Interactive projection optimization method for exhibition halls

Abstract

The present invention relates to an interactive projection optimization method for an exhibition hall, and the following technical solutions are adopted: before starting the projection, using sensors to sense changes in ambient light in the exhibition hall and evaluate lighting conditions in different areas; using spatial modeling technology to obtain the spatial layout in the exhibition hall and identify wall and object occlusion information in the projection area; using cameras or infrared sensing devices to capture audience gestures, movements and gaze directions in real time; dynamically adjusting projection content according to the audience's movements and positions; including when the audience approaches a certain exhibit, amplifying related projection content or switching to a detailed display mode; collecting behavioral data of each audience member, including the audience's stay time and number of interactions in front of the exhibit; predicting the audience's points of interest through a machine learning model, and recommending projection content related to the audience's interests; and adjusting the brightness and color parameters of each projector according to the actual projection area and content of each projection device.

Description

Technical Field

[0001] The invention relates to a projection optimization method, in particular to an interactive projection optimization method for an exhibition hall. Background Art

[0002] The current interactive projection optimization method for exhibition halls, especially the system described in Chinese invention patent 2023102436601, although it has certain innovations and can provide specific projection content for different audiences through interactive setting information, exhibition tracking monitoring and interactive target conditions, it can improve the interactivity and richness of the exhibition, but it still has some shortcomings and limitations, which limit its application and effect in more complex exhibition scenarios.

[0003] First, the method mainly relies on preset interactive conditions to trigger the projection content. Although this rule-based logic can load relevant projection data according to the interests of visitors, its flexibility and intelligence level are still low. When judging whether the interactive target conditions are met, the system described in the patent mainly relies on preset conditions or manually input information to track and monitor the exhibition and make interactive judgments, rather than based on the audience's real-time behavior data and behavior prediction. This means that the system does not conduct in-depth analysis and mining of each audience's points of interest and behavior patterns, but only makes judgments based on fixed logical paths, making it difficult to achieve dynamic content adjustment and more personalized projection recommendations. Therefore, changes in audience behavior and sudden changes in points of interest cannot be captured and responded to in a timely manner, which may lead to a mismatch between the projection content and the audience's needs. For example, if a visitor suddenly shows great interest in an exhibit during the exhibition, the system cannot perceive this change in real time, resulting in a delay in the update of the projection content or missing the best display time. Secondly, the system's monitoring of audience behavior is mainly based on whether the interactive target conditions are met to trigger the projection, and this triggering mode is relatively passive. It cannot actively analyze the audience's behavior patterns, emotional states, and action details, thereby achieving a more natural and smooth interactive experience. For example, when a visitor stops in front of an exhibit or observes a display from a specific angle, the patented system can only determine whether a preset interaction condition is met, but cannot perform deeper projection optimization through behavioral analysis, and lacks the perception and analysis of the visitor's gaze direction, body posture, gestures, and other micro-behaviors. Therefore, this interactive mode limits the depth of interaction between the visitor and the exhibits to a certain extent, and it is difficult to achieve personalized and situational projection adjustments.

[0004] In addition, the method described in the public technical document has limited consideration of environmental factors. Light changes, spatial layout and physical environment in the exhibition hall have a direct impact on the projection effect, but the system does not describe in detail how to perceive and optimize these environmental factors. For example, under different lighting conditions, the brightness and contrast of the projected content need to be adjusted in real time to ensure that the audience can clearly see the projected content under different viewing conditions. However, the technical solution of this patent does not involve how to perceive changes in ambient light in the exhibition hall or how to perform spatial modeling to optimize projection angles and area displays. As a result, in actual operation, the projected content may be too dim or distorted at certain angles or light, affecting the audience's viewing experience. Summary of the invention

[0005] The purpose of the present invention is to provide an interactive projection optimization method for an exhibition hall, thereby solving some of the drawbacks pointed out in the background technology.

[0006] The present invention solves the above-mentioned technical problems by adopting the following technical solution, which includes the following steps:

[0007] S1. Initial optimization of projection for ambient light and spatial perception:

[0008] S1.1. Before starting the projection, use sensors to sense changes in ambient light in the exhibition hall and evaluate lighting conditions in different areas;

[0009] S1.2. Use spatial modeling technology to obtain the spatial layout of the exhibition hall and identify the wall and object occlusion information in the projection area;

[0010] S2. Dynamic adjustment of projection content based on audience behavior perception:

[0011] S2.1. Capture the audience's gestures, movements and gaze direction in real time through cameras or infrared sensing devices;

[0012] S2.2, dynamically adjust the projection content according to the audience's movements and positions; including zooming in on the relevant projection content or switching to detailed display mode when the audience approaches a certain exhibit;

[0013] S3. Intelligent content generation and audience preference analysis:

[0014] S3.1. Collect behavioral data of each visitor, including the duration of stay in front of the exhibits and the number of interactions;

[0015] S3.2, predicting the audience's interests through machine learning models, and recommending projection content related to the audience's interests;

[0016] S4, Multi-projection collaborative optimization and edge fusion:

[0017] S4.1. Adjust the brightness and color parameters of each projector according to the actual projection area and content of each projection device;

[0018] S4.2. Use image processing technology to fuse the edge areas of multiple projection devices.

[0019] Furthermore, the method for constructing the dynamically adjusted projection content includes: arranging RGB-D cameras, LiDAR, infrared sensors and ultrasonic distance sensors in the exhibition hall to capture the spatial position and movement of the audience; constructing the audience's real-time position perception matrix M(t,x,y,z) by fusing sensor data, where t represents time and (x,y,z) is the audience's three-dimensional coordinates; predicting the audience's movement trajectory and behavior changes, using the integral formula:

[0020]

[0021] Among them, M(t,x,y,z) represents the position change function of the audience in time t; is the speed at which the viewer's position changes over time, is the change in the acceleration of the audience; the integral limit 0 to t represents the accumulated observation of the audience's behavior in the entire time range; the parameter λ is the acceleration influence coefficient, which is used to balance the influence of speed and acceleration on the system judgment; the result S(t) is the predicted value of the audience's movement behavior at time t, and it is used to judge in real time whether the audience enters the interactive area of ​​a certain exhibit; when the distance D(t) between the audience and the exhibit is less than the predetermined distance D0, it will be judged that the audience enters the interactive area, triggering content adjustment.

