory, and physiological sensor data; The requirements analysis module is used to categorize user needs into multiple dimensions such as cultural exploration, leisure and entertainment, and teaching and research. The real-time intent prediction unit is used to analyze tourists' dwell time, gaze focus, and interaction actions in a scene through an LSTM network, and dynamically adjust the demand weights.

3. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, The travel route ecosystem generation module includes: The demand matching unit is used to parse user cultural preference tags and real-time status data; Ecological constraint modeling unit, used to quantify the carrying capacity of cultural heritage and environmental impact indicators; A multi-objective optimization engine is used to generate a set of route options that balance experience quality and ecological protection; Real-time adaptive units are used to dynamically adjust the route in response to sudden environmental events.

4. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, The intelligent recommendation and feedback optimization module includes: A context-aware recommendation engine is used to generate contextualized recommendations by integrating user real-time location, time period preferences, and relationships with fellow users. An incremental learning feedback mechanism is used to dynamically update the recommendation model based on user ratings of recommended content, skipping behavior, and repeat visit rate. The cross-platform data synchronization unit is used to connect to social media APIs, collect user sentiment and tag usage habits in shared content, and optimize cultural dissemination strategies.

5. The digital cultural tourism management system based on multi-source data analysis according to claim 3, characterized in that, The travel route ecology generation module integrates cultural heritage protection strategies, including: Vulnerable cultural avoidance units are used to automatically avoid sensitive areas with high visitor flow and set dynamic visitor capacity thresholds for vulnerable cultural relics. The cultural heritage enhancement path unit is used to insert interactive nodes of intangible cultural heritage inheritors and AR historical event trigger points in the route; The Sustainable Development Assessment Unit is used to calculate route carbon emissions and provide carbon alternatives.

6. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, The terminal interaction service module includes: Immersive pre-experience units are used to preview key cultural scenes of the route via VR settings, and support users to virtually mark points of interest; A multi-terminal seamless connection unit allows AR tours that are interrupted by users on their mobile devices to be played continuously on VR devices in the venue. The accessibility feedback channel provides a quick language evaluation interface for visually impaired users and a vibration feedback scoring mechanism for hearing-impaired users.

7. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, It also includes a cultural consumption guidance submodule, which includes: A consumer demand forecasting model is used to recommend purchases of cultural products, reservations for specialty restaurants, and performance ticket packages based on tourist profiles. A virtual-physical integrated consumption system is used to support the cross-scenario circulation of digital currency between physical stores and virtual exhibits and collections; The consumer feedback knowledge feedback unit is used to inject purchase behavior data back into the knowledge graph, enhancing the connection between products and cultural IPs.

8. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, The multi-source heterogeneous data acquisition module includes: The sensory computing unit is used to capture tourists' real-time emotional fluctuations through facial expression recognition and voice emotion analysis. The environmental perception enhancement unit is used to scan the building space structure using LiDAR to provide high-precision three-dimensional topology data for route generation. The user-generated content mining unit is used to extract unstructured preference data from short videos and travelogues uploaded by tourists.

9. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, The system also includes a mixed reality navigation optimization unit, which comprises: The lighting adaptive rendering unit is used to adjust the display parameters of the AR cultural relic restoration model according to the real-time ambient lighting. Multi-user collaborative view unit, used to support group visitors to view each other's annotations and markers in a shared AR scene; The physical-virtual interaction mapping unit is used to define the force feedback rule base when visitors touch virtual exhibits.

10. The digital cultural tourism management system based on multi-source data analysis according to claim 1, characterized in that, A cultural value assessment system is also established, which includes: The knowledge transfer effectiveness indicator unit is used to quantify the educational effect by comparing cultural cognition tests before and after the guided tour. The cultural identity analysis unit is used to calculate the breadth of regional cultural dissemination based on sentiment analysis of content shared on social media. The resource utilization health unit is used to generate a protection early warning heat map by combining cultural relic monitoring data with tourist density.

Images