Environment-aware intelligent lighting control system

Abstract

The invention relates to an intelligent lighting control system with environmental perception, and aims to enhance the visual expressive force of artworks and improve the interactive experience of audiences by accurately controlling the lighting condition in an exhibition hall. The system comprises a system environment data acquisition unit, an intelligent control unit, a self-adaptive illumination execution unit and an interactive illumination adjustment unit. The environmental data acquisition unit is used for comprehensively monitoring environmental conditions in the exhibition hall. The intelligent control unit generates an optimized lighting strategy according to the collected data and artwork protection and display requirements. The self-adaptive illumination execution unit comprises a plurality of LED lamp sources capable of being independently adjusted, and light intensity, color temperature and light beam angles are adjusted according to the strategy of the control unit. And the interactive illumination adjusting unit is used for further optimizing the illumination setting. According to the system, the modern sensing technology and intelligent control are comprehensively utilized, an innovative solution is provided for art museums and museums, human cultural heritage is protected and displayed through science and technology, and meanwhile the sense of participation and educational value of the public are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of lighting technology, and in particular to an environment-aware intelligent lighting control system. Background Art

[0002] In the existing technology, lighting control systems for art galleries and museums generally rely on traditional lighting adjustment methods, such as manual adjustment or the use of basic automation systems. These systems often include standard lighting equipment and simple sensors to adjust lighting intensity and color temperature to protect and display artworks. However, these systems usually do not take into account specific environmental changes in the exhibition hall or the behavior and preferences of the audience, limiting the adaptability and interactivity of the lighting system.

[0003] The main problem with existing technologies is that they lack the ability to respond to environmental diversity and integrate audience interaction. Existing lighting systems cannot monitor multiple key environmental parameters in real time, such as changes in lighting in specific areas, which may cause damage to artworks or affect the exhibition experience. In addition, traditional systems usually do not support dynamic lighting adjustments based on the actual behavior of the audience, which limits the personalization and optimization potential of the exhibition.

[0004] Therefore, it is urgent to develop a new type of environment-aware intelligent lighting control system. Summary of the invention

[0005] The present application provides an environment-aware intelligent lighting control system to enhance the visual expression of artworks and improve the audience's interactive experience.

[0006] The present application provides an environment-aware intelligent lighting control system, comprising:

[0007] Environmental data collection unit, including light intensity sensor, used to monitor the light intensity in the exhibition hall; color temperature sensor, used to measure the color temperature in the exhibition hall; sight tracking device, used to monitor the audience's sight and stay time;

[0008] An intelligent control unit, configured to receive the collected data provided by the environmental data collection unit; generate an optimal lighting strategy according to the display requirements of the artworks and the collected data; wherein the optimal lighting strategy is used to enhance the visual expression of the artworks in the exhibition hall and the interactive experience of the audience;

[0009] An adaptive lighting execution unit, comprising a plurality of LED light sources, wherein each LED light source can independently adjust setting parameters including light intensity, color temperature and beam angle; the adaptive lighting execution unit is used to adjust the setting parameters of the plurality of LED light sources according to the optimal lighting strategy provided by the intelligent control unit;

[0010] Interactive lighting adjustment unit, with integrated touch screen and voice recognition module, is used to receive audience input on lighting preferences; adjust the setting parameters of multiple LED light sources based on audience feedback.

[0011] Furthermore, the environmental data collection unit also includes:

[0012] Infrared thermal imaging sensor, used to detect the surface temperature distribution of artworks;

[0013] 3D depth camera, used to capture the movement trajectory and residence position of visitors in the exhibition hall;

[0014] The data collected by the infrared thermal imaging sensor and the three-dimensional depth camera are used to optimize the lighting strategy to prevent local overheating and improve the audience's interactive experience.

[0015] Furthermore, the eye tracking device includes:

[0016] Multiple high-speed cameras located in different locations in the exhibition hall;

[0017] A real-time image processing module that fuses data from multiple high-speed cameras to build a heat map of audience sightlines;

[0018] The sight heat map is used to identify the level of attention paid to artworks, and dynamically adjust the lighting parameters of the corresponding areas accordingly.

[0019] Furthermore, each LED light source of the adaptive lighting execution unit includes an automatic color temperature adjuster, which is used to automatically adjust the color temperature according to the natural light incident in different time periods in the exhibition hall and the lighting requirements of the artworks, so as to maintain the natural feeling of the lighting and the visual appeal of the artworks.

[0020] Furthermore, the LED light sources of the adaptive lighting execution unit are configured to receive adjustment instructions from the intelligent control unit via wireless signals, and each light source can independently report its energy efficiency and performance status back to the intelligent control unit to optimize energy use and reduce maintenance costs.

[0021] Furthermore, each LED light source of the adaptive lighting execution unit is equipped with a photosensitive element for monitoring the performance of the light emitted by the light source in the actual environment, and the collected real-time feedback data is used to adjust the lighting intensity and color temperature to adapt to the light changes in the exhibition hall.

[0022] Furthermore, the adaptive lighting execution unit includes a lighting balance controller for adjusting the light intensity and color temperature differences between adjacent LED light sources to eliminate uneven lighting in the audience's field of view and improve the overall uniformity of lighting and visual comfort.

[0023] Furthermore, the interactive lighting adjustment unit includes a user interface, which automatically adjusts its operating interface and functional options according to the audience's historical preferences and feedback to provide a personalized interactive experience.

[0024] Furthermore, the intelligent control unit generates an optimal lighting strategy using a pre-trained neural network model according to the display requirements of the artwork and the collected data; wherein the neural network model includes an input layer, a feature enhancement layer, a spatiotemporal processing layer, and an output layer;

[0025] The input of the input layer is the display requirements of the artwork and the collected data; the input layer is implemented using a graph attention network structure, which represents the display requirements of the artwork and the collected data as nodes in the graph, establishes edge connections based on the mutual relationship between the data, and uses a multi-head attention mechanism to calculate the weights between the nodes, and outputs a feature representation that integrates the complex relationship between the data;

[0026] The input of the feature enhancement layer is the feature representation output by the output layer; the feature enhancement layer is implemented by a capsule network structure; the feature enhancement layer uses a dynamic routing algorithm to process low-level features to combine into high-level features, wherein each capsule specifically encodes a specific type of lighting attribute, and the output is a set of feature vectors, each feature vector includes the existence probability of a lighting attribute and its parameters; wherein the lighting attributes include light intensity, color temperature, and beam angle;

[0027] The input layer of the spatiotemporal processing layer is the feature vector group output by the feature enhancement layer; the spatiotemporal processing layer is implemented by a spatiotemporal graph convolutional network structure, the feature vector group output by the feature enhancement layer is constructed into a spatiotemporal graph, the graph convolution operation is used to capture the spatial dependency relationship between features, and the time convolution operation is applied to capture the time series mode, and the output is a comprehensive feature representation of the comprehensive spatiotemporal relationship;

[0028] The input of the output layer is the comprehensive feature representation of the output of the spatiotemporal processing layer; the output layer maps the complex high-dimensional input to a two-dimensional discrete grid through a competitive learning process, wherein each two-dimensional discrete grid node represents a set of specific LED light source setting parameters; the output layer outputs the adjusted optimal light intensity, color temperature and beam angle parameters of each LED light source.

