Intelligent travel text big data supervision platform based on virtual reality

Abstract

The invention relates to the technical field of virtual reality, in particular to a virtual reality-based smart travel big data supervision platform, which comprises a data acquisition layer for acquiring multi-source data, a data processing layer for preprocessing the multi-source data to form a data packet, and a data transmission layer for transmitting the data packet, acquiring a data packet and constructing a virtual reality layer of the three-dimensional virtual mirror image space; and the supervision application layer is used for supervising the three-dimensional virtual mirror image space, and respectively issuing a priority transmission instruction of data at the problem to the data transmission layer and issuing a real-time reconstruction instruction of a three-dimensional scene at the problem to the virtual reality layer when the problem is monitored. According to the invention, in the data transmission layer, the redundant data shell of the protection data packet is dynamically adjusted according to the network quality, the temporary dedicated channel is constructed based on the SDN, and the data of the problem area directly reaches the virtual reality layer, so that the high-fidelity restoration and supervision of the virtual scene in the weak network environment are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of virtual reality, in particular to a smart tourism big data supervision platform based on virtual reality. BACKGROUND

[0002] With the vigorous development of smart tourism industry, virtual reality (VR) technology is increasingly widely used in tourism scenarios, and the construction of a three-dimensional virtual mirror space to realize digital supervision of tourism places has become an industry trend. However, the existing tourism supervision system has many challenges in data processing, transmission and three-dimensional scene construction: On the one hand, during data transmission, the uncertainty of the network environment (such as high bit error rate, low signal-to-noise ratio, etc.) will cause data loss or transmission delay, especially in the high-precision three-dimensional scene construction that virtual reality technology relies on, the integrity and real-time requirements of data are extremely high. The existing data transmission technology lacks a dynamic optimization mechanism for the data characteristics of tourism scenarios, and cannot adjust the transmission strategy in real time according to the network channel quality, resulting in limited construction precision and update efficiency of the three-dimensional virtual mirror space; On the other hand, when facing sudden problems such as tourist gathering exceeding the limit, facility abnormal vibration, etc., the traditional supervision platform (real-time display device) is difficult to quickly locate key data and prioritize transmission, and cannot realize real-time reconstruction and accurate supervision of the three-dimensional virtual scene, affecting the supervision efficiency and timeliness of decision-making.

[0003] In view of this, a smart tourism big data supervision platform based on virtual reality is proposed. SUMMARY

[0004] The purpose of the present application is to provide a smart tourism big data supervision platform based on virtual reality, which realizes the effect of "adapting to complex network environment and supporting real-time supervision".

[0005] In order to solve the above technical problems, the present application provides a smart tourism big data supervision platform based on virtual reality, comprising: A data acquisition layer comprising at least an Internet of Things sensor group for collecting multi-source data in a tourism scenario; A data processing layer connected to the data acquisition layer for preprocessing the multi-source data to form a data packet; A data transmission layer connected to the data processing layer for transmitting the data packet; A virtual reality layer connected to the data transmission layer for acquiring the data packet and constructing a three-dimensional virtual mirror space of the tourism scenario; The regulatory application layer is connected with the virtual reality layer and the data transmission layer respectively, and is used for monitoring the three-dimensional virtual mirror space, and when a problem is monitored, respectively issuing a priority transmission instruction of the data at the problem to the data transmission layer, and issuing a real-time reconstruction instruction of the three-dimensional scene at the problem to the virtual reality layer.

[0006] As a further improvement of the technical solution, the Internet of Things sensor group includes temperature sensors, humidity sensors, cameras, positioning beacons, pressure sensors and RFID card readers deployed in the tourism and travel site; the multi-source data includes space environment data, tourist behavior data and facility state data; wherein the temperature sensors and humidity sensors are used to collect space environment data, the cameras and positioning beacons are used to collect tourist behavior data, and the pressure sensors and RFID card readers are used to collect facility state data.

[0007] As a further improvement of the technical solution, the preprocessing includes data cleaning, format unification, data association and data compression of the multi-source data to form a standardized data package.

[0008] As a further improvement of the technical solution, the input end of the data transmission layer is provided with a protection node, and when the data package is input into the data transmission layer, the protection node copies the data package based on a data error correction mechanism, performs feature screening on the copied data package, retains key features and constitutes a redundant data shell, the data package is stored in the redundant data shell to constitute an encoded package.

[0009] As a further improvement of the technical solution, the user end further includes an identification module, and the identification module is used to generate a unique incremental number identification for the sub-file units in the original file order during transmission of the sub-file units.

[0010] As a further improvement of the technical solution, the redundant data shell is divided into an optimization layer and a core layer from outside to inside, wherein the core layer contains the basic features of the tourist position, space environment and facility ID in the data package; the optimization layer is an auxiliary feature based on the tourist behavior and dressing of the tourist position, the environmental parameter details based on the space environment, and the facility state change trend based on the facility ID. The input end of the data transmission layer and the inside of the data transmission layer are both provided with detection nodes, wherein the number of detection nodes at the input end of the data transmission layer is one, and the number of detection nodes inside the data transmission layer is multiple. All detection nodes are connected in sequence, the input end detection node is the first node, the last internal detection node is connected with the output end of the data transmission layer, and a unidirectional transmission link is formed, each detection node only detects the real-time network channel situation of the directly connected transmission link downstream, and according to the real-time network channel situation, performs a dynamic adjustment operation on the optimization layer in the redundant data shell, including reconstructing the optimization layer, retaining the complete optimization layer and deleting the optimization layer.

[0011] As a further improvement of the technical solution, when the optimization layer is deleted, the mutual information value of the auxiliary feature and the basic feature is calculated based on the mutual information formula by counting the probability distribution of the auxiliary feature and the basic feature; the auxiliary feature is summarized to obtain a plurality of condensed features, the correlation weight between the condensed features is calculated, and a condensed feature relationship network is constructed; the mutual information value is used as the correlation weight, and the condensed feature relationship network is anchored on the basic feature to form a new basic feature containing the basic feature and the condensed feature relationship network; When the optimization layer is reconstructed, the condensed feature relationship network and the mutual information value are extracted from the new basic feature; the corresponding auxiliary feature parameters are generated by reverse mapping according to the correlation weight of each condensed feature in the condensed feature relationship network; the auxiliary feature is refined by interpolation algorithm combined with the auxiliary feature parameters and the basic feature; the refined auxiliary feature is normalized according to the mutual information value to restore the auxiliary feature of the optimization layer.