[0022] Furthermore, when the audience enters the interactive area, the projected content is adjusted based on the audience's position and gaze direction; the adaptive optimization of the projected content is achieved through the following equation, which dynamically optimizes the display effect according to the audience's behavior, including content magnification, focus adjustment, and brightness and angle correction:

[0023]

[0024] Where I(θ,φ,p,t) is the brightness or focus distribution function of the projected content at the audience's gaze angles θ and φ; and They represent the second-order derivatives of the projection content relative to the audience’s gaze angles θ and φ, respectively, indicating the changes in the projection content in different directions; the integral region V is the spatial volume of the projection;

[0025] In the second term, κ is the control coefficient, which controls the response sensitivity of content changes, g(p,t) represents the function of the change of projection content with time t, p0 is the reference projection parameter, which is defined as the static reference content; the time decay factor β controls the smoothness of the content adjustment process.

[0026] Furthermore, the method for constructing the dynamic adjustment of the projection content includes: based on a deep learning convolutional neural network, analyzing the audience's movements, gestures and expressions in real time, extracting the movement feature vector V m (t); the feature vector includes the audience's gestures, posture, and body movement information; the impact of the action on the projected content is dynamically calculated by combining the audience's action characteristics and gaze area through high-order tensor integration, and the formula is as follows:

[0027]

[0028] Among them, V m (t) is the action feature vector of the audience at time t, reflecting the characteristics of the audience’s gestures and body movements; W r (t) is a weight function, which defines the audience's gaze weights on different exhibits and determines the degree of influence of the audience's gaze direction on the adjustment of the projection content; is the divergence of the vector field, which is used to measure the correlation between action and gaze;

[0029] The integral area Ω represents the audience's activity space, and the time decay factor γ controls the long-term effect of action adjustment on content; α is the response coefficient of action adjustment, which adjusts the sensitivity of the response to the audience's action; when H(t) exceeds the preset threshold, the change of the projected content is triggered, including enlarged display or switching to detailed display mode.

[0030] Furthermore, the method for collecting the behavioral data of each visitor includes: capturing the visitor's residence time and refining the visitor's behavior pattern, including changes in position and frequency of interaction, through dynamic hierarchical analysis of the collected data; identifying the visitor's emotional state, including curiosity, fatigue, concentration, and deep interest in the exhibits, and then making personalized adjustments to the exhibits; modeling the behavioral data into a three-dimensional time, position, and interaction matrix M(t,x,y), where t is time and (x,y) represents the plane coordinates of the visitor in the exhibition hall; and performing hierarchical analysis on the visitor's stop behavior through the formula:

[0031]

[0032] Among them, S(t) is the cumulative value of the audience's stop behavior at time t; N represents the time rate of change of the viewer's position on the (x, y) plane, and represents the speed at which the viewer moves in the exhibition hall; interact (t′) is the number of interactions of the audience at time t′, recording the frequency of interaction between the audience and the exhibits; α is the weight coefficient of the interaction, which adjusts the influence of the interaction on the behavior analysis.

[0033] Furthermore, the method for collecting the behavior data of each audience member includes: establishing a dynamic two-way coupling relationship between the audience member's behavior and the exhibit content; analyzing the audience's reaction to the exhibit content in real time, and then adjusting the display method; the coupling relationship is expressed by the formula:

[0034]

[0035] Among them, C(t) is the function of the coupling between audience behavior and exhibit content, which represents the influence of exhibit content on audience behavior at time t; is the second-order derivative of the change of the exhibit content over time, indicating the acceleration of the dynamic change of the exhibit content; I(c,t) is the display effect of the exhibit content at time t, where c is the change parameter of the exhibit content;

[0036] N interact (t) is the number of interactions of the audience, which shows the interaction between the audience and the exhibits, and β is the weight of the impact of the number of interactions on the changes in the exhibit content.

[0037] Furthermore, the method for collecting the behavioral data of each visitor includes: analyzing the change pattern of dwell time through nonlinear dynamic modeling, inferring the change point of the visitor's interest, and adjusting the depth or complexity of the exhibit display in real time; the nonlinear dynamic change of dwell time is modeled by the formula:

[0038]

[0039] Among them, T stay (t) represents the nonlinear dwelling time of visitors in front of the exhibits; the integral term Represents the accumulation of behaviors from time t0 to the current time t; is the time decay function, γ is the decay factor, which indicates the weight decay of time on the behavior accumulation;

[0040] And N interact (t′) is the number of interactions, λ is the weight of the position change P(t′,x,y), which represents the influence of the visitor’s position fine-tuning and action frequency on the dwell time in front of the exhibit; δ is a nonlinear adjustment parameter that identifies the mutation or acceleration point in the dwell time.

[0041] Furthermore, the method for collecting the behavior data of each visitor includes: inferring the potential points of interest of the visitor by analyzing the movement path, turning frequency and exhibit switching behavior of the visitor in the exhibition hall, and actively recommending relevant exhibit content to the visitor; the movement path P(t,x,y) of the visitor is used to infer its potential points of interest through the path association integral equation:

[0042]

[0043] Among them, R(t) is the interest function of the visitors in the exhibition hall; the integral term ∫ Ω represents the accumulation of behaviors within the audience activity area Ω; is the divergence of the path, which indicates the frequency of direction changes or turns of visitors in the exhibition hall, capturing the regularity of visitors switching between different exhibits;

[0044] Among them, T stay (t) is the dwell time, which indicates the audience’s interest in a specific exhibit, and κ is the weight coefficient of dwell time, which regulates the effect of dwell time on interest.

