Ice and snow art travel virtual exhibition hall real-time collaborative editing system and method

Abstract

The invention discloses a real-time collaborative editing system and method for an ice and snow art travel virtual exhibition hall, and particularly relates to the technical field of distributed computing, and the system comprises a central end preliminary screening analysis module, a local autonomous editing response module, an edge end fine screening optimization module, an edge collaborative synchronous exhibition arrangement module and a dynamic architecture switching control module. The central end preliminary screening analysis module outputs a first optimization suggestion, the local autonomous editing response module supports local modification and adjustment of exhibition item content, the edge end fine screening optimization module outputs a second optimization suggestion, and the edge collaborative synchronous exhibition arrangement module realizes synchronization between edge devices. According to the method, local network connection and exhibition item optimization are supported, under the condition that part of edge nodes are disconnected, the exhibition item optimization capacity of the remaining edge nodes is reserved, a weak network scene is effectively dealt with, and the robustness of the system and the feasibility of distributed virtual exhibition hall deployment are improved.

Description

Technical Field

[0001] The present invention relates to the field of distributed computing technology, and more specifically, to a real-time collaborative editing system and method for an ice and snow cultural tourism virtual exhibition hall. Background Art

[0002] Real-time collaborative editing of the virtual exhibition hall for ice and snow cultural tourism uses multi-party collaborative participation and real-time editing of the virtual exhibition hall for ice and snow culture and tourism to improve the efficiency and interactivity of content creation, management, and display. Ice and snow cultural tourism is highly temporal and spatial, and the exhibits in the virtual exhibition hall change with the winter solar terms. Ice and snow cultural tourism exhibits are usually located in outdoor scenic areas at high latitudes, high altitudes, and low temperatures, with severe network latency. Real-time editing that relies on cloud centers has problems with slow loading and easy failure of multi-person synchronization in remote exhibition hall terminals. Existing technologies mostly use distributed frameworks to decentralize the real-time collaborative editing system, perform proxy editing based on edge nodes, and limit multi-person collaborative editing to the edge nodes to avoid frequent multi-person synchronization failures caused by weak network connections between edge nodes and cloud centers. However, existing technologies fail to address the weak network problems of local networks between edge nodes. Due to the large exhibition area and numerous natural obstacles, delays and synchronization interruptions may also occur between edge devices. After the synchronization of each edge device fails, how to optimize the adjustment logic of each exhibit in the virtual exhibition hall has become an urgent problem to be solved.

[0003] In order to solve the above defects, a technical solution is now proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a real-time collaborative editing system and method for an ice and snow cultural tourism virtual exhibition hall to solve the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: a real-time collaborative editing system for a virtual exhibition hall for ice and snow cultural tourism, comprising a central-end preliminary screening and analysis module, a local autonomous editing response module, an edge-end fine screening and optimization module, an edge collaborative synchronous exhibition module, and a dynamic architecture switching control module;

[0006] The central end initial screening and analysis module is used to receive the first visitor behavior data from the visitor terminal, analyze the attention of the exhibition item and output the first optimization suggestion, and push the first optimization suggestion to the edge node;

[0007] The local autonomous editing response module is used to support editors to modify and adjust the exhibition content locally through edge devices when offline, and to perform local caching and asynchronous synchronization of the first visitor behavior data;

[0008] The edge-end fine screening and optimization module is used to perform refined exhibition item sorting, hotspot analysis, and second optimization suggestion generation based on edge computing capabilities after receiving real-time second visitor behavior data at the edge node;

[0009] The edge collaborative synchronous exhibition module is used to achieve local area network synchronization, conflict detection and collaborative presentation of exhibition item editing data among multiple edge devices, and perform multi-user and multi-device collaborative exhibition operations;

[0010] The dynamic architecture switching control module is used to monitor the network connectivity status between edge devices, evaluate the integrity of the second visitor behavior data collected by the edge device within the current time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity and node influence, and determine whether to perform architecture switching.

[0011] In a preferred embodiment, the first behavior heat vector is constructed based on the first tourist behavior data to perform feature extraction. After completing a virtual exhibition, the first behavior heat vector of the visitor is constructed , The expression is ,in, For tourists The duration of stay at exhibit i, For tourists Number of clicks on exhibit i, For tourists The number of interactions on exhibit i, including comments, likes, and drag models. n is the total number of exhibits, where , taking the first line as the heat vector as input, the density-based spatial clustering algorithm DBSCAN is used for clustering modeling;

[0012] According to the output results of cluster modeling, a group weight matrix is ​​constructed, tourists are divided into k categories, and a tourist group influence weight vector is constructed. The expression of the tourist group influence weight vector is: ,and , then the calculation expression for the weighted score of attention of each exhibition item is ,in 、 and are the length of stay, number of clicks and number of interactions of the ty-type visitors on exhibit i, 、 and They are 、 and The weight coefficient of 、 and All are positive numbers, and all exhibits are weighted and scored according to their attention. Sort the items from high to low to form a ranking list of exhibition items as the first optimization suggestion and push it to the edge node;

[0013] The length of stay of type ty tourists at exhibit i The calculation expression is , where is the length of stay of tourist u at exhibition item i, and tourist u belongs to category ty of tourists. is the number of tourists in category ty among tourists in category k;

[0014] Number of clicks on exhibit i by visitors of category ty The calculation expression is , where The number of clicks on exhibit i by visitor u;

[0015] Number of interactions of type ty visitors with exhibit i The calculation expression is , where is the number of interactions of visitor u on exhibit i.

[0016] In a preferred embodiment, the cluster modeling method is:

[0017] Calculate the Euclidean distance between the heat vectors in the first line, set the neighborhood radius threshold to Nrth, set the minimum number of neighborhood samples to minPts, and determine the core vector in the heat vector in the first line. If a vector If the neighborhood radius threshold Nrth contains at least the minimum number of neighborhood samples minPts as the first line heat vector, then the vector is determined to be a core vector. If the neighborhoods between core vectors overlap, the first line heat vectors to which the core vector belongs are classified into the same cluster, and the first line heat vectors that cannot be classified into any cluster are marked as outlier vectors. The output clustering result is ,in , k is the total number of clusters, that is, the total number of cluster types, and Ov is the set of outlier vectors.

