Immersive text travel and night travel path planning system and method based on AR virtual interaction

Abstract

The invention relates to an AR virtual interaction-based immersive travel and night travel path planning system and method, provides a man-machine interaction device, and relates to the field of an input device or an input and output combination device for interaction between a user and a computer. The system comprises an AR interaction terminal worn by a current night tour user in a target scenic area; the path planning core module is used for intelligently predicting each optimal planning path for each scenic spot in the target scenic area; and the virtual-real fusion rendering module is used for executing fusion rendering of the picture based on the intelligent prediction result. Aiming at the technical problems that the route planning result of the tourism application scheme of the AR interaction terminal is low in efficiency, single and insufficient in reliability and lacks a night travel route planning strategy, an artificial intelligence model is adopted, and based on multi-source basic data, synchronous planning of each optimal planning route in a night travel environment is completed at the AR interaction terminal at the same time, so that the route planning efficiency is improved. And executing AR virtual interaction, so that the technical problem is solved.

Description

Technical Field

[0001] The human-computer interaction device proposed in this invention relates to the field of input devices or input and output combination devices for interaction between users and computers, and particularly to an immersive cultural tourism night tour route planning system and method based on AR virtual interaction. Background Technology

[0002] Human-computer interaction (HCI) devices are common among input devices or input / output combinations used for interaction between users and computers, especially AR (Augmented Reality) interactive terminals. These terminals have gradually gained a leading position in the HCI device market due to their superior image rendering capabilities, immersive scene effects, and comprehensive interactive options. AR interactive terminals are used in various niche scenarios such as gaming, movie watching, driving, and tourism. They can be customized to meet specific application scenarios, achieving the intended use while minimizing cost and scale.

[0003] For example, Chinese invention patent publication CN110703922A proposes a method for guiding tourists using an electronic map, including the following steps: collecting basic data of the scenic area, establishing a flat map of the scenic area and associated AR scenes; scanning the flat map of the scenic area to obtain associated AR scenes for interface display; adaptively matching gesture operations with the content edited in the AR scene interface; selecting a marked tour route and / or attraction in the AR scene through gesture operations, obtaining real-time crowd monitoring video of the corresponding route and / or attraction, and planning the tour route; traveling to the destination attraction in the scenic area according to the planned tour route; selecting the marked attraction corresponding to the destination attraction in the AR scene through gesture operations, obtaining the corresponding attraction's introductory information for playback. This method can provide tourists with an AR guide map that integrates real-world scenery, allowing them to view routes and attraction scenes before making route choices, and providing tourists with explanations of attractions in parallel with text, sound, and images.

[0004] For example, Chinese invention patent publication CN120047650A proposes a scenic spot navigation system and method based on AR technology and real-scene roaming. The system includes a navigation module, a guide module, and a footprint module, all three connected via a network. The navigation module is used for intelligent route planning and AR virtual tour guide; the guide module is used for quickly obtaining scenic spot information and personalized route recommendations; and the footprint module is used to generate a personal experience path. This application, through the flexible combination and application of AR technology, large language model technology, and big data technology, improves user tour efficiency, preserves user travel experiences, enriches the interactivity, knowledge, and fun of the user tour experience, and meets the personalized needs of each user.

[0005] Therefore, it is evident that in existing technologies, when applying AR interactive terminals to the specific scenario of tourism, the technical solutions involved are limited to scenic spot guidance. The proposed intelligent route planning is based solely on the user's specific location and target scenic spot information, lacking comprehensive and sufficient multi-source basic data. The planned routes are also single-source, resulting in planning results that fail to meet users' expected efficiency and effectiveness requirements. Crucially, there is a lack of intelligent route planning mechanisms for nighttime tourism environments. The dark scenes at night cause the flow patterns of nighttime tourists to differ from those during the day, leading to a lack of planning for nighttime tourist routes. It is impossible to simultaneously plan multiple optimal routes that minimize the total time spent on multiple targets, resulting in low route planning efficiency for nighttime tourists and making them more prone to crowding. Summary of the Invention

[0006] To address the technical problems in existing technologies, this invention provides an immersive cultural tourism night tour route planning system and method based on AR virtual interaction. It addresses the technical issues of inefficient, singular, and unreliable route planning results in tourism applications using AR interactive terminals, as well as the lack of night tour route planning strategies. By customizing artificial intelligence models with different structures for different scenic spots and introducing comprehensive and sufficient multi-source basic data, the system simultaneously plans optimal routes for multiple objectives in a night tour environment, minimizing the total time required, at the AR interactive terminal. Based on the synchronous planning results, corresponding AR virtual interaction operations are executed, thereby improving the efficiency and effectiveness of route planning for night tour users.

[0007] According to one aspect of the present invention, an immersive cultural tourism night tour route planning system based on AR virtual interaction is provided, the system comprising:

[0008] The AR interactive terminal is worn by the current night tour users in the target scenic area and is equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. It includes an optical display module, an interactive input unit, and a wireless communication unit. The wireless communication unit collects the night tour status data of each other night tour user in the target scenic area at the current moment.

[0009] The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module, and an ambient light sensor, which collects scene perception content of the current night tour user at evenly spaced historical moments before the current moment;

[0010] The core path planning module is connected to the AR interactive terminal and the scene perception module respectively. It adopts a night tour path intelligent planning model. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, it intelligently predicts the optimal planned paths from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time.

[0011] The virtual-real fusion rendering module is connected to the AR interactive terminal, the scene perception module, and the path planning core module, respectively, and is used to perform fusion rendering of the screen based on the intelligent prediction results.

[0012] Among them, the night tour status data corresponding to each other night tour user at the current moment is the user-related parameter set of the other night tour user, the customized visualization information of each historical moment at even intervals before the current moment, the on-site location information, the tourist density of the scenic spot, and the scenic spot number of the scenic spot.

[0013] Among them, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age;

[0014] The customized visualization information for each night tour user at each moment includes the grayscale gradient values, coordinate values, and depth values ​​of all constituent pixels of each human body occupying a sub-frame in the scene in front of the night tour user at that moment, as well as the curvature values ​​of all pixel positions on the contour curve of each human body occupying the sub-frame.

[0015] According to another aspect of the present invention, an immersive cultural tourism night tour route planning method based on AR virtual interaction is provided, the method comprising:

[0016] The AR interactive terminal uses a wireless communication unit to collect nighttime status data from other nighttime visitors in the target scenic area at the current moment. The AR interactive terminal is worn by the current nighttime visitors in the target scenic area and is equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. The AR interactive terminal includes an optical display module, an interactive input unit, and a wireless communication unit.

[0017] The scene perception module is used to collect scene perception content of the current night tour user at evenly spaced historical moments before the current moment. The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module, and an ambient light sensor.

[0018] Using a path planning core module connected to both the AR interactive terminal and the scene perception module, the system employs an intelligent night tour path planning model. This model intelligently predicts optimal planned paths from the current night tour user's current location to various attractions within the target scenic area, based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment.

[0019] A virtual-real fusion rendering module, which is connected to the AR interactive terminal, the scene perception module, and the path planning core module respectively, is used to perform fusion rendering of the screen based on the intelligent prediction results.

[0020] Among them, the night tour status data corresponding to each other night tour user at the current moment is the user-related parameter set of the other night tour user, the customized visualization information of each historical moment at even intervals before the current moment, the on-site location information, the tourist density of the scenic spot, and the scenic spot number of the scenic spot.

[0021] Among them, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age;

[0022] The customized visualization information for each night tour user at each moment includes the grayscale gradient values, coordinate values, and depth values ​​of all constituent pixels of each human body occupying a sub-frame in the scene in front of the night tour user at that moment, as well as the curvature values ​​of all pixel positions on the contour curve of each human body occupying the sub-frame.