[0012] As a further improvement of the technical solution, the real-time network channel condition is judged according to the bit error rate and the signal-to-noise ratio, specifically: When the bit error rate is less than or equal to the threshold value alpha and the signal-to-noise ratio is greater than or equal to the threshold value beta, the network is considered excellent, if the optimization layer has been deleted, the reconstruction operation is performed, otherwise the complete optimization layer is retained; When the bit error rate is greater than the threshold value alpha or the signal-to-noise ratio is less than the threshold value beta, the network is considered good, if the optimization layer data volume is greater than the threshold value gamma, the deletion and reconstruction operations are performed in turn; When the bit error rate is greater than the threshold value alpha and the signal-to-noise ratio is less than the threshold value beta, the network is considered poor, and the deletion operation is performed when the optimization layer has not been deleted.

[0013] As a further improvement of the technical solution, the detection node and the data transmission layer output end are further provided with a checking module, which is used to check whether the data of the data packet in the encoded packet is error when the encoded packet is transmitted to the detection node and the data transmission layer output end, and to replace the error data with the basic feature data of the core layer when the checking module finds that the data of the data packet is error.

[0014] As a further improvement of the technical solution, when the supervisory application layer detects that there are problems such as tourist gathering exceeding the limit, facility abnormal vibration and environmental parameter exceeding the standard in the three-dimensional virtual mirror space, it generates a supervision instruction containing the problem coordinate position and data priority identifier, and synchronously issues it to the data transmission layer and the virtual reality layer.

[0015] As a further improvement of the technical solution, after the detection node of the data transmission layer receives the supervision instruction, the current position of the encoded packet is determined based on the data priority identifier, specifically: If the encoded packet is located between the detection node N and N+1, a temporary dedicated channel directly reaching the virtual reality layer is constructed starting from the detection node N+1; If the encoded packet still stays at the detection node N, a temporary dedicated channel is constructed by dynamically configuring a dedicated bandwidth path through the SDN controller, taking the detection node N as the starting point. The temporary dedicated channel is constructed independently of the unidirectional transmission link through bypass logic, and the problem coordinate position corresponding area encoded packet is directly transmitted to the virtual reality layer through the temporary dedicated channel for real-time model construction.

[0016] Compared with the prior art, the beneficial effects of the present application are: In the wisdom travel big data supervision platform based on virtual reality, the data is divided into core layer basic features (such as tourist position, facility ID) and optimization layer auxiliary features (such as tourist behavior details, facility state change trend) by setting protection nodes and redundant data shells in the data transmission layer. The detection node dynamically adjusts the optimization layer according to the error rate and signal-to-noise ratio of the real-time network channel. When the network is excellent, the complete details are retained or reconstructed to ensure the accuracy of the three-dimensional scene. When the network is poor, the optimization layer is deleted to prioritize the transmission of core data, achieving a balance between transmission efficiency and data integrity, and adapting to complex network environments.

[0017] In the wisdom travel big data supervision platform based on virtual reality, when the supervision application layer detects problems such as tourist gathering exceeding the limit and facility abnormalities, it generates supervision instructions containing problem coordinates and priority, triggers the data transmission layer to build a temporary dedicated channel to directly transmit key data, and drives the virtual reality layer to reconstruct the three-dimensional scene of the problem area in real time, achieving dynamic monitoring and rapid response to the travel scene, and improving the real-time and accuracy of supervision. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The structure diagram of the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0020] In recent years, with the vigorous development of wisdom travel industry, virtual reality (VR) technology is increasingly widely used in travel scenes, and digital supervision of travel places through the construction of three-dimensional virtual mirror space has become an industry trend. However, the existing travel supervision system has many challenges in data processing, transmission and three-dimensional scene construction; At present, the existing big data supervision platform for smart tourism based on virtual reality generally has the problems that the uncertainty of network environment (such as high bit error rate and low signal-to-noise ratio) in the data transmission process will cause data loss or transmission delay, the supervision platform (real-time display device) is difficult to quickly locate key data and preferentially transmit when facing sudden problems such as tourist gathering over-limit and abnormal vibration of facilities, and cannot realize real-time reconstruction and accurate supervision of the three-dimensional virtual scene, thereby affecting the supervision efficiency and timeliness of decision-making. In view of this, please refer to Figure 1 The purpose of the present application is to provide a big data supervision platform for smart tourism based on virtual reality, which comprises: A data acquisition layer comprising at least an Internet of Things sensor group for collecting multi-source data in a tourism scene; A data processing layer connected with the data acquisition layer for pre-processing the multi-source data to form a data packet; A data transmission layer connected with the data processing layer for transmitting the data packet; A virtual reality layer connected with the data transmission layer for acquiring the data packet and constructing a three-dimensional virtual mirror space of the tourism scene; A supervision application layer connected with the virtual reality layer and the data transmission layer respectively for supervising the three-dimensional virtual mirror space, and issuing a preferential transmission instruction of problem data to the data transmission layer and a real-time reconstruction instruction of the three-dimensional scene at the problem to the virtual reality layer when a problem is detected.

[0021] In the big data supervision platform for smart tourism based on virtual reality, the data is divided into core layer basic features (such as tourist position and facility ID) and optimization layer auxiliary features (such as tourist behavior details and facility state change trend) by setting protection nodes and redundant data shells in the data transmission layer. The detection node dynamically adjusts the optimization layer according to the real-time network channel bit error rate and signal-to-noise ratio. When the network is good, the complete details are retained or reconstructed to ensure the accuracy of the three-dimensional scene, and when the network is poor, the optimization layer is deleted to preferentially transmit the core data, thereby balancing between transmission efficiency and data integrity, adapting to complex network environment. At the same time, when the supervision application layer detects problems such as tourist gathering over-limit and facility abnormality, it generates a supervision instruction containing problem coordinates and priority to trigger the data transmission layer to build a temporary dedicated channel to directly transmit key data, and drive the virtual reality layer to real-time reconstruct the three-dimensional scene of the problem area, thereby realizing dynamic monitoring and rapid response of the tourism scene and improving the real-time and accuracy of supervision.

