Brightness adjustment method, device and electronic equipment of LED fill light

Abstract

The present application relates to the field of brightness adjustment, and specifically discloses a brightness adjustment method, device and electronic device for an LED fill light. The method first obtains an environmental image captured by a camera, ambient light intensity information values ​​at multiple predetermined time points captured by a light sensor and fill light working information collected by a database, and then uses deep learning technology to perform feature extraction and correlation analysis on the three. Finally, a classifier is used to determine whether the brightness of the LED fill light needs to be adjusted, thereby automatically adjusting the brightness according to the real-time environment and working status, thereby improving the accuracy and efficiency of the brightness adjustment of the LED fill light and making the adjustment more in line with actual needs.

Description

Technical Field

[0001] The present application relates to the field of brightness adjustment, and more specifically, to a brightness adjustment method, device and electronic equipment for an LED fill light. Background Art

[0002] LED fill light is a lighting device that uses LED as a light source. It is usually used to supplement insufficient light environments, such as photography, video shooting, makeup and other fields. LED fill light has the advantages of energy saving, environmental protection, long life, high brightness, etc. It can adjust the brightness and color temperature, provide soft and uniform light, help improve the shooting or makeup effect, and is one of the common and practical lighting equipment in the modern lighting industry.

[0003] Although LED fill lights play an important role in outdoor lighting, pathfinding lighting, plant cultivation and photography auxiliary lights, there are some limitations in the current design. The lighting range and power are fixed, and the light intensity, brightness, irradiation distance and range cannot be adjusted according to the ambient brightness and needs, resulting in energy waste and lamp overheating problems. In addition, excessive brightness and illumination may cause serious light pollution, affecting the flexibility and reliability of the lamp.

[0004] Therefore, a method, device and electronic device for adjusting the brightness of an LED fill light are desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a method, device and electronic device for adjusting the brightness of an LED fill light, which first obtains an environmental image captured by a camera, ambient light intensity information values ​​at multiple predetermined time points captured by a light sensor and fill light working information collected by a database, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally uses a classifier to determine whether the brightness of the LED fill light needs to be adjusted, thereby realizing automatic brightness adjustment according to the real-time environment and working status, thereby improving the accuracy and efficiency of the LED fill light brightness adjustment, and making the adjustment more in line with actual needs.

[0006] According to one aspect of the present application, a method for adjusting the brightness of an LED fill light is provided, comprising:

[0007] Acquire the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database;

[0008] Extracting a fill light environment global feature vector and a fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database;

[0009] Associating the fill light environment global feature vector with the fill light working information semantic feature vector to obtain a fill light brightness adjustment classification feature vector;

[0010] Performing attribute distribution supplementary description enhancement based on regression parameter space on the fill light brightness adjustment classification feature vector to obtain an optimized fill light brightness adjustment classification feature vector;

[0011] The optimized fill light brightness adjustment classification feature vector is passed through a classifier to obtain a classification result, and the classification result is used to determine whether the brightness of the LED fill light needs to be adjusted.

[0012] According to another aspect of the present application, a brightness adjustment device for an LED fill light is provided, comprising:

[0013] A fill light brightness data acquisition module is used to acquire the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database;

[0014] A fill light brightness data extraction module is used to extract a fill light environment global feature vector and a fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database;

[0015] A fill light brightness feature association module is used to associate the fill light environment global feature vector with the fill light working information semantic feature vector to obtain a fill light brightness adjustment classification feature vector;

[0016] A fill light brightness feature optimization module, used for performing attribute distribution supplementary description enhancement on the fill light brightness adjustment classification feature vector based on the regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector;

[0017] The fill light brightness adjustment judgment classification module is used to pass the optimized fill light brightness adjustment classification feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the brightness of the LED fill light needs to be adjusted.

[0018] According to another aspect of the present application, an electronic device is provided, comprising: a processor; and a memory, in which computer program instructions are stored, and when the computer program instructions are executed by the processor, the processor executes the brightness adjustment device of the LED fill light as described above.

[0019] Compared with the prior art, the present application provides a method, device and electronic device for adjusting the brightness of an LED fill light, which first obtains an environmental image captured by a camera, ambient light intensity information values ​​at multiple predetermined time points captured by a light sensor, and fill light working information collected by a database, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally uses a classifier to determine whether the brightness of the LED fill light needs to be adjusted, thereby automatically adjusting the brightness according to the real-time environment and working status, thereby improving the accuracy and efficiency of the brightness adjustment of the LED fill light, and making the adjustment more in line with actual needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0021] Figure 1 The present invention is a flowchart of a method for adjusting the brightness of an LED fill light according to an embodiment of the present application.

[0022] Figure 2 The present invention is a flow chart of extracting features from the environment image captured by the camera to obtain a panoramic feature vector of the fill light environment brightness in the brightness adjustment method of the LED fill light according to the embodiment of the present application.

[0023] Figure 3 The present invention is a flowchart of performing data preprocessing on the environment image captured by the camera to obtain brightness histogram images of multiple environment areas in the brightness adjustment method of the LED fill light according to the embodiment of the present application.