[0029] Furthermore, the spatiotemporal processing layer is implemented by a spatiotemporal graph convolutional network structure, which constructs the feature vector group output by the feature enhancement layer into a spatiotemporal graph, and captures the spatial and temporal dependencies through the following steps to generate a comprehensive feature representation of the comprehensive spatiotemporal relationship:

[0030] The output feature vector group of the feature enhancement layer is constructed as a spatiotemporal graph G = (V, E, A), where V is a node set, representing the feature vectors at different time steps; E is an edge set, representing the spatial and temporal relationship between features; and A is an adjacency matrix.

[0031] On the spatiotemporal graph, the convolution operation provided by the following formula 1 is used to capture the spatial dependencies between features:

[0032]

[0033] Among them, H (l+1) is the node feature matrix of the l+1th layer; H (l) is the node feature matrix of the lth layer, which represents the feature representation of the current layer; A is the adjacency matrix; D is the degree matrix; λ is the adaptive parameter; W (l) is the weight matrix; σ is the activation function; I is the identity matrix;

[0034] According to the following formula 2, multi-scale temporal convolution is applied at each time step to capture the time series pattern of features:

[0035]

[0036] Among them, Z t is the output feature at time step t; K1 represents the number of convolution kernels; τ is the time offset, which represents the time step considered in the time convolution operation; H t-τ is the input feature at time t-τ; W k,τ is the weight of the kth convolution kernel at the time offset τ; ⊙ represents the product operation between elements; b k is the bias term of the kth convolution kernel; τ1 represents the size of the time window;

[0037] Through the following formulas 3 and 4, the feature representations of different time steps are fused through the attention mechanism:

[0038] α t =softmax(q T tanh(Kh t +b)) (3)

[0039]

[0040] Among them, h t is the feature vector at time step t; q and K are learnable parameter matrices; T represents the total number of time steps; b is the bias term; α t is the attention weight at time step t; c represents the comprehensive feature representation of the comprehensive spatiotemporal relationship.

[0041] This application has the following beneficial technical effects:

[0042] (1) Through integrated light intensity and color temperature sensors, the system can monitor and adjust the lighting conditions in the exhibition hall in real time. This sophisticated environmental control helps to extend the shelf life of artworks and ensure their long-term physical and visual quality.

[0043] (2) Gaze tracking devices and interactive lighting adjustment units enable the system to dynamically adjust lighting settings, such as light intensity and color temperature, based on the audience’s behavior and preferences. This not only enhances the visitor’s experience, but also increases the interactivity of the exhibition, allowing the audience to participate more deeply in the exhibition.

[0044] (3) The intelligent control unit calculates the optimal lighting strategy based on the data collected from the environmental data acquisition unit, and can fine-tune the light intensity, color temperature, and beam angle of each LED light source to ensure the best visual display of the artwork. Such lighting optimization is not limited to protecting the artwork, but also aims to enhance the overall visual effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of an environment-aware intelligent lighting control system provided in the first embodiment of the present application. DETAILED DESCRIPTION

[0046] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0047] The first embodiment of the present application provides an environment-aware intelligent lighting control system. Figure 1 , which is a schematic diagram of the first embodiment of the present application. Figure 1 The first embodiment of the present application provides an environment-aware intelligent lighting control system for detailed description.

[0048] The intelligent water resources management system includes an environmental data collection unit 101 , an intelligent control unit 102 , an adaptive lighting execution unit 103 and an interactive lighting adjustment unit 104 .

[0049] The environmental data acquisition unit 101 includes a light intensity sensor for monitoring the light intensity in the exhibition hall; a color temperature sensor for measuring the color temperature in the exhibition hall; a sight tracking device for monitoring the sight and dwell time of the audience; and a temperature and humidity sensor for monitoring the temperature and humidity in the exhibition hall.

[0050] The environmental data acquisition unit 101 is a key component of an environment-aware intelligent lighting control system, which is designed to provide real-time environmental monitoring for exhibition halls in art galleries or museums. The unit contains a variety of sensors, each with a specific function to ensure that the system can collect comprehensive data about the exhibition hall environment.

[0051] First, the light intensity sensor is used to monitor the light intensity in the exhibition hall. This sensor can accurately measure the light intensity in the exhibition hall, expressed in lux. The light intensity sensor is sensitive to changes in light and can detect a wide range of light, from weak ambient light to direct sunlight. The data from this sensor is crucial for adjusting the lighting inside the exhibition hall to ensure that the artworks are not damaged by too strong or too weak light.

[0052] The color temperature sensor is used to measure the color temperature of light sources in the exhibition hall in Kelvin (K). The sensor can evaluate whether the light color emitted by the light source is warm (yellow) or cool (blue) and provide accurate color temperature readings. This is particularly important for adjusting lighting equipment to best display the original colors of artworks, because color temperature directly affects the audience's perception of the color of the artwork and the preservation state of the artwork.

[0053] Gaze tracking equipment is another component of the unit, which monitors visitors’ gazes and the time they spend in front of each artwork. This equipment collects data on which artworks attract more attention by tracking the movement and resting points of visitors’ eyes. This information is very useful for understanding visitors’ interests and preferences and adjusting lighting accordingly to enhance display effects and audience interaction.

[0054] The environmental data acquisition unit 101 transmits the data collected by these sensors to the intelligent control unit 102, which generates and implements the best lighting strategy based on the data, aiming to protect and display artworks through scientific methods while improving the audience's viewing experience. This detailed environmental monitoring and data collection ensures the efficiency and practicality of the lighting control system, making it an important tool in the environmental management of art galleries and museums.

[0055] Furthermore, the environmental data collection unit also includes:

[0056] Infrared thermal imaging sensor, used to detect the surface temperature distribution of artworks;

[0057] 3D depth camera, used to capture the movement trajectory and residence position of visitors in the exhibition hall;

[0058] The data collected by the infrared thermal imaging sensor and the three-dimensional depth camera are used to optimize the lighting strategy to prevent local overheating and improve the audience's interactive experience.

[0059] The environmental data acquisition unit further integrates an infrared thermal imaging sensor and a three-dimensional depth camera to improve the overall performance of the intelligent lighting control system. The role of the infrared thermal imaging sensor is to monitor the temperature distribution on the surface of the artwork in real time. This sensor can detect local temperature changes on the artwork, especially when the light is strong or the ambient temperature is high, some areas may be locally overheated. By obtaining the temperature information on the surface of the artwork, the system can adjust the lighting intensity and direction in time to avoid long-term exposure of local areas to high temperatures, thereby effectively protecting the artwork from overheating damage, especially temperature-sensitive materials such as paper, paint or fabric.

[0060] In addition, three-dimensional depth cameras are used to capture the movement trajectory and residence position of visitors in the exhibition hall. Through depth sensing technology, such cameras can accurately perceive the location information of visitors in the space and the time they stay in front of different exhibits. These data provide key inputs for the system, which can analyze the behavior patterns of visitors and understand their points of interest and interactive needs in the exhibition hall. Based on these behavioral data, the lighting system can automatically adjust the lighting strategy. For example, when the system detects that the audience stays in front of a certain artwork for a long time, the lighting can be gradually enhanced to highlight the details of the artwork and enhance the audience's viewing experience. Conversely, when it is detected that the audience leaves a certain area, the system can reduce the lighting intensity of that area to save energy.