[0045] The interactive projection optimization method of the exhibition hall of the present invention has the following beneficial effects:

[0046] By capturing the audience's behavior, movements, and gaze direction in real time, the present invention can dynamically adjust the projected content and achieve accurate interactive responses. This personalized content display method enhances the audience's sense of immersion, allowing them to interact with the exhibits more naturally during the exhibition, thereby improving the overall visiting experience. Based on the collection and analysis of audience behavior data, the present invention can intelligently predict the audience's points of interest based on factors such as dwell time, interaction frequency, and movement trajectory, and actively adjust the displayed content. The system can push content related to each audience member's interests in real time, ensuring that the projected content always meets the audience's needs and expectations, thereby increasing the attractiveness and participation of the exhibition.

[0047] Through spatial perception and lighting condition analysis, the present invention can automatically optimize the projection content under different environmental conditions to ensure the clarity and consistency of the projection effect. This flexible projection adaptability enables the display content to maintain high-quality visual presentation at any angle and lighting conditions, thereby enhancing the visual effect of the exhibition. By introducing deep learning and nonlinear dynamic modeling, through a detailed analysis of audience behavior, the system can identify the audience's emotional state (such as curiosity, concentration, fatigue) and adjust the projection content in a timely manner. This efficient data feedback mechanism not only improves the flexibility of the projection content, but also effectively enhances the audience's concentration and sense of participation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The figure is a flow chart of the interactive projection optimization method of the exhibition hall of the present invention.

[0049] Figure 2 The present invention is a flow chart of a construction method for dynamically adjusting projection content.

[0050] Figure 3 This is a flow chart of the method for collecting behavioral data of each audience member in the present invention. DETAILED DESCRIPTION

[0051] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0052] The initial optimization of the projection S1 for ambient light and spatial perception is achieved through sensors and spatial modeling technology:

[0053] Before starting the projection, the system will first use sensors (such as ambient light sensors or photosensors) to sense the ambient light in the exhibition hall in real time and obtain the lighting conditions of different areas. This step is very important because the ambient lighting in the exhibition hall is usually uneven and is interfered by external light sources, such as changes in natural light outside the window or adjustments to the lights in the exhibition hall. Therefore, the system must evaluate the lighting conditions in each area in real time to ensure that the projected content can be presented to the audience with the best effect under different lighting conditions.

[0054] The internal layout of the exhibition hall is accurately modeled using spatial modeling technology. The modeling process not only includes the detection of the exhibition hall walls, but also identifies other objects or structures in the exhibition hall that affect the projection, such as exhibits, furniture, decorations, etc. These objects block the projected light or affect the integrity of the projection area. Through three-dimensional spatial modeling, the system is able to understand the complex layout of the exhibition hall and plan the optimal projection angle and area for the projector. In addition, spatial modeling can also help identify obstructions or interferences in the projection area to ensure that the projected content is not covered by unnecessary shadows or objects.

[0055] Dynamic adjustment of projection content based on audience behavior S2 is achieved through cameras or infrared sensing devices:

[0056] These devices can capture the audience's gestures, movements, and gaze direction in real time. First, the camera will monitor the audience in the exhibition hall throughout the whole process, capturing body movements, gestures, and the gaze direction of the head or eyes; while the infrared sensing device can detect the audience's position changes and specific movements. These sensing devices work together to enable the system to accurately identify the audience's behavioral status. Based on these behavioral data, the system can determine the audience's current attention to the exhibits and their interactive needs. Then, the system will dynamically adjust the projection content according to the audience's movements and positions.

[0057] For example, when the system recognizes that the audience is approaching an exhibit, it will immediately adjust the projection content, automatically amplify the relevant information of the exhibit or switch to a more detailed display mode, so that the audience can view and understand the details of the exhibit more clearly. If the audience's gesture points to a specific content area, the system can also provide additional information related to the area or trigger specific interactive functions based on the gesture. The entire process is based on real-time changes in the audience's behavior, and the system will continue to adjust the projection content to ensure that each audience's interactive experience can be personalized and optimized.

[0058] Intelligent content generation and audience preference analysis S3 collects and analyzes the behavioral data of each audience to achieve:

[0059] First, the system will monitor and record in real time how long the audience stays in front of the exhibits and how many times they interact with the exhibits. These data can reflect the audience's interest in a certain exhibit. For example, the longer the stay time or the more times they interact, it usually means that the audience has shown a greater interest in the exhibit. Based on these behavioral data, the system will use machine learning models for analysis. Machine learning models can identify audience interest patterns and behavioral habits through training on a large amount of audience behavior data. It can predict the audience's points of interest through algorithms, that is, determine which exhibits or display content the audience will be interested in. On this basis, the system can recommend projection content related to each audience's interests and dynamically adjust the exhibition display method. For example, if the system finds that a certain audience's stay time and interaction frequency for historical exhibits are significantly higher than those for other categories of exhibits, then when the audience is close to other similar exhibits, the system can give priority to pushing historical-related projection content to him, providing more detailed background information or interaction options.

[0060] Multi-projection collaborative optimization and edge fusion S4 achieves seamless projection effects by coordinating the parameters and content display of multiple projection devices:

[0061] First, according to the actual projection area and display content of each projection device, the system will make personalized adjustments to the brightness, color and other parameters of each projector. Due to differences in location, ambient lighting and projection area, different projection devices will lead to inconsistent display effects, such as too high or too low brightness, color distortion, etc. Therefore, the system will automatically detect the current status of each projector and adjust the brightness and color parameters of these devices as needed, so that multiple projection areas can maintain consistency and visual harmony.