[0024] Figure 4 The present invention is a flow chart of performing feature encoding on the brightness histogram images of the plurality of environmental areas in the brightness adjustment method of the LED fill light according to the embodiment of the present application to obtain the panoramic feature vector of the fill light environmental brightness.

[0025] Figure 5 4 is a block diagram of a brightness adjustment device for an LED fill light according to an embodiment of the present application.

[0026] Figure 6 is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0028] Figure 1 FIG. 1 is a flow chart of a method for adjusting the brightness of an LED fill light according to an embodiment of the present application. Figure 1 As shown, the brightness adjustment method of the LED fill light according to the embodiment of the present application includes: S110, acquiring the environment image captured by the camera, the environment light intensity information values ​​at multiple predetermined time points captured by the light sensor, and the fill light working information collected by the database; S120, extracting the fill light environment global feature vector and the fill light working information semantic feature vector from the environment image captured by the camera, the environment light intensity information values ​​at multiple predetermined time points captured by the light sensor, and the fill light working information collected by the database; S130, associating the fill light environment global feature vector with the fill light working information semantic feature vector to obtain a fill light brightness adjustment classification feature vector; S140, enhancing the fill light brightness adjustment classification feature vector with a supplementary description of the attribute distribution based on the regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector; S150, passing the optimized fill light brightness adjustment classification feature vector through a classifier to obtain a classification result, and the classification result is used to determine whether the brightness of the LED fill light needs to be adjusted.

[0029] In the above-mentioned method for adjusting the brightness of the LED fill light, the step S110 obtains the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database. It should be understood that the LED fill light is a lighting device that uses LED as a light source, and is mainly used in environments where additional light is required, such as photography, video shooting, and makeup. Its significant advantages include energy saving, environmental protection, long life, and high brightness. The LED fill light can flexibly adjust the brightness and color temperature, provide soft and uniform light, and help optimize the shooting and makeup effects. Therefore, it is widely used in the modern lighting industry. Although the LED fill light plays an important role in outdoor lighting, pathfinding lighting, plant cultivation, and photography assistance, its design has some limitations. At present, the lighting range and power settings of most fill lights are fixed, and cannot be flexibly adjusted according to the ambient light and specific needs, which leads to low energy efficiency and may cause overheating of the lamp. In addition, excessive brightness and illumination may cause serious light pollution, affecting the reliability and adaptability of the lamp. Therefore, in the technical solution of the present application, by acquiring the environmental image captured by the camera, the ambient light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database, and combining with deep learning technology, it is judged whether the brightness of the LED fill light needs to be adjusted, so as to automatically adjust the brightness of the LED fill light by real-time perception of the environment and monitoring of the working status, thereby improving the accuracy and efficiency of the adjustment, and ensuring that the brightness adjustment can better meet actual usage needs.

[0030] Specifically, the environmental images collected by the camera provide visual information of the current environment. These images can be used to analyze the visual features of the environment, such as brightness, color, and shadow, so as to understand the actual lighting needs indoors or outdoors. Through image processing and computer vision technology, the panoramic feature vector of the ambient brightness required by the fill light can be extracted. The ambient light intensity information values ​​collected by the light sensor at multiple time points record the changes in the intensity of the ambient light in different time periods. These data can reflect the impact of factors such as daytime, nighttime, and light changes on the ambient light intensity, and are an important basis for judging whether fill light is needed. By extracting the time-series correlation feature vector, the trend and periodic changes of the ambient light intensity can be analyzed, and the brightness of the LED fill light can be further accurately adjusted. The fill light working information collected by the database includes the model, number, power, position and other operating parameters of the fill light, as well as its historical working status, irradiation angle and energy consumption data. This information can help the system understand the current working status and performance characteristics of the fill light, and provide a decision-making basis for intelligent regulation. By extracting the semantic feature vector of the fill light working information, the use of the fill light can be effectively managed and optimized, thereby saving energy and extending the life of the equipment. In this way, the lighting effect and user experience can be improved, energy can be effectively saved, and it is in line with the concept of sustainable lighting management.

[0031] In the above-mentioned method for adjusting the brightness of the LED fill light, the step S120 extracts the fill light environment global feature vector and the fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database. In this way, the intelligent management of the LED fill light can be realized through a data-driven method, which can improve the lighting quality and user experience, and can also effectively save energy and maintain equipment, meeting the sustainable development needs of modern lighting management.

[0032] In a specific embodiment of the present application, a global feature vector of the fill light environment and a semantic feature vector of the fill light working information are extracted from the environmental image captured by the camera, the ambient light intensity information values ​​at multiple predetermined time points captured by the light sensor, and the fill light working information collected by the database, including: performing feature extraction on the environmental image captured by the camera to obtain a panoramic feature vector of the fill light environment brightness; performing feature extraction on the ambient light intensity information values ​​at multiple predetermined time points captured by the light sensor to obtain a time-series correlation feature vector of the ambient light intensity; fusing the panoramic feature vector of the fill light environment brightness and the time-series correlation feature vector of the ambient light intensity to obtain the global feature vector of the fill light environment; performing feature extraction on the fill light working information collected by the database to obtain the semantic feature vector of the fill light working information.