[0022] Considering the need for comprehensive monitoring of spatial environment, tourist activities, and facility operation status in the context of cultural and tourism scenarios, and the difficulty of a single type of sensor to cover complex monitoring needs, in the data collection layer, the Internet of Things sensor group includes temperature sensors, humidity sensors, cameras, positioning beacons, pressure sensors, and RFID card readers deployed in cultural and tourism sites; multi-source data includes spatial environment data, tourist behavior data, and facility status data; by deploying different functional sensors, temperature / humidity sensors monitor environmental parameters in real time, cameras / positioning beacons capture tourist trajectories and gathering status, and pressure sensors / RFID card readers track facility load and equipment identification, achieving classified and accurate collection of multi-source data, specifically: Spatial environment data collection: through temperature / humidity sensor technology, real-time acquisition of temperature and humidity changes in cultural and tourism sites provides basic data for environmental comfort assessment and extreme weather warning, avoiding the lag of traditional manual inspection; Tourist behavior data collection: with the help of camera image recognition and indoor navigation technology of positioning beacons, tourist flow lines can be accurately tracked and gathering density can be counted, solving the problem of missed or misjudged traditional passenger flow statistics and providing real-time dynamic basis for tourist diversion strategy formulation; Facility status data collection: using pressure sensors to monitor facility structure load (such as bridge and trestle load bearing), and RFID card readers to identify equipment asset status, realizing full-cycle digital monitoring of cultural and tourism facility operation status, and early warning of structural safety hazards and equipment failure risks; Through multi-sensor collaborative networking, a three-dimensional data collection system covering “environment-people-facilities” is constructed, not only solving the problem of one-sidedness of single sensor data, but also through accurate matching of data types and functional zoning, providing standardized input interfaces for multi-source fusion analysis in the subsequent data processing layer, significantly improving the data completeness and scene adaptability of smart cultural and tourism supervision, enabling the supervision platform to achieve more intelligent risk prediction and decision support based on multi-dimensional data correlation.

[0023] Due to the problems of multi-source data collected by Internet of Things sensors in cultural and tourism scenarios, such as noise interference, format heterogeneity, weak correlation, and large data volume, direct transmission or application will lead to waste of computing resources, low analysis efficiency, and even decision bias, therefore, in the data processing layer, preprocessing includes data cleaning, format unification, data correlation, and data compression to form standardized data packets, specifically: Data cleaning technology: using algorithms such as outlier detection (such as Z-score method) and missing value interpolation (such as K-nearest neighbor filling), to eliminate error data caused by sensor failure and repair missing data caused by transmission interruption, solving the problem of “dirty data” in raw data and improving data accuracy, providing reliable data foundation for subsequent analysis; Format unification technology: Establish a standardized data dictionary, define the unified data structure and coding rules (such as JSON format standardization) of space environment data (temperature and humidity, light intensity), tourist behavior data (positioning coordinates, stay duration), and facility status data (device ID, load pressure), eliminate protocol differences of different sensor manufacturers, and improve the compatibility of cross-type data, greatly reducing the data interface cost; Data correlation technology: Based on the space-time dimension (such as the space-time matching of tourist positioning beacons and camera video streams) and business logic (such as the device ID binding of facility RFID tags and pressure sensors), a data correlation model is constructed to mine the potential relationship between data (such as the correlation between tourist gathering areas and facility load), forming a ternary data mapping network of "environment-people-facilities", and providing multi-dimensional correlation features for risk modeling of the supervision application layer; Data compression technology: For time-space sequence data (such as real-time monitoring data of temperature and humidity), differential coding (DPCM) is used, and for image and video data, H.265 compression algorithm is used, which significantly reduces the data transmission bandwidth demand on the premise of retaining most of the effective features, laying a foundation for efficient transmission of the subsequent data transmission layer; Through the pipeline processing of "cleaning-conversion-association-compression", the multi-source heterogeneous raw data is converted into standardized data packets with unified structure, complete features and light weight, which not only solves the problem of data quality and format incompatibility, but also improves the business value of data through data correlation mining, provides "clean, usable, and easy to process" data raw materials for the construction of high-precision three-dimensional mirror space in virtual reality layer, and at the same time reduces the load pressure of the data transmission layer, improves the data processing efficiency of the entire supervision platform, and provides a solid data support for real-time supervision decision-making.

[0024] Considering that the tourism supervision data faces network noise interference, data packet loss, transmission delay and other reliability risks during transmission, especially the integrity of key data such as tourist location and facility status directly affects the construction accuracy of three-dimensional virtual mirror space and the accuracy of supervision decision-making, traditional transmission protocols are difficult to guarantee the stable transmission of core data in complex network environment, therefore, the input end of the data transmission layer is provided with a protection node, the protection node copies the data packet based on the data error correction mechanism when the data packet is input into the data transmission layer, the copied data packet is feature screened, the key features are retained and constitute a redundant data shell, the data packet is stored in the redundant data shell to form an encoded packet, specifically: Data error correction and feature screening technology: The protection node first replicates the pre-processed data packet, screens the replicated data based on feature importance evaluation algorithm (such as information gain method), separates the basic features (constitute the redundant data shell core layer) such as tourist location coordinates, facility ID, spatial environment basic parameters, and auxiliary features (constitute the redundant data shell optimization layer) such as tourist behavior details, environmental parameter fluctuation trend, facility state change gradient. This technology improves the protection priority of key basic features to the highest level through data redundancy replication and feature layering, ensuring that the core data has anti-packet loss capability in the transmission link; Redundant data shell layering encoding technology: A double-layer protection structure of "core layer + optimization layer" is constructed, the core layer uses high robustness encoding (such as low density parity check code LDPC) for encryption and packaging, ensuring that the basic features can still recover the original data through error correction algorithm in poor network with high error rate; the optimization layer dynamically encodes according to the business relevance of auxiliary features, retains complete details when the network condition is good, and can be discarded or compressed flexibly when the network is congested. This layering mechanism makes the encoded packet meet the demand of virtual reality layer for high-precision data, and at the same time, it can prioritize the "minimum effective transmission" of core data when the network fluctuates; Data integrity check preposition technology: The protection node adds a digital fingerprint to each data packet through hash check (such as SHA-256) before generating the encoded packet, forming a double-check mechanism with the core features in the redundant data shell. When the transmission layer detection node finds data errors, it can directly use the core layer basic features for error correction, avoiding the delay problem caused by traditional retransmission mechanism, and shortening the data error correction response time.