[0062] Next, image processing technology is used to seamlessly blend the edge areas of each projection device. Since multiple projection devices are used in the exhibition hall to jointly display a large area of ​​projection content, the image stitching of the edge areas often has visible dividing lines or color inconsistencies. Through image processing technology, the system can adjust and modify the image edges of each projection device to ensure that there will be no obvious seams and color differences in these edge areas when spliced. This edge fusion technology uses smoothing and edge color correction to make the content displayed by different projection devices visually appear as a complete and coherent picture. The audience in the exhibition hall will not notice that the projection content comes from different projection devices, thereby improving the overall quality and immersion of the projection display.

[0063] Embodiment 1:

[0064] In this exhibition hall, the designer hopes to enhance the audience's understanding and interactive experience of historical relics through intelligent projection content adjustment. The museum's exhibits include several important historical relics booths, which display ancient sculptures, calligraphy works and restored models of battle sites. RGB-D cameras, LiDAR, infrared sensors and ultrasonic distance sensors are arranged in front of each exhibit. The main task of these devices is to capture the audience's spatial position, movement and distance information, and input the data into the system for fusion analysis.

[0065] Assuming a visitor approaches an ancient sculpture exhibition stand, the system will fuse the data from these sensors to construct a real-time position perception matrix M(t,x,y,z), where t represents time and (x,y,z) is the visitor's three-dimensional coordinates in the exhibition hall. At this time, the camera captures the visitor's body posture and head direction, LiDAR detects the distance between the visitor and the exhibit in real time, the infrared sensor senses the ambient temperature change and the thermal signal of the movement, and the ultrasonic sensor accurately monitors the distance change.

[0066] Use the following integral formula to predict the audience's movement trajectory and behavior changes:

[0067]

[0068] In this formula, M(t,x,y,z) is the function of the change in the position of the visitor in the exhibition hall. Indicates the audience's moving speed, is the change in acceleration. Through integration, the system can make cumulative observations of the audience's movement behavior at time t to help determine the audience's behavior pattern and subsequent action trajectory. Here, the parameter λ is the acceleration influence coefficient and can be between 0.1 and 0.5 to balance the impact of speed and acceleration on system judgment. The specific value can be determined based on the spatial layout of the exhibition hall, the average movement speed of the audience, and the required interaction sensitivity.

[0069] At a certain moment t = 10 seconds, the audience moves from 3 meters away from the sculpture booth to within 1 meter. The system determines that the audience is gradually approaching the exhibit by calculating the predicted value of the movement behavior S(t). At the same time, the distance function D(t) between the audience and the exhibit is calculated. When D(t) is less than the set distance threshold D0, such as 1 meter, the system will trigger the judgment of entering the interactive zone.

[0070] For example, let the initial position M(0,x,y,z) be (3,0,1). Through sensor monitoring, the system observes M(10,x,y,z)=(1,0,1) at t=10 seconds. At this time, the system calculates the speed of the audience as If the acceleration 0.05 m / s 2, and take λ = 0.3, then the motion behavior prediction value S(t) is:

[0071]

[0072] By calculating the motion behavior prediction value S(10)=2.15, the system determines that the audience is about to enter the interactive area and begins to prepare to dynamically adjust the projection content. At this time, the system detects that the distance between the audience and the sculpture exhibit is D(10)=1 meter, which is less than the set interactive distance threshold I0=1 meter, and the system will immediately start the interactive content display.

[0073] In the interactive area, the projector can automatically enlarge the projected content or switch to a more detailed display mode according to the audience's proximity and viewing angle. For example, when the audience approaches the sculpture, the projection content will switch from a simple historical description to more detailed content such as the story behind the sculpture, the display of carving technology, and the 3D restoration model.

[0074] It is assumed that the audience has entered the interactive area of ​​the sculpture booth, and the system has triggered dynamic adjustment of the projection content based on their position and behavior. At this stage, the system will adaptively optimize the projection content according to the audience's gaze angle and action to enhance the audience's immersion and interactive experience.

[0075] When the audience approaches the sculpture and looks at a specific area, the system uses the RGB-D camera to capture the audience's head direction and sight angle in real time. This information will be used to calculate the audience's gaze angles θ and φ, where θ is the horizontal gaze angle and φ is the vertical gaze angle. At this time, the projected content will adjust the display effect according to the audience's gaze direction and distance to ensure that the content area that the audience is paying attention to presents the best brightness and focus. This dynamic optimization is achieved by the following formula:

[0076]

[0077] Let's start by explaining the first part of the formula This part integrates the second-order derivative of the brightness I of the projected content with respect to the angles θ and φ to ensure that the projected content can be smoothly adjusted when the audience's gaze angle changes.

[0078] Assume that the audience's horizontal gaze angle θ = 30° and the vertical gaze angle φ = 15°, and the system will calculate the current brightness distribution based on these angles. If the audience turns their head from the front to the side (for example, θ increases from 30° to 45°), the system will calculate the current brightness distribution based on the second-order derivative. Automatically adjust the brightness to ensure that the focus of the projected content remains clear and visible at different viewing angles. Set the brightness change rate to 0.02, and the system will dynamically adjust the brightness according to the audience's angle changes to maintain high-quality projection effects.

[0079] Next, let’s analyze the second part of the formula. Here, κ is a control coefficient, which is used to control the response sensitivity of content changes. Setting κ = 0.5 is used to adjust the amplitude of content changes so that the system does not react too drastically or too slowly to the audience's behavior.

[0080] The function g(p,t) represents the change function of the projection content at time t, indicating the dynamic adjustment of the projection content over time, such as the gradual enlargement of the projection image or the gradual appearance of content details. Assume that g(p,t) gradually changes from 0 to 1 (indicating that the details change from blurry to clear) as time t increases, and p0 is the baseline projection parameter, defined as the initial static content. When the audience enters the interactive area, the system will gradually enlarge the sculpture's detail area and provide additional information according to the viewing angle. The adjustment time process is represented by e -βt Controls , where β is a time decay factor that controls the smoothness of content adjustments.