[0023] Therefore, it can be seen that the present invention has at least the following outstanding substantive features:

[0024] Substantive Feature A: Addressing the lack of real-time performance, interactivity, and immersion in traditional nighttime navigation, and especially the critical issue that the dark environment at night causes different patterns of pedestrian flow compared to daytime, leading to a lack of planned routes for nighttime users, we employ a smart nighttime route planning model with a customized structure for the target scenic area. Based on targeted multi-source basic data, we simultaneously plan the optimal routes for multiple targets in the nighttime environment with the minimum total time consumption at the AR interactive terminal, and execute corresponding AR virtual interactive operations, thereby improving the efficiency and effectiveness of route planning for nighttime users.

[0025] Substantive Feature B: It overlays various virtual information, including the optimal planning paths obtained from synchronous planning, onto the current night tour user's current scene in front of them for fusion rendering, and pushes the fused rendering image to the optical display module for display. It also allows the current night tour user to select a route through the interactive input unit of the AR interactive terminal they are wearing, in order to determine the specific attractions they want to visit. This completes specific AR virtual interactive operations based on the synchronous planning results, providing a reliable and convenient route planning mode for different night tour users in different scenic spots.

[0026] Substantive Feature C: To intelligently predict the optimal planned routes for nighttime tourists to reach various attractions within the target scenic area from their current location within the current time segment, minimizing the total travel time, a customized intelligent planning model for nighttime routes is designed for the target scenic area. This model is a convolutional neural network (CNN) trained multiple times. The CNN includes an input layer, an activation layer, and multiple parallel convolutional layers located between the input and activation layers. The number of parallel convolutional layers shows the same numerical trend as the area of ​​the target scenic area, and the number of training iterations of the CNN shows the same numerical trend as the peak number of nighttime tourists in the target scenic area. This allows for the design of customized intelligent planning models for different scenic areas, ensuring the stability and effectiveness of the intelligent prediction results.

[0027] Substantive Feature D: During each training iteration of the convolutional neural network, the optimal planned paths from the starting point of a historical nighttime user within a specific historical time segment to various attractions within the target scenic area, minimizing the total path travel time, are used as the output of the convolutional neural network. The inputs are the time interval between two adjacent historical moments, the user-related parameter set of the historical nighttime user, the nighttime status data of other nighttime users within the target scenic area corresponding to the historical nighttime user at the starting point of the specific historical time segment, the number of nighttime users in the target scenic area on each day prior to the date of the specific historical time segment, and the scene perception content of the historical nighttime user at evenly spaced historical moments before the starting point of the specific historical time segment. This process ensures the training effectiveness of the convolutional neural network in each iteration.

[0028] Substantive Feature E: To intelligently predict the optimal planned paths for a current nighttime visitor to reach various attractions within the target scenic area from their current location within the current time segment while minimizing the total path travel time, targeted selection of multi-source basic data has been employed. This includes the time interval between two adjacent historical moments, the user-related parameter set of the current nighttime visitor, the nighttime status data of other nighttime visitors within the target scenic area at the current moment, the number of nighttime visitors in the target scenic area on previous days, and the scene perception content of the current nighttime visitor at evenly spaced historical moments before the current moment. The targeted selection of the above multi-source basic data further ensures the stability and effectiveness of the intelligent prediction results.

[0029] Substantive Feature F: In the multi-source basic data, specifically, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age. The night tour status data for each other night tour user at the current moment includes the user-related parameter set for the other night tour user, customized visualization information of each historical moment at even intervals before the current moment, on-site location information, tourist density of the scenic spot, and scenic spot number. The customized visualization information for each night tour user at each moment includes the grayscale gradient values, coordinate values, and depth values ​​of all pixels constituting the sub-image of each human body in the scene in front of the night tour user at that moment, as well as the curvature values ​​of all pixel positions on the contour curve of each human body occupying the sub-image. This provides a customized data structure design for the multi-source basic data used for intelligent prediction.

[0030] Substantive Feature G: In the multi-source basic data, specifically, the number of historical moments at even intervals before the current moment is proportional to the total number of attractions in the target scenic area, and the number of days before the current day in the target scenic area shows the same numerical trend as the total number of attractions in the target scenic area. This ensures that the amount of data in the targeted multi-source basic data matches the scale of the target scenic area, guaranteeing the sufficiency and comprehensiveness of the multi-source basic data used for intelligent prediction. Attached Figure Description

[0031] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein:

[0032] Figure 1 This is a schematic diagram of the technical process of the immersive cultural tourism night tour route planning system and method based on AR virtual interaction according to the present invention.

[0033] Figure 2 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the first embodiment of the present invention.

[0034] Figure 3 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the second embodiment of the present invention.

[0035] Figure 4 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the third embodiment of the present invention.

[0036] Figure 5 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the fourth embodiment of the present invention.

[0037] Figure 6 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the fifth embodiment of the present invention.

[0038] Figure 7 The following is a flowchart illustrating the steps of an immersive cultural tourism night tour route planning method based on AR virtual interaction according to the sixth embodiment of the present invention. Detailed Implementation

[0039] like Figure 1 The diagram illustrates the technical flow of an immersive cultural tourism nighttime tour route planning system and method based on AR virtual interaction according to the present invention. The human-computer interaction device proposed in this invention relates to the field of input devices or input-output combination devices for interaction between users and computers.

[0040] The specific technical process of this invention is as follows:

[0041] Technical Process 1: Build the hardware platform for the AR-based immersive cultural tourism night tour route planning system of the present invention. The hardware platform is used to implement the AR-based immersive cultural tourism night tour route planning method of the present invention.

[0042] Specifically, the hardware platform of the immersive cultural tourism night tour route planning system based on AR virtual interaction of the present invention includes the following components: AR interactive terminal, scene perception module, route planning core module and virtual-real fusion rendering, wherein the AR interactive terminal is worn by the current night tour user in the target scenic spot, and the scene perception module, route planning core module and virtual-real fusion rendering module are mounted on the AR interactive terminal. The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module and an ambient light sensor.

[0043] More specifically, the AR interactive terminal worn by the current night tour user includes an optical display module, an interactive input unit, and a wireless communication unit. The wireless communication unit is used to collect the night tour status data of other night tour users in the target scenic area at the current moment. The optical display module is used to display the fused and rendered image for the current night tour user to view immersively. The fused and rendered image is the image obtained by superimposing various virtual information onto the current night tour user's front scene at the current moment. The various virtual information includes the optimal planning paths obtained by subsequent intelligent analysis. The interactive input unit is used for the AR interactive terminal worn by the current night tour user to perform path selection based on the viewed optimal planning paths, so as to determine the specific attractions to visit, thereby completing the specific AR virtual interactive operation based on the synchronous planning results.

[0044] Technical Process Two: Design a customized artificial intelligence model for the target scenic area, namely the intelligent planning model for the night tour route corresponding to the target scenic area. This model is used to intelligently predict the main body of each optimal planned route that minimizes the total travel time from the current night tour user's current location to each attraction in the target scenic area within the current time segment.

[0045] Specifically, intelligent planning models for night tour routes with different customized structures can be designed for different scenic spots. The customized structure design of the intelligent planning model for the night tour route corresponding to the target scenic spot is mainly reflected in the following aspects:

[0046] Aspect A: The intelligent planning model for the night tour route corresponding to the target scenic spot is a convolutional neural network that has been trained multiple times. The number of training times of the convolutional neural network shows the same numerical change trend as the peak number of night tour users in the target scenic spot on a single day.

