Intelligent exhibition hall guiding path guiding system and method

Abstract

The invention discloses an intelligent exhibition hall guiding path guiding system and method, and relates to the technical field of intelligent exhibition halls. The method comprises the following steps: constructing a three-dimensional intelligent exhibition hall model according to an intelligent exhibition hall construction drawing and data acquired by unmanned aerial vehicle photography; internet-of-things data in the intelligent exhibition hall are acquired in real time; constructing a space-time weight matrix, and marking exhibit value weights and path passing costs; constructing a user interest dynamic model, and predicting user interest migration; a user selects a user interest exhibition stand recommended by the system through the mobile terminal; the system performs path planning according to the exhibition stand selected by the user, and selects a path recommendation user with the minimum passing cost; and the user watches the recommended path in the three-dimensional intelligent exhibition hall model through the mobile terminal. According to the method, the regional population density is calculated through the equipment in the exhibition hall, the time-space weight matrix is constructed to calculate the passage cost of the user visiting the path, and the lowest passage cost is recommended to the user, so that the interactivity of audiences and the guide efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart exhibition halls, and in particular relates to a smart exhibition hall guided path guidance system and method. Background Art

[0002] Traditional exhibition guides rely primarily on static display boards, manual explanations, and simple electronic guides. However, these methods have significant drawbacks. Static display boards offer very limited information and are slow to update, failing to meet the dynamic demands of exhibits. While manual explanations can provide relatively detailed information, they are costly, and the quality and content of the explanations vary significantly depending on the guide. Simple electronic guides are limited in functionality and lack intelligent and personalized services, making it difficult to provide accurate information tailored to the diverse needs of visitors.

[0003] In recent years, with the development of technology, some intelligent tour guide systems have gradually emerged, such as positioning systems based on GPS, Wi-Fi fingerprints, or Bluetooth beacons, as well as tour guide systems combined with path planning algorithms. However, these systems still have many problems. The complex indoor environment, with the presence of multipath effects and occlusions, leads to low positioning accuracy and the inability to accurately determine the visitor's location. Path planning algorithms are mostly based on preset routes and cannot be dynamically adjusted based on factors such as real-time crowd flow within the exhibition hall and the status of exhibits, resulting in less than optimized tour guide routes. In addition, most existing tour guide systems lack interactivity with the audience and the ability to provide personalized recommendations, and are unable to provide customized tour guide services based on factors such as the audience's interests, hobbies, and visiting time. Summary of the Invention

[0004] The purpose of the present invention is to provide a smart exhibition hall tour path guidance system and method, which calculates the travel cost of the user's visiting path by constructing a spatiotemporal weight matrix and recommends the lowest travel cost to the user, solving the problem that the existing guide system cannot be based on the audience's interests and hobbies and has a low level of intelligence.

[0005] To solve the above technical problems, the present invention is achieved through the following technical solutions: The present invention is a method for guiding a tour path in a smart exhibition hall, comprising the following steps: Step S1: Construct a 3D smart pavilion model based on the smart pavilion construction drawings and data collected by drone photography; Step S2: Obtain IoT data in the smart exhibition hall in real time; Step S3: Construct a spatiotemporal weight matrix, marking the value weight of the exhibits and the path travel cost; Step S4: Build a user interest dynamic model to predict user interest migration; Step S5: The user selects a user-interested booth recommended by the system through the mobile terminal; Step S6: The system plans a route based on the booth selected by the user and recommends the route with the lowest travel cost to the user; Step S7: The user views the recommended path in the 3D smart pavilion model through a mobile terminal.

[0006] As a preferred technical solution, in step S1, the process of constructing a 3D smart exhibition hall is as follows: Step S11: Verify the completeness of the construction drawings of the smart exhibition hall, including floor plans, elevations, sections, and structural drawings. Mark the data on the drawings and import them into the modeling software, noting key data such as dimensions, elevations, material markings, and door and window locations. The modeling software uses AutoCAD + Revit (BIM modeling), and the imported drawings are in the CAD drawing format (DWG / DXF). Step S12: Use the extrude tool to convert the plane outline into a three-dimensional structure (directly link to BIM data in Revit), and use the component library to insert standard components; use the component library (such as SketchUp 3D Warehouse) to quickly insert standard components, and use Blender's "subdivision surface" tool to refine complex shapes; Step S13: Assign materials, textures, and lighting to the 3D building interior model; add material maps (Revit material library or Substance Painter customization) according to the drawing annotations, set ambient light and shadows, and enhance the realism of the model; Step S14: setting rendering parameters to render the three-dimensional building interior model; Step S15: Install ground laser scanning equipment inside the building and allocate drones in the air to take photos, respectively collecting point cloud data of equipment inside the exhibition hall; Step S16: After pre-processing the point cloud data, the scanned point cloud data of multiple different acquisition areas are spliced, registered, and fused.

