Exhibition display system based on three-dimensional virtual visualized wisdom exhibition hall
Patent Information
- Application Number
- CN202610673742.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-15
- Publication Date
- 2026-08-18
AI Technical Summary
[0003]然而，这些现有技术在实际应用中存在一个显著的共性技术缺陷：展示过程与用户实时行为状态的脱节，导致关键特征信息的展示效率与可靠性降低
该基于三维虚拟可视化智慧展厅的展品展示系统中，通过用户状态定位模块实时对用户的观察特征进行监控，利用响应匹配数据库构建的用户观察状态与特征显现状态的特征映射集进行适应性特征显示，实时根据用户的观察视角与路线位置进行显示规划，为不同观察状态下的用户匹配最佳的特征观察视角，使得用户在观察路线中的实时位置、行进速度和观察姿态形成动态、自适应的协同，保证整体展品观察效果。
Smart Images
Figure CN122597640A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of virtual exhibition technology, and more specifically, to an exhibition system based on a three-dimensional virtual visualization smart exhibition hall. Background Technology
[0002] In the fields of digital exhibitions and cultural heritage preservation, 3D virtual visualization smart exhibition hall systems have become an important technological tool. These systems digitize physical exhibits or scenes through high-precision 3D modeling, real-time rendering, and interactive technologies, reconstructing them in virtual space. This breaks through physical time and space limitations, providing users with an immersive viewing experience. A typical system usually includes a 3D model library, a virtual roaming module, a human-computer interaction interface, and a backend management unit, allowing users to move along preset or free paths and observe virtual exhibits from different angles.
[0003] However, these existing technologies share a significant common technical flaw in practical applications: the display process is disconnected from the user's real-time behavior, leading to reduced efficiency and reliability in displaying key feature information. Specifically, this manifests as follows: First, there is a contradiction between the fixed time sequence of displays and the dynamic changes in user behavior: In fixed animation or fixed route display modes, the flipping, rotating, or zooming of feature areas of exhibits are performed at a preset time and speed. However, the speed at which users move through the virtual exhibition hall is greatly uncertain and variable, influenced by their personal operating habits, network latency, differences in points of interest, and even hardware performance. This means that by the time a user reaches the optimal viewing position, the feature display animation of the exhibit may have already finished playing or has not yet started, causing them to "miss" the core display content.
[0004] Secondly, the one-way display logic lacks awareness and adaptation to user attention: Most existing systems follow a one-way process of "system playback - user viewing," lacking the ability to judge in real time whether the user is currently in an effective observation state. Even if the user happens to be at the observation point, their focus may not be on the exhibit or the currently displayed feature area, and the system will still execute the display action as planned, resulting in important display behaviors not being effectively received.
[0005] In summary, the core problem with existing 3D virtual exhibit display systems lies in the fact that the triggering and execution of display commands are pre-programmed and time-fixed, failing to achieve dynamic and adaptive coordination with the user's real-time position, speed, and viewing posture along the observation route. Therefore, there is an urgent need for a novel intelligent display method and system—an exhibit display system based on a 3D virtual visualization smart exhibition hall that can perceive the user's status in real time. Summary of the Invention
[0006] The purpose of this invention is to provide an exhibit display system based on a three-dimensional virtual visualization smart exhibition hall to solve the problems mentioned in the background art.
[0007] To achieve the above objectives, an exhibit display system based on a 3D virtual visualization smart exhibition hall is provided, including a model integration module, which is used to collect the appearance information of the exhibits to be exhibited, construct a 3D model, and extract the model feature area. It also includes an exhibition area planning module, a user status positioning module, a response matching database, and a feature display change module. The exhibition area planning module is used to plan the display area of exhibits and to divide the display positions of the overall three-dimensional model of the exhibits and the model feature areas. The user status positioning module is equipped with a camera monitoring device to capture the user's position and viewing angle in the exhibit observation path; The response matching database is used to collect perspective observation verification sample data, obtain the optimal display area and model feature area display angle corresponding to different positions and different perspectives of the observation path, and construct a feature mapping set between the user's observation state and the feature display state. The feature display change module combines real-time monitoring of the user's location and viewing angle with the feature mapping set to provide feedback on the display area of exhibits, and works with the display area planning module to adjust the matching model feature area for responsive display.