[0018] In a preferred embodiment, the second behavior heat vector is constructed based on the real-time second visitor behavior data received by the edge node for feature extraction. For each exhibit item i, during its virtual exhibition, the second behavior heat vector of the exhibit item is constructed. , The expression is ,in, is the concentration of visitors’ gaze at exhibit i, is the crowd density in the adjacent area of ​​exhibition item i in unit time, is the ambient temperature of the area where exhibit i is located. An adaptive adjustment mechanism is used to dynamically adjust the weights of each factor in the heat vector of the second behavior according to the current time window and the current flow pressure of the exhibition area, and calculate the edge heat score. The calculation expression of the edge heat score is: ,in 、 and are respectively the concentration of visitors’ gaze at exhibit i, the crowd density in the adjacent area of ​​exhibit i per unit time, and the adaptive adjustment coefficient of the ambient temperature in the area where exhibit i is located, and 、 and are all positive numbers,

[0019] According to the marginal heat score of each exhibit Sort from high to low to form a list of edge exhibits’ attention ranking as a refined exhibit ranking, and push it to each edge node to calculate the average edge heat score of each exhibit. The calculation expression of the average edge heat score is: , calculate the standard deviation of the edge heat score of each exhibit. The standard deviation calculation expression of the edge heat score of each exhibit is , conduct hot spot analysis on the exhibits based on the average edge heat score and the standard deviation of the edge heat score of each exhibit. , then the hot spot analysis determines that the exhibit is a hot spot. If the edge heat score of the exhibit is , then the hot spot analysis determines that the exhibit is a low-activity exhibit, where, is the adjustment coefficient, and is a positive number;

[0020] Mark the hot exhibits as three-star exhibits, mark the low-activity exhibits as one-star exhibits, and mark the remaining exhibits except the hot exhibits and low-activity exhibits as two-star exhibits. Combined with the theme and exhibition area layout of each exhibit, they are packaged and sent to each edge node as the second optimization suggestion for editors.

[0021] In a preferred embodiment, the adaptive adjustment coefficient is adaptively adjusted according to the current time window and the current crowd pressure in the exhibition area, wherein , where p is the current crowd pressure in the exhibition area, that is, the crowd density index is normalized to , is the optimal visual focus value, is the standard deviation of the average number of clicks on exhibits by visitors in historical exhibitions, e is the natural base, and the adaptive adjustment coefficient , t is the current time window, with hours as the time unit, is the peak time window, the adaptive adjustment coefficient , where is the empirical control coefficient, and is a positive number, is the ambient temperature of the location of exhibit i, is the average ambient temperature of all exhibits, and the calculation expression is: , is the standard deviation of the ambient temperature at all exhibit locations, and the calculation expression is: , adjust the weight according to different exhibition time periods and crowd pressure.

[0022] In a preferred embodiment, the method for obtaining the concentration of a visitor's gaze focus at exhibit i is as follows: based on the visitor's sight interaction trajectory obtained by the eye tracking component and the spatial positioning component in the AR head-mounted device, the AR head-mounted device worn by the visitor includes a binocular camera, an infrared eye sensor and an IMU inertial navigation module, which are used to record the visitor's gaze point coordinates and head posture in real time; based on the AR spatial mapping, the visitor's gaze vector is matched with the boundary area of ​​the exhibit i in space; if the visitor's gaze point coordinates coincide with the projection model of the exhibit i, a focus hit event is recorded; if the visitor's gaze point coordinates deviate from the projection model of the exhibit i, a focus departure event is recorded; the time from each focus hit event to the focus departure event is defined as the witnessing duration; the number of visitors who visit the exhibit i within the unit time Ut is un; and the witnessing duration of each visitor's attention to the exhibit i is , then the calculation expression of the concentration of visitors' gaze at exhibit i is ;

[0023] The method for obtaining the crowd density in the adjacent area of ​​exhibit i within unit time is as follows: based on the identification and counting of Wi-Fi probe devices and Bluetooth beacons deployed around the exhibit, several Wi-Fi probe nodes and BLE Bluetooth receiving modules are deployed within the exhibition range of each exhibit. The Wi-Fi probe periodically scans the MAC signal emitted by the visitor terminal, and uses MAC address deduplication, hash anonymization and time window sliding algorithm to aggregate multiple scan records of the same visitor. Suppose the number of unique visitor devices detected in the exhibition area of ​​exhibit i within unit time Ut is Nvd, and the detection radius is Dr. Then the calculation expression for the crowd density in the adjacent area of ​​exhibit i within unit time is: .

[0024] In a preferred embodiment, the method for monitoring the network connectivity status between edge devices is:

[0025] The edge device periodically sends heartbeat status requests to other edge devices, and sets the maximum response time threshold to Maxrtt. If the edge device replies to the heartbeat status request within the maximum response time threshold Maxrtt, it is a valid response, otherwise it is an invalid response. If the edge device does not receive a valid response for Ret consecutive times, it is determined that the sending edge device of the heartbeat status request and the receiving edge device are in a disconnected state, and the receiving edge device of the heartbeat status request is defined as a disconnected node. Otherwise, it is determined that the sending edge device of the heartbeat status request and the receiving edge device are in a connected state, and the connectivity status of the edge device is calibrated. The expression is ;

[0026] The method for calculating the integrity of the second visitor behavior data collected by the edge device within the time window is:

[0027] According to the perceived activity and normalized data density in the time window Uwt, the degree of integrity is calculated. The calculation expression of the degree of integrity is: ,in, For the degree of completeness, To perceive activity, is the normalized data density, are the proportional coefficients of perceived activity and normalized data density, respectively, and All are positive numbers;

[0028] Perceived activity The method of obtaining is: suppose the number of sensing devices attached to the edge node is When the number of effective data reports of the sensing device within the time window Uwt is higher than the active threshold Act, the sensing device is determined to be an active device, and the number of active devices is calibrated as , then the perceived activity within the time window Uwt The calculation expression is ;

[0029] The method for obtaining normalized data density is as follows: let the number of second visitor behavior data items obtained within the time window Uwt be , is the number of displays controlled by the edge device, then the calculation expression of normalized data density is: , where is the reference standard density;

[0030] According to the impact of historical editing of edge devices, the method of calculating node influence is as follows: the optimization suggestions for each round of deployment are recorded locally on the edge node. The optimization suggestions include the first optimization suggestion and the second optimization suggestion. The optimization suggestions are recorded as a set , where is the number of the edge device, and ,in is the total number of edge devices, For edge devices In the The optimization suggestions generated in the time window are compared with the actual layout results and the adopted optimization suggestions are marked as ,and , optimization suggestions that were adopted Analyze the feedback effect and define the feedback improvement rate vector as ,in, is the visitor behavior feedback vector before the exhibit is adopted, is the tourist behavior feedback vector after the exhibit is adopted. The feedback vector includes the three dimensions of the exhibit’s stay time, click times, and interaction times. The threshold vector including the three dimensions of stay time, click times, and interaction times is set as , using the logical decision function Perform feedback judgment and calibrate the feedback effect to reach the threshold vector of the adopted optimization suggestions are hit suggestions ,and , then the calculation expression of node influence is .

