Visual image-based scene light deployment method and system

Abstract

The invention relates to the field of data processing, in particular to a scene light deployment method and system based on a visual image. The method comprises the following steps: acquiring first log information of a monitored object; inputting the first log information into a risk early warning model, and identifying whether the monitoring object has a security risk so as to obtain a first identification result; acquiring second log information of an associated object matched with the monitoring object; inputting the first log information and the second log information into an associated early warning model, and identifying whether a security risk exists in an interaction relationship of the monitored object to obtain a second identification result; and if the first identification result or the second identification result indicates that the monitoring object has the security risk, generating early warning information of a corresponding type to trigger a security risk management process of the monitoring object. According to the method, the light deployment time can be shortened, and the overall working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for deploying scene lighting based on visual images. Background Art

[0002] Set lighting design is a comprehensive art and technology that shapes the scene atmosphere, highlights the main body, guides the audience's line of sight, and enhances the overall spatial expressiveness by skillfully using light.

[0003] In related technologies, set lighting design not only requires lighting design knowledge but also involves knowledge in multiple fields such as architecture, interior design, optics, and electronics. Designers need to understand the impact of building structures on light reflection and refraction, master interior design principles to coordinate lighting with space decoration, be familiar with optical knowledge to accurately control light, and understand electronic technology to ensure the normal operation of lighting equipment. From a technical perspective, designers need to consider multiple factors comprehensively. In terms of spatial layout, in an irregular drama exhibition hall or art exhibition hall, for example, it is necessary to cleverly utilize the projection angle and range of light to overcome the lighting problems brought by the spatial structure and ensure that each display area can receive appropriate light, which is difficult to implement. In addition, incorrect solutions can lead to waste of resources, increased implementation costs, redeployment due to poor effects, and reduced success rate and stability of lighting project deployment.

[0004] Therefore, there is an urgent need to design a new technical solution to solve the above technical problems. Summary of the Invention

[0005] In view of the technical problems existing in the prior art, the present invention provides a method and system for deploying scene lighting based on visual images, which are used to shorten the cumbersome on-site debugging time in traditional lighting deployment, improve work efficiency, avoid the risk of redeployment due to poor effects, and increase the success rate and stability of deployment.

[0006] In a first aspect, an embodiment of the present application provides a method for deploying scene lighting based on visual images, including:

[0007] Scanning the set image of the target scene;

[0008] Based on the set image and the pre-set virtual light source positions, constructing a virtual set model of the target scene; the virtual light source positions correspond one-to-one to the positions of each lighting device in the target scene;

[0009] Identifying the set style of the target scene;

[0010] Generating a corresponding candidate scene lighting scheme in the virtual set model based on the set style, for the user to select the target scene lighting scheme for final display from the candidate scene lighting schemes;

[0011] Generate entity deployment instructions for each lighting device in the target scene based on the target scene lighting scheme, and complete the lighting deployment of the target scene.

[0012] In a second aspect, an embodiment of the present application provides a scene lighting deployment system based on visual images, and the system includes the following units:

[0013] A scanning unit for scanning the scenery image of the target scene;

[0014] A construction unit for constructing a virtual scenery model of the target scene based on the scenery image and the pre-set virtual light source positions; the virtual light source positions correspond one-to-one to the positions where each lighting device is located in the target scene;

[0015] An identification unit for identifying the scenery style of the target scene;

[0016] A generation unit for generating a corresponding candidate scene lighting scheme in the virtual scenery model based on the scenery style, for the user to select the target scene lighting scheme for final display from the candidate scene lighting schemes;

[0017] An execution unit for generating entity deployment instructions for each lighting device in the target scene based on the target scene lighting scheme, and completing the lighting deployment of the target scene.

[0018] In a third aspect, an embodiment of the present application provides an electronic device, and the electronic device includes:

[0019] At least one processor, a memory, and an input / output unit;

[0020] Wherein, the memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the scene lighting deployment method based on visual images in the first aspect.

[0021] In a fourth aspect, a computer-readable storage medium is provided, which includes instructions, and when the instructions are run on a computer, the computer is made to execute the scene lighting deployment method based on visual images in the first aspect.

[0022] The beneficial effects of the present invention are as follows: A method and system for deploying scene lighting based on visual images are provided. In this technical solution, first, the set scenery image of the target scene is scanned. Furthermore, based on the set scenery image and the pre-set virtual light source positions, a virtual set model of the target scene is constructed; the virtual light source positions correspond one-to-one to the positions where each lighting device is located in the target scene. Then, the set style of the target scene is identified. Next, corresponding candidate scene lighting schemes are generated in the virtual set model based on the set style for the user to select the target scene lighting scheme for final display from the candidate scene lighting schemes. Finally, based on the target scene lighting scheme, physical deployment instructions corresponding to each lighting device in the target scene are generated to complete the lighting deployment of the target scene.

[0023] In the embodiments of the present application, by scanning the set scenery image to construct a virtual set model, multiple candidate scene lighting schemes can be quickly generated, greatly shortening the cumbersome on-site debugging time in traditional lighting deployment and improving work efficiency. In the embodiments of the present application, a model is constructed using virtual light sources corresponding to the actual device positions, which can accurately simulate the scene, predict the lighting effects in advance, and ensure accurate and reasonable deployment. In the embodiments of the present application, the set style can be identified and an adapted scheme can be generated to enhance the visual expressiveness of the scene. In the embodiments of the present application, multiple scheme options are provided for the user, fully respecting their wishes and improving satisfaction. In the embodiments of the present application, by generating physical deployment instructions, the virtual-real conversion is realized, ensuring the feasibility of the scheme. In addition, in the embodiments of the present application, the virtual environment design is further optimized, reducing actual errors and resource waste, lowering costs, avoiding the risk of redeployment due to poor effects, and improving the project success rate and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flowchart of a method for deploying scene lighting based on visual images according to an embodiment of the present application;

[0025] Figure 2 is a schematic structural diagram of a system for deploying scene lighting based on visual images according to an embodiment of the present application;

[0026] Figure 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application;

[0027] Figure 4 is a schematic structural diagram of a medium device according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments in the present application without creative efforts belong to the scope of protection of the present application.

[0029] The embodiments of the present application provide a method and system for deploying scene lights based on visual images. In this technical solution, first, the set image of the target scene is scanned. Then, based on the set image and the pre-set virtual light source positions, a virtual set model of the target scene is constructed; the virtual light source positions correspond one-to-one to the positions of each lighting device in the target scene. Next, the set style of the target scene is identified. Then, a corresponding candidate scene lighting scheme is generated in the virtual set model based on the set style for the user to select the final target scene lighting scheme to be displayed from the candidate scene lighting schemes. Finally, based on the target scene lighting scheme, physical deployment instructions corresponding to each lighting device in the target scene are generated to complete the lighting deployment of the target scene.