[0061] By integrating data from infrared thermal imaging sensors and 3D depth cameras, the intelligent lighting control system can not only automatically optimize lighting settings based on ambient temperature, but also dynamically adjust the light distribution in the exhibition hall according to the behavior patterns of the audience. This lighting strategy that combines temperature protection and audience behavior optimization can not only better protect artworks, but also provide a more personalized exhibition experience, ensuring that the system is highly practical and intelligent in actual applications.

[0062] Furthermore, the eye tracking device includes:

[0063] Multiple high-speed cameras located in different locations in the exhibition hall;

[0064] A real-time image processing module that fuses data from multiple high-speed cameras to build a heat map of audience sightlines;

[0065] The sight heat map is used to identify the level of attention paid to artworks, and dynamically adjust the lighting parameters of the corresponding areas accordingly.

[0066] The eye tracking equipment of the intelligent lighting control system includes multiple high-speed cameras distributed in different locations in the exhibition hall and a real-time image processing module. The deployment positions of the high-speed cameras are carefully designed to ensure that they can cover various areas of the exhibition hall from multiple angles and capture the direction of the audience's sight. These cameras work at a high frame rate and can accurately track the audience's eye movements and determine their sight focus.

[0067] The real-time image processing module is responsible for receiving and fusing the data from these high-speed cameras. To ensure the accuracy of eye tracking, the module uses advanced image processing algorithms to process and analyze the images captured by the camera and accurately identify the direction of each visitor's gaze by comparing multi-angle views. With this data, the system can construct a gaze heat map in the exhibition hall. The heat map visualizes the audience's attention to different areas in the exhibition hall. The brighter areas in the heat map represent artworks or exhibition areas that the audience pays more attention to, while the darker areas indicate that the audience pays less attention.

[0068] By analyzing this sight heat map, the intelligent lighting control system can identify the attention level of each area in the exhibition hall. Based on this analysis, the system can dynamically adjust the lighting parameters, for example, increasing the light intensity around the artworks that the audience pays more attention to, or adjusting the color temperature and beam angle to highlight the details of the artworks. At the same time, for areas where the audience pays less attention, the system can appropriately reduce the lighting intensity to reduce energy consumption and avoid overexposure of the artworks.

[0069] The collaborative work of the real-time image processing module and the eye tracking system can not only effectively improve the exhibition experience, but also provide better display conditions for artworks by optimizing lighting strategies. This automated and dynamic lighting adjustment mechanism ensures the intelligence and personalization of the exhibition, which not only enhances the audience's sense of interaction, but also protects the artworks from unnecessary light effects.

[0070] The intelligent control unit 102 is used to receive the collected data provided by the environmental data collection unit; generate an optimal lighting strategy based on the specific protection and display requirements of the artwork and the collected data; wherein the optimal lighting strategy is used to enhance the visual expression of the artwork in the exhibition hall and the interactive experience of the audience.

[0071] The intelligent control unit 102 is the core component of the environment-aware intelligent lighting control system, responsible for integrating and processing various sensor data from the environmental data acquisition unit 101, and generating adaptive lighting strategies based on these data. The design of this unit allows it to receive and analyze multiple key parameters such as light intensity, color temperature, sight tracking, temperature and humidity, etc., to ensure the responsiveness and dynamic adjustment capabilities of the lighting system.

[0072] The intelligent control unit uses advanced data processing technology, such as machine learning algorithms, to identify and predict the lighting needs within the exhibition hall. By analyzing the protection needs of the artworks and the behavior patterns of the audience, the unit is able to generate precise lighting configurations, including light intensity, color temperature and the direction of the light beam, thereby optimizing the visual performance of the artworks and the viewing experience of the audience.

[0073] In addition, the intelligent control unit can also be implemented based on pre-defined rules. For example, in an art gallery dedicated to exhibiting ancient Chinese paintings and calligraphy and precious ceramics. Each exhibit has unique preservation and display requirements, so the lighting system needs to be precisely controlled to protect these priceless treasures. The art gallery uses environmental data acquisition units to monitor key parameters, and the intelligent control unit generates the best lighting strategy based on this data.

[0074] First, consider an exhibition area displaying exquisite porcelain from the Song Dynasty. The color and glaze of these ceramics are particularly eye-catching under the right light. If the light intensity sensor detects that the light intensity is lower than the recommended 200 lux, the intelligent control unit will increase the brightness of the LED light to ensure the visibility of details. At the same time, if the color temperature sensor shows that the color temperature is higher than 5000K, the control unit will adjust the color temperature of the light to a warmer tone, because warm light can better highlight the warm texture of ceramics.

[0075] On the other hand, there are also a number of exhibition areas displaying Tang Dynasty calligraphy works. These paper artifacts are very sensitive and easily damaged by light. If the light intensity sensor detects that the light intensity exceeds 100 lux, the intelligent control unit immediately reduces the lighting intensity to protect these fragile papers from UV and strong light damage. At the same time, the eye tracking device may find that certain works are particularly attractive to the audience, so the control unit will fine-tune the lighting in these areas to make the viewing experience more comfortable while ensuring that the light will not cause damage to the artifacts due to long-term exposure.

[0076] Through these simple but precise "if-then" rules, the intelligent lighting system can dynamically adjust the lighting settings according to the real-time monitored environmental data and the specific needs of the artworks, protecting the integrity of precious artworks while providing the best viewing experience. This approach takes into account the needs of both artwork protection and display, and embodies the efficiency and intelligence of modern museum environmental management.

[0077] The intelligent control unit has a high degree of programming flexibility, which enables it to automatically adjust the lighting strategy according to different exhibition needs and environmental conditions. The unit works closely with the adaptive lighting execution unit 103 to guide it to adjust the specific parameters of multiple LED light sources, such as adjusting the color temperature to match the color attributes of the exhibits, or changing the light intensity to adapt to the changes in the light intensity in the exhibition hall.

[0078] In addition, the intelligent control unit also processes audience input from the interactive lighting adjustment unit 104. Through the integrated touch screen and voice recognition technology, the unit can receive real-time feedback from the audience about lighting preferences and quickly adjust the lighting settings, enhancing interactivity and improving the audience's personalized experience.

[0079] Overall, the intelligent control unit 102 is the key technology that enables the context-aware intelligent lighting control system to achieve accurate and adaptive lighting management.

[0080] Furthermore, the intelligent control unit generates an optimal lighting strategy using a pre-trained neural network model according to the display requirements of the artwork and the collected data; wherein the neural network model includes an input layer, a feature enhancement layer, a spatiotemporal processing layer, and an output layer;

[0081] The input of the input layer is the display requirements of the artwork and the collected data; the input layer is implemented using a graph attention network structure, which represents the display requirements of the artwork and the collected data as nodes in the graph, establishes edge connections based on the mutual relationship between the data, and uses a multi-head attention mechanism to calculate the weights between the nodes, and outputs a feature representation that integrates the complex relationship between the data;

[0082] The input of the feature enhancement layer is the feature representation output by the output layer; the feature enhancement layer is implemented by a capsule network structure; the feature enhancement layer uses a dynamic routing algorithm to process low-level features to combine into high-level features, wherein each capsule specifically encodes a specific type of lighting attribute, and the output is a set of feature vectors, each feature vector includes the existence probability of a lighting attribute and its parameters; wherein the lighting attributes include light intensity, color temperature, and beam angle;

[0083] The input layer of the spatiotemporal processing layer is the feature vector group output by the feature enhancement layer; the spatiotemporal processing layer is implemented by a spatiotemporal graph convolutional network structure, the feature vector group output by the feature enhancement layer is constructed into a spatiotemporal graph, the graph convolution operation is used to capture the spatial dependency relationship between features, and the time convolution operation is applied to capture the time series mode, and the output is a comprehensive feature representation of the comprehensive spatiotemporal relationship;

[0084] The input of the output layer is the comprehensive feature representation of the output of the spatiotemporal processing layer; the output layer maps the complex high-dimensional input to a two-dimensional discrete grid through a competitive learning process, wherein each two-dimensional discrete grid node represents a set of specific LED light source setting parameters; the output layer outputs the adjusted optimal light intensity, color temperature and beam angle parameters of each LED light source.