[0033] It should be understood that the environmental images captured by the camera are one of the important data sources in the system. These images record visual features such as indoor or outdoor lighting conditions, shadow distribution, color temperature, etc. By extracting features from the images, key information can be obtained, such as the overall light intensity distribution, light uniformity, and brightness changes in specific areas. These features reflect the ambient brightness perceived by the human eye and are the basis for the fill light control system to adjust the brightness according to actual conditions. By analyzing the light intensity and uniformity of different areas, the system can dynamically adjust the amount and direction of the fill light to reduce shadows and increase the overall brightness of the environment. This intelligent feedback mechanism enables the lighting system to adapt to changes in light at different times and weather conditions, thereby improving the user's visual comfort and work efficiency.

[0034] Furthermore, feature extraction is performed on the ambient light intensity information values ​​collected by the light sensor at multiple predetermined time points, and the brightness and color temperature of the lighting equipment can be adjusted according to time changes to provide optimal visual comfort and energy efficiency management. Here, the light sensor can accurately measure the light intensity information in the environment. Collecting these data at multiple predetermined time points can capture the changing trend of ambient light in different time periods of the day. Through feature extraction, the system can extract key statistical features from these time series data, such as average light intensity, maximum light intensity, minimum light intensity, light intensity change rate, etc. These features reflect the temporal and spatial distribution of ambient light and provide a basis for subsequent intelligent control. For example, considering that the light intensity changes during the day and at night are significantly different, the feature vector can capture this periodic change pattern. By analyzing these related features, the system can predict the future trend of light intensity changes, so as to adjust the brightness and color temperature of the lighting equipment in advance to maintain the comfort and visual quality of the indoor environment.

[0035] Furthermore, by integrating the panoramic feature vector of the ambient brightness of the fill light and the time-series correlation feature vector of the ambient light intensity, the advantages of visual perception and time series analysis are combined, and the brightness and color temperature of the fill light can be dynamically adjusted according to the overall lighting conditions and changing trends of the environment, thereby improving the lighting effect and energy utilization efficiency. Among them, the panoramic feature vector of the ambient brightness of the fill light reflects the brightness distribution and lighting uniformity in the environmental image collected by the camera. These feature vectors include information such as light intensity, color temperature, and shadows in different areas, which are the key basis for the system to adjust the fill light according to actual lighting conditions. The time-series correlation feature vector of ambient light intensity is based on the light intensity data collected by the light sensor at different time points. These data reflect the changing pattern of light in a day, including the effects of sunrise, sunset, and weather changes on light intensity. By analyzing these time series data, the system can understand the periodic changes and trends of light, predict future lighting needs, and adjust the brightness and color temperature of the fill light according to real-time conditions to maintain the stability of ambient light and the comfort of users. The fusion of the panoramic feature vector of the fill light ambient brightness and the temporal correlation feature vector of the ambient light intensity means that the system can comprehensively consider the static and dynamic characteristics of the environment to achieve smarter and more precise lighting control.

[0036] In particular, feature extraction of the fill light working information collected in the database can extract key features from a large amount of operation and performance data, helping the system to deeply understand the operating status and working characteristics of the fill light, thereby optimizing its control and maintenance strategy. Among them, the fill light working information in the database contains various key parameters and indicators, such as the model, number, power consumption, brightness adjustment range, working time, irradiation angle, etc. of the fill light. This information records the operation history and performance of the fill light under different conditions. These feature vectors not only include the basic parameters of the fill light, but also can express its behavior patterns and performance characteristics under different working conditions. By analyzing these semantic feature vectors, the system can identify abnormal behavior, predict the risk of equipment failure, and optimize the overall performance of the lighting system.

[0037] Figure 2 The present invention is a flow chart of extracting features from the environment image captured by the camera to obtain a panoramic feature vector of the environment brightness of the LED fill light in the brightness adjustment method of the LED fill light according to the embodiment of the present application. Figure 2 As shown, in a specific embodiment of the present application, feature extraction is performed on the environmental image captured by the camera to obtain a panoramic feature vector of the fill light environment brightness, including: S210, data preprocessing is performed on the environmental image captured by the camera to obtain multiple environmental area brightness histogram images; S220, feature encoding is performed on the multiple environmental area brightness histogram images to obtain the panoramic feature vector of the fill light environment brightness.

[0038] It should be understood that the environmental images collected by the camera contain rich visual information, such as light intensity, color temperature, and shadow distribution. This information changes under different environmental conditions, affecting people's visual perception and work efficiency. Through data preprocessing, the system can divide the original image into multiple regions and extract a brightness histogram for each region. The brightness histogram reflects the distribution of pixels with different brightness levels in each region, helping the system understand the lighting status and uniformity of the area.

[0039] Secondly, the generation of brightness histogram images of multiple environmental areas enables the system to grasp the local characteristics of the environment in more detail. By analyzing these images, the system can identify which areas have large lighting differences or shadow coverage, and then adjust the direction and intensity of the fill light to eliminate uneven lighting distribution and improve the overall lighting quality and visual comfort.