[0030] In the embodiments of the present application, first, by scanning the set scene image of the target scene and constructing a virtual set model based on it, and then generating a candidate scene lighting scheme, the cumbersome on-site debugging and repeated attempts in traditional lighting deployment are reduced. Compared with manually adjusting lighting devices one by one, this method can quickly generate multiple candidate schemes, and the user only needs to select a suitable scheme from them, greatly shortening the time of lighting deployment and improving the overall work efficiency. Secondly, by using the pre-set virtual light source positions (corresponding one by one to the positions of each lighting device in the target scene) to construct the virtual set model, the actual situation of the target scene can be accurately simulated. Generating and adjusting the lighting scheme in a virtual environment can predict the lighting effect in advance, avoid the situation of having to re-adjust after the actual deployment if the effect is not good, and ensure the accuracy and rationality of the lighting deployment. Thirdly, it can identify the set style of the target scene and generate the corresponding candidate scene lighting scheme based on this style. This enables the lighting deployment to better fit the overall atmosphere and design intention of the scene. Whether it is a modern minimalist style, a retro style or a scene with other special styles, it can provide a matching lighting effect, enhancing the visual expressiveness and attractiveness of the scene. Providing multiple candidate scene lighting schemes for the user enables the user to make a choice according to their own preferences and needs and determine the final target scene lighting scheme to be displayed. This way fully respects the subjective will of the user, gives the user more sense of participation and decision-making power in the lighting deployment process, and improves the user's satisfaction with the final lighting effect. Finally, after determining the target scene lighting scheme from the virtual set model, generate the entity deployment instructions corresponding to each lighting device in the target scene to complete the conversion from virtual to entity. This virtual-real combination method not only utilizes the flexibility and operability of the virtual environment, but also can accurately convert the virtual scheme into an actual lighting deployment, ensuring the feasibility and consistency of the scheme. In addition, designing and optimizing the lighting scheme in a virtual environment reduces the errors and resource waste that may occur in actual operations, reduces the cost of lighting deployment. At the same time, predicting the lighting effect in advance and making adjustments avoids risks such as re-deployment due to poor lighting effects, and improves the success rate and stability of the project.

[0031] The scene lighting deployment scheme based on visual images provided by the embodiments of the present application can also be executed by an electronic device, and the electronic device can be a server, a server cluster, or a cloud server. The electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a dedicated device (such as a dedicated terminal device with a scene lighting deployment system based on visual images, etc.). The above-mentioned chips introduced in the above embodiments can also be installed in these electronic devices. Or, these electronic devices can also install a service program for executing the scene lighting deployment scheme based on visual images.

[0032] Figure 1Schematic diagram of a method for deploying scene lights based on visual images provided by an embodiment of the present application, as Figure 1 shown, the method includes the following steps:

[0033] 101. Scan the set image of the target scene;

[0034] 102. Based on the set image and the pre-set virtual light source positions, construct a virtual set model of the target scene; the virtual light source positions correspond one-to-one to the positions of each lighting device in the target scene;

[0035] 103. Identify the set style of the target scene;

[0036] 104. Generate corresponding candidate scene lighting schemes in the virtual set model based on the set style for the user to select the target scene lighting scheme for final display from the candidate scene lighting schemes;

[0037] 105. Based on the target scene lighting scheme, generate entity deployment instructions for each lighting device in the target scene to complete the lighting deployment of the target scene.

[0038] In the embodiment of the present application, by scanning the set image to construct a virtual set model, multiple candidate scene lighting schemes are quickly generated, greatly shortening the cumbersome on-site debugging time in traditional lighting deployment and improving work efficiency. Using virtual light sources corresponding to the actual device positions to construct the model can accurately simulate the scene, predict the lighting effects in advance, and ensure accurate and reasonable deployment. It can identify the set style and generate an adapted scheme to enhance the visual expressiveness of the scene. Provide multiple scheme options for the user, fully respect their wishes, and improve satisfaction. By generating entity deployment instructions, the virtual-real conversion is realized, ensuring the feasibility of the scheme. Optimize the design in the virtual environment, reduce actual errors and resource waste, reduce costs, avoid the risk of redeployment due to poor effects, and improve the project success rate and stability.

[0039] As an alternative embodiment, in 101, scanning the set image of the target scene includes:

[0040] Obtain the site size and orientation of the target scene; the site orientation is associated with the distribution of the audience positions; set the scanning perspective based on the site size and orientation so that the scanning perspective covers all visible angles in the target scene; control the image acquisition device to obtain the set image based on the scanning perspective.

[0041] In an actual scenario, for example, a circular theater stage serves as the target scene. The radius of the theater venue is 20 meters, and the venue orientation is due south towards the auditorium. Based on the venue size and orientation, technicians calculate that with the stage center as the origin, a scanning perspective is set every 30 degrees, and a total of 12 scanning perspectives are set, so as to cover all visible angles in the target scene. Then, the image acquisition device is installed on a rotatable bracket. According to the set scanning perspectives, through a remote control device, the bracket is rotated to each angle in turn to control the image acquisition device to obtain the set scenery images at different perspectives.

[0042] Through the above steps, the scenery images of the target scene can be obtained comprehensively and accurately. By considering the relationship between the venue size, venue orientation and the distribution of the audience positions, the set scanning perspectives can ensure that all scene pictures that the audience can see are collected, avoiding visual blind spots. The comprehensive scenery images provide a rich and accurate data basis for subsequent construction of the virtual scenery model, which helps to generate a candidate scene lighting scheme that is more in line with the actual needs and more comprehensive in consideration, thereby improving the quality and effect of the entire scene lighting deployment.

[0043] As an alternative embodiment, in 101, controlling the image acquisition device to obtain the scenery image based on the scanning perspective includes:

[0044] Generating a scanning route map based on the scanning perspective; converting the scanning route map into an image acquisition instruction sequence, and loading the image acquisition instruction sequence into the image acquisition device to control the image acquisition device to perform image acquisition operations based on the scanning route map; and / or, generating an interaction instruction based on the scanning route map, and pushing the interaction instruction to the user, so that the user controls the image acquisition device to scan the target scene based on the image acquisition angles and acquisition routes in the scanning route map.

[0045] For example, assume that the target scene is a large exhibition hall, 100 meters long, 80 meters wide and 10 meters high. The scanning perspective is set to collect a horizontal perspective every 10 meters, and a vertical perspective every 2 meters from the ground to the ceiling. Based on these scanning perspectives, the system generates a scanning route map around the exhibition hall, which details the positions and orientations that the image acquisition device should be at different time points.

[0046] The system converts this scanning route map into an image acquisition instruction sequence. For example, the instructions include "move forward 10 meters, rotate 30 degrees clockwise, take a picture", etc. Then, these instruction sequences are loaded into a drone equipped with a high-definition camera. The drone flies along the planned route in the exhibition hall and takes pictures according to the preset instruction sequence to complete the image acquisition work of the entire exhibition hall.