[0025] Through the integrated processing of "replication-screening-layering-checking", a high-reliability transmission link suitable for tourism supervision scenarios is constructed; the redundant protection of core layer basic features improves the transmission success rate of key data in complex electromagnetic environment, network attack and other extreme scenarios, completely solves the pain point of "key data loss leading to three-dimensional scene distortion" in traditional transmission, and improves the anti-interference ability of data; the dynamic encoding mechanism of auxiliary features automatically reduces the volume of encoded packet when the network is congested, reduces the bandwidth occupation while avoiding link congestion, ensures the real-time response of supervision instructions (such as priority transmission of problem area data), optimizes the transmission efficiency; the standardized structure of redundant data shell adapts to various transmission protocols such as 5G, Wi-Fi, narrowband Internet of Things (NB-IoT), making the platform seamlessly access the existing network infrastructure in tourism places, reducing the system deployment cost, and enhancing the compatibility; Considering that different features in the tourism supervision data have different importance for the construction of three-dimensional virtual mirror and real-time supervision, basic features such as tourist location and facility ID are the core elements for constructing the spatial relationship of the scene and positioning the facilities, and once lost, it will lead to distortion of the virtual scene or "unanchoring" of the supervision object; while auxiliary features such as tourist behavior details and environmental parameter fluctuations can improve the scene refinement level, but transmitting full-quantity data in network congestion will exacerbate the delay, and the traditional "indiscriminate transmission" mode is difficult to balance between data integrity and transmission efficiency, therefore, the redundant data shell is divided into an optimization layer and a core layer from the outside to the inside, wherein the core layer contains the basic features of the tourist location, spatial environment and facility ID in the data packet; the optimization layer is the auxiliary features of the tourist behavior and clothing based on the tourist location, the environmental parameter details based on the spatial environment, and the facility state change trend based on the facility ID; through the use of feature grading and hierarchical packaging technology in the design of redundant data shell, differentiated protection and dynamic transmission of data are realized, specifically: Feature importance grading technology: use information gain algorithm (such as mutual information value calculation) to quantify the contribution of each data feature to the supervision decision, define tourist location coordinates (error needs to be ≤1 meter), facility unique identification (RFID code), and spatial environment basic parameters (temperature and humidity threshold) as core layer basic features (importance weight ≥0.8), the transmission reliability of which directly determines the "recognizability" of the three-dimensional scene; divide tourist clothing features (used for crowd density differentiation), environmental parameter gradient (temperature and humidity spatial distribution heat map), and facility load change rate (pressure sensor time series data) into optimization layer auxiliary features (importance weight 0.3-0.7), the value of which lies in improving the "refinement level" of the scene. This technology realizes the scientific division of data protection priority by establishing a feature grading model; Dual-layer heterogeneous coding technology: strong error correction coding (such as low-density parity-check code LDPC, coding redundancy 30%) is used for the core layer, for example, even if 30% of the data bits are damaged in transmission, the original features can still be recovered through decoding algorithm, ensuring that the transmission success rate of basic data such as tourist location and facility ID is ≥99.9% in a poor network with a bit error rate of up to 20%; lightweight elastic coding (such as principal component analysis PCA dimension reduction + differential pulse coding DPCM) is used for the optimization layer, which retains complete details (such as tourist moving speed vector, facility surface temperature distribution) when the network is good, and automatically compresses auxiliary features (such as reducing the environmental data sampling rate from 200Hz to 20Hz) when the network is congested, so that the data volume of the optimization layer can be dynamically adjusted; Network-aware dynamic adaptation technology: a real-time network quality evaluation module is deployed in the data transmission layer, based on bit error rate (BER) and signal-to-noise ratio (SNR) parameters, the "retention-compression-deletion" strategy of the optimization layer is driven through a pre-set rule engine: Network excellent (BER≤α and SNR≥β): Trigger the "detail first" strategy, if the optimization layer was deleted due to poor network before, restore the complete auxiliary features (such as tourist behavior trajectory details, environmental parameter fluctuation curve) through feature reconstruction technology to ensure that the virtual reality layer renders the scene with the highest accuracy; if the optimization layer has been completely retained, maintain the current transmission state to avoid unnecessary computational overhead; Network good (BER>α or SNR<β): Start the "efficiency balance" strategy, first detect whether the optimization layer data volume exceeds the threshold γ (such as 1MB), if it exceeds, then execute in turn: Delete the optimization layer: strip auxiliary features to reduce the volume of encoded packets (compression rate can reach 70%), prioritize fast transmission of core data through congested links; Feature condensation and relationship network anchoring: abstract auxiliary features into condensed feature relationship networks, embed them into the core layer to form new basic features, and reserve key association information for subsequent network recovery reconstruction; Network poor (BER>α and SNR<β): execute the "core bottom" strategy, immediately delete the optimization layer and only transmit the core layer basic features, while marking the link as "high-risk channel" through the SDN controller to trigger the standby path switching, ensuring that "supervision lifeline data" such as tourist location and facility abnormal state reaches with the lowest delay; It should be noted that the α, β, γ thresholds can be dynamically adjusted based on historical network data (such as the BER / SNR distribution of scenic spots at different times) through machine learning models (such as random forests), and can be dynamically adjusted according to specific scenarios, such as during the peak tourist season of holidays, automatically reduce α to 0.05% to improve data accuracy requirements; in indoor weak signal areas, increase β to 30dB to adapt to higher noise environments, so that the strategy engine always fits the actual network characteristics; Through the "core bottom, optimization enhancement" layered mechanism, the following significant benefits are brought: First, key data reliability is improved: the strong error correction protection of the core layer reduces the transmission error rate of basic features such as tourist location and facility ID, solving the core pain points of "tourist location drift in virtual scene" and "facility state display misalignment", providing "what you see is what you get" accurate scene mapping for supervisors; Second, intelligent allocation of network resources: the dynamic adjustment mechanism of the optimization layer reduces the average volume of encoded packets, and during the 5G network congestion period (such as the peak of holiday travel), more bandwidth resources can be released for emergency supervision instruction transmission (such as problem area data priority channel), shortening the supervision response time; Third, the scene construction is flexibly adapted: the hierarchical design is compatible with different precision virtual reality application requirements - on PC level high-performance terminals, super-realistic scenes containing tourist clothing details and environmental light changes can be presented; in mobile terminal lightweight applications, only the core layer data is loaded to quickly render the basic scene, realizing cross-device adaptation of "one code for multiple terminals", and reducing the terminal computing power threshold; Fourth, the fault tolerance capability is enhanced: when the detection node detects data errors (such as the spatio-temporal misalignment of camera video stream and positioning beacon data), the "anchor data" such as tourist position and facility ID in the core layer can be directly used for error correction, avoiding the delay caused by traditional full data retransmission, improving the virtual reality layer scene update frame rate, and ensuring the smoothness of supervision operation; In summary, the feature grading and hierarchical encapsulation technology, through the deep deconstruction of data value, builds an intelligent transmission system with "unbreakable basic features and flexible auxiliary features", which not only solves the problem of transmission reliability of key data in complex network environment, but also provides flexible data support for the fine construction of virtual reality scene, becoming the core technical support for the realization of "real-time, accurate and reliable" supervision of the intelligent tourism supervision platform.