[0040] Furthermore, the brightness histogram images of multiple environmental areas reflect the distribution of light intensity at different spatial locations. Through feature encoding, the system can extract key statistical features from each image, such as average brightness, lighting uniformity, kurtosis and skewness of brightness distribution, etc. These features can describe the lighting characteristics of each area and help the system understand and analyze the lighting conditions of the overall environment. In addition, the generation of the panoramic feature vector of the ambient brightness of the fill light enables the system to comprehensively consider the lighting conditions of the entire space, rather than just being limited to a single area. This panoramic feature vector not only includes the brightness characteristics of each area, but also reflects the spatial relationship and mutual influence between them, providing a basis for achieving more intelligent and personalized lighting adjustment.

[0041] Figure 3 The present invention is a flow chart of performing data preprocessing on the environment image captured by the camera to obtain multiple environment area brightness histogram images in the brightness adjustment method of the LED fill light according to the embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, data preprocessing is performed on the environmental image captured by the camera to obtain multiple environmental area brightness histogram images, including: S211, performing region segmentation on the environmental image captured by the camera to obtain multiple fill light environmental area images; S212, extracting brightness histograms from the multiple fill light environmental area images to obtain the multiple environmental area brightness histogram images.

[0042] It should be understood that the environmental image collected by the camera covers the visual information of the entire lighting space, including light intensity, color temperature, shadows, and lighting uniformity. Through regional segmentation, these complex environmental data can be decomposed into multiple smaller regional images. Each regional image represents a local area in the environment, such as each work area in an office, different corners in a hall, or different areas in a room. This segmentation enables the system to analyze and understand the lighting requirements and characteristics of each specific area more finely. The generation of multiple fill light environmental regional images provides higher flexibility and accuracy for the intelligent lighting control system. By performing independent lighting analysis on each regional image, the system can adjust the brightness, color temperature, and light distribution of the fill light according to the actual needs of each area to ensure that each area has the best visual comfort and working environment. For example, increase the brightness of the fill light in areas where the light intensity is insufficient, or adjust the color temperature of the lamp in areas where color warmth is required.

[0043] Furthermore, the brightness histogram is extracted from multiple fill light environment area images in order to more accurately understand and analyze the lighting conditions in different areas, thereby providing data support for the control and optimization of the intelligent lighting system. It should be understood that the fill light environment area images reflect the specific lighting conditions of different areas. These images capture information such as light distribution, brightness changes, and lighting uniformity in each area through the camera. By extracting these images, the system can decompose the environment into multiple discrete lighting areas, each with unique lighting characteristics and requirements. The brightness histogram is an effective tool for describing the distribution of light intensity in each area. By converting each fill light environment area image into a corresponding brightness histogram, the system can understand the lighting status of each area from a higher dimensional perspective. The brightness histogram expresses the distribution of pixels at different brightness levels, so that the uniformity, intensity, and change trend of the lighting in the area can be quantified.

[0044] Figure 4 The present invention is a flow chart of performing feature encoding on the brightness histogram images of the plurality of environmental areas in the brightness adjustment method of the LED fill light according to the embodiment of the present application to obtain the panoramic feature vector of the fill light environmental brightness. Figure 4 As shown, in a specific embodiment of the present application, feature encoding is performed on the multiple ambient area brightness histogram images to obtain the fill light ambient brightness panoramic feature vector, including: S221, passing the multiple ambient area brightness histogram images through an ambient brightness convolutional neural network as a feature encoder to obtain a fill light ambient brightness feature map; S222, performing global mean pooling on the fill light ambient brightness feature map to obtain the fill light ambient brightness panoramic feature vector.

[0045] It should be understood that the ambient area brightness histogram images reflect the light intensity and distribution of each area. These images are usually generated by the aforementioned data preprocessing and feature extraction steps, and contain detailed statistical information about the illumination, such as average brightness, illumination uniformity, the shape of the brightness distribution, etc. However, the histogram image itself is a low-level representation and cannot directly reflect the structure and global information in the image. The ambient brightness convolutional neural network, as a feature encoder, can convert these low-level brightness histogram images into more expressive and abstract feature maps. Here, the convolutional neural network can gradually extract the spatial relationships and high-level features in the image through multi-layer convolution and pooling operations. For example, the network can learn the illumination change patterns between different regions, edge information, and brightness gradients inside and outside the region. Specifically, each layer of the ambient brightness convolutional neural network used as a feature encoder performs convolution processing, mean pooling processing based on the local feature matrix, and nonlinear activation processing on the input data in the forward pass of the layer, so that the last layer of the ambient brightness convolutional neural network used as a feature encoder outputs the fill light ambient brightness feature map, wherein the input of the ambient brightness convolutional neural network used as a feature encoder is the brightness histogram images of the multiple ambient areas.