[0047] Similarly, continuing with the above-mentioned exhibition hall scenario as an example, after generating the scanning route map, the interaction instructions are pushed to the operator's tablet computer. The interaction instructions are presented in an intuitive interface, including a dynamic demonstration of the scanning route, detailed information about each image acquisition angle, and the corresponding acquisition route. According to the information displayed on the tablet computer, the operator holds a portable image acquisition device and walks in the exhibition hall according to the instructions of the route map, performing image acquisition operations at the specified positions and angles. For example, at a certain position indicated by the route map, the operator adjusts the image acquisition device to a specific angle and then presses the shooting button.

[0048] Thus, whether it is an automated or interactive method, image acquisition based on the scanning route map can ensure that no key area is missed, guaranteeing that the captured scene images are comprehensive and complete, providing an accurate data basis for subsequent scene analysis and lighting deployment. In the automated method, the image acquisition device operates automatically according to the preset instructions without frequent manual operations by the operator, greatly saving time and labor costs and enabling rapid image acquisition of large-scale scenes. The interactive method provides clear operation guidelines for the operator, reducing repeated or incorrect acquisitions caused by blind operations and improving the acquisition efficiency. The interactive method allows the operator to flexibly adjust the acquisition process according to the actual situation. For example, when encountering temporary obstacles or areas that need to be focused on, appropriate changes can be made on the basis of following the overall scanning route map. The automated method is suitable for scenarios with relatively stable environments and extremely high requirements for acquisition accuracy and efficiency. The two methods meet the image acquisition needs in different scenarios.

[0049] As an optional embodiment, in 102, based on the scene image and the pre-set virtual light source positions, a virtual scene model of the target scene is constructed, including:

[0050] Extracting visual feature information from the scene image; the visual feature information at least includes: color distribution, material, object shape, and spatial layout; automatically learning significant feature points in the target scene based on the visual feature information, and generating corresponding first feature descriptors for the significant feature points; generating second feature descriptors corresponding to each pixel feature point in the target scene based on the visual feature information; mapping the first feature descriptors and the second feature descriptors from the two-dimensional image space to the three-dimensional image space using a neural radiance field network, and constructing the geometric shapes and appearance information corresponding to each scene in the target scene; generating an initial scene model corresponding to the target scene based on the geometric shapes and appearance information corresponding to each scene; associating the positions of each lighting device in the target scene with the initial scene model to obtain the virtual scene model including each virtual light source position.

[0051] In the embodiment of the present application, the first feature descriptor is obtained using a feature point detection and description network based on deep learning.

[0052] Specifically, taking an indoor drama scene as an example, assume that a virtual set model of a living room is to be constructed, and multiple set images of the living room are obtained through an image acquisition device. Here, a classic ORB (Oriented FAST and Rotated BRIEF) algorithm combined with deep learning can be selected (for example, in some improved ORB algorithm variants based on convolutional neural networks). The ORB algorithm is a fast feature point detection and description algorithm that combines the advantages of the FAST (Features from Accelerated Segment Test) feature point detection and the BRIEF (Binary Robust Independent Elementary Features) descriptor, and has characteristics such as rotation invariance and scale invariance. Before inputting the obtained living room set images into the network, some preprocessing operations need to be carried out. For example, the size of the images is uniformly adjusted to a suitable size (such as 640x480 pixels), normalized, and the pixel value range is adjusted to between [0,1], etc., to improve the training and running efficiency of the network. The ORB algorithm is used to detect feature points in the preprocessed images. In the living room images, the algorithm will automatically find some significant feature points, which are usually located at the edges, corners of objects, or places with large texture changes. For example, the corners of the sofa, the edges of the coffee table, and the four corners of the TV screen may all be detected as feature points. In this process, the network will calculate the position coordinates (x,y) of each feature point based on the grayscale information and local structure of the image. For each detected feature point, the corresponding descriptor is generated using the ORB algorithm. The descriptor is a binary vector that contains the feature information of the local area around the feature point and is used to describe the uniqueness of the feature point. In the deep learning-based implementation, further feature extraction and encoding may be performed on the image patches around the feature points through a convolutional neural network (CNN) to generate more discriminative descriptors.

[0053] For example, for a feature point at the corner of the sofa, the descriptor will include the texture information around the point (such as the fabric texture of the sofa), color information, and the relative position relationship with other surrounding feature points, etc. The generated descriptor is a vector of a fixed length (such as a 128-dimensional binary vector), and this vector is the first feature descriptor corresponding to the feature point.

[0054] To improve the quality of feature points and the accuracy of descriptors, the detected feature points and the generated descriptors may be further screened and optimized. For example, some feature points that are too close to the edge or in a noisy area are removed, and the descriptors are normalized or optimized in other forms to ensure that the finally obtained first feature descriptors can accurately represent the significant feature points in the living room scene.

[0055] Through the above steps, the deep learning-based feature point detection and description network can extract effective first feature descriptors from the setting image of the living room, and these descriptors will be used in the subsequent process of constructing the virtual setting model to help accurately locate and describe the key positions and object features in the scene.

[0056] In the embodiment of this application, the second feature descriptor is constructed by using a deep learning-based dense feature description network.

[0057] Taking a restaurant scene in a play as an example, for the scene lighting deployment of a restaurant, the scenery images of all angles of the restaurant have been obtained through an image acquisition device. Next, it is necessary to construct a second feature descriptor. Use a high-definition camera to take multiple images around the restaurant to ensure that all areas of the restaurant are covered, including dining tables, dining chairs, wall decorations, ceiling lights, etc. Uniformly adjust the collected images to the same size, such as 512×512 pixels, and at the same time perform normalization processing to scale the pixel values to between 0 and 1 to improve the stability of subsequent network training and feature extraction. Select a dense feature description network based on the convolutional neural network (CNN) architecture, such as an improved version of the SuperPoint network. This network structure has multiple convolutional layers and pooling layers, which can effectively extract local and global features in the image. In addition to the images of the restaurant, some images of similar scenes (such as other restaurants, cafes, etc.) can be collected as training data to enhance the generalization ability of the network. Input the preprocessed images into the dense feature description network for training. The network gradually adjusts its own weight parameters by learning the relationships between pixels in the image to generate descriptors that can accurately describe the features of each pixel. Use a suitable loss function to measure the difference between the generated descriptor and the real feature, such as a contrast loss function, to prompt the descriptor generated by the network to have a smaller distance between similar pixels and a larger distance between different pixels. Input the preprocessed restaurant images into the trained dense feature description network for forward propagation. The network will process each pixel point in the image and generate a corresponding feature descriptor for it. The generated second feature descriptor is usually a high-dimensional vector, such as 256 dimensions. This vector contains rich information about the local area around the pixel point, such as features like color, texture, and edges. When constructing a virtual scenery model later, the second feature descriptor can be used for feature matching. For example, when stitching restaurant images taken from different angles, by comparing the second feature descriptors of pixel points, corresponding matching points can be found, thus achieving accurate stitching of the images. The second feature descriptor can also help better understand the structure and layout of the restaurant scene. By analyzing the descriptors of pixel points in different regions, different objects in the restaurant (such as dining tables, chairs, walls, etc.) can be identified, and their spatial positions and mutual relationships can be understood.