[0031] In a preferred embodiment, the method for dynamic evaluation and judgment based on connectivity status, integrity level and node influence is:

[0032] The modeling fitness of edge nodes is calculated based on connectivity status, integrity level and node influence. The calculation expression of modeling fitness is: ,in, and Connectivity status , degree of completeness and node influence The weight coefficient of and Both are positive numbers, and the preset fitness threshold is Fith. If the calculated modeling fitness Mof is less than the fitness threshold Fith, the edge device is marked as a closed node. If the calculated modeling fitness Mof is greater than or equal to the fitness threshold Fith, the edge device is marked as an open node. For closed nodes, only local data is used for independent heat modeling, and for open nodes, data is synchronized for collaborative heat modeling.

[0033] A real-time collaborative editing method for a virtual exhibition hall of ice and snow cultural tourism, comprising the following steps: collecting first visitor behavior data through a visitor terminal, analyzing the attention paid to an exhibit and outputting a first optimization suggestion, and pushing the first optimization suggestion to an edge node;

[0034] Editors can modify and adjust the exhibit content locally through edge devices while offline, and perform local caching and asynchronous synchronization of the first visitor behavior data;

[0035] The edge node receives real-time second visitor behavior data and performs refined exhibition item sorting, hot spot analysis, and second optimization suggestion generation based on edge computing capabilities;

[0036] Conflict detection and collaborative presentation of exhibition item editing data are synchronized across multiple edge devices via the local area network, enabling collaborative exhibition operations for multiple users and multiple devices.

[0037] Monitor the network connectivity status between each edge device, evaluate the integrity of the second visitor behavior data collected by the edge device within the time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity and node influence, and determine whether to perform architecture switching.

[0038] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0039] By supporting local network connections and exhibition item optimization, the present invention retains the modeling capabilities of the remaining edge nodes when the central end is unreachable or some edge nodes lose network connection, ensuring that the exhibition item content can still be edited in an offline state. The cache and delay synchronization mechanism ensures the eventual consistency of data. The dynamic architecture switching mechanism evaluates the independent decision-making modeling mode locally through the edge node, without relying on the central server, effectively coping with weak network scenarios.

[0040] Based on integrated behavioral data such as gaze tracking, regional crowd density perception, and regional temperature characteristics, it is far superior to traditional click behavior statistics. The adaptive weight adjustment mechanism automatically adjusts the weight of each factor according to the time window and traffic pressure, making heat analysis more dynamic and adaptable to the environment. The edge device directly completes the identification of hot exhibits and the adjustment of low-activity exhibits, which helps to improve exhibition accuracy and interactive experience.

[0041] Adaptively switch between single-node and multi-node modeling, dynamically optimize computing resource allocation based on node connectivity status and data quality. Even if some edge nodes are offline, the remaining edge nodes still maintain functional operation, improving the robustness of the system and the feasibility of distributed exhibition hall deployment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0043] Figure 1 It is a system module diagram of the present invention.

[0044] Figure 2 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0045] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0046] Example 1: Please refer to Figure 1 As shown, the present invention is a real-time collaborative editing system for a virtual exhibition hall of ice and snow tourism, including a central-end preliminary screening and analysis module, a local autonomous editing response module, an edge-end fine screening and optimization module, an edge collaborative synchronous exhibition module, and a dynamic architecture switching control module;

[0047] The central end initial screening and analysis module is used to receive the first visitor behavior data from the visitor terminal, analyze the attention of the exhibition item and output the first optimization suggestion, and push the first optimization suggestion to the edge node;

[0048] The local autonomous editing response module is used to support editors to modify and adjust the exhibition content locally through edge devices when offline, and to perform local caching and asynchronous synchronization of the first visitor behavior data;

[0049] The edge-end fine screening and optimization module is used to perform refined exhibition item sorting, hotspot analysis, and second optimization suggestion generation based on edge computing capabilities after receiving real-time second visitor behavior data at the edge node;

[0050] The edge collaborative synchronous exhibition module is used to achieve local area network synchronization, conflict detection and collaborative presentation of exhibition item editing data among multiple edge devices, and perform multi-user and multi-device collaborative exhibition operations;

[0051] The dynamic architecture switching control module is used to monitor the network connectivity status between edge devices, determine whether the heat model synchronization conditions are met, evaluate the integrity of the second visitor behavior data collected by the edge device within the current time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity level and node influence, and determine whether to perform architecture switching.

[0052] The central-side initial screening and analysis module is deployed on the central server side of the virtual exhibition hall to identify visitors' hot spots, provide initial optimization suggestions for exhibits, and improve the basic accuracy and generalization capabilities of subsequent edge-side optimization.

[0053] The central-side initial screening and analysis module collects the first visitor behavior data within the historical time window Th from multiple visitor terminals through SDK integration or data collection interface, and caches it in the central database. The first visitor behavior data includes the number of click events and frequency of exhibit items, the duration of visitor viewing, visitor interaction behavior, and tour path trajectory;

[0054] According to the first tourist behavior data, the first behavior heat vector is constructed for feature extraction. After completing a virtual exhibition, the first behavior heat vector of the visitor is constructed , The expression is ,in, For tourists The duration of stay at exhibit i, For tourists Number of clicks on exhibit i, For tourists The number of interactions on exhibit i, including comments, likes, and drag models. n is the total number of exhibits, where , taking the first line as the heat vector as input, the density-based spatial clustering algorithm DBSCAN is used for cluster modeling. The cluster modeling method is:

[0055] Calculate the Euclidean distance between the heat vectors in the first line, set the neighborhood radius threshold to Nrth, set the minimum number of neighborhood samples to minPts, and determine the core vector in the heat vector in the first line. If a vector If the neighborhood radius threshold Nrth contains at least the minimum number of neighborhood samples minPts as the first line heat vector, then the vector is determined to be a core vector. If the neighborhoods between core vectors overlap, the first line heat vectors to which the core vector belongs are classified into the same cluster, and the first line heat vectors that cannot be classified into any cluster are marked as outlier vectors. The output clustering result is ,in , k is the total number of clusters, that is, the total number of cluster types, Ov is the set of outlier vectors;