[0007] As a preferred technical solution, in step S15, the UAV flight planning and design is performed in combination with the generated three-dimensional building indoor model. By selecting the overlap values ​​of the lateral overlap and the heading overlap, the UAV is caused to shoot and obtain oblique images according to the pre-set route. The collected image data is analyzed and transmitted to the computer and the cloud. The calculation formulas for the lateral overlap and the heading overlap are as follows: ; Where, Represent the heading overlap and lateral overlap respectively, Respectively represent the projection of the image width and image height on the ground, Represent the width and height of the sensor respectively, Represents the exposure spacing and the distance between flight lines, Respectively represent the number of pixels in the rows and columns of the digital image, Indicates flight altitude. Indicates the focal length of the drone camera, Indicates the pixel size.

[0008] As a preferred technical solution, in step S16, the specific process of stitching the point cloud data is as follows: Step S161: performing hierarchical neural feature registration on the pre-processed point cloud data; Step S162: Learn the implicit deformation field from the source point cloud to the target point cloud. The specific calculation formula is as follows: ; Where, represents the deformation field parameterized by the neural network, represents the network weight, Represents the three-dimensional coordinates of a point in the input point cloud, represents the coordinate offset predicted by the neural network, represents the implicit features of the source point cloud, Represent the implicit features of the target point cloud; Step S163: Perform differentiable probabilistic registration optimization and calculate the probability of point pair matching. The specific calculation formula is as follows: ; Where, represents the probability of point pair matching, represents the Sigmoid activation function, represents the learnable temperature coefficient, Indicates the first The feature vector of a point, Indicates the first The feature vector of a point; Step S164: Construct a spatiotemporal continuous energy function. The specific formula is as follows: ; Where, represents the rigid transformation energy term, represents the non-rigid deformation regularization term, represents the event camera data consistency term, represents the deformation weight coefficient, Represents the event data weight coefficient; Step S165: Use the semantic graph neural network to identify key landmarks in the scene, and fuse the asynchronous data of the camera to correct the registration drift caused by dynamic objects.

[0009] As a preferred technical solution, in step S2, the IoT data in the smart exhibition hall includes WIFI probes, temperature and humidity sensors, carbon dioxide concentration sensors, and light intensity sensors; the WIFI probes are installed around the booths and are used to calculate the population density of the current area in real time; The calculation formula of the Wi-Fi probe for the population density of the current area is as follows: ; Where, Indicates the total number of active mobile devices in the booth area. Indicates the effective space area around the booth. Represents the mobile activity coefficient.

[0010] As a preferred technical solution, in step S3, the calculation formula of the spatiotemporal weight matrix is ​​as follows: ; Where, Indicates the comprehensive tour guide priority of booth s at time t, They represent the weight of traffic, the weight of exhibit popularity and the weight of event response respectively. They represent real-time traffic indicators, booth popularity, and event response values ​​respectively; The exhibit value weight includes cultural value dimension, educational value dimension and commercial value dimension, and the specific calculation formula is: ; Where, They represent cultural value dimension, educational value dimension and commercial value dimension respectively; The calculation formula for the path cost is as follows: ; Where, represents the path cost from point a to point b, represents the physical distance from point a to point b, t represents the timestamp, represents the congestion coefficient from point a to point b at time t, Represents the cost of traveling through a barrier-free passage.