[0058] Through the dense feature description network based on deep learning, a second feature descriptor with rich information can be generated for each pixel point in the restaurant scene. These descriptors play an important role in subsequent tasks such as scene modeling, image stitching, and scene understanding, providing strong support for accurately constructing a virtual scenery model and achieving precise lighting deployment.

[0059] In the foregoing steps, by way of example, assume that the target scenario is a rectangular meeting room with a length of 10 meters, a width of 8 meters, and a height of 4 meters. Based on this assumption, after obtaining the setting image of the meeting room through an image acquisition device, visual feature information is extracted from the image. For example, it is observed that the wall color is grayish white, the conference table is made of wood, the shape is rectangular, the tables and chairs are neatly arranged in the center of the room in terms of spatial layout, and the projector is installed on the front wall, etc. Furthermore, a feature point detection and description network based on deep learning is used to automatically learn the significant feature points in the meeting room scene. For example, the four corners of the conference table, the junction of the wall and the ceiling, and other significant positions are used to generate corresponding first feature descriptors. These descriptors contain information such as the positions of the significant feature points and the surrounding environmental features, and are used to accurately mark the key positions in the scene. Then, a dense feature description network based on deep learning is adopted to generate second feature descriptors for each pixel feature point in the meeting room image. It can meticulously describe the local features around each pixel point, enabling the model to have a more comprehensive understanding of the details of the scene, such as the details of the wood grain on the conference table and the subtle texture of the wall. Then, the first feature descriptors and the second feature descriptors are input into the neural radiance field network to map from the two-dimensional image space to the three-dimensional image space. In this process, the network constructs the geometric shapes and appearance information corresponding to each object in the meeting room (such as the conference table, chairs, projector, etc.) according to the information of the feature descriptors. For example, determine the length, width, and height dimensions of the conference table and the appearance presentation of its wood texture. Based on the geometric shapes and appearance information corresponding to each object, an initial setting model corresponding to the meeting room is generated. At this time, the model has initially presented the three-dimensional structure of the meeting room and the general form of the internal objects. Finally, since the positions of the lighting devices (such as chandeliers and wall lights) in the meeting room are known, these positions are associated with the initial setting model to obtain a virtual setting model containing the positions of each virtual light source. Thus, a complete virtual setting model of the meeting room containing lighting position information is constructed.

[0060] In this way, by extracting rich visual feature information, including color distribution, material, object shape, and spatial layout, etc., and using an advanced deep learning network to generate feature descriptors at different levels, the true situation of the target scene can be restored with high accuracy. Whether it is the shape of large objects or the subtle material texture, they can all be reflected in the virtual set model, providing a very realistic and reliable basis for subsequent lighting simulation and design. The accurate marking of significant feature points by the first feature descriptor and the detailed description of pixel feature points by the second feature descriptor enable the model to grasp the key positions in the scene and present rich details. This is crucial for accurately arranging lights in a virtual environment and simulating the reflection and scattering effects of lights on different object surfaces, which can greatly improve the accuracy and authenticity of lighting simulation. The neural radiance field network realizes an efficient mapping from two dimensions to three dimensions, quickly constructs the geometric shape and appearance information of the scenery, and generates an initial set model. This deep learning-based method greatly improves the modeling efficiency compared with traditional modeling methods, and the accuracy and quality of the model are higher. Associating the positions of lighting devices with the initial set model to obtain a virtual set model containing the positions of virtual light sources ensures the integrity and accuracy of lighting deployment in a virtual environment. Subsequently, based on this model, the scene lighting effects under different lighting settings can be accurately simulated, providing scientific guidance for actual lighting deployment.

[0061] As an optional embodiment, in 103, identifying the set style of the target scene includes:

[0062] Identifying the visual feature information in the set image; the visual feature information at least includes: color distribution, material, object shape, spatial layout; collecting the scene types and scene functions to which each scene in the target scene belongs; classifying the set style of the target scene based on the visual feature information, the scene types, and the scene functions to obtain candidate set styles; correcting the candidate set styles based on the drama style in the target scene to determine the set style; the drama style is input by the user or extracted from the theater information corresponding to the target scene.

[0063] Exemplarily, assume that the theater stage is preparing for a performance, and the image acquisition device obtains the stage set image. Identify the visual feature information from the image, such as the color of the stage background curtain is dark red, the material of the stage floor is wood, there is a semi-circular wooden stage opening, and there are some tables and chairs props on the stage. The overall spatial layout presents an open structure tilted towards the auditorium.

[0064] The scene types on the stage include curtains, stage openings, tables and chairs, etc. The curtain is used to create an atmosphere and separate scenes. The stage opening is the area boundary for actors to enter and perform. The tables and chairs may be used for the daily life scenes of characters in the play, etc.

[0065] Based on the above visual feature information, scene types, and functions, candidate set design styles are initially classified. A dark red curtain, wooden materials, and an open - space layout, combined with props such as tables and chairs, suggest that the candidate set design style might be a retro style. This style often uses warm colors and natural materials to create a nostalgic atmosphere.

[0066] At this time, the user inputs that the play style of this performance is "Shakespeare's classic tragedy". Considering that Shakespeare's classic tragedies often require a solemn and serious atmosphere, the previous candidate set design style is revised. Although the retro style has a certain sense of nostalgia, it may not be solemn and serious enough. Therefore, the set design style is further determined to be a "classical and solemn retro style". On the basis of being retro, this style emphasizes the creation of a more solemn and solemn atmosphere, meeting the requirements of the play style of Shakespeare's classic tragedies.

[0067] Thus, by comprehensively identifying the visual feature information in the set image, collecting scene types and functions, the set design style can be analyzed from multiple dimensions. This multi - dimensional analysis method avoids the one - sidedness of single - factor judgment and greatly improves the accuracy of set design style recognition. For example, judging only from color may not accurately identify the style. Combining information such as materials, object shapes, and space layouts can more precisely locate the candidate style. Introducing the play style to revise the candidate set design style makes the finally determined set design style closely fit the play theme. This helps to adjust the lighting effects according to the atmosphere required by the play during lighting deployment. For example, for a set with a classical and solemn retro style, soft and warm - yellow - toned lights may be more suitable in terms of lighting to highlight the solemn and nostalgic atmosphere and enhance the overall artistic appeal of the performance. Allowing the user to input the play style or extracting the play style from theater information increases user participation. Users can intervene and adjust the recognition results of the set design style according to the specific needs and creativity of the performance, making the technical application more flexible and better able to adapt to different performance scenarios and artistic creation requirements.