[0085] The intelligent lighting control system provided in this embodiment generates the best lighting strategy according to the display requirements of the artwork and environmental data through the intelligent control unit. The intelligent control unit adopts a pre-trained neural network model, which is divided into multiple key layers: input layer, feature enhancement layer, spatiotemporal processing layer and output layer. Each layer works together to achieve dynamic and intelligent lighting adjustment to ensure that the artwork is displayed under the best lighting conditions.

[0086] First, the input layer receives various data collected by the system, including the light intensity, color temperature, audience eye tracking data in the exhibition hall environment, and the specific display requirements of each artwork. These data are represented as nodes of a graph structure. The input layer uses a graph attention network structure to model the correlation between nodes as edges, and calculates the weights between each node through a multi-head attention mechanism. The weights represent the complex relationship between the data. Finally, the output of this layer is a feature representation that has been weighted and contains the relationship between the data, which serves as the basis for subsequent processing.

[0087] Next, the feature enhancement layer processes the feature representation output by the input layer. This layer uses a capsule network structure to process low-level features through a dynamic routing algorithm and combine them into higher-level features. Each capsule encodes a specific lighting attribute, such as light intensity, color temperature, and beam angle, and each output feature vector represents the probability of existence of these lighting attributes and their specific parameters. In this way, the feature enhancement layer can extract more detailed and targeted lighting attributes, providing more accurate input for subsequent decision-making.

[0088] The spatiotemporal processing layer is then responsible for processing the feature vector group output by the feature enhancement layer. This layer uses a spatiotemporal graph convolutional network structure to construct the feature vector group into a spatiotemporal graph. Through the graph convolution operation, this layer is able to capture the spatial dependencies between features, while applying the temporal convolution operation to capture the time series pattern. This combination allows the system to comprehensively consider the information in the spatial and temporal dimensions, thereby generating a comprehensive feature representation that includes spatiotemporal dependencies.

[0089] Finally, the output layer processes the comprehensive feature representation output by the spatiotemporal processing layer. This layer maps the complex high-dimensional input to a two-dimensional discrete grid through a competitive learning process. Each grid node represents a specific set of LED light source setting parameters, such as light intensity, color temperature, and beam angle. Based on these mapping results, the output layer automatically adjusts the lighting parameters of each LED light source to ensure that each artwork in the exhibition hall is displayed under optimal lighting conditions.

[0090] Through this multi-level neural network model, the system can automatically generate lighting strategies that meet the needs of artwork display, and adjust the lighting parameters of LED light sources in real time to achieve intelligent and dynamic lighting control and improve the artwork display effect.

[0091] The following is the reference implementation code of the above neural network model:

[0092]

[0093]

[0094]

[0095]

[0096]

[0097] When training this neural network model, the following steps need to be followed to ensure that the model can generate the best lighting strategy based on the collected data and the artwork display requirements. The training process involves key steps such as model initialization, data preparation, forward propagation, loss calculation, back propagation, and optimization.

[0098] First, the input of the model includes the environmental data of the exhibition hall and the display requirements of the artworks. The environmental data may include information such as light intensity, color temperature, and audience eye tracking, while the display requirements involve the requirements of the artworks on light intensity, color temperature, beam angle, etc. Through the Graph Attention Network (GAT), these input data will be represented as nodes in the graph, and edge connections will be built based on the relationship between the nodes. The multi-head attention mechanism will calculate the weights between each node, thereby outputting a feature representation that integrates the complex relationships between the data.

[0099] The first step in training the model is to initialize the model parameters, including the input layer, feature enhancement layer (capsule network), spatiotemporal processing layer (spatial-temporal graph convolutional network), and output layer (competitive learning of two-dimensional discrete grids). Weight and bias initialization can use common Xavier initialization or He initialization to ensure that the network has a reasonable parameter distribution at the beginning.

[0100] Next, prepare the training data. The data set should contain different lighting conditions in various exhibition hall scenes, artwork display requirements, and corresponding lighting strategies as labels. Divide the data into training sets and validation sets to ensure that the model can be cross-validated during training to avoid overfitting. Each training data contains input features (such as light intensity, color temperature, and audience position) and corresponding target outputs (such as the optimal LED light source setting parameters).

[0101] In each training iteration, forward propagation is performed first. The input layer processes the input environmental data and display requirements through the graph attention network and outputs the feature representation of each node. Next, the feature enhancement layer processes these features through the capsule network, combines low-level lighting attributes (such as light intensity, color temperature, etc.) into high-level features, and generates the existence probability and parameters of each attribute. These high-level features are input into the spatiotemporal processing layer, which uses the spatiotemporal graph convolutional network to capture the spatial and temporal dependencies of the features and outputs the feature representation of the comprehensive spatiotemporal relationship. Finally, the output layer maps the high-dimensional input to a two-dimensional discrete grid through a competitive learning mechanism to generate the optimal setting parameters for each LED light source.

[0102] After the forward propagation is completed, the loss function is calculated. The loss function should measure the difference between the LED light source parameters generated by the model and the real lighting settings in the label. Common loss functions include mean squared error (MSE) or cross entropy loss, and the specific choice depends on the format of the label and the task requirements. By comparing the difference between the model output and the target lighting strategy, the loss function provides feedback on the model performance.

[0103] Next, backpropagation is performed. Backpropagation calculates the gradient of the loss function with respect to the model parameters and updates the model's weights and biases through an optimization algorithm such as Adam or SGD. This process is performed using the chain rule, calculating the gradient layer by layer, starting from the output layer until the input layer. The model's weights and biases are adjusted in each iteration according to the direction of the gradient, gradually optimizing the resulting lighting strategy to make it closer to the target value.

[0104] In order to ensure the training effect of the model, the validation set should be used for performance evaluation at each training stage. Through the performance of the validation set, the generalization ability of the model can be observed, hyperparameters (such as learning rate, batch size, etc.) can be adjusted, and the model can be prevented from overfitting. The validation process does not involve weight updates, but it will provide a reference for model adjustments.

[0105] After multiple rounds of iterations, when the loss function on the validation set converges or reaches the set threshold, the training ends. At this point, the model is able to generate the optimal lighting settings for each LED light source based on the environmental data of the exhibition hall and the display requirements of the artwork. After the training is completed, the inference efficiency of the model can be further optimized through tuning and pruning to ensure sufficient real-time performance and computing efficiency in practical applications.