[0011] As a preferred technical solution, in step S4, a user interest dynamic model is constructed based on the user's behavior process of visiting the smart exhibition hall. The user interest dynamic model is trained to predict the user's favorite exhibits and push them to the user's mobile terminal. The specific process of constructing the user interest dynamic model is as follows: the user's explicit behavior, implicit behavior and background tags are obtained; the explicit behavior includes the user's stay time in front of the booth, the number of AR interactions and the frequency of taking photos; the implicit behavior includes eye tracking through VR glasses worn by the user and heart rate changes obtained through a smart bracelet; the background tags include age, occupation and visit records; the dynamic interest vector calculation formula of user u at time t is as follows: ; Where, Indicates the total number of user behavior categories, represents the dynamic weight of the k-th type of behavior, represents the quantitative value of the k-th type of behavior of user u at time k, represents the event decay function, represents the spatial attention of user u at time t.

[0012] As a preferred technical solution, in step S6, the calculation process of minimizing the travel cost introduces dynamic composite parameters, including: a distance-time weight distribution ratio that is automatically adjusted based on real-time crowd density, a path priority value that is positively correlated with the user's demand for interactive exhibits, and a compensation correction coefficient for travel time based on the operating status of exhibition hall equipment.

[0013] As a preferred technical solution, in step S7, the exhibit information of each exhibition hall and the booths in the exhibition hall is marked in the three-dimensional smart exhibition hall model; the exhibit information includes the exhibit name, exhibit age, exhibit introduction, current visitor flow of the exhibit, exhibit popularity and exhibit location information.

[0014] The present invention is a smart exhibition hall guide path guidance system, including acquisition equipment, a smart exhibition hall platform and a mobile terminal; the acquisition equipment includes a ground laser scanning device, a drone device, a WIFI probe, a temperature and humidity sensor, a carbon dioxide concentration sensor and a light intensity sensor; the smart exhibition hall platform includes a three-dimensional building model production module, a point cloud splicing module, a regional population density calculation module, a spatiotemporal weight matrix calculation module, a user interest dynamic model construction module, a user interest recommendation module, a path travel cost calculation module and an exhibition hall guide module; the three-dimensional building model production module is used to construct a three-dimensional building model of the smart exhibition hall according to the smart exhibition hall construction drawings The point cloud stitching module is used to stitch point cloud data collected by ground laser scanning equipment and drone equipment; the regional population density calculation module is used to calculate the population density of the area around the booth based on the WIFI probe; the spatiotemporal weight matrix calculation module is used to calculate the comprehensive navigation priority of the booth at any time; the user interest dynamic model construction module is used to construct a user interest dynamic model based on user behavior; the user interest recommendation module is used to recommend booths of interest to users based on user information; the path travel cost calculation module is used to generate the lowest path travel cost based on the booth selected by the user; the exhibition hall navigation module is used to generate a visual navigation path; The mobile terminal is used to access the smart exhibition hall platform to view the guided path.

[0015] The present invention has the following beneficial effects: (1) The present invention collects indoor and outdoor data of the smart exhibition hall to construct a three-dimensional smart exhibition hall model, uses the equipment in the exhibition hall to calculate the regional population density, constructs a spatiotemporal weight matrix to calculate the travel cost of the user's visiting path, and recommends the lowest travel cost to the user. It also recommends guided tour routes based on the flow of people, thereby improving the audience's interactivity and tour efficiency; (2) The present invention constructs a user interest dynamic model based on the user's behavior process of visiting the smart exhibition hall. By training the user interest dynamic model, the user's favorite exhibits are predicted and pushed to the user's mobile terminal, thereby improving the system's personalized recommendation capabilities and providing customized guided tour services based on the audience's interests and hobbies.

[0016] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 This is a flow chart of a smart exhibition hall tour path guidance method of the present invention. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0021] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1 It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application.

[0022] See also Figure 1 As shown, the present invention is a smart exhibition hall guide path guidance method, comprising the following steps: Step S1: Construct a 3D smart pavilion model based on the smart pavilion construction drawings and data collected by drone photography; Step S2: Obtain IoT data in the smart exhibition hall in real time; Step S3: Construct a spatiotemporal weight matrix, marking the value weight of the exhibits and the path travel cost; Step S4: Build a user interest dynamic model to predict user interest migration; Step S5: The user selects a user-interested booth recommended by the system through the mobile terminal; Step S6: The system plans a route based on the booth selected by the user and recommends the route with the lowest travel cost to the user; Step S7: The user views the recommended path in the 3D smart pavilion model through a mobile terminal.