[0046] Furthermore, the fill light ambient brightness feature map is globally averaged and pooled in order to transform the complex feature map data into a compact and representative global description for higher-level lighting control and optimization decisions. This process utilizes the pooling operation, which can effectively reduce the dimension of the data while retaining important spatial information and lighting features. Among them, the fill light ambient brightness feature map reflects the lighting conditions of different areas in the entire lighting space. These feature maps are extracted from the original image through a convolutional neural network, which contains the light intensity, distribution pattern and spatial relationship of each area. However, feature maps usually have high dimensions and complex structures, and need to be further processed to achieve efficient real-time response and resource utilization of the system. Global average pooling is a common pooling operation that compresses the entire feature map into a global feature vector by calculating the average value of all pixels in the feature map. This vector represents the overall lighting characteristics of the entire environment, including important information such as average brightness, overall light intensity distribution and spatial uniformity. This global description can effectively capture the comprehensive characteristics of the entire lighting space, thereby reducing the system's computational burden and latency while retaining key information, making lighting control more responsive.

[0047] In a specific embodiment of the present application, feature extraction is performed on the ambient light intensity information values ​​at multiple predetermined time points collected by the light sensor to obtain an ambient light intensity time series associated feature vector, including: constructing the ambient light intensity information values ​​at the multiple predetermined time points collected by the light sensor into multiple ambient light intensity input vectors; passing the multiple ambient light intensity input vectors through an ambient light intensity time series encoder to obtain the ambient light intensity time series associated feature vector.

[0048] It should be understood that the ambient light intensity information value collected by the light sensor is real-time data reflecting the ambient light level at the current time point. These data may include information in multiple dimensions such as light intensity, color temperature, and spectral distribution, which are used to describe the lighting conditions in or around the building at the current moment. By collecting data at multiple predetermined time points, the system can obtain the changes in lighting at different time periods during the day. Constructing multiple ambient light intensity input vectors allows the system to perform in-depth analysis and processing in the time dimension. Each input vector represents the ambient light intensity data at a specific time point (such as morning, noon, evening, etc.). Such time segmentation can help the system better understand the periodic changes, peaks and valleys of light in different time periods, and directional changes in light (such as changes in the direction of sunlight). The generation of multiple ambient light intensity input vectors provides a basis for the response speed and accuracy of the intelligent lighting system. The system can update the working status of the lighting equipment in real time based on these vectors, so that it can make timely adjustments according to the current ambient light conditions to ensure that the indoor light level always meets the requirements.

[0049] Furthermore, the ambient light intensity input vector represents the light intensity data at different time points. These vectors may contain real-time data collected from light sensors, such as light intensity measured at intervals (such as a few minutes or hours). Through vectors at multiple time points, the system can capture the overall change trend and periodic pattern of light over a period of time. In the technical solution of the present application, the ambient light intensity time series encoder, as a key component of time series data processing, can effectively convert these input vectors into time series related feature vectors. Among them, the encoder is usually composed of models such as recurrent neural networks (RNNs), long short-term memory networks (LSTMs) or transformers, which can process sequence data and capture long-term dependencies and dynamic changes therein. In this way, the encoder can extract key features in the time series from continuous light intensity input vectors. These features may involve daily change patterns of light, special light patterns on holidays, and even the impact of weather conditions on light intensity. This dynamic modeling can help the system understand and predict possible changes in future light, so as to make more intelligent and timely lighting adjustments. Specifically, the multiple ambient light intensity input vectors are fully-connected encoded using the fully-connected layer of the ambient light intensity temporal encoder to extract high-dimensional implicit features of the eigenvalues ​​at each position in the multiple ambient light intensity input vectors; and, the multiple ambient light intensity input vectors are one-dimensionally encoded using the one-dimensional convolutional layer of the ambient light intensity temporal encoder to extract high-dimensional implicit correlation features of the correlation between the eigenvalues ​​at each position in the multiple ambient light intensity input vectors.

[0050] In a specific embodiment of the present application, feature extraction is performed on the fill light working information collected from the database to obtain a semantic feature vector of the fill light working information, including: word segmentation of the fill light working information collected from the database to obtain a plurality of fill light working information data items; and passing the plurality of fill light working information data items through a converter-based fill light working information semantic feature encoder to obtain the fill light working information semantic feature vector.

[0051] It should be understood that the working information of the fill light usually includes various data about the operating status of the lamp, power usage, brightness adjustment, color temperature setting, etc. This information is often stored in the database in the form of a large amount of text or structured records, but the direct use of this information for analysis and application may be limited by the complexity and difficulty of data processing. By segmenting the working information of the fill light, the overall information can be disassembled into multiple independent data items or data fields. The data items after word segmentation allow the system to more accurately monitor and manage the operation of each fill light, while providing a more flexible and convenient operation method for further data analysis and processing.