[0056] According to the output results of cluster modeling, a group weight matrix is ​​constructed, tourists are divided into k categories, and a tourist group influence weight vector is constructed. The expression of the tourist group influence weight vector is: ,and , then the calculation expression for the weighted score of attention of each exhibition item is ,in 、 and are the length of stay, number of clicks and number of interactions of the ty-type visitors on exhibit i, 、 and They are 、 and The weight coefficient of 、 and All are positive numbers, and all exhibits are weighted and scored according to their attention. Sort the items from high to low to form a ranking list of exhibition items as the first optimization suggestion and push it to the edge node;

[0057] The length of stay of type ty tourists at exhibit i The calculation expression is , where is the length of stay of tourist u at exhibition item i, and tourist u belongs to category ty of tourists. is the number of tourists in category ty among tourists in category k;

[0058] Number of clicks on exhibit i by visitors of category ty The calculation expression is , where The number of clicks on exhibit i by visitor u;

[0059] Number of interactions of type ty visitors with exhibit i The calculation expression is , where is the number of interactions of visitor u on exhibit i.

[0060] The visitor terminal is a user-side interactive device used by visitors when browsing or participating in the content of the virtual exhibition hall, including smartphones, tablets, PC web browsers, head-mounted display devices, touch-screen all-in-one machines, and head-mounted display devices such as VR / AR devices installed with client applications.

[0061] When the edge device loses network connection with the central end, the local autonomous editing response module is used to support editors to perform local offline editing operations on exhibition layout, content resources, and interactive configurations. It stores the first visitor behavior data locally in timestamp order in the offline state and triggers asynchronous data upload when the network is restored;

[0062] The local autonomous editing response module enters the local editing mode by loading the most recently synchronized exhibit resources and exhibition configuration, supporting editors to adjust the exhibit location and partitions, replace exhibit content resources, and edit interactive logic and environment parameters in the local interface. During the local editing process, a version snapshot is generated, recording the exhibit ID, modified fields, operator ID, and timestamp of each modification. The first visitor behavior data received locally is cached in the local database in a time series. The cached data structure includes visitor ID, exhibit number, behavior type, behavior timestamp, and stay time. When the edge device restores the network connection with the center, the cached first visitor behavior data is checked for integrity and uploaded in batches according to the timestamp.

[0063] Through the above mechanism, even in a weak network or offline environment, the continuity of exhibition editing and data security can be guaranteed, and synchronization can be completed efficiently after the network is restored, ensuring the consistency of exhibition content and ensuring that the optimization chain is not interrupted.

[0064] The edge-end fine screening optimization module is used to perform refined exhibition item sorting, hot spot analysis and second optimization suggestion generation based on edge computing capabilities after the edge node receives real-time second visitor behavior data.

[0065] In the real-time collaborative editing system of the ice and snow cultural tourism virtual exhibition hall, the first visitor behavior data is centrally collected by the center end, and the first visitor behavior data is data within the historical time window. The first visitor behavior data is independently collected by the edge end and obtained in real time during the exhibition;

[0066] According to the real-time second visitor behavior data received by the edge node, the second behavior heat vector is constructed for feature extraction. For each exhibit item i, during its virtual exhibition process, the second behavior heat vector of the exhibit item is constructed. , The expression is ,in, is the concentration of visitors’ gaze at exhibit i, is the crowd density in the adjacent area of ​​exhibition item i in unit time, is the ambient temperature of the area where exhibit i is located. An adaptive adjustment mechanism is used to dynamically adjust the weights of each factor in the heat vector of the second behavior according to the current time window and the current flow pressure of the exhibition area, and calculate the edge heat score. The calculation expression of the edge heat score is: ,in 、 and are respectively the concentration of visitors’ gaze at exhibit i, the crowd density in the adjacent area of ​​exhibit i per unit time, and the adaptive adjustment coefficient of the ambient temperature in the area where exhibit i is located, and 、 and are all positive numbers,

[0067] According to the marginal heat score of each exhibit Sort from high to low to form a list of edge exhibits’ attention ranking as a refined exhibit ranking, and push it to each edge node to calculate the average edge heat score of each exhibit. The calculation expression of the average edge heat score is: , calculate the standard deviation of the edge heat score of each exhibit. The standard deviation calculation expression of the edge heat score of each exhibit is , conduct hot spot analysis on the exhibits based on the average edge heat score and the standard deviation of the edge heat score of each exhibit. , then the hot spot analysis determines that the exhibit is a hot spot. If the edge heat score of the exhibit is , then the hot spot analysis determines that the exhibit is a low-activity exhibit, where, is the adjustment coefficient, and is a positive number;

[0068] Mark the hot exhibits as three-star exhibits, mark the low-activity exhibits as one-star exhibits, and mark the remaining exhibits except the hot exhibits and low-activity exhibits as two-star exhibits. Combined with the theme and exhibition area layout of each exhibit, they are packaged and sent to each edge node as the second optimization suggestion for editors.

[0069] It should be noted that the adjustment coefficient It is set by professional technicians in this field based on the calculation accuracy of the edge device.

[0070] The adaptive adjustment coefficient is adaptively adjusted according to the current time window and the current exhibition area flow pressure, where , where p is the current crowd pressure in the exhibition area, that is, the crowd density index is normalized to , is the optimal visual focus value, is the standard deviation of the average number of clicks on exhibits by visitors in historical exhibitions, e is the natural base, and the adaptive adjustment coefficient , t is the current time window, with hours as the time unit, is the peak time window, the adaptive adjustment coefficient , where is the empirical control coefficient, and is a positive number, is the ambient temperature of the location of exhibit i, is the average ambient temperature of all exhibits, and the calculation expression is: , is the standard deviation of the ambient temperature at all exhibit locations, and the calculation expression is: , adjust the weight according to different exhibition time periods and crowd pressure, for example, increase , during the peak period of holidays, increase , in the afternoon or during hot weather, increase ;

[0071] Optimal visual focus value In order to ensure that visitors pay the most attention to visual focus when the flow of people is moderate, it is set by professional technicians in this field according to the path width and spatial distribution of the exhibition area. The standard deviation of the average number of clicks on exhibits by visitors in historical exhibitions The expression is ,in, is the number of clicks on exhibit item i by visitors in the history exhibition, is the average number of clicks on all exhibits by visitors in the historical exhibition, and the calculation expression is: , peak time window The experience control coefficient is set by the exhibition staff according to the browsing history of the exhibition area. It is set by the staff according to the environment of the exhibition area;