[0047] For example, when the peak number of nighttime visitors to the target scenic spot is 500, the selected convolutional neural network is trained 1000 times; when the peak number of nighttime visitors to the target scenic spot is 600, the selected convolutional neural network is trained 1200 times; when the peak number of nighttime visitors to the target scenic spot is 700, the selected convolutional neural network is trained 1400 times; when the peak number of nighttime visitors to the target scenic spot is 800, the selected convolutional neural network is trained 1600 times, and so on.

[0048] Aspect B: The convolutional neural network used includes an input layer, an activation layer, and multiple convolutional layers in parallel. These multiple convolutional layers are located between the input layer and the activation layer, and the number of convolutional layers in parallel shows the same numerical trend as the area of ​​the target scenic spot.

[0049] For example, when the target scenic area covers an area of ​​10,000 hectares, the number of parallel convolutional layers selected is 2; when the target scenic area covers an area of ​​20,000 hectares, the number of parallel convolutional layers selected is 4; when the target scenic area covers an area of ​​30,000 hectares, the number of parallel convolutional layers selected is 6; when the target scenic area covers an area of ​​40,000 hectares, the number of parallel convolutional layers selected is 8, and so on.

[0050] Aspect C: During each training iteration of the convolutional neural network, the optimal planned paths from the starting point of a historical nighttime user within a specific historical time segment to various attractions within the target scenic area, minimizing the total path travel time, are used as the output of the convolutional neural network. The time interval between two adjacent historical moments, the user-related parameter set of the historical nighttime user, the nighttime status data of other nighttime users within the target scenic area corresponding to the historical nighttime user at the starting point of the specific historical time segment, the number of nighttime users in the target scenic area on each day prior to the date of the specific historical time segment, and the scene perception content of the historical nighttime user at evenly spaced historical moments before the starting point of the specific historical time segment are used as the input of the convolutional neural network to complete this training, thereby ensuring the training effect of the convolutional neural network in each iteration.

[0051] In this way, through the customized structural designs mentioned above, it is possible to design intelligent planning models for night tour routes with different customized structures for different scenic spots, ensuring the stability and effectiveness of intelligent prediction results.

[0052] Technical Process 3: To intelligently predict the optimal planned routes for nighttime tourists to reach various attractions within the target scenic area from their current location within the current time segment while minimizing the total travel time, multi-source basic data was specifically selected.

[0053] Specifically, the multi-source basic data used for intelligent prediction includes the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data corresponding to each other night tour user in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before the current day, and the scene perception content of each historical moment at even intervals before the current night tour user.

[0054] More specifically, in the multi-source basic data, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age. The night tour status data for each other night tour user at the current moment includes the user-related parameter set for the other night tour user, customized visualization information of each historical moment at even intervals before the current moment, on-site location information, tourist density of the scenic spot, and scenic spot number. The customized visualization information for each night tour user at each moment includes the grayscale gradient values, coordinate values, and depth values ​​of all constituent pixels of each human body occupying the sub-frame in the scene in front of the night tour user at the current moment, as well as the curvature values ​​of all pixel positions on the contour curve of each human body occupying the sub-frame. This provides a customized data structure design for the multi-source basic data used for intelligent prediction.

[0055] More specifically, in the multi-source basic data, the number of historical moments at even intervals before the current moment is proportional to the total number of attractions in the target scenic area, and the number of days before the current day in the target scenic area shows the same numerical trend as the total number of attractions in the target scenic area. This ensures that the amount of data in the targeted multi-source basic data matches the scale of the target scenic area, guaranteeing the sufficiency and comprehensiveness of the multi-source basic data used for intelligent prediction.

[0056] For example, if the total number of attractions in the target scenic area is 5, the number of historical moments selected at even intervals before the current time is 10; if the total number of attractions in the target scenic area is 6, the number of historical moments selected at even intervals before the current time is 12; if the total number of attractions in the target scenic area is 7, the number of historical moments selected at even intervals before the current time is 14; if the total number of attractions in the target scenic area is 8, the number of historical moments selected at even intervals before the current time is 16, and so on.

[0057] For example, if the total number of attractions in the target scenic area is 5, and the number of days prior to the selected target scenic area is 5; if the total number of attractions in the target scenic area is 10, and the number of days prior to the selected target scenic area is 10; if the total number of attractions in the target scenic area is 15, and the number of days prior to the selected target scenic area is 15; if the total number of attractions in the target scenic area is 20, and the number of days prior to the selected target scenic area is 20, and so on;

[0058] In this way, the targeted selection of the above-mentioned multi-source basic data further ensures the stability and effectiveness of the intelligent prediction results;

[0059] Technical Process 4: Using the intelligent planning model for night tour routes with a customized structure designed for the target scenic area based on Technical Process 2, and based on the multi-source basic data selected in Technical Process 3, the model intelligently predicts the optimal planned routes from the current night tour user's current location at the current time segment to each attraction in the target scenic area, minimizing the total travel time.

[0060] Specifically, each optimal planned route corresponds to a different attraction within the same target scenic area. Subsequent AR rendering can help nighttime visitors choose the optimal planned route to visit within the current time segment. In reality, the selected attraction is the one corresponding to the optimal planned route.

[0061] Technical Process 5: Based on the intelligent prediction results of Technical Process 4, the screen is fused and rendered, and the fused and rendered screen is pushed to the optical display module of the AR interactive terminal worn by the current night tour user for display, so that the current night tour user can perform path selection and attraction selection through the interactive input unit of the AR interactive terminal worn by the user.

[0062] In this way, specific AR virtual interactive operations based on the synchronous planning results can be completed, providing a reliable and convenient route planning mode for different nighttime tourists in different scenic spots;

[0063] Therefore, through the coordinated operation of the above five technical processes, this invention addresses the technical problems of traditional nighttime navigation lacking real-time performance, interactivity, and immersion. Crucially, the dark environment at night causes different patterns of pedestrian flow compared to daytime, leading to a lack of planned routes for nighttime visitors. By employing an intelligent nighttime route planning model with a customized structure for the target scenic area, and based on selectively filtered multi-source data, the invention simultaneously plans optimal routes for multiple targets in the nighttime environment with minimal total time consumption at the AR interactive terminal, and executes corresponding AR virtual interactive operations. This improves the efficiency and effectiveness of route planning for nighttime visitors, overcoming the aforementioned technical problems.

[0064] The key points of this invention are: synchronous intelligent analysis of the optimal planned paths corresponding to each scenic spot in the same scenic area within future time segments; targeted hardware platform construction of the immersive cultural tourism night tour path planning system based on AR virtual interaction in the night tour environment; directional customized structure design of intelligent planning models for different night tour paths corresponding to different scenic spots; customized data structure design of multi-source basic data; and matching selection of data scale.

[0065] The following will describe in detail the immersive cultural tourism night tour route planning system and method based on AR virtual interaction of the present invention through embodiments.

[0066] First Embodiment

[0067] Figure 2 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the first embodiment of the present invention.

[0068] like Figure 2 As shown, the immersive cultural tourism night tour route planning system based on AR virtual interaction includes the following components:

[0069] The AR interactive terminal is worn by the current night tour users in the target scenic area and is equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. It includes an optical display module, an interactive input unit, and a wireless communication unit. The wireless communication unit collects the night tour status data of each other night tour user in the target scenic area at the current moment.

[0070] Here, the AR interactive terminal itself serves as a data collection terminal for some basic data, as well as a display terminal for the fused rendering screen based on intelligent prediction results, and a corresponding interactive terminal for the fused rendering screen.