[0106] In summary, the core of model training is to generate the optimal lighting strategy through forward propagation, continuously optimize parameters through back propagation, and ensure the generalization ability and efficiency of the model through cross-validation and tuning.

[0107] Furthermore, the spatiotemporal processing layer is implemented by a spatiotemporal graph convolutional network structure, which constructs the feature vector group output by the feature enhancement layer into a spatiotemporal graph, and captures the spatial and temporal dependencies through the following steps to generate a comprehensive feature representation of the comprehensive spatiotemporal relationship:

[0108] The output feature vector group of the feature enhancement layer is constructed as a spatiotemporal graph G = (V, E, A), where V is a node set, representing the feature vectors at different time steps; E is an edge set, representing the spatial and temporal relationship between features; and A is an adjacency matrix.

[0109] On the spatiotemporal graph, the convolution operation provided by the following formula 1 is used to capture the spatial dependencies between features:

[0110]

[0111] Among them, H (l+1) is the node feature matrix of the l+1th layer; H (l) is the node feature matrix of the lth layer, which represents the feature representation of the current layer; A is the adjacency matrix; D is the degree matrix; λ is the adaptive parameter; W (l) is the weight matrix; σ is the activation function;

[0112] According to the following formula 2, multi-scale temporal convolution is applied at each time step to capture the time series pattern of features:

[0113]

[0114] Among them, Z t is the output feature at time step t; K1 represents the number of convolution kernels; τ is the time offset, which represents the time step considered in the time convolution operation; H t-τ is the input feature at time t-τ; W k,τ is the weight of the kth convolution kernel at the time offset τ; ⊙ represents the product operation between elements; b k is the first bias term; τ1 represents the size of the time window;

[0115] Through the following formulas 3 and 4, the feature representations of different time steps are fused through the attention mechanism:

[0116] α t =softmax(q T tanh(Kh t +b)) (3)

[0117]

[0118] Among them, h t is the feature vector of time step t; q and K are learnable parameter matrices; T represents the total number of time steps; b is the second bias term; αt is the attention weight at time step t; c represents the comprehensive feature representation of the comprehensive spatiotemporal relationship.

[0119] The intelligent lighting control system provided in this embodiment implements the operation of the spatiotemporal processing layer through the spatiotemporal graph convolutional network structure, which is responsible for processing the feature vector group output by the feature enhancement layer and constructing it into a spatiotemporal graph G = (V, E, A). Among them, V is a node set, which represents the feature vectors at different time steps; E is an edge set, which represents the spatial and temporal relationship between feature vectors; A is an adjacency matrix, which is used to represent the connection relationship between nodes.

[0120] On this spatiotemporal graph, the spatial dependencies between nodes are captured by convolution operations. Specifically, Formula 1 defines this convolution operation:

[0121]

[0122] In formula 1, H (l+1) Represents the node feature matrix of the l+1th layer, which contains the updated node features after the convolution operation of the current layer; \A (l) is the node feature matrix of the lth layer, which represents the feature representation of the current layer. A is the adjacency matrix, which is used to describe the connection between nodes and reflects the relationship between data. The degree matrix D is a diagonal matrix whose elements are the degree of each node (that is, the number of edges connected to the node). It is used to normalize the adjacency matrix to ensure that the feature information is not affected by the number of node connections when it is transmitted between different nodes. λ is an adaptive parameter that adjusts the weight between the adjacency matrix A and the identity matrix I to properly balance the self-connection of the node (that is, the connection between the node and itself). This balance helps to maintain the local information of the node. W (l) is a weight matrix, which contains learnable parameters and performs linear transformation on input features through training process. σ is an activation function, and nonlinear functions such as ReLU or Sigmoid are usually selected to introduce nonlinear characteristics and improve the expressive power of the model.

[0123] Next, the spatiotemporal processing layer also captures the time series pattern of features through multi-scale temporal convolution operations. This process is defined by Formula 2:

[0124]

[0125] In Formula 2, Z t represents the output feature at time step t, which is the result of multi-scale time convolution at time t. K1 represents the number of convolution kernels, which are learnable parameter groups used to extract different feature patterns. The time offset τ represents the time step, indicating the features extracted by the model at different time delays. H t-τis the input feature at time step t-τ, representing the input data at that moment. k,τ is the weight of the kth convolution kernel at the time offset τ, and is the parameter matrix used to weight the input features at a specific time step. ⊙ represents the product operation between elements, namely the Hadamard product, which is used to calculate the weighted sum of each input feature. k is the bias term of the kth convolution kernel, which is used to adjust the output after the convolution operation to ensure that the model has sufficient flexibility. T1 is the size of the time window, which indicates the length of historical time considered by the model in each time step, thereby capturing feature dependencies at different time scales.

[0126] After the convolution operation in the time dimension, the system fuses the feature representations of different time steps through the attention mechanism. This process is defined by Formula 3 and Formula 4:

[0127] α t =softmax(q T tanh(Kh t +b)) (3)

[0128]

[0129] In formula 3, q T represents the transpose of the parameter matrix q; α t h represents the attention weight at time step t, which is normalized by the softmax function to ensure that the sum of the weights at each time step is 1. t is the feature vector at time step t, representing the input features at that moment. Attention weight α t The calculation of h is realized through the learnable parameter matrices q and K. First, h t After the linear transformation of matrix K, it is activated by the hyperbolic tangent function (tanh), and finally multiplied by matrix q, and the bias term b is added. The attention weight of each time step is calculated by the softmax function. The c in formula 4 represents the final comprehensive feature representation, which is calculated by taking the feature vector h of each time step t Multiply it by its corresponding attention weight α t , and then sum the weighted features of all time steps to obtain a comprehensive spatiotemporal feature representation. This process ensures that the system can flexibly integrate information from different time steps and focus on those time steps that are most important for the final lighting decision.

[0130] Through the collaborative work of the above steps, the spatiotemporal processing layer can not only capture the spatial dependencies between features, but also make full use of the information in the time series to generate a feature representation of the comprehensive spatiotemporal relationship. This feature representation provides an important input for the output layer of the system, enabling the system to generate an optimized lighting strategy based on the environmental changes in the exhibition hall and the display requirements of the artworks.

[0131] The adaptive lighting execution unit 103 includes multiple LED light sources, wherein each LED light source can independently adjust setting parameters including light intensity, color temperature and beam angle; the adaptive lighting execution unit is used to adjust the setting parameters of the multiple LED light sources according to the optimal lighting strategy provided by the intelligent control unit.

[0132] The adaptive lighting execution unit 103 is a key component designed to respond to the lighting strategy generated by the intelligent control unit 102 and accurately adjust the exhibition hall lighting to maximize the visual display effect of the artwork and the audience experience. This unit contains multiple LED light sources, each of which has a high degree of adjustment flexibility and can independently adjust the light intensity, color temperature and beam angle.

[0133] Each LED light source is equipped with a corresponding driver and control interface, so that it can receive instructions from the intelligent control unit. These instructions are based on the real-time data collected by the environmental data acquisition unit 101, such as the light intensity and color temperature in the exhibition hall, and the audience's sight and stay time information.

[0134] Adjustment capability of LED light source:

[0135] 1. Light intensity adjustment: Each LED lamp can adjust the brightness according to the specific lighting needs of the exhibition hall. For example, if the light intensity of a certain artwork area is detected to be too low, the intelligent control unit will instruct the relevant LED light source to increase the light intensity to ensure that the details of the artwork can be clearly displayed.