[0008] As a further improvement to this technical solution, the three-dimensional model construction method in the model integration module is as follows: S101. Use laser scanning equipment to set up multiple scanning stations around the exhibits; S102. Extract the raw point cloud data of each station; S103. Synchronously record the three-dimensional parameters of a single site using laser scanning equipment; S104. Using 3D model building software, multiple site clouds are merged into a complete point cloud based on feature matching, and outlier noise is removed and uniform sampling is performed to optimize the amount of data. S105. Using the Poisson reconstruction algorithm, a continuous triangular mesh surface is generated based on the complete point cloud. Texture mapping is then used to attach the scanned high-resolution photos of the exhibits onto the network model, generating a solid 3D model with realistic textures.
[0009] As a further improvement to this technical solution, the model feature region extraction method in the model integration module is as follows: S1011. The three-dimensional model features are extracted using a multi-modal feature definition method, specifically including geometric feature extraction, semantic feature extraction, and global feature extraction. Regarding the geometric feature extraction method, based on the rate of change of the normal of the mesh vertex, high curvature regions are identified. At the same time, regions with similar normal directions are clustered to extract the triangular skin ID set of the above feature regions as the geometric features of the current 3D model. For semantic feature extraction, a combination of manual annotation and AI assistance is used. Specific regions of the model are directly selected by human and associated with text descriptions. AI is then used to train the 3D segmentation model and automatically identify the corresponding feature regions. For global feature extraction, a viewpoint quality algorithm is used to calculate the model's metrics from multiple perspectives and extract the best viewing perspective for the current feature region. S1014. All feature regions and their attributes are associated with and stored in XML format with the 3D model file to form a structured feature library.
[0010] As a further improvement to this technical solution, the method for dividing the display area by the display area planning module includes the following steps: S201. Set parameters for the display area to divide the display area into a main image display area and multiple feature view display areas; S202. Multi-camera collaborative rendering is used for display processing. Different projection areas are assigned to different cameras, and display parameters are configured.
[0011] As a further improvement to this technical solution, the method for capturing the user's position and viewing angle in the exhibit observation path in the user status positioning module includes the following steps: S301. Configure video surveillance equipment and deploy it in various locations to ensure that its monitoring field of view covers the observation area corresponding to the current display area. S302. Using the MediaPipe Pose solution, 33 3D key points of the human body are inferred and captured from a single frame of RGB image; S303. Calculate the relative positions of each key point to obtain the current user's posture; S304. Capture reference objects on the observation route, calculate the positional relationship between the 3D key points of the human body and the key points of the reference objects, and obtain the current position of the user on the observation route. S305. Using the MediaPipe Face Mesh solution, 468 key points on the user's face are detected and outlined. S306. Combining the detected facial key points with the user's posture, the three-dimensional vector formed by the tip of the nose and the center point of the face is used as the user's viewing perspective.
[0012] As a further improvement to this technical solution, the method for constructing a feature mapping set between user observation state and feature display state in the response matching database includes the following steps: S401. Divide the observation route into equal areas, allowing users to switch perspectives and adjust their viewpoints within the observation range. S402. Divide the observation field of view into specific view ranges, adjust for different specific view ranges, adjust the display parameters of the feature view display area and the model feature view in real time, and mark it as the feature display status. S403. Obtain user feedback, define the best display parameters and the best model feature perspective for the current user position within a specific viewpoint range based on the user feedback, and construct a feature mapping set between the user's observation state and the feature display state.
[0013] As a further improvement to this technical solution, the viewing angle range in S401 is 0°-90°, where 0° indicates that the user's viewing angle is parallel to the display area, and 90° indicates that the user's viewing angle is perpendicular to the display area.
[0014] As a further improvement to this technical solution, the specific viewing angle range in S402 includes , as well as .