[0081] The β value range is set to 0.05 to 0.3, depending on the complexity of the projected content and the rhythm of the display. For example, if β = 0.1, the system will gradually amplify the relevant information in a relatively smooth manner while the audience is looking at the sculpture, and maintain the coherence and natural transition of the content.

[0082] To further specify this process, it is assumed that when the audience is looking at the details of the sculpture's head, the system recognizes the change in the direction of the audience's head, and the current interaction has lasted for 10 seconds. At this point, the projection system calculates the impact of the audience's gaze behavior on the brightness and detail display according to the formula. The audience's gaze angle gradually changes from θ = 30° and φ = 15° to θ = 45°, which results in an adjustment rate of 0.02 for brightness and focus. At the same time, the system gradually amplifies the historical background information and detail display of the sculpture through g(p, t) for 10 seconds. In this process, the time decay factor β = 0.1 ensures that the dynamic changes of the projection content are smooth and natural. The audience will not feel abrupt adjustments, but will experience a gradual visual transition, allowing them to focus on the projection content and enjoy the interactive experience.

[0083] Next, in order to further enhance the interactive experience, the system will use a deep learning-based convolutional neural network to analyze the audience's gestures, posture, body movements and other behaviors in real time and automatically adjust the projection content.

[0084] When the audience is looking at the sculpture, they will make gestures to point to a certain part of the sculpture, or change their posture, such as standing to approach the exhibit or raising their hands to touch the virtual projection area. At this time, the system extracts the audience's action feature vector V through the convolutional neural network m(t), the feature vector includes the audience's gestures, postures and body movements, reflecting the audience's body dynamics at different time points. This vector will be used to analyze the impact of the audience's specific actions on the projected content.

[0085] For example, if the audience raises their hand and points to the head of the sculpture, the system first extracts the audience's gesture and arm movement information V m (t), and combined with the weight function W of the gaze area r (t) to calculate the impact of the action on the projected content. The weight function W r (t) defines the priority of the exhibit area that the audience is looking at. If the audience's gaze is concentrated on the head area of ​​the sculpture, the system will give this area a higher weight value, such as W r (t)=0.8, while the weights for other regions are lower, such as W r (t) = 0.2. Based on this analysis, the system will calculate the correlation between action features and gaze weights using the following formula:

[0086]

[0087] In this formula, V m W (t) is the action feature vector of the audience at time t, reflecting information such as gestures and body movements. r (t) is the audience's gaze weight on the sculpture head area, indicating the importance of this area to the adjustment of the projection content. Represents the divergence of the feature vector field, which is used to measure the correlation between action and gaze. Through this divergence term, the system can calculate how the audience's gestures and their gaze direction synergistically affect the projected content. Set the audience's gesture frequency to 0.1 m / s and the gaze weight W r (t) is 0.8, and the system calculates the divergence value based on these parameters Indicates a strong correlation between gesture and gaze area.

[0088] In addition, α in the formula is the response coefficient of motion adjustment, which indicates the system's sensitivity to the audience's motion. α can be set to 0.5, at which point the system's response to motion is more balanced, neither too sensitive nor ignoring the audience's tiny motions. Indicates the rate of change of the audience's action. Set the raising speed of the audience's arm to 0.05 m / s. The system will analyze the action in real time and calculate its impact on the adjustment of the projected content. The time decay factor γ controls the long-term effect of the action on the content adjustment. You can set γ = 0.1, which means that the system will gradually reduce the weight of the earlier actions and give priority to responding to the current gestures and actions.

[0089] The audience's movements and gaze behaviors are assumed to last for 5 seconds, during which the system gradually accumulates the H(t) value calculated by the high-order tensor integral. If H(5) reaches the threshold at the 5th second, for example, H(5) = 0.5, the system will trigger a change in the projected content. At this time, the projected content will automatically zoom in on the sculpture head area where the audience is looking at, and provide more detailed displays and background information. This dynamic change not only responds to the audience's gestures, but also ensures the core influence of the gaze direction on the content adjustment through the weight function.

[0090] In a specific example, when the audience points to the head of the sculpture, the system calculates the feature vector of the action and the weight of the gaze area through the above formula, and then obtains the cumulative value of H(t). When this value exceeds the preset threshold, the system will automatically enlarge the head of the sculpture to 1.5 times its original size, while displaying the historical background information behind the sculpture and providing virtual reconstructed images related to the sculpture to help the audience understand the exhibits more deeply. Due to the time decay factor γ = 0.1, the system can gradually adjust the impact on the audience's early actions within a certain period of time to ensure that subsequent actions and gazes can obtain timely feedback.

[0091] Embodiment 2:

[0092] In the exhibition hall of a historical museum, the system collects the behavioral data of each visitor and implements personalized projection optimization to enhance the interactive experience. The audience is set to stop in front of the sculpture booth and interact with the exhibits multiple times. At this time, the system captures the audience's stay time, behavior pattern, position change, and interaction frequency through the sensor network, analyzes the audience's emotional state, such as curiosity, fatigue, or concentration, and dynamically adjusts the content of the exhibits.