[0068] As an alternative embodiment, in 104, generating a corresponding candidate scene lighting plan in the virtual set model based on the set design style includes:

[0069] The corresponding lighting parameters are set based on the scenery style; the lighting parameters include at least: light source type, lighting intensity, lighting color, lighting angle, and lighting change style; the constraint conditions corresponding to the target scene are set according to the scenery style and scene attributes, and a plurality of candidate scene lighting schemes matching the lighting parameters are constructed using a random search model based on the constraint conditions; virtual lighting parameters corresponding to each virtual light source position are set in each candidate scene lighting scheme; based on the virtual lighting parameters corresponding to each virtual light source position, the corresponding candidate scene lighting effects are rendered in the virtual scenery model; in response to the user's selection instruction for the candidate scene lighting scheme, the corresponding candidate scene lighting effects are displayed to the user so that the user can select any candidate scene lighting scheme as the target scene lighting scheme.

[0070] The specific contents of the lighting parameters in the embodiment of the present application are as follows:

[0071] Light source type: In view of the need to create a soft and natural atmosphere for the classical and solemn retro style, warm-toned spotlights and soft lights are chosen as the main light sources. Spotlights can be used to highlight the actor's performance area, while soft lights are used to evenly illuminate the entire stage background and reduce shadows.

[0072] Light intensity: For the entire stage, set the light intensity of the soft light to a medium to low level, such as 500 lux, to create a soft and warm atmosphere. In the key areas where the actors perform, increase the light intensity to 1500 lux through spotlights to highlight the actors' movements and expressions.

[0073] Lighting color: Use warm yellow as the main color, with a color temperature of about 3000K-3500K, to simulate the effect of traditional candlelight or oil lamps and enhance the retro feel. At the same time, add some dark red auxiliary light behind the background curtain to further enhance the solemn atmosphere.

[0074] Lighting angle: The spotlight is set at an angle of about 45 degrees, illuminating the actors from above on both sides of the stage, which can highlight the three-dimensional sense of the actors while avoiding too heavy shadows. The soft light is evenly distributed above the stage at a horizontal angle to ensure that the background curtain and the stage floor are evenly illuminated.

[0075] Lighting change style: The lighting change style is designed to be slow and smooth. For example, when the scene is switched, the light intensity and color gradually transition to avoid sudden changes in light and dark from causing visual impact on the audience, which is in line with the calm and solemn rhythm pursued by the classical and solemn retro style.

[0076] Based on the above lighting parameters, due to the classical and solemn retro style, the lighting scheme must highlight the solemn and nostalgic atmosphere, and cannot use overly bright, dazzling colors or overly exaggerated lighting effects. The space size of the theater stage is 30 meters long, 20 meters wide and 10 meters high. It is necessary to ensure that the light can evenly cover the entire stage area and will not produce obvious light spots or shadows due to lighting angles or intensity problems. Based on these constraints, a random search model is used to generate multiple candidate scene lighting schemes. For example, in one scheme, the number and position of spotlights and soft lights are adjusted to make the light intensity and color distribution in different areas of the stage slightly different to highlight different performance focuses; in another scheme, the time interval and amplitude of lighting changes are changed to explore different atmosphere creation effects. Each candidate scheme clearly sets the virtual lighting parameters corresponding to each virtual light source position. According to the virtual lighting parameters corresponding to the virtual light source position in each candidate scene lighting scheme, rendering is performed in the virtual set model. For example, in a candidate solution, when the spotlight shines from the left side of the stage at a 45-degree angle, with an intensity of 1500 lux and a warm yellow color (3200K), and the soft light shines horizontally from the top of the stage with an intensity of 500 lux, the rendered stage effect shows that the actors are performing in a bright warm yellow light spot, the background curtain is evenly illuminated by soft warm light, and the dark red auxiliary light is faintly visible behind the curtain, creating a solemn and nostalgic atmosphere. Multiple candidate scene lighting solutions and their rendering effects are presented to the user. After the user views the effects of different solutions through the operation interface, he issues a selection instruction for a candidate scene lighting solution. For example, if the user selects a solution with slower lighting changes and a more stable overall atmosphere, the system immediately shows the user the detailed candidate set lighting effects corresponding to the solution, including the lighting status at different time points, the lighting conditions of various areas of the stage, etc., so that the user can confirm whether to use it as the final target scene lighting solution.

[0077] Thus, by setting the lighting parameters according to the scene style, the generated candidate scene lighting schemes can perfectly match the style of the target scene. In the case of the classical and retro style of an old manor, the warm-colored lighting, soft lighting intensity, and slow lighting change style accurately set off a solemn and nostalgic atmosphere, enhancing the artistic appeal of the scene. Considering the constraints of scene attributes to ensure the feasibility and rationality of the lighting scheme in the actual scene. For example, setting the coverage range and intensity distribution of the lighting according to the space size of the theater stage avoids uneven lighting or visual problems caused by improper lighting settings, ensuring the viewing experience of the audience. Using a random search model to construct multiple candidate scene lighting schemes provides users with a rich selection space. Different schemes show different effects in the subtle differences of lighting parameters. Users can choose the most suitable scheme according to their own creativity and understanding of the performance, enhancing the user's sense of participation and creative autonomy. Rendering the candidate scene lighting effects in the virtual scene model and presenting them to the user in response to user instructions enables the user to intuitively see the presentation effects of different lighting schemes in the actual scene. This visual preview method greatly reduces the difficulty of user decision-making, improves the accuracy and efficiency of selection, and helps to quickly determine the final target scene lighting scheme.

[0078] In the minimalist style, a simple ceiling light is used as the main light source to provide overall uniform lighting, and floor lamps and table lamps with simple shapes are used as auxiliary light sources for local key lighting, such as the sofa reading area and the coffee table display area. The main light source is set to a medium intensity of about 800 - 1000 lux to create a bright and open sense of space. The floor lamps and table lamps of the auxiliary light sources have a lighting intensity of 300 - 500 lux, meeting the local functional requirements while avoiding strong light stimulation. The lighting color is selected as cool white light with a color temperature between 5000K - 6000K, showing the freshness and simplicity of the modern minimalist style and making the space appear brighter and more transparent. The lighting angle is that the ceiling light shines vertically downward to evenly illuminate the living room space. The floor lamp can be adjusted to focus the light on the sofa area for convenient reading; the table lamp can focus the light on the coffee table to illuminate the placed items or decorations. The lighting change style is to use constant lighting without setting complex change effects to maintain a simple and stable atmosphere.

[0079] The fashionable and trendy style in the commercial setting uses track spotlights as the main light source, which can flexibly adjust the position and angle to highlight the clothing on the display rack. It is matched with linear light strips to outline the store outline and the edge of the shelf to increase the sense of space hierarchy. The light intensity of the track spotlight is high, at 1500-2000lux, which strongly highlights the details and colors of the clothing. The linear light strip has a moderate intensity of about 500-800lux, which is used to create an overall atmosphere. The light color chooses white light, with a color temperature of about 4000K-4500K, which truly restores the color of the clothing. At the same time, some colored spotlights, such as light blue and pink, can be added in specific display areas, such as new product display stands, to create a fashionable and dynamic atmosphere. The lighting angle is that the track spotlight illuminates the clothing at different angles to highlight the three-dimensional sense and texture of the clothing. The linear light strip is installed horizontally along the edge of the store and the bottom of the shelf, emitting light upward or downward to enhance the sense of space hierarchy. The lighting change style sets dynamic lighting effects, such as timing the switching of the spotlight brightness in different areas, or letting the linear light strip flash and change, to attract customers' attention and add a sense of fashion and trend.