[0015] As a further improvement to this technical solution, the method for providing feedback on the display area of exhibits in the feature display change module includes the following steps: S501. Based on posture tracking, capture the current user's body posture and current location area; S502, in conjunction with eye-tracking, captures the user's viewing angle and matches the corresponding specific viewing angle range; S503. Obtain the current user's observation posture and combine it with the feature mapping set as a condition for the dynamic scheduling engine to respond and control the corresponding feature view display area. S504. Configure the best display parameters and the best model feature perspective for the current feature view display area.
[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This exhibit display system based on a 3D virtual visualization smart exhibition hall monitors users' observation characteristics in real time through a user status positioning module. It uses a feature mapping set of user observation status and feature display status constructed by a response matching database to perform adaptive feature display. The system plans the display in real time according to the user's observation perspective and route position, matching the best feature observation perspective for users in different observation states. This enables the user's real-time position, speed, and posture in the observation route to form a dynamic and adaptive coordination, ensuring the overall exhibit observation effect. Attached Figure Description
[0017] Figure 1 This is a block diagram of the overall system structure of the present invention; Figure 2 This is a schematic diagram illustrating the steps of constructing a 3D model of the exhibits according to the present invention; Figure 3 This is a schematic diagram of the display area division process of the present invention.
[0018] The meanings of the labels in the diagram are as follows: 10. Model Integration Module; 20. Exhibition Area Planning Module; 30. User status positioning module; 40. Response matching database; 50. Feature Display Change Module. Detailed Implementation
[0019] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0020] Please see Figure 1 As shown, an exhibit display system based on a 3D virtual visualization smart exhibition hall is provided, including a model integration module 10, which is used to collect the appearance information of the exhibits to be exhibited, construct a 3D model, and extract the model feature area. It also includes an exhibition area planning module 20, a user status positioning module 30, a response matching database 40, and a feature display change module 50. The exhibition area planning module 20 is used to plan the display area of exhibits and to divide the display positions of the overall 3D model of the exhibits and the model feature areas. The user status positioning module 30 is equipped with a camera monitoring device to capture the user's position and viewing angle in the exhibit observation path; The response matching database 40 is used to collect perspective observation verification sample data, obtain the optimal display area and model feature area display angle corresponding to different positions and different perspectives of the observation path, and construct a feature mapping set between the user's observation state and the feature display state. The feature display change module 50 combines real-time monitoring of the user's location and viewing angle with the feature mapping set to provide feedback on the display area of exhibits, and works with the display area planning module 20 to adjust the matching model feature area for responsive display.
[0021] The specific plan is as follows: To ensure the effective display of exhibits, this solution uses the model integration module 10 to collect the appearance information of the exhibits, construct a 3D model, and extract the model's feature areas, such as... Figure 2 As shown, the specific construction and extraction methods are as follows: First, multiple scanning stations are set up around the exhibit using laser scanning equipment (mechanical light or LiDAR scanner) to perform a comprehensive scan, extracting raw point cloud data from each station (each station contains millions to tens of millions of 3D coordinate points). The laser scanning equipment simultaneously records 3D parameters such as RGB color information, reflectivity, and surface normal vectors for each station. Based on these parameters, a 3D model is constructed. 3D modeling software (such as Geomagic Wrap or RealityCapture) is used to merge the point clouds from multiple stations into a complete point cloud based on feature matching, removing outliers and optimizing data volume through uniform sampling. Then, a continuous triangular mesh surface is generated from the complete point cloud using the Poisson reconstruction algorithm. Texture mapping is used to overlay high-resolution photographs of the exhibit onto the network model, generating a realistic 3D model. To facilitate more intuitive observation of exhibit features and extract and display details, this solution employs a multi-modal feature definition approach to extract features from the 3D model, specifically including geometric feature extraction, semantic feature extraction, and global feature extraction.
[0022] Regarding the geometric feature extraction method, based on the rate of change of the normal of the mesh vertices, high curvature regions (such as edges, patterns, etc.) are identified. At the same time, regions with similar normal directions are clustered to identify feature regions such as planes and cylinders, and the triangular skin ID set of the above feature regions is extracted as the geometric features of the current 3D model.