[0093] The system models the collected data into a three-dimensional time, position and interaction matrix M(t,x,y), where t is time and (x,y) is the plane coordinate of the visitor in the exhibition hall, reflecting the changes in their behavior at different locations. In order to perform a hierarchical analysis of the visitor's stop behavior, the system uses the following formula:

[0094]

[0095] In this formula, S(t) represents the cumulative value of the visitor's stop behavior at time t. The system uses this value to judge the visitor's interest in the exhibit. First, It indicates the time rate of change of the visitor's position (x, y) in the exhibition hall plane, reflecting the visitor's moving speed. For example, assume that the visitor moves from the left side of the booth (x=2, y=1) to the front (x=0, y=1), the distance is 2 meters, and the movement occurs within 10 seconds, then the visitor's speed is:

[0096]

[0097] Next, the system will also calculate the audience's interaction frequency N interact (t′) is used to capture the number of interactions with the exhibit. Assuming that the audience interacts with the exhibit three times in the same time period (such as clicking on the screen to view the sculpture details or triggering information display), then N interact (t′) = 3. The system will adjust the influence of interaction on behavior analysis according to the interaction weight coefficient α.

[0098] Set α = 0.5, so that the system can reasonably pay attention to the audience's stay time and will not over-adjust the display content due to accidental frequent interactions.

[0099] By substituting these data into the formula, the cumulative value of the audience's stop behavior can be calculated. Assuming that the audience enters the booth from the starting point t0 = 0 seconds to the end of observation t = 30 seconds, the system will calculate their stop behavior:

[0100] s(30)=∫0 30 (0.2+0.5×3)dt′=30×(0.2+1.5)=51

[0101] Therefore, within 30 seconds, the cumulative value of the audience's stop behavior S(30) = 51. The system uses this value to judge the audience's high interest in the exhibits, and believes that the audience's stop time and number of interactions at the sculpture booth are relatively long, indicating that they have a high degree of in-depth interest in the exhibits. Based on this analysis, the system will make personalized adjustments to the projection content, add more historical background information related to the sculpture, or provide a detailed display of the craftsmanship behind the sculpture to further attract the audience's attention.

[0102] At the same time, the system can also identify the emotional state of the audience. If the cumulative value of stop behavior is high and the interaction frequency remains stable, the system will determine that the audience is in a state of concentration or curiosity, and then push more detailed content.

[0103] In the previous steps, the system has gradually established the audience's interest model by capturing the audience's behavior patterns in real time (such as location changes, dwell time, and interaction frequency). Next, the system needs to further optimize the display effect of the projection content based on the dynamic two-way coupling relationship between the audience and the exhibits to enhance interactivity.

[0104] At this point, the coupling relationship between audience behavior and exhibit content is modeled through a formula, and the system will analyze the audience's reaction to the exhibit content in real time and then adjust the projection method. The formula is as follows:

[0105]

[0106] In this formula, C(t) represents the coupling function between audience behavior and exhibit content, which is used to measure the dynamic impact of exhibit content on audience behavior. First, It is the second-order derivative of the change of the exhibit content over time, indicating the acceleration of the dynamic change of the exhibit content. The sculpture projection content in the exhibition hall is initially static. As the audience approaches and gradually interacts, the projection content begins to change, for example, switching from a simple sculpture outline to the historical background behind it. In this process, the speed of change of the projection content increases over time. The acceleration of the content is set to 0.03 seconds -2 , and as time t increases, the display content transitions from static to dynamic.

[0107] Next, N interact (t) represents the number of interactions between the audience and the exhibits, reflecting the actual interaction between the audience and the exhibits. Assume that the audience has performed 5 interactions in the past 30 seconds, including clicking on the screen to learn about the background information of the sculpture and triggering the rotation display of the 3D model. At this time, N interact (t) = 5. The weight coefficient β is used to adjust the impact of the number of interactions on the change of exhibit content. β is set to 0.7 to indicate the importance of the number of interactions on the change of content. At this point, the system can judge the depth of the audience's interest based on the frequency of interaction and adjust the display content in a timely manner.

[0108] Assuming that the visitor stayed in front of the booth for 45 seconds, the system performed coupled calculations on their behavior. Taking into account the acceleration of the transition from static to dynamic exhibits and the frequent interactive behaviors of the visitor, the system substituted the above data into the formula for calculation:

[0109] c(45)=∫ V (0.03+0.7·5)dV

[0110] at this time, Indicates that the rate of change of the exhibit content is relatively stable; while N interact (t) = 5 indicates high frequency interactive behavior of the audience. The system substitutes these values ​​into the formula to calculate:

[0111] c(45)=∫ V (0.03+3.5)dV=3.53×V

[0112] The calculation result C(45) indicates that the coupling strength between the audience behavior and the exhibit content is 3.53×V, that is, the high-frequency interaction of the audience significantly enhances the dynamic change effect of the exhibit content. This shows that the system analyzes the audience's behavioral response and believes that the dynamic display of the projection content should be further enhanced. For example, the production process of the sculpture, the restoration process or its important role in history can be further demonstrated.

[0113] To make this process more realistic, the system dynamically adjusts the details of the projected content. As the coupling strength increases, the system not only enlarges the 3D model of the sculpture, but also begins to show deeper details, such as the tiny textures on the surface of the sculpture, the historical usage scenarios, and the background information of the relevant characters. The audience can continue to interact with the projected content through gestures or clicks, and these new interactions will further deepen the depth of the content.

[0114] In front of the museum's sculpture booth, the system monitors the behavior of visitors in real time through sensors and collects data on their dwell time, position changes, and number of interactions. In order to further accurately analyze the changes in the audience's interests, the system needs to analyze the audience's dwell time patterns through nonlinear dynamic modeling. Through this modeling, the system can infer the audience's interest change points and dynamically adjust the depth or complexity of the exhibits based on these change points.

[0115] Assume that the visitor stayed in front of the sculpture booth for a period of time and interacted with it many times, such as clicking to trigger the 3D rotating model and viewing background information. The system will use the following formula to perform nonlinear dynamic modeling of the dwell time:

[0116]

[0117] First, T stay (t) represents the nonlinear dwell time of visitors in front of the exhibits. The system calculates this value by monitoring the changes in the behavior of visitors at different times. represents the accumulation of all behavioral data from the time t0 when the visitor arrives at the booth to the current time t. In this process, the time decay function It is used to reduce the impact of earlier behaviors on current results, ensuring that the system pays more attention to the audience's recent behaviors. The attenuation factor γ is set to 0.05, indicating that the behavior will slowly weaken over time.