[0080] In the romantic dream style, a large number of string lights and lanterns are used as the main decorative light sources to create a romantic atmosphere. Some high-power soft lights are used to illuminate the entire venue to ensure that the guests have a clear view. The light intensity of string lights and lanterns is low, about 100-200lux, creating a soft and warm atmosphere. The soft light intensity is moderate, about 600-800lux, providing overall lighting. String lights and lanterns use warm yellow light with a color temperature between 2500K-3000K to create a warm and romantic atmosphere. Soft lights can choose white light with a color temperature of 3500K-4000K, which is matched with warm yellow light to make the overall tone soft and natural. String lights can be hung along the edges of the venue, trees, tents, etc., and emit light at multiple angles. Soft lights are installed around the venue and shine upward at a 45-degree angle to avoid shadows while illuminating the entire venue. The lighting change style sets a slowly changing lighting effect. As the wedding ceremony progresses, the light intensity and color gradually change. For example, when the newlyweds take their vows, the lights are slightly dimmed to highlight the romantic atmosphere; during the celebration, the lights are slightly brightened to increase the joyful atmosphere.

[0081] As an optional embodiment, in 105, after the corresponding candidate scene lighting scheme is generated in the virtual scenery model based on the scenery style, the movement trajectory of the target object in the target scene can also be monitored in real time; a dynamic light tracking model is used to generate a corresponding dynamic fill light scheme based on the movement trajectory and the identity information of the target object; the fill light angle and fill light intensity of the light source in the dynamic fill light scheme are related to the behavior type of the target object; the behavior type is inferred from the movement trajectory and the identity information; the dynamic fill light scheme is added to the target scene lighting scheme to improve the visual effect of the target object in the target scene.

[0082] Suppose in the living room, the target object is an elderly person at home. The elderly person often walks between the sofa and the TV and occasionally picks up items on the coffee table. The actions of the elderly person are captured in real time by a camera installed in the corner of the living room. The system analyzes the video stream and determines the action trajectory of the elderly person, who gets up from the sofa, walks towards the TV to adjust the volume, then returns to the sofa and sits down, with a brief stop at the coffee table in between. Based on the action trajectory and the identity information of the elderly person (for example, considering that the elderly person has relatively poor eyesight), the dynamic light chasing model generates a dynamic supplementary lighting scheme. When the elderly person gets up and walks towards the TV, the floor lamp on one side of the sofa automatically adjusts the supplementary lighting angle, tilts towards the walking path, and the supplementary lighting intensity increases to 400 lux to illuminate the walking route. When the elderly person stops at the coffee table, the supplementary lighting angle of the table lamp above the coffee table focuses on the coffee table area, and the supplementary lighting intensity increases to 350 lux to facilitate the elderly person to clearly see the items. The inferred behavior types are "walking" and "operating items", and the supplementary lighting scheme is adapted accordingly. Thus, adding the dynamic supplementary lighting scheme to the original target scene lighting scheme greatly improves the visual effect when the elderly person is moving in the living room. The elderly person can see the road and items more clearly, reducing the risk of bumping due to insufficient light, while also maintaining the simplicity and stability of the overall lighting in the modern minimalist style.

[0083] In another example, in a clothing store, the target object is a customer who is trying on clothes. The customer frequently walks between the fitting room and the display mirror and has different actions in front of the display mirror, such as turning around and posing. A visual monitoring system composed of multiple cameras installed in the store is used to accurately track the action trajectory of the customer. The process of the customer walking out of the fitting room, walking towards the display mirror, moving around the display mirror, and then returning to the fitting room is recorded. The dynamic light chasing model generates a dynamic supplementary lighting scheme based on the action trajectory and the identity information of the customer (such as the customer is a young female who pays more attention to the clothing display effect). When the customer walks towards the display mirror, the track spotlights installed around the display mirror adjust the supplementary lighting angle, shine light from the side and above, and the supplementary lighting intensity increases to 1800 lux to highlight the three-dimensional sense and style details of the clothing. When the customer turns around and poses in front of the display mirror, the spotlight angle changes with the customer's actions, and the supplementary lighting intensity is adjusted appropriately according to the action rhythm, such as slightly decreasing the intensity when turning slowly and increasing the intensity when posing quickly. The inferred behavior types are "displaying clothes" and "moving", and the supplementary lighting scheme fits them. Thus, after integrating the dynamic supplementary lighting scheme into the original target scene lighting scheme, during the process of the customer trying on clothes, the clothing can be displayed in the best effect, improving the customer's satisfaction and purchase desire for the clothing, while also enhancing the fashionable and trendy atmosphere of the clothing store and attracting more customers' attention.

[0084] In another example, at a wedding ceremony, the target objects are the bride and groom. They move among the stage, the aisle, and the dining tables, performing activities such as taking vows, dancing, and toasting. With the help of multiple high-definition cameras arranged around the venue, the movement trajectories of the bride and groom are monitored comprehensively. Their movement paths are recorded, including walking from the aisle to the stage, taking the vow ceremony on the stage, then walking off the stage to the dining tables to toast, and finally dancing on the dance floor. According to the movement trajectories and the identity information of the bride and groom (the protagonists of the wedding, whose romantic images need to be highlighted), the dynamic spotlight model generates a dynamic supplementary lighting scheme. When the newlyweds walk from the aisle to the stage, the supplementary lighting angles of the string lights and lanterns on both sides of the aisle tilt towards the newlyweds, and the supplementary lighting intensity slightly increases, creating an effect of focusing on the newlyweds. When taking the vow on the stage, high-power soft lights illuminate from the front side at an angle of 30 degrees, and the intensity is stabilized at 700 lux, highlighting the facial expressions of the newlyweds. When dancing on the dance floor, the light color changes with the rhythm of the music, and the supplementary lighting angle and intensity are also dynamically adjusted accordingly. For example, the supplementary lighting is soft during slow music and strong with variable angles during lively music. The behavior types include "entering the venue", "taking vows", "dancing", etc., and the supplementary lighting schemes are precisely matched. Thus, after adding the dynamic supplementary lighting scheme to the original target scene lighting scheme, the newlyweds can be in the best lighting effects in all aspects of the wedding, highlighting the leading role of the newlyweds, enhancing the romantic and dreamy atmosphere of the wedding, and bringing a more unforgettable visual experience to the guests.