[0136] 2. Color temperature adjustment: Color temperature is crucial to the visual impact of artworks. The adaptive lighting actuator can adjust the color temperature according to the display requirements of the artwork. For example, in order to enhance the color saturation of antique paintings, a warmer color temperature may be required.

[0137] 3. Beam Angle Adjustment: Adjusting the angle of the beam can change the range and intensity of light projection, which is very useful for highlighting specific parts of artworks or creating specific visual effects. For example, for sculptures, by adjusting the beam angle, the three-dimensional sense and texture details of the work can be enhanced.

[0138] To achieve the above functions, each LED light source is equipped with a microprocessor and sensors that can accurately execute the instructions of the intelligent control unit. In addition, these light sources are connected to the intelligent control unit through a communication network (such as Wi-Fi or ZigBee) to ensure the timely transmission and execution of instructions.

[0139] The intelligent control unit calculates the specific settings required for each light source based on the collected environmental data and the preset protection and display logic. For example, if visitors stay in an exhibition area for a long time, the intelligent control unit may instruct to adjust the color temperature and brightness of the lights in that area to optimize the visitor experience while protecting the exhibits.

[0140] Through such configuration and technical implementation, the adaptive lighting execution unit 103 not only improves the display quality of artworks, but also responds to the behavior of the audience by dynamically adjusting the lighting, thereby maximizing the efficiency and effect of the lighting system. This highly adjustable LED lighting system provides a flexible lighting solution for art galleries that protects artworks and meets the needs of the audience.

[0141] Furthermore, each LED light source of the adaptive lighting execution unit includes an automatic color temperature adjuster, which is used to automatically adjust the color temperature according to the natural light incident in different time periods in the exhibition hall and the lighting requirements of the artworks, so as to maintain the natural feeling of the lighting and the visual appeal of the artworks.

[0142] The adaptive lighting execution unit of the intelligent lighting control system is equipped with an automatic color temperature adjuster for each LED light source, ensuring dynamic lighting adjustment in the exhibition hall according to the ambient light conditions and the specific needs of the artwork. The core function of the automatic color temperature adjuster is to automatically adjust the color temperature of the LED light source by sensing the changes in natural light in the exhibition hall in real time, so as to ensure that the entire exhibition hall always maintains the most natural and comfortable lighting environment, while enhancing the visual expression of the artwork.

[0143] Specifically, the automatic color temperature adjuster detects the incoming natural light in real time through a built-in light sensor. The change of natural light can be very significant throughout the day, especially when there are a large number of windows or skylights in the exhibition hall. The intensity and color temperature of the external light will fluctuate significantly with the change of time, weather or season. The automatic color temperature adjuster can accurately sense the intensity and color temperature of these external lights and make corresponding adjustments to coordinate artificial lighting with natural light, keeping the light transition in the entire exhibition hall natural and smooth.

[0144] At the same time, the automatic color temperature adjuster will also adjust the color temperature according to the lighting requirements of specific artworks. Different types of artworks have different requirements for the color temperature of lighting. For example, ancient paintings or textiles can better show their details and texture under warmer light, while modern artworks or sculptures may require cooler light to highlight their unique texture and form. The automatic color temperature adjuster will automatically select the most appropriate color temperature based on the preset lighting requirements of the artwork and the natural light conditions to ensure that the audience can appreciate the details and colors of the artwork under the most ideal lighting conditions.

[0145] Through this real-time color temperature adjustment mechanism, the intelligent lighting control system can achieve efficient light management during the exhibition, which not only improves the overall visual experience in the exhibition hall, but also ensures the best display effect of the artwork. In addition, the design of the automatic color temperature adjuster allows the system to seamlessly integrate natural light and artificial lighting, providing stable and continuously optimized lighting effects even when natural light changes dramatically.

[0146] Furthermore, the LED light sources of the adaptive lighting execution unit are configured to receive adjustment instructions from the intelligent control unit via wireless signals, and each light source can independently report its energy efficiency and performance status back to the intelligent control unit to optimize energy use and reduce maintenance costs.

[0147] Each LED light source in the adaptive lighting execution unit of the intelligent lighting control system has wireless communication capability and can receive adjustment instructions from the intelligent control unit through wireless signals. The wireless signal transmission system enables the light source to seamlessly communicate with the intelligent control unit to achieve efficient and dynamic lighting adjustment.

[0148] Specifically, each LED light source is connected to the intelligent control unit through a wireless module. When the intelligent control unit generates a lighting strategy based on real-time collected environmental data (such as light intensity, color temperature, audience behavior, etc.), it sends specific adjustment instructions to each light source through wireless signals. These instructions may involve the adjustment of parameters such as light intensity, color temperature, and beam angle. Through wireless communication, the system can achieve fast and flexible adjustments to ensure that the lighting settings in the entire exhibition hall match the current environment and artwork display requirements.

[0149] In addition to receiving control instructions, each LED light source also has the function of self-monitoring and reporting energy efficiency and performance status. Each light source is embedded with sensors and energy efficiency monitoring modules, which can detect its own operating status in real time, including power consumption, luminous efficiency, temperature and possible faults. These monitoring data are sent back to the intelligent control unit regularly or on demand via the wireless network. After receiving this data, the intelligent control unit will analyze the energy efficiency of the entire lighting system and make further optimization adjustments based on the status of the light source. For example, the system can reduce unnecessary lighting intensity in a timely manner according to the energy consumption of the light source, or arrange maintenance or replacement in a timely manner when certain light sources are detected to be faulty or energy efficiency is reduced.

[0150] Through this wireless communication and feedback system, the intelligent lighting control system can effectively optimize energy use, ensuring that the light source meets the needs of art display while avoiding unnecessary energy waste. At the same time, by obtaining the performance status of each light source in real time, the system can perform preventive maintenance and reduce downtime caused by light source failure, thereby reducing maintenance costs and extending the service life of the light source.

[0151] Furthermore, each LED light source of the adaptive lighting execution unit is equipped with a photosensitive element for monitoring the performance of the light emitted by the light source in the actual environment, and the collected real-time feedback data is used to adjust the lighting intensity and color temperature to adapt to the light changes in the exhibition hall.

[0152] Each LED light source of the adaptive lighting execution unit is equipped with a photosensitive element, the main function of which is to monitor the actual performance of the light emitted by each light source in the exhibition hall environment in real time. The design of this function ensures that the light source can dynamically adapt to the changes in the light conditions in the exhibition hall, thereby providing the most suitable lighting conditions under different external environments and display requirements.

[0153] Specifically, the photosensitive element is installed near each LED light source, which can accurately sense the light intensity, color temperature and other parameters emitted by the light source, and collect real-time feedback data of the lighting environment in the exhibition hall. The lighting conditions of the exhibition hall environment may change constantly due to changes in natural light, interference from other lights or movement of the audience. By continuously monitoring these changes, the photosensitive element can provide accurate feedback information to the system. This feedback data includes the actual effect of the light emitted by the light source in the exhibition hall space, such as whether the light intensity is sufficient, whether the color temperature meets the display requirements of the artwork, whether the light is uneven or overexposed, etc.