[0026] Considering the long data transmission path in the tourism scene, the complex and variable network environment (such as signal fluctuation indoors and outdoors in scenic spots, and multi-device concurrent interference), a single detection point cannot realize real-time sensing of the network quality change of the whole link, and traditional end-to-end detection has the problems of high feedback delay and inability to optimize locally, therefore, detection nodes are arranged at the input end of the data transmission layer and inside the data transmission layer, the number of detection nodes at the input end of the data transmission layer is one, and the number of detection nodes inside the data transmission layer is multiple; all detection nodes are connected in sequence, the input end detection node is the first node, the last internal detection node is connected with the output end of the data transmission layer, forming a one-way transmission link, each detection node only detects the real-time network channel situation of the directly connected transmission link downstream, and according to the real-time network channel situation, performs dynamic adjustment operation on the optimization layer in the redundant data shell, including reconstructing the optimization layer, retaining the complete optimization layer, and deleting the optimization layer, specifically: Multi-point cascaded network sensing technology: the first detection node (S0) is deployed at the input end of the transmission layer, responsible for initial network quality evaluation of the incoming link encoded packet; multiple relay detection nodes (S1-Sn) are deployed at equal intervals inside the transmission link, each node only monitors the real-time bit error rate (BER) and signal-to-noise ratio (SNR) parameters of the directly connected link downstream (such as S1 monitors the S1→S2 link). This distributed deployment enables the system to capture local anomalies in the link (such as signal attenuation of a certain section of Wi-Fi), and compared with traditional single-point detection, the network state sensing accuracy is improved, and the response speed is shortened from seconds to milliseconds; One-way link coordination mechanism: All detection nodes establish one-way communication links through a lightweight heartbeat protocol (such as CoAP), and each node only needs to feed back the decision result (such as "delete optimization layer" and "reconstruction complete") to the upstream node, avoiding the complex consensus mechanism of traditional distributed systems. This design reduces the decision delay and improves the link throughput in the case of 10 detection nodes, effectively solving the "butterfly effect" problem (i.e., single-point failure causing the collapse of the entire link) in long-link transmission. Through the architectural innovation of "distributed sensing-local decision-global coordination", the following significant effects are brought about: Transmission reliability is improved: Multi-point detection improves the transmission success rate of critical data (such as tourist gathering area coordinates), ensuring the continuity of virtual reality layer scene construction in complex electromagnetic environments (such as metal building-intensive areas) in scenic areas; Optimization of bandwidth utilization: The dynamic deletion / reconstruction optimization layer mechanism reduces the link bandwidth occupancy rate, and in the busy period of 5G network (such as holiday tourist peak), limited bandwidth can be preferentially allocated to emergency supervision instructions (such as facility anomaly alarm), shortening the instruction response time; System fault tolerance is enhanced: When a detection node fails (such as hardware damage), adjacent nodes can automatically take over its monitoring function, maintain the basic transmission capacity of the link through the gradient decision model, and realize "detection node-level hot backup", with strong system availability; Reduced operation and maintenance costs: The distributed architecture avoids the single-point pressure of traditional central nodes, reduces the hardware cost of detection nodes, and reduces network retransmission energy consumption through local optimization of the link, meeting the "low-carbon" construction needs of travel sites; In summary, the distributed multi-point detection and gradient optimization technology sinks the network sensing and decision-making capabilities to each node of the transmission link, building an intelligent transmission system with "self-sensing, self-decision, and self-optimization", which not only solves the reliability problem of long-link transmission in complex travel scenarios, but also maximizes the use of limited network resources, providing a solid network infrastructure support for the real-time application of virtual reality technology in travel supervision.