[0023] In step S1, the process of constructing a 3D smart pavilion is as follows: Step S11: Determine the integrity of the construction drawings of the smart exhibition hall, including floor plans, elevations, sections, structural drawings, etc., ensure that the drawing information is complete, accurate, and clearly marked. After marking the data on the drawings, import them into the modeling software, marking key data such as: dimensions, elevations, material markings, and door and window positions; the modeling software uses AutoCAD+Revit (BIM modeling), and the imported drawing format is CAD drawings (DWG / DXF). Set the units to ensure consistency with the units of the drawings to avoid dimensional errors; usually, you can use the "Import" or "Open" function to select the corresponding drawing file format for import. After importing, adjust the position, scale, and angle of the drawing as needed to make its coordinate system and spatial position in the software consistent with the actual situation; Step S12: Use the extrude tool to convert the plane outline into a three-dimensional structure (directly link to BIM data in Revit), and use the component library to insert standard components; use the component library (such as SketchUp 3D Warehouse) to quickly insert standard components, and use Blender's "subdivision surface" tool to refine complex shapes; The specific operation method is as follows: Use the drawing tools in the modeling software, such as lines, curves, rectangles, and polygons, to draw along the edges and outlines of the building on the drawing to outline the basic shape of the building. Based on the building's height information, use the modeling software's stretch or extrude function to stretch or extrude the drawn 2D outline along the Z axis to make it a 3D model with a certain height. Refer to the section and elevation drawings. Based on the basic model, gradually add building details such as doors, windows, balconies, stairs, decorative moldings, etc. You can use the various modeling tools in the modeling software, such as Boolean operations, lofting, chamfering, etc., to create these details.

[0024] Step S13: Assign materials to the three-dimensional building interior model and add textures and lighting; add material maps (Revit material library or Substance Painter customization) according to the drawing annotations, set ambient light and shadows, and enhance the realism of the model; the specific operations are as follows: according to the actual material of the building, assign corresponding material properties to each part of the model, such as concrete, wood, glass, metal, etc. In modeling software, there is usually a material library to choose from, and you can also customize the material's color, glossiness, transparency and other parameters; in order to make the model more realistic, you can add texture maps to the material. By pasting texture images that match the actual material on the surface of the model, such as brick wall texture, wooden floor texture, marble texture, etc., the realism of the model is enhanced; Step S14: Set rendering parameters to render the three-dimensional building interior model; in the modeling software, set appropriate lighting according to the actual lighting conditions of the building and the effect to be expressed. For example, natural light can be simulated by setting parallel light to simulate the direction and intensity of sunlight, and indoor light can be simulated by adding point light sources, spotlights, etc.; after completing the model construction, adding material textures and setting lights, you can perform the rendering operation. Set rendering parameters as needed, such as resolution, rendering quality, anti-aliasing, etc., and then render the model into high-quality pictures or videos to better display the three-dimensional effect of the building; Since the above steps all construct a 3D exhibition hall model based on the smart exhibition hall drawings, the 3D exhibition hall model only shows the exhibition hall framework. The drawings do not include the exhibition hall's internal booths, decorations, electrical equipment, furniture and other facilities. Therefore, after the exhibition hall is deployed, it is necessary to use ground laser scanning equipment and drone equipment to collect the data after the exhibition hall is deployed, generate a 3D facility model and place it in the 3D exhibition hall model; Due to the complex environment inside the smart exhibition hall, there are certain guardrails, slopes, billboards, and obstructions such as safety equipment and fire-fighting equipment. A single measurement method cannot complete high-precision data collection. A terrestrial laser scanner combined with drone imaging is needed to improve the accuracy and integrity of data collection. The terrestrial laser scanner uses the SmartScan VR800 3D laser scanner with a collection resolution of 1mm. The collection time of a single terrestrial laser scanner is set to 5 minutes, and the laser emission frequency is , frame rate 30 To ensure full coverage of the entire survey area and avoid data overlap and blind spots, 36 stations were deployed, with data collection taking place from 10 PM to 3 AM, during the evening hours of the museum's closing time. After the terrestrial laser scanning survey was completed, the generated point cloud data was imported into the corresponding programs for processing, and engineering collaboration was facilitated through a web browser.