[0154] The collected real-time feedback data will be transmitted back to the intelligent control unit, which will analyze the current lighting status based on the data. If the system detects that the light in the exhibition hall does not meet the preset conditions, the intelligent control unit will promptly issue adjustment instructions to adjust the lighting intensity and color temperature of each LED light source. For example, if the photosensitive element detects that the light in a specific area is too strong or the color temperature is too low, the intelligent control unit will automatically reduce the light intensity or adjust the color temperature in that area to ensure the best display effect of the artwork. At the same time, if the light in certain areas of the exhibition hall is weak, the system can also ensure the uniform distribution of light by increasing the light intensity of the corresponding light source.

[0155] This feedback adjustment mechanism enables the lighting system to respond intelligently to changes in light conditions, avoiding the problems that may arise from fixed lighting settings. This not only improves the visual effects in the exhibition hall, but also provides more refined protection for the artworks to prevent them from being damaged by excessive light. In addition, since the system can dynamically adjust the lighting settings according to the actual environmental conditions, this approach also optimizes energy use and avoids unnecessary energy waste.

[0156] Furthermore, the adaptive lighting execution unit includes a lighting balance controller for adjusting the light intensity and color temperature differences between adjacent LED light sources to eliminate uneven lighting in the audience's field of view and improve the overall uniformity of lighting and visual comfort.

[0157] The adaptive lighting execution unit integrates a lighting balance controller specifically designed to adjust the light intensity and color temperature differences between adjacent LED light sources to ensure that the lighting effect of the entire exhibition hall achieves visual uniformity and comfort.

[0158] The role of the lighting balance controller is to coordinate the lighting settings between adjacent light sources by analyzing the light characteristics of each LED light source in the exhibition hall in real time, especially the light intensity and color temperature. When the light intensity or color temperature between adjacent light sources is very different, obvious bright and dark areas or inconsistent color temperatures will appear in the audience's field of vision. This uneven lighting will not only affect the visual effect of the artwork, but may also cause visual discomfort to the audience. To avoid these problems, the lighting balance controller can dynamically adjust the light output of each light source to make it more consistent with the light properties of the surrounding light sources.

[0159] Specifically, when the system detects that the light intensity in a certain area is higher or lower than that in an adjacent area, the lighting balance controller will automatically issue an adjustment command to reduce the light intensity in the overly bright area or increase the light output in the darker area to balance the light intensity in each area. Similarly, if a large difference in color temperature is detected in different areas, the lighting balance controller will also adjust the color temperature of adjacent light sources accordingly to ensure that the light in the exhibition hall remains consistent in color tone. This dynamic light balance mechanism can significantly improve the audience's visual experience, allowing the artworks in the exhibition hall to be displayed under more uniform and harmonious light conditions, reducing visual interference caused by uneven light.

[0160] In addition, the lighting balance controller can also make intelligent adjustments based on the distribution of visitors in the exhibition hall. For example, when the number of visitors in a certain area is relatively high, the system can appropriately increase the lighting intensity in that area, while in areas with fewer visitors or no one, the brightness of the light source can be reduced to save energy. This flexible lighting adjustment method not only improves the overall viewing effect of the exhibition, but also improves the energy efficiency of the lighting system and avoids energy waste.

[0161] The interactive lighting adjustment unit 104 integrates a touch screen and a voice recognition module, and is used to receive the audience's input on lighting preferences; and adjust the setting parameters of the multiple LED light sources according to the audience's feedback.

[0162] The interactive lighting adjustment unit 104 is a key component in the context-aware intelligent lighting control system, which is designed to enhance the interactive experience between visitors and the exhibition hall lighting. This unit includes an integrated touch screen and voice recognition module, allowing visitors to directly influence and adjust the lighting settings in the exhibition hall, adjusting the light intensity, color temperature and beam angle according to personal preferences or specific needs.

[0163] The interactive lighting adjustment unit 104 allows viewers to input their lighting preferences through a highly user-friendly interface. This can be done through simple sliders on the touch screen to adjust light intensity and color temperature, or through voice commands to quickly adjust lighting conditions. For example, viewers can say "increase light" or "reduce blue light", and the system interprets these commands through the voice recognition module and adjusts the lighting in real time.

[0164] The touch screen displays a simple graphical user interface with a slider to control the illumination of the artwork, color temperature adjustment options, and selectable settings for the beam angle. Visitors can directly adjust the lighting by sliding to make it more suitable for personal viewing habits or to enhance the visual effect of the artwork.

[0165] To enhance accessibility and convenience, the voice recognition module enables visitors to control the lighting through simple voice commands. Designed with natural language processing capabilities, the system can understand and execute common lighting-related commands, such as adjusting brightness, changing color temperature or the intensity of light in a specific area.

[0166] The interactive lighting adjustment unit 104 is tightly integrated with the adaptive lighting execution unit 103 and the intelligent control unit 102 to ensure that the audience's input can be converted into specific lighting adjustments in real time. The unit contains the necessary hardware and software components to support the reception, processing and sending of data:

[0167] The interactive lighting adjustment unit 104 includes a touch screen display, a microphone array (for capturing sound input), a processor (for parsing touch and sound data), and a network interface (for communicating with other parts of the system).

[0168] The interactive lighting adjustment unit 104 includes interface design software, a speech-to-text conversion engine, and a command transmission protocol to ensure that the audience's input is quickly and accurately converted into lighting control commands.

[0169] The interactive lighting adjustment unit 104 not only enhances the technical interactivity of the exhibition hall, but also greatly enriches the visitor's visiting experience, allowing each visitor to adjust the exhibition lighting according to personal preferences, thereby appreciating and understanding the artwork more deeply.

[0170] Furthermore, the interactive lighting adjustment unit includes a user interface, which automatically adjusts its operating interface and functional options according to the audience's historical preferences and feedback to provide a personalized interactive experience.

[0171] The interactive lighting adjustment unit is equipped with a user interface that automatically adjusts the operating interface and function options based on the audience's historical preferences and feedback, thereby providing a personalized interactive experience. This personalized design is designed to enable the lighting system to better respond to the unique needs of each audience by analyzing the audience's behavioral data.

[0172] The user interface records and analyzes historical interaction data of visitors, including the time they stay in front of different exhibits, the lighting settings they choose, and any feedback information. The system builds a personalized preference model for each visitor. As the visitor continues to interact with the system, the user interface gradually learns and updates these preferences, ensuring that the next time the visitor uses it, the interface can automatically adjust to present the operating options that best suit the visitor's habits and preferences. For example, if a visitor prefers to use warmer light to view artworks, then when he uses the system again, the interface will prioritize displaying adjustment options related to warm-toned light, simplifying the visitor's operating steps and improving the interactive experience.

[0173] In addition, the functional options of the user interface can also be dynamically adjusted based on the feedback of different viewers. For example, when viewers report that certain lighting modes do not meet their expectations, the system will record this information and avoid recommending similar settings in future interactions. The system can also optimize the overall layout and functional logic of the interface by analyzing the feedback of multiple viewers to make it more in line with the usage habits of most people, thereby improving the ease of use of the system.

[0174] This design of personalized adjustment based on historical data and feedback ensures that the user interface is highly intelligent and flexible, allowing each visitor to find their preferred lighting settings in the shortest possible time, thereby enhancing the visiting experience. This function is particularly suitable for visitors with different backgrounds and interests in the exhibition, providing each visitor with tailored lighting control options.