[0085] By monitoring the movement trajectories in real time, inferring the behavior types in combination with the identity information, and thus generating a matching dynamic supplementary lighting scheme, it is possible to precisely optimize the visual effects of the target objects in different scenarios. Whether it is to ensure the safety of the elderly in the home scene, improve the display effect of goods in the commercial scene, or highlight the images of the newlyweds in the wedding scene, targeted visual optimization can be achieved. The dynamic supplementary lighting scheme enables the lights to change in real time according to the actions of the target objects, increasing the interactivity between the scene and the target objects. In a clothing store, the lights change with the movements of customers, bringing a unique shopping experience to the customers; at a wedding ceremony, the lights are adjusted according to the activities of the newlyweds, creating a more vivid wedding atmosphere. In each scenario, the dynamic supplementary lighting scheme significantly improves the experiences of the target objects and relevant personnel. The elderly can move more safely and comfortably at home, customers can shop more pleasantly in the clothing store, and guests can feel a stronger romantic atmosphere at the wedding ceremony, comprehensively enhancing the user experiences in different scenarios. The embodiments of this application achieve intelligent lighting control, and at the same time take into account the identity information of the target objects, achieving personalized customization. Target objects with different identities can obtain dynamic supplementary lighting that meets their needs in the same scenario, reflecting the perfect combination of intelligence and personalization.

[0086] In 105, based on the target scene lighting scheme, entity deployment instructions corresponding to each lighting device in the target scene are generated to complete the lighting deployment of the target scene.

[0087] Suppose the target scene lighting plan is determined as follows: Use a ceiling light as the main light source to provide overall lighting, and pair it with a floor lamp and a table lamp for local lighting. The ceiling light emits cold white light at 6000K with a light intensity of 1000 lux; the floor lamp beside the sofa emits neutral white light at 4000K with a light intensity of 300 lux; the table lamp on the coffee table emits warm white light at 3500K with a light intensity of 250 lux. Then, the instructions can be that the ceiling light instruction clearly states that the ceiling light is installed in the exact center of the living room ceiling, wired to the main lighting circuit of the living room, ensuring that the lamp is firmly installed and the lampshade is directed vertically downward. The floor lamp is placed 0.5 meters to the left of the sofa, and the power cord is laid along the corner of the wall and connected to a nearby socket. Adjust the angle of the lamp post so that the light is focused on the reading area of the sofa. The table lamp is placed in the upper left corner of the coffee table, and the power cord is led out from the back of the coffee table and connected to the socket under the coffee table. Adjust the angle of the lampshade to illuminate the coffee table top. The construction workers follow the physical deployment instructions, first install the ceiling light, ensuring that it is firm and the circuit connection is correct. Then place the floor lamp and the table lamp, and complete the power cord laying and angle adjustment. Finally, the entire living room presents a bright and reasonably partitioned lighting effect, in line with the modern minimalist style.

[0088] Thus, through precise physical deployment instructions, the virtual target scene lighting plan is accurately transformed into an actual lighting layout, ensuring that the expected effect of the lighting design is achieved. Clear and explicit instructions reduce the communication cost and operation errors of construction workers, making the lighting deployment process more efficient and shortening the construction time. The detailed requirements in the instructions for the installation location of lamps, circuit connection, etc. ensure the safe installation and use of lighting equipment, reducing potential safety risks.

[0089] The target scene lighting plan in a romantic and dreamy style is as follows: A large number of string lights and lanterns create a romantic atmosphere, and high-power soft lights provide overall lighting. The string lights and lanterns emit warm yellow light at 2700K, and the soft lights emit white light at 3500K with a soft light intensity of 700 lux. Based on the above assumed plan, the physical deployment instructions generated are that the string lights and lanterns instruction stipulates that the string lights are hung along the trees and tent frames around the wedding venue, with a group hung every 0.3 meters and fixed firmly by hooks or ropes. The lanterns are hung on both sides of the aisle, one every 1.5 meters, and fixed with iron wires or special hooks. The power cords of all string lights and lanterns are connected to a unified low-voltage power supply system to ensure electrical safety. The soft lights are installed on the lamp stands around the venue, with the lamp stand height being 2 meters, and the angle of the soft lights is adjusted to shine upward at 45 degrees to cover the entire venue. The power cord of each soft light is connected to an independent power distribution box to ensure stable operation of the lights. The construction workers follow the instructions, first hang the string lights and lanterns, and then install the soft lights and adjust the angles and circuit connections. The wedding scene presents a romantic and dreamy lighting effect, creating a warm atmosphere for the newlyweds and guests.

[0090] In this way, by accurately executing the entity deployment instructions, string lights, lanterns, and soft lights together create a romantic and dreamy wedding atmosphere, meeting the special needs of the wedding scene and adding a touch of romance to the wedding. The instructions are tailored to the characteristics of outdoor venues, with detailed planning for the installation method, fixing method, and circuit connection of the lighting fixtures, ensuring that the lighting equipment can operate stably in the complex outdoor environment and resist the influence of natural factors. The reasonable lighting deployment not only provides good lighting effects but also guarantees the smooth progress of the wedding ceremony and activities, allowing the newlyweds and guests to enjoy the wedding in a comfortable lighting environment.

[0091] In the embodiments of this application, by scanning the scenery image to construct a virtual scenery model, multiple candidate scene lighting schemes are quickly generated, greatly shortening the cumbersome on-site debugging time in traditional lighting deployment and improving work efficiency. In the embodiments of this application, a model is constructed using virtual light sources corresponding to the actual device positions, which can accurately simulate the scene, predict the lighting effects in advance, and ensure accurate and reasonable deployment. The embodiments of this application can identify the scenery style and generate an adapted scheme to enhance the visual expressiveness of the scene. The embodiments of this application provide multiple scheme options for users, fully respecting their wishes and enhancing satisfaction. The embodiments of this application realize the conversion between virtual and real by generating entity deployment instructions, ensuring the feasibility of the scheme. In addition, the embodiments of this application further optimize the design in the virtual environment, reduce actual errors and resource waste, lower costs, avoid the risk of redeployment due to poor effects, and improve the success rate and stability of the project.

[0092] Figure 2 It is a schematic structural diagram of a scene lighting deployment system based on visual images provided by the embodiments of this application. As Figure 2 shown, the system includes the following steps:

[0093] A scanning unit for scanning the scenery image of the target scene;

[0094] A construction unit for constructing a virtual scenery model of the target scene based on the scenery image and the pre-set virtual light source positions; the virtual light source positions correspond one-to-one to the positions of each lighting device in the target scene;

[0095] An identification unit for identifying the scenery style of the target scene;

[0096] A generation unit for generating corresponding candidate scene lighting schemes in the virtual scenery model based on the scenery style for the user to select the target scene lighting scheme for the final display from the candidate scene lighting schemes;

[0097] An execution unit for generating entity deployment instructions corresponding to each lighting device in the target scene based on the target scene lighting scheme to complete the lighting deployment of the target scene.