[0025] Step S15: Install ground laser scanning equipment inside the building and allocate drones in the air to take photos, respectively collecting point cloud data of equipment inside the exhibition hall; In step S15, the UAV flight planning is performed in combination with the generated 3D building indoor model. By selecting the overlap values ​​of the lateral overlap and the heading overlap, the UAV is directed to shoot and acquire oblique images according to the pre-set route. The quality of the acquired image data is analyzed to check for problems such as missing shots, image blur, and color distortion. Retakes are performed in a timely manner, the acquired image data is analyzed, and the acquired image data is transmitted to a computer and the cloud. In actual operation, the lateral overlap and heading overlap coverage areas are required to exceed the measurement boundary by no less than 20%. The calculation formulas for lateral overlap and heading overlap are as follows: ; Where, Represent the heading overlap and lateral overlap respectively, Respectively represent the projection of the image width and image height on the ground, Represent the width and height of the sensor respectively, Represents the exposure spacing and the distance between flight lines, Respectively represent the number of pixels in the rows and columns of the digital image, Indicates flight altitude. Indicates the focal length of the drone camera, Indicates the pixel size.

[0026] The preprocessing of point cloud data includes: ground laser scanning data preprocessing and UAV image data preprocessing; The preprocessing process for terrestrial laser scanning data is as follows: First, the terrestrial laser scanning data is cleaned to remove abnormal points, noise, and invalid data. Next, step S16 is executed. To avoid missed scans, each of the eight stations is scanned four times. After fully automatic registration of point clouds from the same station, eight point cloud groups are generated. Adjacent point cloud groups are then fully automatically registered. The registered data is then filtered to remove abnormal data and outliers, improving data quality and smoothness. Finally, the scan data from different acquisition areas or at different times is spliced ​​together to generate point cloud data for the entire building.

[0027] The preprocessing process for drone images is as follows: First, geometric correction is performed to eliminate image distortion in the drone imagery, ensuring accurate object positions and shapes. Next, step S16 is performed to perform data registration, eliminating image offsets and deformations and ensuring data consistency. Finally, the registered data is fused into a single image with higher resolution and richer colors, enhancing detail and resolution, and enhancing the three-dimensionality and visual quality. The fused image data is processed using PIX4d software into a point cloud in the .LAS format and saved.

[0028] Step S16: After pre-processing the point cloud data, the scanned point cloud data of multiple different acquisition areas are spliced, registered, and fused.

[0029] In step S16, the specific process of stitching the point cloud data is as follows: Step S161: performing hierarchical neural feature registration on the pre-processed point cloud data; Step S162: Learn the implicit deformation field from the source point cloud to the target point cloud. The specific calculation formula is as follows: ; Where, represents the deformation field parameterized by the neural network, represents the network weight, Represents the three-dimensional coordinates of a point in the input point cloud, represents the coordinate offset predicted by the neural network, represents the implicit features of the source point cloud, Represent the implicit features of the target point cloud; Step S163: Perform differentiable probabilistic registration optimization and calculate the probability of point pair matching. The specific calculation formula is as follows: ; Where, represents the probability of point pair matching, represents the Sigmoid activation function, represents the learnable temperature coefficient, Indicates the first The feature vector of a point, Indicates the first The feature vector of a point; Step S164: Construct a spatiotemporal continuous energy function. The specific formula is as follows: ; Where, represents the rigid transformation energy term, represents the non-rigid deformation regularization term, represents the event camera data consistency term, represents the deformation weight coefficient, Represents the weight coefficient of event data; can propose dynamically moving point clouds to avoid interference from moving objects; Step S165: Use the semantic graph neural network to identify key landmarks in the scene, and fuse the asynchronous data of the camera to correct the registration drift caused by dynamic objects.

[0030] In step S2, the IoT data in the smart pavilion includes Wi-Fi probes, temperature and humidity sensors, carbon dioxide concentration sensors, and light intensity sensors; the Wi-Fi probes are installed around the booths and are used to calculate the population density of the current area in real time; The calculation formula of the Wi-Fi probe for the current area population density is as follows: ; Where, Indicates the total number of active mobile devices in the booth area. The Bluetooth beacon / WiFi probes deployed in the exhibition hall capture devices with RSSI > -70dBm and data updated in the last 5 seconds. Indicates the effective space area around the booth. Vector calculation based on GIS map, excluding inaccessible areas, Indicates the mobile activity coefficient (0.1-0.3), Calculated through the standard deviation of device displacement (reflecting the discreteness of crowd movement).