[0027] Although deleting the optimization layer can improve the efficiency of core data transmission when the network is congested or fails, it will cause the loss of auxiliary features, and the virtual reality layer needs complete details to reconstruct a high-precision scene after the network recovers. The traditional "delete and discard" mode cannot balance transmission efficiency and data recoverability. Therefore, when the optimization layer is deleted, the mutual information value of the auxiliary features and the basic features is calculated based on the mutual information formula by statistically analyzing the probability distribution of the auxiliary features and the basic features. The auxiliary features are summarized to obtain a plurality of condensed features, the correlation weight between each condensed feature is calculated, and a condensed feature relationship network is constructed. The condensed feature relationship network is anchored to the basic features by taking the mutual information value as the correlation weight, and a new basic feature containing the basic features and the condensed feature relationship network is formed. When the optimization layer is reconstructed, the condensed feature relationship network and the mutual information value are extracted from the new basic feature. According to the correlation weight of each condensed feature in the condensed feature relationship network, the corresponding auxiliary feature parameters are generated by reverse mapping. The auxiliary features are refined by interpolation algorithm combined with the auxiliary feature parameters and the basic features. The refined auxiliary features are normalized according to the mutual information value to restore the auxiliary features of the optimization layer. The specific technical implementation path is as follows: First, the deletion stage: feature condensation and relationship network anchoring Mutual information quantifies correlation: use mutual information formula Calculate the information sharing degree of auxiliary features (such as tourist behavior trajectory ) and basic features (such as tourist location ), and select auxiliary features (mutual information value ≥0.6) that are strongly related to the core layer to ensure that key associations are not lost during condensation; Feature summarization and relationship network construction: K-means clustering is used to discretize continuous auxiliary features (such as environmental parameter fluctuation curve) to generate 5-10 condensed features (such as "high / medium / low" density aggregation state), and then the correlation weight between condensed features is calculated based on Pearson correlation coefficient (threshold ≥0.5) to construct a directed graph containing feature dependency relationships (such as "tourist stay duration → facility load change" causal link); Anchoring fusion to form new basic features: embed the condensed feature relationship network into the core layer basic features (such as adding "aggregation density level → facility load risk" mapping rules to tourist location coordinates) with mutual information value as weight, so that the new basic features which only account for 20% of the original data still retain more than 80% of the feature association information; Second, the reconstruction stage: reverse mapping and detail recovery Relationship network analysis and parameter generation: Decode the previously anchored condensed feature nodes (such as "high-density aggregation" and "facility load warning") and mutual information values (such as the correlation between tourist aggregation density and facility stress of 0.85) from the new basic features, and use graph traversal algorithms (such as depth-first search) to analyze the dependency between features (such as the causal link "tourist stay time → facility contact stress change"). Based on the pre-set feature mapping library (such as "aggregation level L3 → corresponding 15-20 people / ㎡" and "facility load level H → pressure sensor value ≥800N"), convert condensed features into initial auxiliary feature parameters (such as tourist moving speed interval and environmental parameter fluctuation range), ensuring that the restored feature parameters have a distribution error of ≤5% from the original data; Interpolation refinement and detail filling: For spatial continuous type data (such as temperature and humidity distribution, tourist trajectory coordinates), use cubic spline interpolation algorithm to dynamically fill intermediate values between condensed feature nodes (such as points A and B), so that the data sampling density is restored to 90% of the original (for example, expand the discrete 3 aggregation density points to 10 continuous gradient values). For multi-dimensional auxiliary features (such as tourist behavior speed, direction, and clothing color), combine the time and space reference in the basic features (tourist position coordinates and time stamp) to supplement missing details (such as estimating the current turning angle based on the moving direction at the previous and next time) through Gaussian process interpolation algorithm, ensuring that the reconstructed feature sequence is smooth and continuous. Mutual information driven normalization processing: According to the mutual information value, prioritize the auxiliary features (such as tourist moving trajectory details with mutual information value >0.7 are set as high priority), and retain the original accuracy of high-weight features (error ≤2%) during normalization, and moderately simplify low-weight features (such as individual tourist clothing RGB color values) (error ≤10%). Through Z-score standardization method, map the refined auxiliary feature parameters to the original data distribution interval (such as limiting the reconstructed temperature and humidity values within [-20℃, 40℃] and [0%, 100%]), ensuring dimensional consistency with the core layer basic features, and avoiding data conflicts during virtual reality layer rendering; Through the above technical means, in poor network conditions, the encoding package volume is compressed by deleting the optimization layer, the core data transmission speed is improved, and the "scene skeleton" data such as tourist position and facility state is ensured to reach the virtual reality layer in real time, avoiding "scene fragments"; when the network recovers, relying on the condensed feature relationship network and interpolation algorithm, the optimization layer is reconstructed in a short time, the "scene flesh and blood" data such as tourist behavior details (such as moving trajectory curve) and environmental parameter fluctuation (such as temperature and humidity gradient change) are quickly recovered, ensuring the coherence and authenticity of the three-dimensional scene; Secondly, the mutual information calculation and correlation weight construction technology ensures that the deletion operation only discards low-value independent features (such as slight differences in the clothing color of individual tourists), and retains high-value associated features (such as the association between the clothing commonality of a crowd and the regional safety risk), so that the regulatory application layer can still perform multi-dimensional association analysis of "tourist gathering-facility load-environmental impact" based on new basic features, avoiding the business logic break caused by traditional compression technology, thereby making the feature association be losslessly inherited; At the same time, in the stable outdoor 5G signal area of the scenic spot, the optimization layer is completely retained, supporting the virtual reality layer to render a hyper-realistic scene (with a precision of millimeter level) containing tourist facial expressions and facility surface textures. When entering an indoor weak signal area, the optimization layer is automatically deleted and the feature condensation mechanism is activated, only transmitting the core data of "tourist position + gathering level + facility basic state", ensuring that the mobile terminal device can still load the basic scene (the loading time is shortened from 10 seconds to 2 seconds), and at the same time, the complete scene is reconstructed within 1 second after the signal is restored, solving the pain point of "weak network black screen" in the traditional scheme; Moreover, when detecting an emergency event such as tourist gathering exceeding the limit, even in a network congestion state, the condensed new basic features can still provide the regulatory application layer with key information of "problem coordinates + risk level + associated facilities", supporting the rapid triggering of emergency plans (such as automatically planning a diversion path); after the event is handled, the complete data is restored through the optimization layer reconstruction technology, facilitating post-mortem analysis (such as the correlation between tourist behavior patterns and environmental parameters in the gathering event), realizing the "monitoring-response-review" closed loop of the regulatory process; In summary, the feature condensation and elastic reconstruction technology based on mutual information, through "intelligent slimming" and "precise recovery" of auxiliary features, builds a dynamic balance mechanism of "basic features real-time online, auxiliary features loaded on demand" in complex network environments, which not only solves the efficiency problem of "data transmission not moving", but also avoids the precision loss of "details not coming back", providing core technical support for the stable operation and accurate decision-making of the smart tourism regulatory platform in extreme network conditions, enabling the virtual reality layer to always present a "usable, reliable, and analyzable" three-dimensional scene mirror.