[0072] The method for obtaining the concentration of the visitor's gaze focus at exhibit i is as follows: the visitor's gaze interaction trajectory is obtained based on the eye tracking component and spatial positioning component in the AR head-mounted device. The AR head-mounted device worn by the visitor includes a binocular camera, an infrared eye sensor and an IMU inertial navigation module, which are used to record the visitor's gaze point coordinates and head posture in real time. According to the AR space mapping, the visitor's gaze vector is matched with the boundary area of ​​exhibit i in space. If the visitor's gaze point coordinates coincide with the projection model of exhibit i, a focus hit event is recorded. If the visitor's gaze point coordinates deviate from the projection model of exhibit i, a focus departure event is recorded. The time from each focus hit event to the focus departure event is defined as the witnessing duration. The number of visitors who visit exhibit i within the unit time Ut is un, and the witnessing duration of each visitor's attention to exhibit i is , then the calculation expression of the concentration of visitors' gaze at exhibit i is ;

[0073] The method for obtaining the crowd density in the adjacent area of ​​exhibit i within unit time is as follows: based on the identification and counting of Wi-Fi probe devices and Bluetooth beacons deployed around the exhibit, several Wi-Fi probe nodes and BLE Bluetooth receiving modules are deployed within the exhibition range of each exhibit. The Wi-Fi probe periodically scans the MAC signal emitted by the visitor terminal, and uses MAC address deduplication, hash anonymization and time window sliding algorithm to aggregate multiple scan records of the same visitor. Suppose the number of unique visitor devices detected in the exhibition area of ​​exhibit i within unit time Ut is Nvd, and the detection radius is Dr. Then the calculation expression for the crowd density in the adjacent area of ​​exhibit i within unit time is: .

[0074] By analyzing real-time secondary visitor behavior data for more detailed data granularity screening, and integrating multi-source behavior sensing methods such as AR eye tracking and Wi-Fi probe sensing, it accurately identifies local heat dynamics and improves the response sensitivity to individual behavior and local crowd flow dynamics. This makes up for the shortcomings of coarse-grained data analysis at the center in terms of exhibit item accuracy and spatial resolution. Even if the center is delayed or fails due to network reasons, the edge can still independently complete effective optimization, improving the continuous optimization capability in weak network environments.

[0075] In terms of heat modeling methods, the center relies on cluster analysis combined with weighted averaging, and the weighting logic is static weight distribution. The edge introduces an adaptive weight mechanism to dynamically adjust the weight in real time to respond to the actual flow of people in the exhibition area and changes in the time window. Through the analysis of the crowd density in the adjacent areas, it avoids the misjudgment that an exhibit has a large number of people and is therefore popular. In addition, it can quickly identify the exhibits that are truly the focus of attention in high-density areas of the exhibition area, thereby improving the accuracy of heat analysis.

[0076] The edge collaborative synchronization exhibition module is used to synchronize the exhibition item editing content between multiple edge devices through the multicast communication protocol in a local area network environment. Based on the operation timestamp and editing source identifier, it detects the exhibition item editing conflicts during the synchronization process, and adopts the priority merge strategy for version fusion, and updates the exhibition item status according to the fused editing results.

[0077] Each edge device runs in the local network of the virtual exhibition hall. Through the multicast communication protocol deployed in the local area network, the locally generated exhibition item editing operations are broadcast to all edge devices in the form of data packets. Each data packet contains the editing operation type, the unique identifier of the exhibition item, the operation source device ID and the operation timestamp.

[0078] It should be noted that the display item editing operations include modifying the display item location, replacing resources, and adding annotations;

[0079] When different edge devices perform editing operations on the same display item, the operation timestamp and device ID are compared to determine whether there is overlap or overwriting of operations. The conflict threshold is set as Coth. If the difference between different operation timestamps is lower than the conflict threshold Coth and the operation content conflicts, the merge strategy is triggered. The merge strategy determines the retained version based on the priority of the operation source device or the weight of the edit history. The non-retained version is recorded in the conflict log for manual review and recovery.

[0080] After the retained version is determined, it is pushed to all edge devices to update the exhibition item content in real time.

[0081] Conflict detection is performed on the edge device side to ensure data consistency between edge devices in weak network environments, avoid asynchronous editing versions due to poor connection to the central end, and determine the retained version for version merging based on the priority of the operation source device or the weight of the editing history. This improves the collaborative efficiency and user experience of multi-person editing and avoids confusion or loss of exhibit content due to editing conflicts.

[0082] The dynamic architecture switching control module is used to monitor the network connectivity status between edge devices, evaluate the integrity of the second visitor behavior data collected by the edge device within the current time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity and node influence, and determine whether to perform architecture switching.

[0083] The edge device periodically sends heartbeat status requests to other edge devices, and sets the maximum response time threshold to Maxrtt. If the edge device replies to the heartbeat status request within the maximum response time threshold Maxrtt, it is a valid response, otherwise it is an invalid response. If the edge device does not receive a valid response for Ret consecutive times, it is determined that the sending edge device of the heartbeat status request and the receiving edge device are in a disconnected state, and the receiving edge device of the heartbeat status request is defined as a disconnected node. Otherwise, it is determined that the sending edge device of the heartbeat status request and the receiving edge device are in a connected state, and the connectivity status of the edge device is calibrated. The expression is ;

[0084] The method for calculating the integrity of the second visitor behavior data collected by the edge device within the current time window is:

[0085] According to the perceived activity and normalized data density in the time window Uwt, the degree of integrity is calculated. The calculation expression of the degree of integrity is: ,in, For the degree of completeness, To perceive activity, is the normalized data density, are the proportional coefficients of perceived activity and normalized data density, respectively, and All are positive numbers;

[0086] Perceived activity The method of obtaining is: suppose the number of sensing devices attached to the edge node is When the number of effective data reports of the sensing device within the time window Uwt is higher than the active threshold Act, the sensing device is determined to be an active device, and the number of active devices is calibrated as , then the perceived activity within the time window Uwt The calculation expression is ;

[0087] The method for obtaining normalized data density is as follows: let the number of second visitor behavior data items obtained within the time window Uwt be , is the number of displays controlled by the edge device, then the calculation expression of normalized data density is: , where is the reference standard density;

[0088] It should be noted that the reference standard density It is set by professionals in this field;

[0089] The activity of the sensing device reflects the normal observation capabilities of the edge device from the perspective of the hardware working status. When the amount of data received decreases, if the edge device is still active, it is judged that the behavior is reduced rather than the system abnormality. The activity of the sensing device reflects the system health of the node sensing hardware in real time. If the activity decreases, it indicates that the node is in a functionally restricted state even if it is not completely disconnected.