[0031] Temperature and humidity sensors are used to calculate the visitor comfort index and construct the comfort index CI=0.6T+0.4H. When the temperature and humidity thresholds are exceeded, visitors are preferentially guided to the CI∈[18,22] area. The carbon dioxide concentration sensor detects that the concentration exceeds 800ppm, triggering the fresh air system linkage, and the safe evacuation path is automatically activated. The light intensity sensor is used to dynamically adjust the brightness of the AR interface, and the adaptive algorithm is used to reduce the energy consumption of the equipment by more than 40%.

[0032] In step S3, the calculation formula of the spatiotemporal weight matrix is ​​as follows: ; Where, Indicates the comprehensive tour guide priority of booth s at time t, They represent the weight of traffic flow (to prevent congestion, take 0.6), the weight of exhibit popularity (to enhance the value of the visit, take 0.3), and the weight of event response (such as temporary closure, take 0.1). They represent real-time traffic indicators, booth popularity, and event response values ​​respectively. The methods for obtaining traffic, exhibit popularity, and event response data are as follows:

[0033] The weight of the exhibit value includes the cultural value dimension (expert rating 0-10 points), the educational value dimension (student ticket visit frequency) and the commercial value dimension (corresponding to the sales volume of cultural and creative products). The specific calculation formula is: ; Where, Respectively represent the cultural value dimension, educational value dimension and commercial value dimension; the calculated score is such as: Along the River During the Qingming Festival 、Spring and Autumn Lotus and Crane Square Pot etc; The formula for calculating the path travel cost is as follows: ; Where, represents the path cost from point a to point b, represents the physical distance from point a to point b, t represents the timestamp, It represents the congestion coefficient from point a to point b at time t (real-time passenger flow / maximum carrying capacity), Represents the cost of accessing an accessible route (wheelchair ramp = 0, stairs = +0.5).

[0034] In step S4, a user interest dynamic model is constructed based on the user's behavior process of visiting the smart exhibition hall. After training, the user interest dynamic model is completed to predict the user's favorite exhibits and push them to the user's mobile terminal. The specific process of constructing the user interest dynamic model is as follows: the user's explicit behavior, implicit behavior and background tags are obtained; explicit behavior includes the user's stay time in front of the booth, the number of AR interactions and the frequency of taking photos; implicit behavior includes eye tracking through VR glasses worn by the user and heart rate changes obtained through smart bracelets; background tags include age, occupation and visit history; the dynamic interest vector calculation formula of user u at time t is as follows: ; Where, Indicates the total number of user behavior categories, represents the dynamic weight of the k-th type of behavior, represents the quantitative value of the k-th type of behavior of user u at time k, represents the event decay function, represents the spatial attention of user u at time t.

[0035] In step S6, the calculation process for minimizing the travel cost introduces dynamic composite parameters, including: a distance-time weight distribution ratio that is automatically adjusted based on real-time crowd density, a path priority value that is positively correlated with the user's desire to interact with the exhibits selected, and a compensation correction coefficient for the travel time based on the operating status of the exhibition hall equipment.

[0036] In step S7, the exhibit information of each exhibition hall and booth in the exhibition hall is marked in the 3D smart exhibition hall model; the exhibit information includes the exhibit name, exhibit age, exhibit introduction, current exhibit visitor flow, exhibit popularity and exhibit location information.

[0037] The present invention is a smart pavilion guide path guidance system, including acquisition equipment, a smart pavilion platform and a mobile terminal; the acquisition equipment includes a ground laser scanning device, a drone device, a WIFI probe, a temperature and humidity sensor, a carbon dioxide concentration sensor and a light intensity sensor; the smart pavilion platform includes a three-dimensional building model making module, a point cloud splicing module, a regional population density calculation module, a time-space weight matrix calculation module, a user interest dynamic model construction module, a user interest recommendation module, a path travel cost calculation module and a pavilion guide module; the three-dimensional building model making module is used to construct a three-dimensional building model of the smart pavilion according to the construction drawings of the smart pavilion. Building model; point cloud stitching module is used to stitch point cloud data collected by ground laser scanning equipment and drone equipment; regional population density calculation module is used to calculate the population density of the area around the booth based on Wi-Fi probes; spatiotemporal weight matrix calculation module is used to calculate the comprehensive guide priority of the booth at any time; user interest dynamic model construction module is used to build a user interest dynamic model based on user behavior; user interest recommendation module is used to recommend booths of user interest based on user information; path travel cost calculation module is used to generate the lowest path travel cost based on the booth selected by the user; exhibition hall guide module is used to generate a visual guide path; Mobile terminals are used to access the smart exhibition hall platform to view the guided tour path.