[0028] Since the data transmission process is easily affected by electromagnetic interference, link failure and other factors, bit-level errors or data packet damage may occur. Therefore, the detection node and the output end of the data transmission layer are also provided with a checking module. The checking module is used to check whether the data of the data packet in the encoded packet is erroneous when the encoded packet is transmitted to the detection node and the output end of the data transmission layer. When the checking module finds that the data of the data packet is erroneous, the basic feature data of the core layer is used to replace the erroneous data. Specifically: Hash fingerprint rapid detection: Deploy SHA-256 hash algorithm in the inspection module to generate 128-bit unique fingerprints for the core layer's basic features of the encoded package (such as tourist location coordinates and facility ID). Compare these fingerprints with real-time fingerprints calculated dynamically during transmission. Errors caused by data tampering or bit errors can be located within 0.1 ms (accuracy ≥ 99.99%). Error localization and isolation: When the inspection module detects an error (such as the output detecting that the tourist location coordinates do not pass the verification), immediately mark the error data block location and trigger the core layer's "cold start" mechanism, which directly calls the latest valid basic features stored in the redundant data shell core layer (such as the correct location coordinates cached 30 ms ago). Through the "rapid detection-accurate replacement" protection system, the "last line of defense" for data transmission is constructed. This technology not only solves the problem of bit errors in complex electromagnetic environments, but also achieves "zero-delay repair" of error data through efficient use of core layer redundant features, providing "absolutely reliable" basic data input for the virtual reality layer. In the smart tourism supervision scene, this technology makes the three-dimensional virtual mirror space a reliable "digital twin", allowing supervisors to make decisions without worrying about data errors, truly achieving the goal of "what you see is what you get, and what you get is correct" intelligent supervision.

[0029] Considering the suddenness and high risk of problems such as tourist gathering exceeding limits, facility abnormal vibration, and environmental parameter exceeding standards in the tourism scene, traditional supervision systems lack precise positioning and real-time response mechanisms for abnormal events. The priority of data transmission is often chaotic, and the three-dimensional scene update is lagging, leading to delayed disposal and even safety accidents. Therefore, when the supervision application layer detects problems such as tourist gathering exceeding limits, facility abnormal vibration, and environmental parameter exceeding standards in the three-dimensional virtual mirror space, it generates supervision instructions containing problem coordinate positions and data priority identifiers, and synchronously issues them to the data transmission layer and virtual reality layer. Specifically: First, real-time anomaly detection technology for three-dimensional scenes: Construct a multi-modal detection system that includes a spatial density model (tourist gathering degree ≥ 50 people / ㎡ triggers an over-limit warning), facility vibration spectrum analysis (identify abnormal frequency components through Fourier transform), and environmental parameter threshold engine (alarm when temperature and humidity / noise exceed GB / T26354 tourism place standards). Use computer vision (such as YOLO to detect tourist gathering) and machine learning (LSTM to predict facility vibration trends) algorithms to achieve 24-hour all-angle monitoring of the virtual mirror space, improving anomaly detection accuracy. When an anomaly is detected, automatically extract the three-dimensional coordinates (precision to 0.1 meters) of the problem area, event type (such as "facility abnormal vibration" corresponding to priority P1), and impact range (such as tourists and facilities within a radius of 50 meters), and generate structured supervision instructions, for example: {event ID, coordinates (x, y, z), priority, impact range, trigger time}; Second, priority-driven dual-link synchronization mechanism: Data transmission layer directional optimization: The "data priority identifier" (P1-P3) contained in the instruction triggers the SDN controller of the transmission layer to perform the encoding package for the problem coordinate position: If the encoding package is located between detection nodes N and N+1, a temporary dedicated channel of the direct virtual reality layer is constructed starting from detection node N+1 (the bandwidth is increased to 3 times that of the ordinary channel); If the encoding package is still at detection node N, a temporary dedicated channel is constructed by dynamically configuring a dedicated bandwidth path through the SDN controller starting from detection node N; If the encoding package is not sent, it is forcibly inserted into the head of the transmission queue through a priority queue scheduling algorithm (such as SPQ strict priority queue), ensuring that the scheduling is completed within a short time; Virtual reality layer real-time reconstruction: The problem coordinates carried by the instruction trigger the regional-level dynamic update of the scene rendering engine, and only the three-dimensional model of the problem area (such as within a 50-meter radius) is reconstructed, saving computing power compared to full-scene reconstruction, while completing model refreshing (including tourist dynamics, facility status, and environmental parameter visualization) within a very short time.

[0030] Third, cross-layer instruction verification and closed-loop feedback After the supervisory application layer issues an instruction, it verifies the temporary dedicated channel establishment state (3-way handshake confirmation) of the data transmission layer and the reconstruction completion degree (receiving scene hash value verification) of the virtual reality layer through the heartbeat mechanism. If a certain link is timed out (such as 200ms), the standby link (such as a satellite communication channel) is automatically triggered to ensure the success of instruction execution.

[0031] In summary, through the "accurate detection-priority driving-cross-layer cooperation" three-layer architecture, an "emergency nerve center" for intelligent tourism supervision is constructed. This technology not only solves the core problems of traditional systems such as "slow discovery, slow response, and chaotic processing", but also realizes "detection response, response visualization, and visualization control" through the deep coupling of three-dimensional scenes and data transmission. This technology reduces the incidence of safety accidents in tourism sites, improves the efficiency of supervision, and becomes a key technological breakthrough in promoting intelligent tourism from "digitalization" to "intelligence".