[0118] Next, N interact (t′) is the number of interactions of the audience at a certain moment, recording each interactive behavior of the audience such as clicks or gestures. It is assumed that the audience has some preliminary interactions after time t0 and completed 5 interactions in the next 20 seconds, including zooming in on the 3D model and viewing the historical details behind the sculpture. Set N interact (t′)=5 times.

[0119] In addition, λ is the weight coefficient of position change, which is used to adjust the impact of slight position changes of visitors in front of the exhibits on the dwell time. Assuming that the visitor moves slightly forward during the stay, from 1.5 meters to 1 meter away from the booth, the position change P(t′,x,y) is calculated as 0.5 meters, and λ is taken as 0.3, which means that the system is less sensitive to position changes, but still takes into account the proximity of the visitor to the exhibits.

[0120] Finally, δ is a nonlinear adjustment parameter used to identify mutations or acceleration points in the audience's dwell time. δ = 0.1 is set so that the system can capture subtle changes in audience behavior patterns, especially when audience behavior changes suddenly (such as from simple gaze to frequent interaction). Through this parameter, the system can identify the acceleration points of audience interest and adjust the complexity of the displayed content.

[0121] Assuming that the visitor stays in front of the booth for 30 seconds, the system calculates the nonlinear accumulation of their stop behavior through the formula:

[0122]

[0123] First, calculate the decay function e -0.05×30 =e -1.5 ≈0.2231, then N interact (t′) = 5 and position change P(t′,x,y) = 0.5, so:

[0124]

[0125] The calculation results in:

[0126]

[0127] Therefore, the nonlinear dwelling time T of the visitors in front of the exhibits stay (30) is 8.61. Based on this value, the system can infer that the audience's interest level is high, especially when they interact frequently and are close to the exhibits. Based on this result, the system will adjust the depth and complexity of the projected content in a timely manner. For example, the system can further display the historical background and important details of the sculpture, and add extended content related to the sculpture, such as a multimedia demonstration of the carving process, or provide an interactive restoration display.

[0128] In order to accurately predict the potential points of interest of visitors and proactively recommend relevant exhibits, the system analyzes the visitors' movement paths, turning frequencies, and exhibit switching behaviors to determine the exhibit areas they are interested in. The system calculates the visitor's interest function R(t) in the exhibition hall through the path association integral equation.

[0129] Assume that the visitor stops in front of the sculpture booth for 30 seconds and interacts with the projection frequently, then leaves the booth and moves to the next exhibit (such as a restored model of a battle site). The visitor's motion path P(t,x,y) represents the movement trajectory from a certain coordinate point (x1,y1) in the exhibition hall to the next exhibit point (x2,y2). The following path association integral equation is used to infer the visitor's potential points of interest:

[0130]

[0131] In this formula, R(t) is the audience's interest function, which represents the audience's potential interest in different exhibits in the exhibition hall. Ω represents the cumulative behavior of the audience in the activity area Ω. First, is the divergence of the path, which reflects the frequency of direction changes or turns of visitors in the exhibition hall. When visitors frequently turn or change routes between different exhibits, the path divergence increases, indicating that the visitors are looking for new points of interest. Assuming that the visitors move from the sculpture booth (x1=2, y1=1) to the restoration model booth (x2=5, y2=3), the visitors' path is a straight line trajectory, and the turning frequency is low, which indicates that the visitors have identified the next exhibit of interest, so the path divergence is small.

[0132] Next, T sta y(t) is the time that visitors spend in front of a specific exhibit. The time that visitors spend in front of the sculpture booth is set to 30 seconds, which indicates that the visitors have a greater interest in the exhibit. κ is the weight coefficient of the dwell time, which is used to adjust the impact of the dwell time on the interest. κ can be set to 0.6, indicating that the dwell time has a greater weight in predicting the visitor's interest.

[0133] Substituting these data into the formula, the integral area Ω is set to be the exhibition hall area where the audience is active, and the path from the sculpture stand to the restored model is recorded. Path divergence It can be calculated by the audience's turning rate of change. It is assumed that the audience's direction changes little on this path, so the divergence At the same time, the audience's stay time at the sculpture booth is 30 seconds, and κ = 0.6. Substitute these values ​​into the formula to calculate the audience's interest function:

[0134]

[0135] Assuming the exhibition area in the integral area Ω is 50 square meters, the integral result is:

[0136] R(30)=(0.05+18)×50=18.05×50=902.5

[0137] Therefore, the audience's interest function within 30 seconds is R(30) = 902.5, which reflects the audience's accumulated interest in the current activity area. Based on this interest, the system infers that the audience is highly interested in the sculpture booth and judges that they are interested in similar exhibits. Therefore, when the audience approaches the next exhibit, the battle site restoration model, the system will actively recommend other exhibits related to the sculpture, such as other historical restoration models, historical relics related to the sculpture, etc. As the audience continues to move in the exhibition hall, the system continuously analyzes the audience's path divergence and dwell time, and timely infers the audience's potential points of interest. For example, if the audience stays in front of the next booth for a long time, the system will adjust the recommendation method of the projection content and provide more background information and interactive options for related exhibits.

[0138] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. The interactive projection optimization method of the exhibition hall is characterized by The following steps are involved: S1. Initial optimization of projection for ambient light and spatial perception: S1.

1. Before starting the projection, use sensors to sense changes in ambient light in the exhibition hall and evaluate lighting conditions in different areas; S1.