[0090] The ratio of the number of behavioral data items to the number of exhibits is used to measure the observation support for each exhibit. Low density indicates insufficient sampling of local exhibits, avoiding the risk of meeting the overall data volume requirement but incomplete local data. The normalized data density indicator supports the judgment of node sample support before heat modeling, avoiding premature modeling of exhibits with insufficient data support, and improving the stability of ranking and recommendation.

[0091] Combining perceived activity with normalized data density provides a more comprehensive assessment, avoids ignoring local anomalies due to normal data volume, and improves the sensitivity of modeling quality control.

[0092] According to the impact of historical editing of edge devices, the method of calculating node influence is as follows: the optimization suggestions for each round of deployment are recorded locally on the edge node. The optimization suggestions include the first optimization suggestion and the second optimization suggestion. The optimization suggestions are recorded as a set , where is the number of the edge device, and ,in is the total number of edge devices, For edge devices In the The optimization suggestions generated in the time window are compared with the actual layout results and the adopted optimization suggestions are marked as ,and , optimization suggestions that were adopted Analyze the feedback effect and define the feedback improvement rate vector as ,in, is the visitor behavior feedback vector before the exhibit is adopted, is the tourist behavior feedback vector after the exhibit is adopted. The feedback vector includes the three dimensions of the exhibit’s stay time, click times, and interaction times. The threshold vector including the three dimensions of stay time, click times, and interaction times is set as , using the logical decision function Perform feedback judgment and calibrate the feedback effect to reach the threshold vector of the adopted optimization suggestions are hit suggestions ,and , then the calculation expression of node influence is ;

[0093] The percentage of successful optimization suggestions reflects whether the optimization suggestions previously proposed by edge devices have effectively improved the popularity of exhibits and user behavior. When the network is disconnected or some node data is missing, the importance of the disconnected node is judged based on the influence of the edge device, improving the balance between stability and optimization effect and strengthening the system's fault tolerance.

[0094] Based on connectivity status, integrity level, and node influence, the heat modeling mode of the edge-end fine-screening optimization module is dynamically evaluated to determine whether to switch architectures.

[0095] The method for dynamic evaluation and judgment based on connectivity status, integrity level and node influence is as follows:

[0096] The modeling fitness of edge nodes is calculated based on connectivity status, integrity level and node influence. The calculation expression of modeling fitness is: ,in, and Connectivity status , degree of completeness and node influence The weight coefficient of and Both are positive numbers, and the preset fitness threshold is Fith. If the calculated modeling fitness Mof is less than the fitness threshold Fith, the edge device is marked as a closed node. If the calculated modeling fitness Mof is greater than or equal to the fitness threshold Fith, the edge device is marked as an open node. For closed nodes, only local data is used for independent heat modeling, and for open nodes, data is synchronized for collaborative heat modeling.

[0097] After the dynamic architecture switching control module determines the heat modeling mode of the edge node, it feeds back to the edge-end fine screening optimization module, which performs refined item sorting, hotspot analysis, and second optimization suggestion generation based on the heat modeling mode and edge computing capabilities, forming a loop structure.

[0098] Example 2: Please refer to Figure 2 As shown, the present invention is a real-time collaborative editing method for a virtual exhibition hall of ice and snow cultural tourism, which specifically includes the following steps: collecting first visitor behavior data through a visitor terminal, analyzing the attention of an exhibit item and outputting a first optimization suggestion, and pushing the first optimization suggestion to an edge node;

[0099] Editors can modify and adjust the exhibit content locally through edge devices while offline, and perform local caching and asynchronous synchronization of the first visitor behavior data;

[0100] The edge node receives real-time second visitor behavior data and performs refined exhibition item sorting, hot spot analysis, and second optimization suggestion generation based on edge computing capabilities;

[0101] Conflict detection and collaborative presentation of exhibition item editing data are synchronized across multiple edge devices via the local area network, enabling collaborative exhibition operations for multiple users and multiple devices.

[0102] Monitor the network connectivity status between each edge device, evaluate the integrity of the second visitor behavior data collected by the edge device within the time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity and node influence, and determine whether to perform architecture switching.

[0103] By supporting local network connections and exhibition item optimization, the present invention retains the modeling capabilities of the remaining edge nodes when the central end is unreachable or some edge nodes lose network connection, ensuring that the exhibition item content can still be edited in an offline state. The cache and delay synchronization mechanism ensures the eventual consistency of data. The dynamic architecture switching mechanism evaluates the independent decision-making modeling mode locally through the edge node, without relying on the central server, effectively coping with weak network scenarios.

[0104] Based on integrated behavioral data such as gaze tracking, regional crowd density perception, and regional temperature characteristics, it is far superior to traditional click behavior statistics. The adaptive weight adjustment mechanism automatically adjusts the weight of each factor according to the time window and traffic pressure, making heat analysis more dynamic and adaptable to the environment. The edge device directly completes the identification of hot exhibits and the adjustment of low-activity exhibits, which helps to improve exhibition accuracy and interactive experience.

[0105] Adaptively switch between single-node and multi-node modeling, dynamically optimize computing resource allocation based on node connectivity status and data quality. Even if some edge nodes are offline, the remaining edge nodes still maintain functional operation, improving the robustness of the system and the feasibility of distributed exhibition hall deployment.

[0106] In this application, for the sake of terminology unification and technical expression standardization, edge devices and edge nodes are regarded as synonymous terms. Edge devices are computing terminals deployed at the exhibition site with local computing, perception and control capabilities, including but not limited to AR headsets, edge servers, smart guide terminals, smart interactive screens, mobile workstations, etc.

[0107] An edge node is an edge computing unit that participates in the exhibition hall's distributed collaborative operation system. It can be a single edge device or a collaborative entity composed of multiple edge devices.

[0108] Therefore, edge devices and edge nodes can be understood and used equivalently in this application, and have the following corresponding relationship in actual system deployment and control strategy:

[0109] In a distributed collaborative structure, each edge device with independent operation and decision-making capabilities can be regarded as an edge node;

[0110] An exhibition area subsystem composed of multiple devices, such as multiple collaboratively controlled display screens and guided interactive terminals, can also be logically regarded as a single edge node;

[0111] In dynamic architecture switching control, node integrity assessment, and node influence modeling, the term edge node is used to refer to each edge device or device collection that participates in behavioral data collection, processing, modeling, and response.