[0175] Although the present application is disclosed as above in the form of a preferred embodiment, it is not intended to limit the present application. Any technical personnel in this field may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

Claims

1. An environment-aware intelligent lighting control system, characterized in that: include: An environmental data acquisition unit, including a light intensity sensor, for monitoring the light intensity in the exhibition hall; Color temperature sensor, used to measure the color temperature in the exhibition hall; Eye tracking equipment to monitor the audience’s gaze and dwell time; An intelligent control unit, configured to receive the collected data provided by the environmental data collection unit; generate an optimal lighting strategy according to the display requirements of the artworks and the collected data; wherein the optimal lighting strategy is used to enhance the visual expression of the artworks in the exhibition hall and the interactive experience of the audience; An adaptive lighting execution unit, comprising a plurality of LED light sources, wherein each LED light source can independently adjust setting parameters including light intensity, color temperature and beam angle; the adaptive lighting execution unit is used to adjust the setting parameters of the plurality of LED light sources according to the optimal lighting strategy provided by the intelligent control unit; Interactive lighting adjustment unit, with integrated touch screen and voice recognition module, is used to receive audience input on lighting preferences; adjust the setting parameters of multiple LED light sources based on audience feedback.

2. The intelligent lighting control system according to claim 1, characterized in that: The environmental data acquisition unit also includes: Infrared thermal imaging sensor, used to detect the surface temperature distribution of artworks; 3D depth camera, used to capture the movement trajectory and residence position of visitors in the exhibition hall; The data collected by the infrared thermal imaging sensor and the three-dimensional depth camera are used to optimize the lighting strategy to prevent local overheating and improve the audience's interactive experience.

3. The intelligent lighting control system according to claim 1, characterized in that: The eye tracking device comprises: Multiple high-speed cameras located in different locations in the exhibition hall; A real-time image processing module that fuses data from multiple high-speed cameras to build a heat map of audience sightlines; The sight heat map is used to identify the level of attention paid to artworks, and dynamically adjust the lighting parameters of the corresponding areas accordingly.

4. The intelligent lighting control system according to claim 1, characterized in that: Each LED light source of the adaptive lighting execution unit includes an automatic color temperature adjuster, which is used to automatically adjust the color temperature according to the natural light incident in different time periods in the exhibition hall and the lighting requirements of the artworks, so as to maintain the natural feeling of the lighting and the visual appeal of the artworks.

5. The intelligent lighting control system according to claim 1, characterized in that: The LED light sources of the adaptive lighting execution unit are configured to receive adjustment instructions from the intelligent control unit via wireless signals, and each light source can independently report its energy efficiency and performance status back to the intelligent control unit to optimize energy use and reduce maintenance costs.

6. The intelligent lighting control system according to claim 1, characterized in that: Each LED light source of the adaptive lighting execution unit is equipped with a photosensitive element for monitoring the performance of the light emitted by the light source in the actual environment. The collected real-time feedback data is used to adjust the lighting intensity and color temperature to adapt to the light changes in the exhibition hall.

7. The intelligent lighting control system according to claim 1, characterized in that: The adaptive lighting execution unit includes a lighting balance controller for adjusting the light intensity and color temperature differences between adjacent LED light sources to eliminate uneven lighting in the audience's field of view and improve the overall uniformity of lighting and visual comfort.

8. The intelligent lighting control system according to claim 1, characterized in that: The interactive lighting adjustment unit includes a user interface, which automatically adjusts its operation interface and functional options according to the audience's historical preferences and feedback to provide a personalized interactive experience.

9. The intelligent lighting control system according to claim 1, characterized in that: The intelligent control unit generates an optimal lighting strategy using a pre-trained neural network model according to the display requirements of the artwork and the collected data; wherein the neural network model includes an input layer, a feature enhancement layer, a spatiotemporal processing layer, and an output layer; The input of the input layer is the display requirements of the artwork and the collected data; the input layer is implemented using a graph attention network structure, which represents the display requirements of the artwork and the collected data as nodes in the graph, establishes edge connections based on the mutual relationship between the data, and uses a multi-head attention mechanism to calculate the weights between the nodes, and outputs a feature representation that integrates the complex relationship between the data; The input of the feature enhancement layer is the feature representation output by the output layer; the feature enhancement layer is implemented by a capsule network structure; the feature enhancement layer uses a dynamic routing algorithm to process low-level features to combine into high-level features, wherein each capsule specifically encodes a specific type of lighting attribute, and the output is a set of feature vectors, each feature vector includes the existence probability of a lighting attribute and its parameters; wherein the lighting attributes include light intensity, color temperature, and beam angle; The input layer of the spatiotemporal processing layer is the feature vector group output by the feature enhancement layer; the spatiotemporal processing layer is implemented by a spatiotemporal graph convolutional network structure, the feature vector group output by the feature enhancement layer is constructed into a spatiotemporal graph, the graph convolution operation is used to capture the spatial dependency relationship between features, and the time convolution operation is applied to capture the time series mode, and the output is a comprehensive feature representation of the comprehensive spatiotemporal relationship; The input of the output layer is the comprehensive feature representation of the output of the spatiotemporal processing layer; the output layer maps the complex high-dimensional input to a two-dimensional discrete grid through a competitive learning process, wherein each two-dimensional discrete grid node represents a set of specific LED light source setting parameters; the output layer outputs the adjusted optimal light intensity, color temperature and beam angle parameters of each LED light source.

10. The intelligent lighting control system according to claim 9, characterized in that: The spatiotemporal processing layer is implemented by a spatiotemporal graph convolutional network structure, which constructs the feature vector group output by the feature enhancement layer into a spatiotemporal graph, and captures the spatial and temporal dependencies through the following steps to generate a comprehensive feature representation of the comprehensive spatiotemporal relationship: The output feature vector group of the feature enhancement layer is constructed as a spatiotemporal graph G = (V, E, A), where V is a node set, representing the feature vectors at different time steps; E is an edge set, representing the spatial and temporal relationship between features; and A is an adjacency matrix. On the spatiotemporal graph, the convolution operation provided by the following formula 1 is used to capture the spatial dependencies between features: Among them, H (l+1) is the node feature matrix of the l+1th layer; H (l) is the node feature matrix of the lth layer, which represents the feature representation of the current layer; A is the adjacency matrix; D is the degree matrix; λ is the adaptive parameter; W (l) is the weight matrix; σ is the activation function; I is the identity matrix; According to Formula 2 below, multi-scale temporal convolution is applied at each time step to capture the time series pattern of features: Among them, Z t is the output feature at time step t; K1 represents the number of convolution kernels; τ is the time offset, which represents the time step considered in the time convolution operation; H t-τ is the input feature at time t-τ; W k,τ is the weight of the kth convolution kernel at the time offset τ; ⊙ represents the product operation between elements; b k is the bias term of the kth convolution kernel; T1 represents the size of the time window; Through the following formulas 3 and 4, the feature representations of different time steps are fused through the attention mechanism: α t =softmax(q T tanh(Kh t +b)) (3) Among them, h t is the feature vector at time step t; q and K are learnable parameter matrices; T represents the total number of time steps; b is the bias term; α t is the attention weight at time step t; c represents the comprehensive feature representation of the comprehensive spatiotemporal relationship.

Images