2. Use spatial modeling technology to obtain the spatial layout of the exhibition hall and identify the wall and object occlusion information in the projection area; S2. Dynamic adjustment of projection content based on audience behavior perception: S2.

1. Capture the audience's gestures, movements and gaze direction in real time through cameras or infrared sensing devices; S2.2, dynamically adjust the projection content according to the audience's movements and positions; including zooming in on the relevant projection content or switching to detailed display mode when the audience approaches a certain exhibit; S3. Intelligent content generation and audience preference analysis: S3.

1. Collect behavioral data of each visitor, including the duration of stay in front of the exhibits and the number of interactions; S3.2, predicting the audience's interests through machine learning models, and recommending projection content related to the audience's interests; S4, Multi-projection collaborative optimization and edge fusion: S4.

1. Adjust the brightness and color parameters of each projector according to the actual projection area and content of each projection device; S4.2, using image processing technology to fuse edge areas of multiple projection devices; The method for constructing the dynamically adjusted projection content includes: arranging RGB-D cameras, LiDAR, infrared sensors and ultrasonic distance sensors in the exhibition hall to capture the spatial position and movement of the audience; constructing the audience's real-time position perception matrix M(t,x,y,z) by fusing sensor data, where t represents time and (x,y,z) is the audience's three-dimensional coordinates; predicting the audience's movement trajectory and behavior changes, using integral calculation: Among them, M(t,x,y,z) represents the position change function of the audience in time t; is the speed at which the viewer's position changes over time, is the change in the acceleration of the audience; the integral limit 0 to t represents the accumulated observation of the audience's behavior in the entire time range; the parameter λ is the acceleration influence coefficient, which is used to balance the influence of speed and acceleration on the system judgment; the result S(t) is the predicted value of the audience's movement behavior at time t, and it is used to judge in real time whether the audience enters the interactive area of ​​a certain exhibit; when the distance D(t) between the audience and the exhibit is less than the predetermined distance D0, it will be judged that the audience enters the interactive area, triggering content adjustment.

2. The interactive projection optimization method of the exhibition hall according to claim 1, characterized in that When the audience enters the interactive area, the projected content is adjusted based on the audience's position and gaze direction; the adaptive optimization of the projected content is achieved through the following equation, which dynamically optimizes the display effect according to the audience's behavior, including content magnification, focus adjustment, and brightness and angle correction: Where I(θ,φ,p,t) is the brightness or focus distribution function of the projected content at the audience's gaze angles θ and φ; and They represent the second-order derivatives of the projection content relative to the audience’s gaze angles θ and φ, respectively, indicating the changes in the projection content in different directions; the integral region V is the spatial volume of the projection; In the second term, κ is the control coefficient, which controls the response sensitivity of content changes, g(p,t) represents the function of the change of projection content with time t, p0 is the reference projection parameter, which is defined as the static reference content; the time decay factor β controls the smoothness of the content adjustment process.

3. The interactive projection optimization method of the exhibition hall according to claim 1, characterized in that The method for dynamically adjusting the projection content includes: based on a deep learning convolutional neural network, analyzing the audience's movements, gestures and expressions in real time, extracting the movement feature vector V m (t); the feature vector includes the audience's gestures, posture, and body movement information; the high-order tensor integral combines the audience's action characteristics and gaze area to dynamically calculate the impact of the action on the projected content, as follows: Among them, V m (t) is the action feature vector of the audience at time t, reflecting the characteristics of the audience’s gestures and body movements; W r (t) is a weight function, which defines the audience's gaze weights on different exhibits and determines the degree of influence of the audience's gaze direction on the adjustment of the projection content; is the divergence of the vector field, which is used to measure the correlation between action and gaze; The integral area Ω represents the audience's activity space, and the time decay factor γ controls the long-term effect of action adjustment on content; α is the response coefficient of action adjustment, which adjusts the sensitivity of the response to the audience's action; when H(t) exceeds the preset threshold, the change of the projected content is triggered, including enlarged display or switching to detailed display mode.

4. The interactive projection optimization method of the exhibition hall according to claim 1, characterized in that The method for collecting the behavioral data of each visitor includes: capturing the visitor's stay time and refining the visitor's behavioral patterns, including changes in position and frequency of interaction, through dynamic layered analysis of the collected data; identifying the visitor's emotional states, including curiosity, fatigue, and concentration, and the deep interest in the exhibits, and then making personalized adjustments to the exhibit content.

5. The interactive projection optimization method of the exhibition hall according to claim 4, characterized in that The method for collecting the behavior data of each audience member includes: establishing a dynamic two-way coupling relationship between the audience's behavior and the exhibit content; and analyzing the audience's reaction to the exhibit content in real time, thereby adjusting the display method.

6. The interactive projection optimization method of the exhibition hall according to claim 5, characterized in that The method for collecting the behavior data of each visitor includes: analyzing the change pattern of dwell time through nonlinear dynamic modeling, inferring the change point of the visitor's interest, and adjusting the depth or complexity of the exhibit display in real time; the nonlinear dynamic change of dwell time is modeled by calculation: Among them, T stay (t) represents the nonlinear dwelling time of visitors in front of the exhibits; the integral term Represents the accumulation of behaviors from time t0 to the current time t; is the time decay function, γ is the decay factor, which indicates the weight decay of time on the behavior accumulation; And N interact (t′) is the number of interactions, λ is the weight of the position change P(t′,x,y), which represents the influence of the visitor’s position fine-tuning and action frequency on the dwell time in front of the exhibit; δ is a nonlinear adjustment parameter that identifies the mutation or acceleration point in the dwell time.

7. The interactive projection optimization method of the exhibition hall according to claim 6, characterized in that The method for collecting the behavior data of each visitor includes: inferring the potential interests of the visitor by analyzing the movement path, turning frequency and exhibit switching behavior of the visitor in the exhibition hall, and actively recommending relevant exhibit content to the visitor.

Images