[0098] Please refer toFigure 3 , Figure 3 is a schematic diagram of an embodiment of an electronic device provided by an embodiment of the present application. As Figure 3 shown, an embodiment of the present application provides an electronic device 500, including a memory 510, a processor 520, and a computer program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 511, the foregoing embodiments are implemented.

[0099] Please refer to Figure 4 , Figure 4 is a schematic diagram of an embodiment of a computer-readable storage medium provided by an embodiment of the present application. As Figure 4 shown, this embodiment provides a computer-readable storage medium 600, on which a computer program 611 is stored. When the computer program 611 is executed by a processor, the foregoing embodiments are implemented.

[0100] It should be noted that in the above embodiments, the descriptions of the respective embodiments have their own emphases. For parts not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0101] Those skilled in the art should understand that the embodiments of the present invention may be provided as a method, a system, or a computer program product. Therefore, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0102] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded computers, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the specified functions in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0103] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the instructions in the flowFigure 1 one process or multiple processes and / or blocks Figure 1 the functions specified in one block or multiple blocks.

[0104] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide for implementing the process Figure 1 one process or multiple processes and / or blocks Figure 1 the steps of the functions specified in one block or multiple blocks.

[0105] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0106] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for deploying scene lighting based on visual images, characterized in that, The method at least includes: Scanning the set image of the target scene; Based on the set image and the pre-set virtual light source positions, constructing a virtual set model of the target scene; the virtual light source positions correspond one-to-one to the positions of each lighting device in the target scene; Identifying the set style of the target scene; Generating a corresponding candidate scene lighting scheme in the virtual set model based on the set style for the user to select the final target scene lighting scheme for display from the candidate scene lighting schemes; Generating entity deployment instructions corresponding to each lighting device in the target scene based on the target scene lighting scheme to complete the lighting deployment of the target scene.

2. The method for deploying scene lighting based on visual images according to claim 1, wherein The scanning of the set image of the target scene includes: Obtaining the site size and site orientation of the target scene; the site orientation is associated with the distribution of the audience positions; Setting a scanning perspective based on the site size and site orientation so that the scanning perspective covers all visible angles in the target scene; Controlling an image acquisition device to obtain the set image based on the scanning perspective.

3. The method for deploying scene lighting based on visual images according to claim 2, wherein The controlling the image acquisition device to obtain the set image based on the scanning perspective includes: Generating a scanning route map based on the scanning perspective; Converting the scanning route map into an image acquisition instruction sequence and loading the image acquisition instruction sequence into the image acquisition device to control the image acquisition device to perform image acquisition operations based on the scanning route map; and / or Generating an interaction instruction based on the scanning route map and pushing the interaction instruction to the user so that the user controls the image acquisition device to scan the target scene based on the image acquisition angles and acquisition route in the scanning route map.

4. The method for deploying scene lighting based on visual images according to claim 1, wherein The identifying the set style of the target scene includes: Identifying the visual feature information in the set image; the visual feature information at least includes: color distribution, material, object shape, spatial layout; Collecting the scene types and scene functions to which each scene in the target scene belongs; Classifying the set style of the target scene based on the visual feature information, the scene types, and the scene functions to obtain candidate set styles; Correcting the candidate set styles based on the drama style in the target scene to determine the set style; the drama style is input by the user or extracted from the theater information corresponding to the target scene.

5. The method for deploying scene lighting based on visual images according to claim 1, wherein, The constructing a virtual set model of the target scene based on the set image and the pre-set virtual light source positions includes: Extracting visual feature information from the set image; the visual feature information at least includes: color distribution, material, object shape, spatial layout; Automatically learning the significant feature points in the target scene based on the visual feature information and generating corresponding first feature descriptors for the significant feature points; wherein, the first feature descriptors are obtained by using a feature point detection and description network based on deep learning; Generating second feature descriptors corresponding to each pixel feature point in the target scene based on the visual feature information; wherein, the second feature descriptors are constructed by using a dense feature description network based on deep learning; Map the first feature descriptor and the second feature descriptor from a two-dimensional image space to a three-dimensional image space using a neural radiance field network, and construct the geometric shapes and appearance information corresponding to each scene in the target scene; Generate an initial scene model corresponding to the target scene based on the geometric shapes and appearance information corresponding to each scene; Associate the positions of each lighting device in the target scene with the initial scene model to obtain the virtual scene model including the positions of each virtual light source.

6. The method for deploying scene lighting based on visual images according to claim 1, wherein The generating corresponding candidate scene lighting schemes in the virtual scene model based on the scene style includes: Set corresponding lighting parameters based on the scene style; the lighting parameters at least include: light source type, light intensity, light color, light angle, light change style; Set the constraint conditions corresponding to the target scene according to the scene style and scene attributes, and construct multiple candidate scene lighting schemes matching the lighting parameters based on the constraint conditions using a random search model; each candidate scene lighting scheme is set with virtual lighting parameters corresponding to the positions of each virtual light source; Render the corresponding candidate scene lighting effects in the virtual scene model based on the virtual lighting parameters corresponding to the positions of each virtual light source; In response to the user's selection instruction for the candidate scene lighting scheme, display the corresponding candidate scene lighting effects to the user so that the user can select any candidate scene lighting scheme as the target scene lighting scheme.

7. The method for deploying scene lighting based on visual images according to claim 6, wherein After generating the corresponding candidate scene lighting schemes in the virtual scene model based on the scene style, it further includes: Real-time monitor the action trajectory of the target object in the target scene; Generate a corresponding dynamic fill light scheme based on the action trajectory and the identity information of the target object using a dynamic follow light model; the light source fill light angle and fill light intensity in the dynamic fill light scheme are both related to the behavior type of the target object; the behavior type is inferred from the action trajectory and the identity information; Add the dynamic fill light scheme to the target scene lighting scheme to improve the visual effect of the target object in the target scene.

8. A scene lighting deployment system based on visual images, characterized in that, The system at least includes the following units: A scanning unit for scanning the scene image of the target scene; A construction unit for constructing a virtual scene model of the target scene based on the scene image and the pre-set virtual light source positions; the virtual light source positions correspond one-to-one to the positions of each lighting device in the target scene; An identification unit for identifying the scene style of the target scene; A generation unit for generating corresponding candidate scene lighting schemes in the virtual scene model based on the scene style for the user to select the final displayed target scene lighting scheme from the candidate scene lighting schemes; An execution unit for generating entity deployment instructions corresponding to each lighting device in the target scene based on the target scene lighting scheme to complete the lighting deployment of the target scene.

9. An electronic device, characterized in that, Include a memory for storing computer software programs; A processor for reading and executing the computer software program, thereby implementing the functions of each component part in the scene lighting deployment system based on visual images according to any one of claims 1-7.

10. A non-transitory computer-readable storage medium, characterized in that, A computer software program is stored in the storage medium, and when the computer software program is executed by a processor, it realizes the functions of each component part in the scene lighting deployment system based on visual images as described in any one of claims 1-7.

Images