[0052] Furthermore, multiple fill light working information data items usually contain various detailed parameters and status records, such as brightness setting, color temperature adjustment, energy consumption statistics, etc. Although these data items are already structured, they are essentially in the form of data and need to be further converted into semantic representations that can be understood and processed. The transformer-based fill light working information semantic feature encoder can adopt advanced neural network structures such as the Transformer model. These encoders can not only process sequence data, but also capture semantic associations and contextual information between data items. Through the processing of the encoder, each data item can be converted into a vector, which not only contains the numerical features of the data item itself, but also integrates its semantic information in the overall context, so as to more accurately express the specific meaning and contextual relationship represented by the data item. Specifically, the embedding layer of the converter-based fill light working information semantic feature encoder is used to map each of the multiple fill light working information data items into a word embedding vector to obtain a sequence of word embedding vectors; the converter-based Bert model of the converter-based fill light working information semantic feature encoder is used to perform global context semantic encoding on the sequence of word embedding vectors to obtain multiple feature vectors; and, the multiple feature vectors are cascaded to obtain the fill light working information semantic feature vector.

[0053] In the above-mentioned method for adjusting the brightness of the LED fill light, in step S130, the fill light environment global feature vector and the fill light working information semantic feature vector are associated to obtain a fill light brightness adjustment classification feature vector. It should be understood that the fill light environment global feature vector generally includes comprehensive features such as the overall illumination, temperature, and humidity of the current environment. These features can reflect the actual illumination environment in or around the building, and are one of the important factors affecting the brightness adjustment of the lamp. The fill light working information semantic feature vector contains detailed information such as the specific operating status, brightness setting, and color temperature adjustment of each lamp. These information can express the real-time needs and operating status of the lamp in a more in-depth manner in the form of semantic feature vectors. The two types of vectors are associated, that is, the global features of the environment and the semantic features of the working information are integrated to obtain a fill light brightness adjustment classification feature vector. This vector not only comprehensively considers the lighting conditions of the current environment, but also combines the specific operating status and needs of each lamp, so as to intelligently adjust the brightness level of each lamp to meet different scenarios and user needs, and optimize the lighting energy consumption.

[0054] In the above-mentioned method for adjusting the brightness of LED fill light, the step S140 performs a supplementary description enhancement of the fill light brightness adjustment classification feature vector based on the attribute distribution of the regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector. It should be understood that the enhancement of the fill light brightness adjustment classification feature vector based on the attribute distribution of the regression parameter space is intended to further optimize the expression ability and prediction accuracy of the feature vector. This process involves adding more detailed and accurate information to the feature vector through the attribute distribution of the statistical regression model parameters, thereby improving the intelligent lighting system's understanding and response capabilities to the lighting fixture brightness adjustment requirements.

[0055] Among them, the fill light brightness adjustment classification feature vector is enhanced by supplementary description of attribute distribution based on regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector, including: multiplying the fill light brightness adjustment classification feature vector with the classification weight matrix of the classifier to obtain a first intermediate fill light brightness adjustment classification feature vector; using the concat function to process the fill light brightness adjustment classification feature vector and the first intermediate fill light brightness adjustment classification feature vector to obtain a second intermediate fill light brightness adjustment classification feature vector; adding the fill light brightness adjustment classification feature vector and the first intermediate fill light brightness adjustment classification feature vector by position to obtain a third intermediate fill light brightness adjustment classification feature vector; multiplying the first weight matrix by the second intermediate fill light brightness adjustment classification feature vector plus the first bias vector and then passing it through a sigmoid function to obtain a first activation value; The second weight matrix is ​​multiplied by the third intermediate fill light brightness adjustment classification feature vector and then added with the second bias vector, and then passed through the sigmoid function to obtain a second activation value; the mean of the first activation value and the second activation value is calculated to obtain the activation mean; the difference between the first and second activation values ​​is calculated, and the difference is used as the first weighting coefficient and the activation mean is used as the second weighting coefficient, and the second intermediate fill light brightness adjustment classification feature vector and the fill light brightness adjustment classification feature vector are weighted by position to obtain a fourth intermediate fill light brightness adjustment classification feature vector; each eigenvalue of the backward anchor reference vector is subjected to an exponential operation with a natural constant as the base to obtain a fifth intermediate fill light brightness adjustment classification feature vector; the fourth intermediate fill light brightness adjustment classification feature vector is subtracted from the fifth intermediate fill light brightness adjustment classification feature vector by position, and then passed through the ReLU function to obtain an optimized fill light brightness adjustment classification feature vector.

[0056] In particular, in the technical solution of the present application, it is considered that the fill light brightness adjustment classification feature vector is formed by the fusion of multiple features of different sources and types, including the brightness feature of the environmental image, the temporal feature of the ambient light intensity, and the semantic feature of the fill light working information. These features may have different distribution characteristics and statistical properties in nature, resulting in inconsistent distribution of the fused fill light brightness adjustment classification feature vector in the feature space. The ambient light intensity and the fill light working state may change with time, location and environmental conditions. This dynamic change may cause the distribution of the feature vector in the feature space to be unstable. The ambient brightness convolutional neural network and the ambient light intensity temporal encoder may not be able to fully capture all relevant environmental information, or information loss or noise may be introduced during the encoding process, affecting the consistency of the feature vector. When the fill light brightness adjustment classification feature vector is inconsistently distributed in the feature space, it may be difficult for the classifier to accurately learn the boundaries between different categories, especially in areas with sparse distribution in the feature space. This may cause the classifier to have deviations in the predictions in these areas, affecting the accuracy of the classification results. If the classifier is too sensitive to certain features during the training process, it may lead to overfitting; and if the classifier cannot capture all important feature information, it may lead to underfitting. Based on this, the fill light brightness adjustment classification feature vector is enhanced by supplementary description of attribute distribution based on regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector.