[0038] It is worth noting that in the above system embodiment, the various units included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0039] In addition, those skilled in the art will appreciate that all or part of the steps in the above-mentioned embodiments can be accomplished by instructing related hardware through a program, and the corresponding program can be stored in a computer-readable storage medium.

[0040] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for guiding a tour path in a smart exhibition hall, characterized in that: The steps include: Step S1: Construct a 3D smart pavilion model based on the smart pavilion construction drawings and data collected by drone photography; Step S2: Obtain IoT data in the smart exhibition hall in real time; Step S3: Construct a spatiotemporal weight matrix, marking the value weight of the exhibits and the path travel cost; Step S4: Build a user interest dynamic model to predict user interest migration; Step S5: The user selects a user-interested booth recommended by the system through the mobile terminal; Step S6: The system plans a route based on the booth selected by the user and recommends the route with the lowest travel cost to the user; Step S7: The user views the recommended path in the 3D smart pavilion model through a mobile terminal.

2. The method for guiding a smart exhibition hall tour path according to claim 1, characterized in that: In step S1, the process of constructing a 3D smart pavilion is as follows: Step S11: Determine the integrity of the smart pavilion construction drawings, annotate the data on the drawings, and import them into the modeling software; Step S12: Use the stretching tool to convert the plane outline into a three-dimensional structure, and use the component library to insert standard components; Step S13: assigning materials to the three-dimensional building interior model and adding texture and lighting; Step S14: setting rendering parameters to render the three-dimensional building interior model; Step S15: Install ground laser scanning equipment inside the building and allocate drones in the air to take photos, respectively collecting point cloud data of equipment inside the exhibition hall; Step S16: After pre-processing the point cloud data, the scanned point cloud data of multiple different acquisition areas are spliced, registered, and fused.

3. The method for guiding a smart exhibition hall tour path according to claim 2, characterized in that: In step S15, the UAV flight planning is performed in combination with the generated three-dimensional building indoor model. By selecting the overlap values ​​of the lateral overlap and the heading overlap, the UAV is directed to shoot and acquire oblique images according to the pre-set route. The acquired image data is analyzed and transmitted to the computer and the cloud. The calculation formulas for the lateral overlap and the heading overlap are as follows: ； Where, Represent the heading overlap and lateral overlap respectively, Respectively represent the projection of the image width and image height on the ground, Represent the width and height of the sensor respectively, Represents the exposure spacing and the distance between flight lines, Respectively represent the number of pixels in the rows and columns of the digital image, Indicates flight altitude. Indicates the focal length of the drone camera, Indicates the pixel size.

4. The method for guiding a smart exhibition hall tour path according to claim 2, characterized in that: In step S16, the specific process of stitching the point cloud data is as follows: Step S161: performing hierarchical neural feature registration on the pre-processed point cloud data; Step S162: Learn the implicit deformation field from the source point cloud to the target point cloud. The specific calculation formula is as follows: ； Where, represents the deformation field parameterized by the neural network, represents the network weight, Represents the three-dimensional coordinates of a point in the input point cloud, represents the coordinate offset predicted by the neural network, represents the implicit features of the source point cloud, Represent the implicit features of the target point cloud; Step S163: Perform differentiable probabilistic registration optimization and calculate the probability of point pair matching. The specific calculation formula is as follows: ； Where, represents the probability of point pair matching, represents the Sigmoid activation function, represents the learnable temperature coefficient, Indicates the first The feature vector of a point, Indicates the first The feature vector of a point; Step S164: Construct a spatiotemporal continuous energy function. The specific formula is as follows: ； Where, represents the rigid transformation energy term, represents the non-rigid deformation regularization term, represents the event camera data consistency term, represents the deformation weight coefficient, Represents the event data weight coefficient; Step S165: Use the semantic graph neural network to identify key landmarks in the scene, and fuse the asynchronous data of the camera to correct the registration drift caused by dynamic objects.