[0112] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0113] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the method described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

Claims

1. A real-time collaborative editing system for a virtual exhibition hall of ice and snow tourism, characterized by: It includes the central-end preliminary screening and analysis module, the local autonomous editing and response module, the edge-end fine screening and optimization module, the edge collaborative synchronization module, and the dynamic architecture switching control module; The central end initial screening and analysis module is used to receive the first visitor behavior data from the visitor terminal, analyze the attention of the exhibition item and output the first optimization suggestion, and push the first optimization suggestion to the edge node; The local autonomous editing response module is used to support editors to modify and adjust the exhibition content locally through edge devices when offline, and to perform local caching and asynchronous synchronization of the first visitor behavior data; The edge-end fine screening and optimization module is used to perform refined exhibition item sorting, hotspot analysis, and second optimization suggestion generation based on edge computing capabilities after receiving real-time second visitor behavior data at the edge node; The edge collaborative synchronous exhibition module is used to achieve local area network synchronization, conflict detection and collaborative presentation of exhibition item editing data among multiple edge devices, and perform multi-user and multi-device collaborative exhibition operations; The dynamic architecture switching control module is used to monitor the network connectivity status between edge devices, evaluate the integrity of the second visitor behavior data collected by the edge device within the current time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity and node influence, and determine whether to perform architecture switching.

2. A real-time collaborative editing system for a virtual exhibition hall of ice and snow culture and tourism according to claim 1, characterized in that: According to the first tourist behavior data, the first behavior heat vector is constructed for feature extraction. After completing a virtual exhibition, the first behavior heat vector of the visitor is constructed , The expression is ,in, For tourists The duration of stay at exhibit i, For tourists Number of clicks on exhibit i, For tourists The number of interactions on exhibit i, including comments, likes, and drag models. n is the total number of exhibits, where , taking the first line as the heat vector as input, the density-based spatial clustering algorithm DBSCAN is used for clustering modeling; According to the output results of cluster modeling, a group weight matrix is ​​constructed, tourists are divided into k categories, and a tourist group influence weight vector is constructed. The expression of the tourist group influence weight vector is: ,and , then the calculation expression for the weighted score of attention of each exhibition item is ,in 、 and are the length of stay, number of clicks and number of interactions of the ty-type visitors on exhibit i, 、 and They are 、 and The weight coefficient of 、 and All are positive numbers, and all exhibits are weighted and scored according to their attention. Sort the items from high to low to form a ranking list of exhibition items as the first optimization suggestion and push it to the edge node; The length of stay of type ty tourists at exhibit i The calculation expression is , where is the length of stay of tourist u at exhibition item i, and tourist u belongs to category ty of tourists. is the number of tourists in category ty among tourists in category k; Number of clicks on exhibit i by visitors of category ty The calculation expression is , where The number of clicks on exhibit i by visitor u; Number of interactions of type ty visitors with exhibit i The calculation expression is , where is the number of interactions of visitor u on exhibit i.

3. A real-time collaborative editing system for a virtual exhibition hall of ice and snow tourism according to claim 2, characterized in that: The method for cluster modeling is: Calculate the Euclidean distance between the heat vectors in the first line, set the neighborhood radius threshold to Nrth, set the minimum number of neighborhood samples to minPts, and determine the core vector in the heat vector in the first line. If a vector If the neighborhood radius threshold Nrth contains at least the minimum number of neighborhood samples minPts as the first line heat vector, then the vector is determined to be a core vector. If the neighborhoods between core vectors overlap, the first line heat vectors to which the core vector belongs are classified into the same cluster, and the first line heat vectors that cannot be classified into any cluster are marked as outlier vectors. The output clustering result is ,in , k is the total number of clusters, that is, the total number of cluster types, and Ov is the set of outlier vectors.

4. The real-time collaborative editing system for the ice and snow cultural tourism virtual exhibition hall according to claim 1 is characterized in that: According to the real-time second visitor behavior data received by the edge node, the second behavior heat vector is constructed for feature extraction. For each exhibit item i, during its virtual exhibition process, the second behavior heat vector of the exhibit item is constructed. , The expression is ,in, is the concentration of visitors’ gaze at exhibit i, is the crowd density in the adjacent area of ​​exhibition item i in unit time, is the ambient temperature of the area where exhibit i is located. An adaptive adjustment mechanism is used to dynamically adjust the weights of each factor in the heat vector of the second behavior according to the current time window and the current flow pressure of the exhibition area, and calculate the edge heat score. The calculation expression of the edge heat score is: ,in 、 and are respectively the concentration of visitors’ gaze at exhibit i, the crowd density in the adjacent area of ​​exhibit i per unit time, and the adaptive adjustment coefficient of the ambient temperature in the area where exhibit i is located, and 、 and are all positive numbers, According to the marginal heat score of each exhibit Sort from high to low to form a list of edge exhibits’ attention ranking as a refined exhibit ranking, and push it to each edge node to calculate the average edge heat score of each exhibit. The calculation expression of the average edge heat score is: , calculate the standard deviation of the edge heat score of each exhibit. The standard deviation calculation expression of the edge heat score of each exhibit is , conduct hot spot analysis on the exhibits based on the average edge heat score and the standard deviation of the edge heat score of each exhibit. , then the hot spot analysis determines that the exhibit is a hot spot. If the edge heat score of the exhibit is , then the hot spot analysis determines that the exhibit is a low-activity exhibit, where, is the adjustment coefficient, and is a positive number; Mark the hot exhibits as three-star exhibits, mark the low-activity exhibits as one-star exhibits, and mark the remaining exhibits except the hot exhibits and low-activity exhibits as two-star exhibits. Combined with the theme and exhibition area layout of each exhibit, they are packaged and sent to each edge node as the second optimization suggestion for editors.

5. A real-time collaborative editing system for a virtual exhibition hall of ice and snow tourism according to claim 4, characterized in that: The adaptive adjustment coefficient is adaptively adjusted according to the current time window and the current exhibition area flow pressure, where , where p is the current crowd pressure in the exhibition area, that is, the crowd density index is normalized to , is the optimal visual focus value, is the standard deviation of the average number of clicks on exhibits by visitors in historical exhibitions, e is the natural base, and the adaptive adjustment coefficient , t is the current time window, with hours as the time unit, is the peak time window, the adaptive adjustment coefficient , where is the empirical control coefficient, and is a positive number, is the ambient temperature of the location of exhibit i, is the average ambient temperature of all exhibits, and the calculation expression is: , is the standard deviation of the ambient temperature at all exhibit locations, and the calculation expression is: , adjust the weight according to different exhibition time periods and crowd pressure.