[0071] Meanwhile, the AR interactive terminal is also equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. The scene perception module, the path planning core module, and the virtual-real fusion rendering module can be integrated as independent components on the external control motherboard of the AR interactive terminal, or they can be integrated into the control motherboard of the AR interactive terminal itself.

[0072] The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module, and an ambient light sensor, which collects scene perception content of the current night tour user at evenly spaced historical moments before the current moment;

[0073] For example, the scene perception content collected from the current nighttime user at evenly spaced historical moments before the current moment includes: the time interval between two adjacent historical moments can be selected as 1 minute, and the evenly spaced historical moments before the current moment can be selected as 10 historical moments. When the current moment is 8:00 PM, the evenly spaced historical moments before the current moment are 7:59 PM, 7:58 PM, 7:57 PM, 7:56 PM, 7:55 PM, 7:54 PM, 7:53 PM, 7:52 PM, 7:51 PM, 7:50 PM, etc.

[0074] The core path planning module is connected to the AR interactive terminal and the scene perception module respectively. It adopts a night tour path intelligent planning model. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, it intelligently predicts the optimal planned paths from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time.

[0075] For example, all other nighttime visitors within the target scenic area can wear their own AR interactive terminals. All the AR interactive terminals of the nighttime visitors are of the same model and are delivered to the target scenic area upon entry. The intelligently predicted optimal planning paths correspond to various attractions within the same target scenic area. This provides a data basis for the current nighttime visitors to select attractions within the current time segment as a future time segment by choosing the optimal planning path through the interactive input unit of their AR interactive terminals.

[0076] The virtual-real fusion rendering module is connected to the AR interactive terminal, the scene perception module, and the path planning core module, respectively, and is used to perform fusion rendering of the screen based on the intelligent prediction results.

[0077] Among them, the fusion rendering of the screen based on the intelligent prediction results includes: overlaying various virtual information onto the scene in front of the current night tour user at the current moment for fusion rendering, and pushing the fusion rendered screen to the optical display module for display. The various virtual information includes the optimal planning paths obtained by intelligent analysis.

[0078] Clearly, the virtual information rendered in the scene in front of the current night tour user at the current moment is not limited to each optimal planned path, but also includes other types of virtual information;

[0079] Among them, the night tour status data corresponding to each other night tour user at the current moment is the user-related parameter set of the other night tour user, the customized visualization information of each historical moment at even intervals before the current moment, the on-site location information, the tourist density of the scenic spot, and the scenic spot number of the scenic spot.

[0080] Among them, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age;

[0081] Specifically, the four parameters of other nighttime tourists—body fat percentage, height-to-weight ratio, gender, and age—not only affect their own nighttime travel route preferences, but also, since they are all within the same scenic area, influence other nighttime tourists, such as the current nighttime tourist's nighttime travel route preferences.

[0082] The following section will describe in detail how four parameters—body fat percentage, height-to-weight ratio, gender, and age—influence each nighttime traveler's nighttime travel route preferences from four aspects:

[0083] Aspect 1: The impact of body fat percentage:

[0084] As a type of body fat data, body fat percentage directly determines the length and slope tolerance of night tour routes by affecting basal metabolic efficiency and sustained exercise capacity. For example, users with high body fat tend to choose short circular routes or routes with multiple rest points, actively avoid three-dimensional attractions that require continuous climbing, and prefer flat waterfront walkways or plaza-type spaces. On the other hand, users with low body fat are more likely to choose routes that allow for in-depth exploration, accept linear routes with a one-way distance of more than 1.5 kilometers, and have a higher tolerance for vertical transportation such as steps and ramps.

[0085] Part Two: The Influence of Body Mass Index (BMI)

[0086] BMI influences the lower limb joints' pressure-bearing capacity and center of gravity stability, thus determining users' preferences for terrain complexity and road surface materials. BMI is linearly positively correlated with knee joint load during walking. When obese individuals with a BMI > 28 walk on uneven surfaces (such as cobblestone paths or gravel trails), their compensatory ankle joint activity increases, resulting in energy consumption that is 22-30% higher than that of people with normal BMI. Insufficient light at night further reduces the visual feedforward adjustment ability to terrain, amplifying the exercise risks associated with BMI.

[0087] Users with high BMI show a significant preference for hard, flat surfaces, avoid groups of steps that require high leg movements, and tend to choose ramps with handrails rather than open staircases. They are also more sensitive to the width of the path. Users with normal BMI are more accepting of terrain diversity and are willing to choose "fun paths" with moderate undulations (such as walking through garden paths or stepping stone bridges).

[0088] Thirdly, the influence of gender:

[0089] Gender differences in nighttime walking scenarios are not only reflected in traditional behavioral preferences, but more fundamentally in differences in visual attention allocation mechanisms and risk perception topologies. Neurobehavioral studies have shown that women adopt a "peripheral scanning mode" when walking at night, paying significantly more attention to potentially dangerous areas such as bushes, dark corners, and enclosed spaces on both sides of the path, while men adopt a "focus-oriented mode," concentrating their attention on the path ahead and fixed objects (such as streetlights and signs). This difference is significantly amplified at night. Female users prioritize well-lit main paths, avoid poorly lit shortcuts or easy paths, prefer paths with open views (clear vision, no blind spots), and show significant avoidance of paths with vegetation covering both sides. They also tend to walk in groups. Male users, on the other hand, are more accepting of efficiency-first paths (choosing the shortest route, even if the lighting conditions are average) and have a higher willingness to try exploratory paths (willing to try unmarked trails, construction passages, etc.).

[0090] Fourthly, the impact of age:

[0091] Age influences the efficiency of multisensory integration, dark adaptation, and fatigue recovery speed, thus determining the time structure and rhythm of night tour routes. Young people (18-35 years old) prefer high-density experience routes, accepting routes with a large number of attractions per unit distance, high information density, and fast pace. They are also accepting of "check-in" style movement (quick and continuous visits to multiple attractions) and have lower sensitivity to the end time of the night tour, willing to extend their tour into the late night. Middle-aged people (40-60 years old) focus on functional adaptability, prioritizing routes with clear rest facilities and reasonable distribution of public health facilities, and have a lower tolerance for the duration of the night tour. Elderly people (over 60 years old) strictly follow the main route, have rigid requirements for the width of the passageway and nighttime lighting, prefer familiar routes, and start and end their night tours early.

[0092] Clearly, the four parameters of body fat percentage, height-to-weight ratio, gender, and age for each nighttime tourist do not act independently, but rather form a cumulative constraint effect, which will affect their nighttime travel route preferences as a whole. Since they are all in the same scenic area, this will further affect other nighttime tourists as a whole, such as the current nighttime tourist's nighttime travel route preferences.

[0093] Among them, the customized visualization information of each night tour user at each moment is the grayscale gradient value, coordinate value and depth value of all the constituent pixels of each human body occupying the sub-frame in the scene in front of the night tour user at that moment, as well as the curvature value of all pixel positions on the contour curve of each human body occupying the sub-frame.

[0094] For example, for each constituent pixel of a sub-image occupied by a human body, the mean square difference of each gray value corresponding to each of its surrounding adjacent constituent pixels can be used as its gray gradient value.

[0095] Among them, the RGB-D depth camera is used to collect customized visualization information of the current night tour user at uniform intervals of each historical moment before the current moment; the inertial measurement unit is used to collect the acceleration of the current night tour user at uniform intervals of each historical moment before the current moment; the GPS / BeiDou dual-mode positioning module is used to collect the on-site location positioning information of the current night tour user at uniform intervals of each historical moment before the current moment; and the ambient light sensor is used to collect the on-site illumination of the current night tour user at uniform intervals of each historical moment before the current moment.