5. The method for guiding a smart exhibition hall tour route according to claim 1, characterized in that: In step S2, the IoT data in the smart pavilion includes WIFI probes, temperature and humidity sensors, carbon dioxide concentration sensors, and light intensity sensors; the WIFI probes are installed around the booths and are used to calculate the population density of the current area in real time; The calculation formula of the Wi-Fi probe for the population density of the current area is as follows: ； Where, Indicates the total number of active mobile devices in the booth area. Indicates the effective space area around the booth. Represents the mobile activity coefficient.

6. The method for guiding a smart exhibition hall tour path according to claim 1, characterized in that: In step S3, the calculation formula of the spatiotemporal weight matrix is ​​as follows: ； Where, Indicates the comprehensive tour guide priority of booth s at time t, They represent the weight of traffic, the weight of exhibit popularity and the weight of event response respectively. They represent real-time traffic indicators, booth popularity, and event response values ​​respectively; The exhibit value weight includes cultural value dimension, educational value dimension and commercial value dimension, and the specific calculation formula is: ； Where, They represent cultural value dimension, educational value dimension and commercial value dimension respectively; The calculation formula for the path cost is as follows: ； Where, represents the path cost from point a to point b, represents the physical distance from point a to point b, t represents the timestamp, represents the congestion coefficient from point a to point b at time t, Represents the cost of traveling through a barrier-free passage.

7. The method for guiding a smart exhibition hall tour path according to claim 1, characterized in that: In step S4, a user interest dynamic model is constructed based on the user's behavior process of visiting the smart exhibition hall. The user interest dynamic model is trained to predict the user's favorite exhibits and push them to the user's mobile terminal. The specific process of constructing the user interest dynamic model is as follows: the user's explicit behavior, implicit behavior and background tags are obtained; the explicit behavior includes the user's stay time in front of the booth, the number of AR interactions and the frequency of taking photos; the implicit behavior includes eye tracking through VR glasses worn by the user and heart rate changes obtained through the smart bracelet; the background tags include age, occupation and visit history; the dynamic interest vector calculation formula of user u at time t is as follows: ; Where, Indicates the total number of user behavior categories, represents the dynamic weight of the k-th type of behavior, represents the quantitative value of the k-th type of behavior of user u at time k, represents the event decay function, represents the spatial attention of user u at time t.

8. The method for guiding a smart exhibition hall tour path according to claim 1, characterized in that: In step S6, the calculation process for minimizing the travel cost introduces dynamic composite parameters, including: a distance-time weight distribution ratio automatically adjusted based on real-time crowd density, a path priority value positively correlated with the user's desire for interactive exhibits, and a compensation correction coefficient for travel time based on the operating status of exhibition hall equipment.

9. The method for guiding a smart exhibition hall tour path according to claim 1, characterized in that: In step S7, the exhibit information of each exhibition hall and booth in the exhibition hall is marked in the three-dimensional smart exhibition hall model; the exhibit information includes the exhibit name, exhibit age, exhibit introduction, current exhibit visitor flow, exhibit popularity and exhibit location information.

10. A smart exhibition hall guide path guidance system, characterized in that: The acquisition equipment, smart exhibition hall platform and mobile terminal; the acquisition equipment includes ground laser scanning equipment, drone equipment, WIFI probe, temperature and humidity sensor, carbon dioxide concentration sensor and light intensity sensor; the smart exhibition hall platform includes a three-dimensional building model production module, a point cloud splicing module, a regional population density calculation module, a spatiotemporal weight matrix calculation module, a user interest dynamic model construction module, a user interest recommendation module, a path travel cost calculation module and an exhibition hall guide module; the three-dimensional building model production module is used to construct a three-dimensional building model of the smart exhibition hall according to the construction drawings of the smart exhibition hall; the point cloud splicing module Used to stitch point cloud data collected by ground laser scanning equipment and drone equipment; the regional population density calculation module is used to calculate the population density of the area around the booth based on WIFI probes; the spatiotemporal weight matrix calculation module is used to calculate the comprehensive navigation priority of the booth at any time; the user interest dynamic model construction module is used to construct a user interest dynamic model based on user behavior; the user interest recommendation module is used to recommend booths of interest to users based on user information; the path travel cost calculation module is used to generate the lowest path travel cost based on the booth selected by the user; the exhibition hall navigation module is used to generate a visual navigation path; The mobile terminal is used to access the smart exhibition hall platform to view the guided path.

Images