6. A real-time collaborative editing system for a virtual exhibition hall of ice and snow tourism according to claim 4, characterized in that: The method for obtaining the concentration of the visitor's gaze focus at exhibit i is as follows: the visitor's gaze interaction trajectory is obtained based on the eye tracking component and spatial positioning component in the AR head-mounted device. The AR head-mounted device worn by the visitor includes a binocular camera, an infrared eye sensor and an IMU inertial navigation module, which are used to record the visitor's gaze point coordinates and head posture in real time. According to the AR space mapping, the visitor's gaze vector is matched with the boundary area of ​​exhibit i in space. If the visitor's gaze point coordinates coincide with the projection model of exhibit i, a focus hit event is recorded. If the visitor's gaze point coordinates deviate from the projection model of exhibit i, a focus departure event is recorded. The time from each focus hit event to the focus departure event is defined as the witnessing duration. The number of visitors who visit exhibit i within the unit time Ut is un, and the witnessing duration of each visitor's attention to exhibit i is , then the calculation expression of the concentration of visitors' gaze at exhibit i is ; The method for obtaining the crowd density in the adjacent area of ​​exhibit i within unit time is as follows: based on the identification and counting of Wi-Fi probe devices and Bluetooth beacons deployed around the exhibit, several Wi-Fi probe nodes and BLE Bluetooth receiving modules are deployed within the exhibition range of each exhibit. The Wi-Fi probe periodically scans the MAC signal emitted by the visitor terminal, and uses MAC address deduplication, hash anonymization and time window sliding algorithm to aggregate multiple scan records of the same visitor. Suppose the number of unique visitor devices detected in the exhibition area of ​​exhibit i within unit time Ut is Nvd, and the detection radius is Dr. Then the calculation expression for the crowd density in the adjacent area of ​​exhibit i within unit time is: .

7. The real-time collaborative editing system for the ice and snow cultural tourism virtual exhibition hall according to claim 1 is characterized in that: The method for monitoring the network connectivity status between edge devices is: The edge device periodically sends heartbeat status requests to other edge devices, and sets the maximum response time threshold to Maxrtt. If the edge device replies to the heartbeat status request within the maximum response time threshold Maxrtt, it is a valid response, otherwise it is an invalid response. If the edge device does not receive a valid response for Ret consecutive times, it is determined that the sending edge device of the heartbeat status request and the receiving edge device are in a disconnected state, and the receiving edge device of the heartbeat status request is defined as a disconnected node. Otherwise, it is determined that the sending edge device of the heartbeat status request and the receiving edge device are in a connected state, and the connectivity status of the edge device is calibrated. The expression is ; The method for calculating the integrity of the second visitor behavior data collected by the edge device within the time window is: According to the perceived activity and normalized data density in the time window Uwt, the degree of integrity is calculated. The calculation expression of the degree of integrity is: ,in, For the degree of completeness, To perceive activity, is the normalized data density, are the proportional coefficients of perceived activity and normalized data density, respectively, and All are positive numbers; Perceived activity The method of obtaining is: suppose the number of sensing devices attached to the edge node is When the number of effective data reports of the sensing device within the time window Uwt is higher than the active threshold Act, the sensing device is determined to be an active device, and the number of active devices is calibrated as , then the perceived activity within the time window Uwt The calculation expression is ; The method for obtaining normalized data density is as follows: let the number of second visitor behavior data items obtained within the time window Uwt be , is the number of displays controlled by the edge device, then the calculation expression of normalized data density is: , where is the reference standard density; According to the impact of historical editing of edge devices, the method of calculating node influence is as follows: the optimization suggestions for each round of deployment are recorded locally on the edge node. The optimization suggestions include the first optimization suggestion and the second optimization suggestion. The optimization suggestions are recorded as a set , where is the number of the edge device, and ,in is the total number of edge devices, For edge devices In the The optimization suggestions generated in the time window are compared with the actual layout results and the adopted optimization suggestions are marked as ,and , optimization suggestions that were adopted Analyze the feedback effect and define the feedback improvement rate vector as ,in, is the visitor behavior feedback vector before the exhibit is adopted, is the tourist behavior feedback vector after the exhibit is adopted. The feedback vector includes the three dimensions of the exhibit’s stay time, click times, and interaction times. The threshold vector including the three dimensions of stay time, click times, and interaction times is set as , using the logical decision function Perform feedback judgment and calibrate the feedback effect to reach the threshold vector of the adopted optimization suggestions are hit suggestions ,and , then the calculation expression of node influence is .

8. The real-time collaborative editing system for the ice and snow cultural tourism virtual exhibition hall according to claim 7 is characterized in that: The method for dynamic evaluation and judgment based on connectivity status, integrity level and node influence is as follows: The modeling fitness of edge nodes is calculated based on connectivity status, integrity level and node influence. The calculation expression of modeling fitness is: ,in, and Connectivity status , degree of completeness and node influence The weight coefficient of and Both are positive numbers, and the preset fitness threshold is Fith. If the calculated modeling fitness Mof is less than the fitness threshold Fith, the edge device is marked as a closed node. If the calculated modeling fitness Mof is greater than or equal to the fitness threshold Fith, the edge device is marked as an open node. For closed nodes, only local data is used for independent heat modeling, and for open nodes, data is synchronized for collaborative heat modeling.

9. A real-time collaborative editing method for a virtual exhibition hall of ice and snow culture and tourism, based on the real-time collaborative editing system for a virtual exhibition hall of ice and snow culture and tourism according to any one of claims 1 to 8, characterized in that: The specific steps include: collecting first visitor behavior data through a visitor terminal, analyzing the attention of an exhibit and outputting a first optimization suggestion, and pushing the first optimization suggestion to an edge node; Editors can modify and adjust the exhibit content locally through edge devices while offline, and perform local caching and asynchronous synchronization of the first visitor behavior data; The edge node receives real-time second visitor behavior data and performs refined exhibition item sorting, hot spot analysis, and second optimization suggestion generation based on edge computing capabilities; Conflict detection and collaborative presentation of exhibition item editing data are synchronized across multiple edge devices via the local area network, enabling collaborative exhibition operations for multiple users and multiple devices. Monitor the network connectivity status between each edge device, evaluate the integrity of the second visitor behavior data collected by the edge device within the time window, calculate the node influence based on the impact of the historical editing of the edge device, and dynamically evaluate the heat modeling mode of the edge-end fine screening optimization module based on the connectivity status, integrity and node influence, and determine whether to perform architecture switching.

Images