[0096] For example, the GPS / BeiDou dual-mode positioning module is used to collect the on-site location information of the current night tour user at evenly spaced historical moments before the current moment. This includes: the GPS / BeiDou dual-mode positioning module uses the arithmetic mean of the current night tour user's GPS positioning information and BeiDou positioning information at any moment as the on-site location information of the current night tour user at any moment.

[0097] Among them, the customized visualization information of the current night tour user at uniform intervals before the current moment, the acceleration of the current night tour user at uniform intervals before the current moment, the on-site location information of the current night tour user at uniform intervals before the current moment, and the on-site illumination of the current night tour user at uniform intervals before the current moment are used as the scene perception content of the current night tour user at uniform intervals before the current moment.

[0098] Among them, the intelligent planning model for night tour routes is a convolutional neural network after each training. The convolutional neural network includes an input layer, an activation layer and multiple parallel convolutional layers. The multiple parallel convolutional layers are located between the input layer and the activation layer, and the number of parallel convolutional layers shows the same numerical change trend as the area of ​​the target scenic spot.

[0099] For example, the number of parallel convolutional layers showing the same numerical trend as the area of ​​the target scenic spot includes: when the area of ​​the target scenic spot is 10,000 hectares, the number of parallel convolutional layers selected is 2 layers; when the area of ​​the target scenic spot is 20,000 hectares, the number of parallel convolutional layers selected is 4 layers; when the area of ​​the target scenic spot is 30,000 hectares, the number of parallel convolutional layers selected is 6 layers; when the area of ​​the target scenic spot is 40,000 hectares, the number of parallel convolutional layers selected is 8 layers, and so on.

[0100] The number of times the convolutional neural network was trained showed the same numerical trend as the peak number of nighttime visitors to the target scenic spot.

[0101] For example, the number of training iterations of the convolutional neural network shows the same numerical trend as the peak number of nighttime visitors to the target scenic area: when the peak number of nighttime visitors to the target scenic area is 500, the selected convolutional neural network is trained 1000 times; when the peak number of nighttime visitors to the target scenic area is 600, the selected convolutional neural network is trained 1200 times; when the peak number of nighttime visitors to the target scenic area is 700, the selected convolutional neural network is trained 1400 times; when the peak number of nighttime visitors to the target scenic area is 800, the selected convolutional neural network is trained 1600 times, and so on.

[0102] In each training iteration of the convolutional neural network, the optimal planned paths that minimize the total path travel time from the starting point of a historical nighttime user's location within a specific historical time segment to various attractions within the target scenic area are used as the output of the convolutional neural network. The time interval between two adjacent historical moments, the user-related parameter set of the historical nighttime user, the nighttime status data of other nighttime users within the target scenic area corresponding to the historical nighttime user at the starting point of the specific historical time segment, the number of nighttime users in the target scenic area on each day prior to the date of the specific historical time segment, and the scene perception content of the historical nighttime user at evenly spaced historical moments prior to the starting point of the specific historical time segment are used as the input of the convolutional neural network to complete this training.

[0103] Second Embodiment

[0104] Figure 3 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the second embodiment of the present invention.

[0105] like Figure 3 As shown, compared to Figure 2 The AR-based immersive cultural tourism night tour route planning system also includes:

[0106] The route planning server, wirelessly connected to the route planning core module, is used to receive the optimal planned routes from the current night tour user's current location within the current time segment to each attraction in the target scenic area, minimizing the total route travel time.

[0107] For example, a path planning server can be implemented using big data service network elements, cloud computing service network elements, or blockchain service network elements. This server receives the optimal planned paths from the current night tour user's current location within the current time segment, which minimize the total path travel time to each attraction within the target scenic area.

[0108] The route planning server uses a physical storage address that matches the current night tour user to store each of the received optimal planned routes, meaning that the optimal planned routes stored by different night tour users are all different.

[0109] Third Embodiment

[0110] Figure 4 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the third embodiment of the present invention.

[0111] like Figure 4 As shown, compared to Figure 2 The AR-based immersive cultural tourism night tour route planning system also includes:

[0112] The signal generation module is mounted on the AR interactive terminal and is connected to the AR interactive terminal, the scene perception module, the path planning core module and the virtual-real fusion rendering module respectively.

[0113] The signal generation module is used to provide the reference clock signals required by the AR interactive terminal, the scene perception module, the path planning core module and the virtual-real fusion rendering module respectively.

[0114] Specifically, the signal generation module is used to provide the AR interactive terminal, scene perception module, path planning core module and virtual-real fusion rendering module with their respective reference clock signals. The signal generation module has a built-in quartz oscillation unit to generate a reference pulse signal of a preset frequency, and generates the reference clock signal required by the AR interactive terminal, scene perception module, path planning core module and virtual-real fusion rendering module based on the reference pulse signal of the preset frequency.

[0115] Fourth embodiment

[0116] Figure 5 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the fourth embodiment of the present invention.

[0117] like Figure 5 As shown, compared to Figure 2 The AR-based immersive cultural tourism night tour route planning system also includes:

[0118] Parallel communication interface, mounted on AR interactive terminal and connected to AR interactive terminal, scene perception module, path planning core module and virtual-real fusion rendering module respectively;

[0119] Among them, the parallel communication interface is used to establish parallel data connections between AR interactive terminals, scene perception modules, path planning core modules and virtual-real fusion rendering modules.

[0120] For example, the parallel communication interface is used to establish parallel data connections between the AR interactive terminal, the scene perception module, the path planning core module, and the virtual-real fusion rendering module, including: the parallel communication interface is an 8-bit, 16-bit, or 32-bit parallel communication interface.

[0121] Fifth embodiment

[0122] Figure 6 This is an internal structure diagram of an immersive cultural tourism night tour route planning system based on AR virtual interaction, as shown in the fifth embodiment of the present invention.

[0123] like Figure 6 As shown, compared to Figure 2 The AR-based immersive cultural tourism night tour route planning system also includes:

[0124] The target construction module, connected to the path planning core module, is used to establish an intelligent planning model for night tour paths;

[0125] The target construction module performs training on the convolutional neural network to obtain the convolutional neural network after each training, and sends the convolutional neural network after each training as the intelligent planning model for night tour routes to the path planning core module.

[0126] For example, you can choose to use the MATLAB toolbox to perform each training iteration of the convolutional neural network to simulate and test the modeling process of the convolutional neural network after each training iteration.

[0127] Next, various embodiments of the present invention will be further described.

[0128] Optionally, within the above embodiments, in the AR-based immersive cultural tourism night tour route planning system:

[0129] Each optimal planned path is represented by a location information stream, which is formed by sequentially connecting the location information corresponding to each position at even intervals along the path.

[0130] Specifically, the location information corresponding to each position at even intervals along the path is a numerically normalized representation, for example, a binary value with the same number of bits.

[0131] Specifically, the number of evenly spaced positions is equal in each optimal planning path;

[0132] The current time segment starts at the current moment and its duration is a multiple of the time interval between two adjacent historical moments.

[0133] For example, the current time segment starts at the current moment and the duration of the current time segment is a multiple of the time interval between two adjacent historical moments, including: when the time interval between two adjacent historical moments is 1 minute, the duration of the current time segment is 20 minutes.

[0134] The virtual information also includes the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of each other night tour user in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before today, and the scene perception content of each historical moment at even intervals before the current night tour user.

[0135] And, optionally, within the above embodiments, in the AR-based immersive cultural tourism night tour route planning system:

[0136] The intelligent night tour route planning model is adopted. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, the model intelligently predicts the optimal planned routes from the current night tour user's current location to various attractions in the target scenic area within the current time segment, which minimize the total route time. The model includes inputting the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment into the intelligent night tour route planning model.