[0023] For semantic feature extraction, a combination of manual annotation and AI assistance is used. In the modeling software, specific regions of the model are directly selected by human and associated with text descriptions. With the help of AI, the 3D segmentation model is trained and the corresponding feature regions are automatically identified.
[0024] For global feature extraction, the viewpoint quality algorithm is used to calculate the visible surface area, projection contour saliency, and other indicators of the model under multiple viewpoints, and to extract the best viewing angle of the current feature region.
[0025] All feature regions (defined using triangular facet indexes or 3D bounding boxes) and their attributes are stored in XML format in association with the 3D model file to form a structured feature library, which serves as the basis for subsequent feature display.
[0026] Finally, display conditions are assigned to different feature areas to determine the timing of feature display. In this solution, the display condition is the state of the observing user, that is, the corresponding observation area and observation perspective. When the display condition meets the current timing of feature display, dynamic calculation is performed and the feature focusing instruction is executed (that is, the matching model features are retrieved to the corresponding display area). The corresponding model features and the main perspective (the entire 3D model of the exhibit) are displayed synchronously through picture-in-picture projection.
[0027] Furthermore, due to the limited display area of the exhibits, in order to ensure the display effect of the overall 3D model and feature areas, it is necessary to use the exhibition area planning module 20 to plan the exhibit display area and divide the display positions of the overall 3D model and model feature areas, such as... Figure 3 As shown, the display area is parameter-set, dividing it into a main image display area and multiple feature view display areas. For example, the parameters of the main image display area are (x: 0.15, y: 0.15, width: 0.7, height: 0.7), located in the center of the screen. The parameters of a feature view display area are (x: 0, y: 0, width: 0.7, height: 0.7), located at the edge of the screen. These parameters constitute the rectangular window of the display area, representing the overall 3D model and the display position of model features. After the area division is completed, multi-camera collaborative rendering is used for display processing. Different cameras are assigned corresponding projection areas, i.e., corresponding rectangular windows, and display parameters are configured. For example, the rendering target of the main camera is the main image display area, used to display the overall model, and its corresponding display parameters are (FOV: 45°-60°, normal lighting), where FOV is the scene angle range of the camera projection. The rendering target of the feature cameras is the feature view display area, used to render model features, and its corresponding display parameters are (FOV: 20°-30°, close-up lighting). In the specific display process, as shown... Figure 3 As shown, the exhibit display area is divided into a single central area (main image display area) and multiple edge areas (feature view display areas), with corresponding cameras configured to display the model. The target of the display is the 3D model and feature data constructed by the model integration module 10. The display conditions are the user's observation state, i.e., the user's position and observation view, and are combined with a dynamic scheduling engine for response projection.
[0028] Furthermore, since users' observation positions and perspectives can easily change during exhibit observation, including both static and dynamic observation states, some features may be missed during the display. To avoid this problem, this solution uses a user status positioning module 30 configured with a camera monitoring device to capture the user's position and perspective along the exhibit observation path. The specific steps are as follows: First, configure the video surveillance equipment. This solution uses high-definition network cameras with a frame rate of over 30fps to ensure smooth operation. Position the high-definition network cameras so that their field of view covers the observation area corresponding to the current display area. During the monitoring process, it is necessary to track the user's posture and field of view.
[0029] For pose tracking, the MediaPipe Pose solution is used to infer and capture 33 3D key points of the human body (such as shoulder, elbow, and knee key points) from a single frame of RGB image, and calculate the relative position of each key point to determine the current user's pose, such as sitting, sideways, and standing, providing context for determining the relative relationship between the user and the screen. At the same time, reference objects on the observation path, such as ground markings, are captured, and reference key points are generated for the corresponding positions of the reference objects. The positional relationship between the human body's 3D key points and the reference key points is calculated to obtain the current position of the user on the observation path.