[0057] In particular, the fill light brightness adjustment classification feature vector is enhanced by supplementing the description of the attribute distribution based on the regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector, including:

[0058] The following optimization formula is used to perform attribute distribution supplementary description enhancement based on the regression parameter space on the fill light brightness adjustment classification feature vector, wherein the optimization formula is: ;in, represents the classification feature vector of the fill light brightness adjustment, represents the classification weight matrix of the classifier, represents matrix multiplication, represents a cascade function, is the backward anchor reference vector, represents the first weight matrix, represents the first bias vector, represents the first activation value, represents the second weight matrix, represents the second bias vector, represents the second activation value, represents the logistic function, It means subtracting by position. represents the linear rectification function, represents the natural exponential function, Represents the classification feature vector for optimizing fill light brightness adjustment.

[0059] That is, in the technical solution of the present application, the manifold geometric consistency of the overall feature distribution of the fill light brightness adjustment classification feature vector is poor, resulting in a long-range distribution regression deviation across classifiers when it is classified by the classifier, affecting the accuracy of the classification result. Therefore, in the technical solution of the present application, the fill light brightness adjustment classification feature vector is enhanced by a supplementary description of the attribute distribution based on the regression parameter space, and the classification weight matrix of the classifier is used to perform an auxiliary description of the attribute distribution of the fill light brightness adjustment classification feature vector to support the descriptiveness of the improved fill light brightness adjustment classification feature vector for the different distance feature descriptions of the classification weight matrix of the classifier to the category probability of the preset classification, wherein the backward anchor reference vector is used as an offset and activated by an activation operation to maintain the reinforcement of the distribution description dependency with a positive effect, so that the manifold geometric consistency of the overall attribute distribution of the fill light brightness adjustment classification feature vector can be significantly improved, thereby reducing the class probability distribution regression deviation of the fill light brightness adjustment classification feature vector when it passes through the classifier, so as to improve the accuracy of the classification result.

[0060] In the above-mentioned method for adjusting the brightness of the LED fill light, in step S150, the optimized fill light brightness adjustment classification feature vector is passed through a classifier to obtain a classification result, and the classification result is used to determine whether the brightness of the LED fill light needs to be adjusted. It should be understood that the optimized fill light brightness adjustment classification feature vector already contains a comprehensive expression of the global characteristics of the environment and the fill light working information. These feature vectors are enhanced by the aforementioned attribute distribution supplementary description, which more accurately reflects the current working state of each lamp and the actual lighting requirements of the environment. As a kind of machine learning model, the classifier can learn a prediction model based on the known feature vector and its corresponding classification label (brightness adjustment is required or not), so as to support the system to adjust the brightness of the lamp in time according to the dynamic changes of the environment to maintain the best lighting effect, optimize energy consumption, avoid unnecessary energy waste, and improve the level of energy efficiency management.

[0061] Figure 5 FIG. 1 is a block diagram of a brightness adjustment device for an LED fill light according to an embodiment of the present application. Figure 5As shown, according to the brightness adjustment device 110 of the LED fill light of the embodiment of the present application, it includes: a fill light brightness data acquisition module 110, which is used to obtain the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database; a fill light brightness data extraction module 120, which is used to extract the fill light environment global feature vector and the fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database; the fill light brightness The feature association module 130 is used to associate the global feature vector of the fill light environment with the semantic feature vector of the fill light working information to obtain a fill light brightness adjustment classification feature vector; the fill light brightness feature optimization module 140 is used to enhance the fill light brightness adjustment classification feature vector based on the attribute distribution supplementary description of the regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector; the fill light brightness adjustment judgment classification module 150 is used to pass the optimized fill light brightness adjustment classification feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the brightness of the LED fill light needs to be adjusted.

[0062] As described above, the brightness adjustment device 100 of the LED fill light according to the embodiment of the present application can be implemented in various terminal devices. In one example, the brightness adjustment device 100 of the LED fill light can be integrated into the terminal device as a software module and / or a hardware module. For example, the brightness adjustment device 100 of the LED fill light can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the brightness adjustment device 100 of the LED fill light can also be one of the many hardware modules of the terminal device.

[0063] Alternatively, in another example, the brightness adjustment device 100 of the LED fill light and the terminal device may also be separate devices, and the brightness adjustment device 100 of the LED fill light may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0064] Here, those skilled in the art can understand that the specific operations of each step in the brightness adjustment device of the LED fill light have been described in the above reference. Figures 1 to 4 The brightness adjustment method of the LED fill light has been described in detail, and therefore, its repeated description will be omitted.

[0065] Below, reference Figure 6 To describe an electronic device according to an embodiment of the present application.