[0137] For example, a programmable logic device designed using VHDL language can be selected to realize the synchronous input of the time interval length between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data corresponding to each other night tour user in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before the current day, and the scene perception content of each historical moment at even intervals before the current night tour user.

[0138] Among them, the intelligent planning model for night tour routes is adopted. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before the current day, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, the model intelligently predicts the optimal planned paths from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time. The model also includes: running the intelligent planning model for night tour routes to obtain the optimal planned paths output by the intelligent planning model from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time.

[0139] Among them, the optimal planned paths that minimize the total path time for the current night tour user to reach each attraction in the target scenic area from the current location within the current time segment, the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data corresponding to each other night tour user in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before today, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment are all represented in a numerically normalized form.

[0140] For example, the numerical representation of the normalized value is the numerical representation after binary conversion.

[0141] Sixth Embodiment

[0142] Figure 7 The following is a flowchart illustrating the steps of an immersive cultural tourism night tour route planning method based on AR virtual interaction according to the sixth embodiment of the present invention.

[0143] like Figure 7 As shown, the immersive cultural tourism night tour route planning method based on AR virtual interaction includes the following steps:

[0144] The AR interactive terminal uses a wireless communication unit to collect nighttime status data from other nighttime visitors in the target scenic area at the current moment. The AR interactive terminal is worn by the current nighttime visitors in the target scenic area and is equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. The AR interactive terminal includes an optical display module, an interactive input unit, and a wireless communication unit.

[0145] Here, the AR interactive terminal itself serves as a data collection terminal for some basic data, as well as a display terminal for the fused rendering screen based on intelligent prediction results, and a corresponding interactive terminal for the fused rendering screen.

[0146] Meanwhile, the AR interactive terminal is also equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. The scene perception module, the path planning core module, and the virtual-real fusion rendering module can be integrated as independent components on the external control motherboard of the AR interactive terminal, or they can be integrated into the control motherboard of the AR interactive terminal itself.

[0147] The scene perception module is used to collect scene perception content of the current night tour user at evenly spaced historical moments before the current moment. The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module, and an ambient light sensor.

[0148] For example, the scene perception content collected from the current nighttime user at evenly spaced historical moments before the current moment includes: the time interval between two adjacent historical moments can be selected as 1 minute, and the evenly spaced historical moments before the current moment can be selected as 10 historical moments. When the current moment is 8:00 PM, the evenly spaced historical moments before the current moment are 7:59 PM, 7:58 PM, 7:57 PM, 7:56 PM, 7:55 PM, 7:54 PM, 7:53 PM, 7:52 PM, 7:51 PM, 7:50 PM, etc.

[0149] Using a path planning core module connected to both the AR interactive terminal and the scene perception module, the system employs an intelligent night tour path planning model. This model intelligently predicts optimal planned paths from the current night tour user's current location to various attractions within the target scenic area, based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment.

[0150] For example, all other nighttime visitors within the target scenic area can wear their own AR interactive terminals. All the AR interactive terminals of the nighttime visitors are of the same model and are delivered to the target scenic area upon entry. The intelligently predicted optimal planning paths correspond to various attractions within the same target scenic area. This provides a data basis for the current nighttime visitors to select attractions within the current time segment as a future time segment by choosing the optimal planning path through the interactive input unit of their AR interactive terminals.

[0151] A virtual-real fusion rendering module, which is connected to the AR interactive terminal, the scene perception module, and the path planning core module respectively, is used to perform fusion rendering of the screen based on the intelligent prediction results.

[0152] Among them, the fusion rendering of the screen based on the intelligent prediction results includes: overlaying various virtual information onto the scene in front of the current night tour user at the current moment for fusion rendering, and pushing the fusion rendered screen to the optical display module for display. The various virtual information includes the optimal planning paths obtained by intelligent analysis.

[0153] Clearly, the virtual information rendered in the scene in front of the current night tour user at the current moment is not limited to each optimal planned path, but also includes other types of virtual information;

[0154] Among them, the night tour status data corresponding to each other night tour user at the current moment is the user-related parameter set of the other night tour user, the customized visualization information of each historical moment at even intervals before the current moment, the on-site location information, the tourist density of the scenic spot, and the scenic spot number of the scenic spot.

[0155] Among them, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age;

[0156] Specifically, the four parameters of other nighttime tourists—body fat percentage, height-to-weight ratio, gender, and age—not only affect their own nighttime travel route preferences, but also, since they are all within the same scenic area, influence other nighttime tourists, such as the current nighttime tourist's nighttime travel route preferences.

[0157] The following section will describe in detail how four parameters—body fat percentage, height-to-weight ratio, gender, and age—influence each nighttime traveler's nighttime travel route preferences from four aspects:

[0158] Aspect 1: The impact of body fat percentage:

[0159] As a type of body fat data, body fat percentage directly determines the length and slope tolerance of night tour routes by affecting basal metabolic efficiency and sustained exercise capacity. For example, users with high body fat tend to choose short circular routes or routes with multiple rest points, actively avoid three-dimensional attractions that require continuous climbing, and prefer flat waterfront walkways or plaza-type spaces. On the other hand, users with low body fat are more likely to choose routes that allow for in-depth exploration, accept linear routes with a one-way distance of more than 1.5 kilometers, and have a higher tolerance for vertical transportation such as steps and ramps.

[0160] Part Two: The Influence of Body Mass Index (BMI)

[0161] BMI influences the lower limb joints' pressure-bearing capacity and center of gravity stability, thus determining users' preferences for terrain complexity and road surface materials. BMI is linearly positively correlated with knee joint load during walking. When obese individuals with a BMI > 28 walk on uneven surfaces (such as cobblestone paths or gravel trails), their compensatory ankle joint activity increases, resulting in energy consumption that is 22-30% higher than that of people with normal BMI. Insufficient light at night further reduces the visual feedforward adjustment ability to terrain, amplifying the exercise risks associated with BMI.

[0162] Users with high BMI show a significant preference for hard, flat surfaces, avoid groups of steps that require high leg movements, and tend to choose ramps with handrails rather than open staircases. They are also more sensitive to the width of the path. Users with normal BMI are more accepting of terrain diversity and are willing to choose "fun paths" with moderate undulations (such as walking through garden paths or stepping stone bridges).

[0163] Thirdly, the influence of gender:

[0164] Gender differences in nighttime walking scenarios are not only reflected in traditional behavioral preferences, but more fundamentally in differences in visual attention allocation mechanisms and risk perception topologies. Neurobehavioral studies have shown that women adopt a "peripheral scanning mode" when walking at night, paying significantly more attention to potentially dangerous areas such as bushes, dark corners, and enclosed spaces on both sides of the path, while men adopt a "focus-oriented mode," concentrating their attention on the path ahead and fixed objects (such as streetlights and signs). This difference is significantly amplified at night. Female users prioritize well-lit main paths, avoid poorly lit shortcuts or easy paths, prefer paths with open views (clear vision, no blind spots), and show significant avoidance of paths with vegetation covering both sides. They also tend to walk in groups. Male users, on the other hand, are more accepting of efficiency-first paths (choosing the shortest route, even if the lighting conditions are average) and have a higher willingness to try exploratory paths (willing to try unmarked trails, construction passages, etc.).

[0165] Fourthly, the impact of age:

[0166] Age influences the efficiency of multisensory integration, dark adaptation, and fatigue recovery speed, thus determining the time structure and rhythm of night tour routes. Young people (18-35 years old) prefer high-density experience routes, accepting routes with a large number of attractions per unit distance, high information density, and fast pace. They are also accepting of "check-in" style movement (quick and continuous visits to multiple attractions) and have lower sensitivity to the end time of the night tour, willing to extend their tour into the late night. Middle-aged people (40-60 years old) focus on functional adaptability, prioritizing routes with clear rest facilities and reasonable distribution of public health facilities, and have a lower tolerance for the duration of the night tour. Elderly people (over 60 years old) strictly follow the main route, have rigid requirements for the width of the passageway and nighttime lighting, prefer familiar routes, and start and end their night tours early.