[0032] The above shows and describes the basic principles, main features, and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above examples, and the above examples and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1.A virtual reality-based smart travel and leisure big data supervision platform, characterized in that, The application relates to a data collection layer, a data processing layer, a data transmission layer, a virtual reality layer and a supervision application layer. The data collection layer comprises at least an Internet of Things sensor group for collecting multi-source data in a travel scene. The data processing layer is connected with the data collection layer and is used for preprocessing the multi-source data to form a data package. The data transmission layer is connected with the data processing layer and is used for transmitting the data package. The virtual reality layer is connected with the data transmission layer, acquires the data package and builds a three-dimensional virtual mirror space of the travel scene. The supervision application layer is connected with the virtual reality layer and the data transmission layer respectively, is used for supervising the three-dimensional virtual mirror space, and respectively issues a priority transmission instruction of problem data to the data transmission layer and a real-time reconstruction instruction of a three-dimensional scene at a problem position to the virtual reality layer when a problem is monitored. 2.The virtual reality-based smart travel and leisure big data supervision platform according to claim 1, characterized in that: The Internet of Things sensor group comprises temperature sensors, humidity sensors, cameras, positioning beacons, pressure sensors and RFID card readers arranged in a travel site. 3.The virtual reality-based smart travel and leisure big data supervision platform of claim 1, wherein, The multi-source data comprises space environment data, tourist behavior data and facility state data. 4.The virtual reality-based smart travel and leisure big data supervision platform of claim 1, characterized in that: The preprocessing comprises data cleaning, format unification, data correlation and data compression of the multi-source data to form a standardized data package. 5.The virtual reality-based smart travel and leisure big data supervision platform according to claim 4, characterized in that: The input end of the data transmission layer is provided with a protection node. The protection node copies the data package based on a data error correction mechanism when the data package is input into the data transmission layer, performs feature screening on the copied data package, retains key features and forms a redundant data shell, the data package is stored in the redundant data shell to form an encoded package. The redundant data shell is divided into an optimization layer and a core layer from outside to inside. 6.The virtual reality-based smart travel and leisure big data supervision platform according to claim 5, characterized in that, The core layer contains basic features of tourist positions, space environments and facility IDs in the data package. The optimization layer is an auxiliary feature based on tourist behavior and dressing of the tourists, environmental parameter details of the space environment and facility state change trends of the facility ID. The input end of the data transmission layer and the data transmission layer are both provided with detection nodes. The detection nodes are sequentially connected, the input end detection node is a first node, the last internal detection node is connected with the output end of the data transmission layer, a one-way transmission link is formed, each detection node only detects real-time network channel conditions of a directly connected transmission link downstream of the detection node, and dynamic adjustment operations of the optimization layer in the redundant data shell are performed according to the real-time network channel conditions, including reconstructing the optimization layer, retaining the complete optimization layer and deleting the optimization layer. When the optimization layer is deleted, mutual information values of the auxiliary feature and the basic feature are calculated based on a mutual information formula by counting probability distributions of the auxiliary feature and the basic feature. The auxiliary feature is summarized to obtain a plurality of condensed features, correlation weights between the condensed features are calculated, and a condensed feature relationship network is built. The condensed feature relationship network is anchored on the basic feature with the mutual information values as the correlation weights to form a new basic feature containing the basic feature and the condensed feature relationship network. The optimization layer is reconstructed, the condensed feature relationship network and mutual information value are extracted from the new basic features, the corresponding auxiliary feature parameters are generated according to the correlation weight of each condensed feature in the condensed feature relationship network, the auxiliary features are refined by interpolation algorithm in combination with the auxiliary feature parameters and the basic features, and the refined auxiliary features are normalized according to the mutual information value to restore the auxiliary features of the optimization layer. 7.The virtual reality-based smart travel and leisure big data supervision platform according to claim 5, characterized in that, The real-time network channel condition is judged according to the bit error rate and the signal-to-noise ratio, and specifically: When the bit error rate is less than or equal to a threshold value alpha and the signal-to-noise ratio is greater than or equal to a threshold value beta, the network is considered excellent, if the optimization layer has been deleted, the reconstruction operation is performed, otherwise the complete optimization layer is retained; When the bit error rate is greater than the threshold value alpha or the signal-to-noise ratio is less than the threshold value beta, the network is considered good, if the optimization layer data volume is greater than a threshold value gamma, the deletion and reconstruction operations are performed in turn; When the bit error rate is greater than the threshold value alpha and the signal-to-noise ratio is less than the threshold value beta, the network is considered poor, and when the optimization layer has not been deleted, the deletion operation is performed. 8.The virtual reality-based smart travel and leisure big data supervision platform according to claim 5, characterized in that: The detection node and the data transmission layer output end are further provided with a checking module, which is used to check whether the data of the data packet in the encoded packet is error when the encoded packet is transmitted to the detection node and the data transmission layer output end, and to replace the error data with the basic feature data of the core layer when the checking module finds that the data of the data packet is error. 9.The virtual reality-based smart travel and leisure big data supervision platform according to claim 5, characterized in that: When the supervisory application layer detects that there are problems of tourist gathering exceeding the limit, facility abnormal vibration and environmental parameter exceeding the standard in the three-dimensional virtual mirror space, it generates a supervision instruction containing the problem coordinate position and data priority identifier, and synchronously issues it to the data transmission layer and the virtual reality layer. 10.The virtual reality-based smart travel and leisure big data supervision platform according to claim 9, characterized in that: After the detection node of the data transmission layer receives the supervision instruction, it determines the current position of the encoded packet based on the data priority identifier, and specifically: If the encoded packet is located between the detection nodes N and N+1, a temporary dedicated channel directly reaching the virtual reality layer is constructed starting from the detection node N+1; If the encoded packet still stays in the detection node N, a temporary dedicated channel is constructed starting from the detection node N by dynamically configuring a dedicated bandwidth path through the SDN controller; The temporary dedicated channel is constructed independently of the one-way transmission link through bypass logic, and the region encoded packet corresponding to the problem coordinate position is directly transmitted to the virtual reality layer through the temporary dedicated channel for real-time model construction.

Images