[0030] For eye-tracking, the MediaPipe Face Mesh solution was used to detect 468 key points on the user's face, outlining key points such as the user's eyelids, iris contours, lips, and facial contours. Combining the detected facial key points with the user's posture, the three-dimensional vector formed by the tip of the nose and the center point of the face was used as the user's viewing perspective.
[0031] After completing the user observation status monitoring, in order to display the features of the display area in real time, it is necessary to collect perspective observation verification sample data in advance through the response matching database 40 to obtain the optimal display area and model feature area display angle corresponding to different positions and perspectives of the observation path. This will construct a feature mapping set between the user observation status and the feature display status. Specifically, during the construction process, state calibration is first required. The observation route is divided into equal areas, allowing the user to sequentially switch perspectives in different areas. Perspective adjustments are made within the observation perspective range, which is 0°-90°. 0° indicates that the user's perspective is parallel to the display area, 90° indicates that the user's perspective is perpendicular to the display area, and less than 0° or greater than 90° indicates that the current user's perspective cannot observe the display area. To further refine this, specific perspective ranges are divided within the observation perspective range. In this solution, three specific perspective ranges are defined as follows: , as well as The decimal places are rounded off and adjusted for different specific viewpoint ranges. The display parameters of the feature viewpoint display area and the model feature viewpoint are adjusted in real time and marked as feature display status. User feedback is obtained, and the best display parameters and the best model feature viewpoint for the current user position under a specific viewpoint range are defined based on the user feedback. A feature mapping set between the user observation state and the feature display state is constructed as the response condition for the dynamic scheduling engine.
[0032] Finally, during the actual demonstration, the feature display change module 50, combined with real-time monitoring of the user's position and viewing angle, and the feature mapping set, provides feedback on the display area of the exhibits. This, along with the display area planning module 20, adjusts the matching model feature area for responsive display. Specifically, based on posture tracking, the module captures the user's current body posture and location, and with vision tracking, captures the user's viewing angle, matches the corresponding specific viewing angle range, and obtains the user's current viewing posture. This, combined with the feature mapping set, serves as the condition for the dynamic scheduling engine to respond and adjust the corresponding feature view display area. Furthermore, it configures the optimal display parameters and optimal model feature view for the current feature view display area, enabling adaptive feature display.
[0033] This invention monitors the user's observation characteristics in real time through the user status positioning module 30, and uses the feature mapping set of user observation status and feature display status constructed by the response matching database 40 to perform adaptive feature display. It plans the display in real time according to the user's observation perspective and route position, and matches the best feature observation perspective for users in different observation states. This enables the user's real-time position, travel speed and observation posture in the observation route to form a dynamic and adaptive coordination, ensuring the overall observation effect of the exhibits.
[0034] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.
Claims
1. An exhibit display system based on a three-dimensional virtual visualization smart exhibition hall, comprising a model integration module (10), wherein the model integration module (10) is used to collect the appearance information of the exhibits to be exhibited, construct a three-dimensional model, and extract the model feature areas, characterized in that: It also includes a display area planning module (20), a user status positioning module (30), a response matching database (40), and a feature display change module (50); The exhibition area planning module (20) is used to plan the display area of exhibits and divide the display positions of the overall three-dimensional model of the exhibits and the model feature area; The user status positioning module (30) is equipped with a camera monitoring device to capture the user's position and viewing angle in the exhibit observation path; The response matching database (40) is used to collect perspective observation verification sample data, obtain the best display area and model feature area display angle corresponding to different positions and different perspectives of the observation path, and construct a feature mapping set of user observation state and feature display state; The feature display change module (50) combines real-time monitoring of the user's location and viewing angle, and combines the feature mapping set to display feedback of the exhibit display area, and cooperates with the display area planning module (20) to adjust the matching model feature area for response display.
2. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 1, characterized in that: The three-dimensional model construction method in the model integration module (10) is as follows: S101. Use laser scanning equipment to set up multiple scanning stations around the exhibits; S102. Extract the raw point cloud data of each station; S103. Synchronously record the three-dimensional parameters of a single site using laser scanning equipment; S104. Using 3D model building software, multiple site clouds are merged into a complete point cloud based on feature matching, and outlier noise is removed and uniform sampling is performed to optimize the amount of data. S105. Using the Poisson reconstruction algorithm, a continuous triangular mesh surface is generated based on the complete point cloud. Texture mapping is then used to attach the scanned high-resolution photos of the exhibits onto the network model, generating a solid 3D model with realistic textures.
3. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 2, characterized in that: The model feature region extraction method in the model integration module (10) is as follows: S1011. The three-dimensional model features are extracted using a multi-modal feature definition method, specifically including geometric feature extraction, semantic feature extraction, and global feature extraction. Regarding the geometric feature extraction method, based on the rate of change of the normal of the mesh vertex, high curvature regions are identified. At the same time, regions with similar normal directions are clustered to extract the triangular skin ID set of the above feature regions as the geometric features of the current 3D model. For semantic feature extraction, a combination of manual annotation and AI assistance is used. Specific regions of the model are directly selected by human and associated with text descriptions. AI is then used to train the 3D segmentation model and automatically identify the corresponding feature regions. For global feature extraction, a viewpoint quality algorithm is used to calculate the model's metrics from multiple perspectives and extract the best viewing perspective for the current feature region. S1014. All feature regions and their attributes are associated with and stored in XML format with the 3D model file to form a structured feature library.
4. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 1, characterized in that: The method for dividing the display area by the display area planning module (20) includes the following steps: S201. Set parameters for the display area to divide the display area into a main image display area and multiple feature view display areas; S202. Multi-camera collaborative rendering is used for display processing. Different projection areas are assigned to different cameras, and display parameters are configured.
5. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 1, characterized in that: The method for capturing the user's position and viewing angle in the exhibit observation path in the user status positioning module (30) includes the following steps: S301. Configure video surveillance equipment and deploy it in various locations to ensure that its monitoring field of view covers the observation area corresponding to the current display area. S302. Using the MediaPipe Pose solution, 33 3D key points of the human body are inferred and captured from a single frame of RGB image; S303. Calculate the relative positions of each key point to obtain the current user's posture; S304. Capture reference objects on the observation route, calculate the positional relationship between the 3D key points of the human body and the key points of the reference objects, and obtain the current position of the user on the observation route. S305. Using the MediaPipe Face Mesh solution, 468 key points on the user's face are detected and outlined. S306. Combining the detected facial key points with the user's posture, the three-dimensional vector formed by the tip of the nose and the center point of the face is used as the user's viewing perspective.
6. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 1, characterized in that: The method for constructing a feature mapping set between user observation state and feature display state in the response matching database (40) includes the following steps: S401. Divide the observation route into equal areas, allowing users to switch perspectives and adjust their viewpoints within the observation range. S402. Divide the observation field of view into specific view ranges, adjust for different specific view ranges, adjust the display parameters of the feature view display area and the model feature view in real time, and mark it as the feature display status. S403. Obtain user feedback, define the best display parameters and the best model feature perspective for the current user position within a specific viewpoint range based on the user feedback, and construct a feature mapping set between the user's observation state and the feature display state.
7. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 6, characterized in that: The viewing angle range in S401 is 0°-90°, where 0° means the user's viewing angle is parallel to the display area and 90° means the user's viewing angle is perpendicular to the display area.
8. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 6, characterized in that: The specific viewing angle range in S402 includes , as well as .
9. The exhibit display system based on a three-dimensional virtual visualization smart exhibition hall according to claim 1, characterized in that: The method for providing feedback on the display area of exhibits in the feature display change module (50) includes the following steps: S501. Based on posture tracking, capture the current user's body posture and current location area; S502, in conjunction with eye-tracking, captures the user's viewing angle and matches the corresponding specific viewing angle range; S503. Obtain the current user's observation posture and combine it with the feature mapping set as a condition for the dynamic scheduling engine to respond and control the corresponding feature view display area. S504. Configure the best display parameters and the best model feature perspective for the current feature view display area.