[0167] Clearly, the four parameters of body fat percentage, height-to-weight ratio, gender, and age for each nighttime tourist do not act independently, but rather form a cumulative constraint effect, which will affect their nighttime travel route preferences as a whole. Since they are all in the same scenic area, this will further affect other nighttime tourists as a whole, such as the current nighttime tourist's nighttime travel route preferences.

[0168] Among them, the customized visualization information of each night tour user at each moment is the grayscale gradient value, coordinate value and depth value of all the constituent pixels of each human body occupying the sub-frame in the scene in front of the night tour user at that moment, as well as the curvature value of all pixel positions on the contour curve of each human body occupying the sub-frame.

[0169] For example, for each constituent pixel of a sub-image occupied by a human body, the mean square difference of each gray value corresponding to each of its surrounding adjacent constituent pixels can be used as its gray gradient value.

[0170] Among them, the RGB-D depth camera is used to collect customized visualization information of the current night tour user at uniform intervals of each historical moment before the current moment; the inertial measurement unit is used to collect the acceleration of the current night tour user at uniform intervals of each historical moment before the current moment; the GPS / BeiDou dual-mode positioning module is used to collect the on-site location positioning information of the current night tour user at uniform intervals of each historical moment before the current moment; and the ambient light sensor is used to collect the on-site illumination of the current night tour user at uniform intervals of each historical moment before the current moment.

[0171] For example, the GPS / BeiDou dual-mode positioning module is used to collect the on-site location information of the current night tour user at evenly spaced historical moments before the current moment. This includes: the GPS / BeiDou dual-mode positioning module uses the arithmetic mean of the current night tour user's GPS positioning information and BeiDou positioning information at any moment as the on-site location information of the current night tour user at any moment.

[0172] Among them, the customized visualization information of the current night tour user at uniform intervals before the current moment, the acceleration of the current night tour user at uniform intervals before the current moment, the on-site location information of the current night tour user at uniform intervals before the current moment, and the on-site illumination of the current night tour user at uniform intervals before the current moment are used as the scene perception content of the current night tour user at uniform intervals before the current moment.

[0173] Among them, the intelligent planning model for night tour routes is a convolutional neural network after each training. The convolutional neural network includes an input layer, an activation layer and multiple parallel convolutional layers. The multiple parallel convolutional layers are located between the input layer and the activation layer, and the number of parallel convolutional layers shows the same numerical change trend as the area of ​​the target scenic spot.

[0174] For example, the number of parallel convolutional layers showing the same numerical trend as the area of ​​the target scenic spot includes: when the area of ​​the target scenic spot is 10,000 hectares, the number of parallel convolutional layers selected is 2 layers; when the area of ​​the target scenic spot is 20,000 hectares, the number of parallel convolutional layers selected is 4 layers; when the area of ​​the target scenic spot is 30,000 hectares, the number of parallel convolutional layers selected is 6 layers; when the area of ​​the target scenic spot is 40,000 hectares, the number of parallel convolutional layers selected is 8 layers, and so on.

[0175] The number of times the convolutional neural network was trained showed the same numerical trend as the peak number of nighttime visitors to the target scenic spot.

[0176] For example, the number of training iterations of the convolutional neural network shows the same numerical trend as the peak number of nighttime visitors to the target scenic area: when the peak number of nighttime visitors to the target scenic area is 500, the selected convolutional neural network is trained 1000 times; when the peak number of nighttime visitors to the target scenic area is 600, the selected convolutional neural network is trained 1200 times; when the peak number of nighttime visitors to the target scenic area is 700, the selected convolutional neural network is trained 1400 times; when the peak number of nighttime visitors to the target scenic area is 800, the selected convolutional neural network is trained 1600 times, and so on.

[0177] In each training iteration of the convolutional neural network, the optimal planned paths that minimize the total path travel time from the starting point of a historical nighttime user's location within a specific historical time segment to various attractions within the target scenic area are used as the output of the convolutional neural network. The time interval between two adjacent historical moments, the user-related parameter set of the historical nighttime user, the nighttime status data of other nighttime users within the target scenic area corresponding to the historical nighttime user at the starting point of the specific historical time segment, the number of nighttime users in the target scenic area on each day prior to the date of the specific historical time segment, and the scene perception content of the historical nighttime user at evenly spaced historical moments prior to the starting point of the specific historical time segment are used as the input of the convolutional neural network to complete this training.

[0178] Furthermore, in the immersive cultural tourism night tour route planning system and method based on AR virtual interaction according to the present invention:

[0179] The number of historical moments at even intervals before the current moment is proportional to the total number of attractions in the target scenic area, and the number of days before the current day in the target scenic area shows the same numerical trend as the total number of attractions in the target scenic area.

[0180] For example, the number of historical moments evenly spaced before the current moment that is proportional to the total number of attractions in the target scenic area includes: if there are 5 attractions in the target scenic area, the number of historical moments evenly spaced before the current moment is 10; if there are 6 attractions in the target scenic area, the number of historical moments evenly spaced before the current moment is 12; if there are 7 attractions in the target scenic area, the number of historical moments evenly spaced before the current moment is 14; if there are 8 attractions in the target scenic area, the number of historical moments evenly spaced before the current moment is 16, and so on.

[0181] For example, the following scenarios show the same numerical trend in the number of days prior to the selected target scenic area as the total number of attractions within the target scenic area: 5 attractions in the target scenic area, 5 days prior to the selected target scenic area; 10 attractions in the target scenic area, 10 days prior to the selected target scenic area; 15 attractions in the target scenic area, 15 days prior to the selected target scenic area; 20 attractions in the target scenic area, 20 days prior to the selected target scenic area; and so on.

[0182] The statement that the number of historical moments at even intervals before the current moment is proportional to the total number of attractions in the target scenic area, and that the number of days in the target scenic area before the current day shows the same numerical trend as the total number of attractions in the target scenic area, includes: using a first numerical conversion function to represent the numerical conversion relationship that the number of historical moments at even intervals before the current moment is proportional to the total number of attractions in the target scenic area, and using a second numerical conversion function to represent the numerical conversion relationship that the number of days in the target scenic area before the current day shows the same numerical trend as the total number of attractions in the target scenic area.

[0183] The foregoing description of exemplary embodiments of the invention is provided for illustrative and descriptive purposes. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations will obviously be apparent to those skilled in the art. Exemplary embodiments were chosen and described in order to best illustrate the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the various embodiments and variations of the invention suitable for the contemplated particular purpose. The scope of the invention is intended to be defined by the appended claims and their equivalents.

Claims

1. An immersive cultural tourism night tour route planning system based on AR virtual interaction, characterized in that, The system includes: The AR interactive terminal is worn by the current night tour users in the target scenic area and is equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. It includes an optical display module, an interactive input unit, and a wireless communication unit. The wireless communication unit collects the night tour status data of each other night tour user in the target scenic area at the current moment. The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module, and an ambient light sensor, which collects scene perception content of the current night tour user at evenly spaced historical moments before the current moment; The core path planning module is connected to the AR interactive terminal and the scene perception module respectively. It adopts a night tour path intelligent planning model. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, it intelligently predicts the optimal planned paths from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time. The virtual-real fusion rendering module is connected to the AR interactive terminal, the scene perception module, and the path planning core module, respectively, and is used to perform fusion rendering of the screen based on the intelligent prediction results. Among them, the night tour status data corresponding to each other night tour user at the current moment is the user-related parameter set of the other night tour user, the customized visualization information of each historical moment at even intervals before the current moment, the on-site location information, the tourist density of the scenic spot, and the scenic spot number of the scenic spot. Among them, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age; The customized visualization information for each night tour user at each moment includes the grayscale gradient values, coordinate values, and depth values ​​of all constituent pixels of each human body occupying a sub-frame in the scene in front of the night tour user at that moment, as well as the curvature values ​​of all pixel positions on the contour curve of each human body occupying the sub-frame.

2. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in claim 1, characterized in that: The fusion rendering of the screen based on intelligent prediction results includes: overlaying various virtual information onto the scene in front of the current night tour user at the current moment for fusion rendering, and pushing the fusion-rendered screen to the optical display module for display. The various virtual information includes the optimal planning paths obtained by intelligent analysis. Among them, the RGB-D depth camera is used to collect customized visualization information of the current night tour user at uniform intervals of each historical moment before the current moment; the inertial measurement unit is used to collect the acceleration of the current night tour user at uniform intervals of each historical moment before the current moment; the GPS / BeiDou dual-mode positioning module is used to collect the on-site location positioning information of the current night tour user at uniform intervals of each historical moment before the current moment; and the ambient light sensor is used to collect the on-site illumination of the current night tour user at uniform intervals of each historical moment before the current moment. Specifically, the customized visualization information of the current night tour user at uniform intervals before the current moment, the acceleration of the current night tour user at uniform intervals before the current moment, the on-site location information of the current night tour user at uniform intervals before the current moment, and the on-site illumination of the current night tour user at uniform intervals before the current moment are used as the scene perception content of the current night tour user at uniform intervals before the current moment.

3. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in claim 2, characterized in that: The intelligent planning model for night tour routes is a convolutional neural network that has been trained multiple times. The convolutional neural network includes an input layer, an activation layer, and multiple parallel convolutional layers. The multiple parallel convolutional layers are located between the input layer and the activation layer, and the number of parallel convolutional layers shows the same numerical trend as the area of ​​the target scenic spot. The number of times the convolutional neural network was trained showed the same numerical trend as the peak number of nighttime visitors to the target scenic spot. In each training iteration of the convolutional neural network, the output of the network consists of the known optimal planned paths from the starting point of a historical nighttime user within a specific historical time segment to various attractions within the target scenic area, minimizing the total path travel time. The input of the network includes the time interval between two adjacent historical moments, the user-related parameter set for the historical nighttime user, the nighttime status data of other nighttime users within the target scenic area corresponding to the historical nighttime user at the starting point of the historical time segment, the number of nighttime users in the target scenic area on each day prior to the date of the historical time segment, and the scene perception data of the historical nighttime user at evenly spaced intervals before the starting point of the historical time segment.

4. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in claim 3, characterized in that, The system also includes: The route planning server, wirelessly connected to the route planning core module, is used to receive the optimal planned routes from the current night tour user's current location within the current time segment to each attraction in the target scenic area, minimizing the total route travel time. The route planning server uses a physical storage address that matches the current nighttime user to store each of the received optimal planned routes.

5. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in claim 3, characterized in that, The system also includes: The signal generation module is mounted on the AR interactive terminal and is connected to the AR interactive terminal, the scene perception module, the path planning core module and the virtual-real fusion rendering module respectively. The signal generation module is used to provide the reference clock signals required by the AR interactive terminal, the scene perception module, the path planning core module, and the virtual-real fusion rendering module, respectively.

6. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in claim 3, characterized in that, The system also includes: Parallel communication interface, mounted on AR interactive terminal and connected to AR interactive terminal, scene perception module, path planning core module and virtual-real fusion rendering module respectively; The parallel communication interface is used to establish parallel data connections between AR interactive terminals, scene perception modules, path planning core modules, and virtual-real fusion rendering modules.

7. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in claim 3, characterized in that, The system also includes: The target construction module, connected to the path planning core module, is used to establish an intelligent planning model for night tour paths; The target construction module performs training on the convolutional neural network to obtain the trained convolutional neural network, and sends the trained convolutional neural network as the intelligent planning model for night tour routes to the path planning core module.

8. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in any one of claims 3-7, characterized in that: Each optimal planned path is represented by a location information stream, which is formed by sequentially connecting the location information corresponding to each position at even intervals along the path. The current time segment starts at the current moment and its duration is a multiple of the time interval between two adjacent historical moments. The various types of virtual information also include the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of each other night tour user in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before today, and the scene perception content of each historical moment at even intervals before the current night tour user.

9. The immersive cultural tourism night tour route planning system based on AR virtual interaction as described in any one of claims 3-7, characterized in that: The intelligent night tour route planning model is adopted. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, the model intelligently predicts the optimal planned routes from the current night tour user's current location to various attractions in the target scenic area within the current time segment, which minimize the total route time. The model includes inputting the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment into the intelligent night tour route planning model. Among them, the intelligent planning model for night tour routes is adopted. Based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on each day before the current day, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment, the model intelligently predicts the optimal planned paths from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time. The model also includes: running the intelligent planning model for night tour routes to obtain the optimal planned paths output by the intelligent planning model from the current night tour user's current location to each attraction in the target scenic area within the current time segment, which minimizes the total path time. Among them, the optimal planned paths that minimize the total travel time of the current night tour user from their current location to various attractions in the target scenic area within the current time segment, the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data corresponding to each other night tour user in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment are all represented in a numerically normalized form.

10. An immersive cultural tourism night tour route planning method based on AR virtual interaction, characterized in that, The method includes: The AR interactive terminal uses a wireless communication unit to collect nighttime status data from other nighttime visitors in the target scenic area at the current moment. The AR interactive terminal is worn by the current nighttime visitors in the target scenic area and is equipped with a scene perception module, a path planning core module, and a virtual-real fusion rendering module. The AR interactive terminal includes an optical display module, an interactive input unit, and a wireless communication unit. The scene perception module is used to collect scene perception content of the current night tour user at evenly spaced historical moments before the current moment. The scene perception module includes an RGB-D depth camera, an inertial measurement unit, a GPS / BeiDou dual-mode positioning module, and an ambient light sensor. Using a path planning core module connected to both the AR interactive terminal and the scene perception module, the system employs an intelligent night tour path planning model. This model intelligently predicts optimal planned paths from the current night tour user's current location to various attractions within the target scenic area, based on the time interval between two adjacent historical moments, the user-related parameter set of the current night tour user, the night tour status data of other night tour users in the target scenic area at the current moment, the number of night tour users in the target scenic area on previous days, and the scene perception content of the current night tour user at evenly spaced historical moments before the current moment. A virtual-real fusion rendering module, which is connected to the AR interactive terminal, the scene perception module, and the path planning core module respectively, is used to perform fusion rendering of the screen based on the intelligent prediction results. Among them, the night tour status data corresponding to each other night tour user at the current moment is the user-related parameter set of the other night tour user, the customized visualization information of each historical moment at even intervals before the current moment, the on-site location information, the tourist density of the scenic spot, and the scenic spot number of the scenic spot. Among them, the user-related parameter set for each night tour user includes the user's body fat data, height-to-weight ratio, gender, and age; The customized visualization information for each night tour user at each moment includes the grayscale gradient values, coordinate values, and depth values ​​of all constituent pixels of each human body occupying a sub-frame in the scene in front of the night tour user at that moment, as well as the curvature values ​​of all pixel positions on the contour curve of each human body occupying the sub-frame.

Images