[0066] like Figure 6 As shown, the electronic device 10 includes an input device 11, an input interface 12, a central processing unit 13, a memory 14, an output interface 15, an output device 16, and a bus 17. The input interface 12, the central processing unit 13, the memory 14, and the output interface 15 are connected to each other through the bus 17, and the input device 11 and the output device 16 are connected to the bus 17 through the input interface 12 and the output interface 15, respectively, and then connected to other components of the electronic device 10.

[0067] Specifically, the input device 11 receives input information from the outside and transmits the input information to the central processing unit 13 through the input interface 12; the central processing unit 13 processes the input information based on the computer executable instructions stored in the memory 14 to generate output information, temporarily or permanently stores the output information in the memory 14, and then transmits the output information to the output device 16 through the output interface 15; the output device 16 outputs the output information to the outside of the electronic device 10 for user use.

[0068] In one embodiment, Figure 6 The electronic device 10 shown can be implemented as a network device, which may include: a memory configured to store programs; a processor configured to run the programs stored in the memory to execute any one of the brightness adjustment methods of the LED fill light described in the above embodiments.

[0069] According to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network, and / or installed from a removable storage medium.

[0070] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0071] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present application, but the present application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of the present application, and these modifications and improvements are also considered to be within the scope of protection of the present application.

Claims

1. A method for adjusting the brightness of an LED fill light, characterized in that: include: Acquire the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database; Extracting a fill light environment global feature vector and a fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database; Associating the fill light environment global feature vector with the fill light working information semantic feature vector to obtain a fill light brightness adjustment classification feature vector; Performing attribute distribution supplementary description enhancement based on regression parameter space on the fill light brightness adjustment classification feature vector to obtain an optimized fill light brightness adjustment classification feature vector; Passing the optimized fill light brightness adjustment classification feature vector through a classifier to obtain a classification result, wherein the classification result is used to determine whether the brightness of the LED fill light needs to be adjusted; The method of performing attribute distribution supplementary description enhancement based on the regression parameter space on the fill light brightness adjustment classification feature vector to obtain an optimized fill light brightness adjustment classification feature vector includes: Multiplying the fill light brightness adjustment classification feature vector by the classification weight matrix of the classifier to obtain a first intermediate fill light brightness adjustment classification feature vector; Using a concat function, the fill light brightness adjustment classification feature vector and the first intermediate fill light brightness adjustment classification feature vector are processed to obtain a second intermediate fill light brightness adjustment classification feature vector; Add the fill light brightness adjustment classification feature vector and the first intermediate fill light brightness adjustment classification feature vector according to position to obtain a third intermediate fill light brightness adjustment classification feature vector; The first weight matrix is ​​multiplied by the second intermediate fill light brightness adjustment classification feature vector and then added with the first bias vector and then passed through a sigmoid function to obtain a first activation value; The second weight matrix is ​​multiplied by the third intermediate fill light brightness adjustment classification feature vector and then added with the second bias vector and passed through a sigmoid function to obtain a second activation value; Calculating a mean of the first activation value and the second activation value to obtain an activation mean; Calculate a difference value minus the activation mean, use the difference value as a first weighting coefficient and the activation mean value as a second weighting coefficient, and weight the second intermediate fill light brightness adjustment classification feature vector and the fill light brightness adjustment classification feature vector by position to obtain a fourth intermediate fill light brightness adjustment classification feature vector; Performing an exponential operation with a natural constant as the base on each eigenvalue of the backward anchor reference vector to obtain a fifth intermediate fill light brightness adjustment classification eigenvector; The fourth intermediate fill light brightness adjustment classification feature vector is subtracted from the fifth intermediate fill light brightness adjustment classification feature vector by position and then passed through a ReLU function to obtain an optimized fill light brightness adjustment classification feature vector.

2. The brightness adjustment method of the LED fill light according to claim 1, characterized in that: Extracting a fill light environment global feature vector and a fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database, including: Extracting features from the environment image captured by the camera to obtain a panoramic feature vector of the fill light environment brightness; Performing feature extraction on the ambient light intensity information values ​​at a plurality of predetermined time points collected by the light sensor to obtain a time series associated feature vector of the ambient light intensity; The fill light environment brightness panoramic feature vector and the ambient light intensity time series correlation feature vector are combined to obtain the fill light environment global feature vector; Feature extraction is performed on the fill light working information collected from the database to obtain a semantic feature vector of the fill light working information.

3. The brightness adjustment method of the LED fill light according to claim 2, characterized in that: Extracting features from the environment image captured by the camera to obtain a panoramic feature vector of the fill light environment brightness includes: Performing data preprocessing on the environment image captured by the camera to obtain a plurality of environment area brightness histogram images; Feature encoding is performed on the multiple environment area brightness histogram images to obtain the fill light environment brightness panoramic feature vector.

4. The brightness adjustment method of the LED fill light according to claim 3, characterized in that: Performing data preprocessing on the environment image captured by the camera to obtain a plurality of environment area brightness histogram images, including: Performing region segmentation on the environment image captured by the camera to obtain a plurality of fill light environment region images; A brightness histogram is extracted from the multiple fill light environment area images to obtain the multiple environment area brightness histogram images.

5. The brightness adjustment method of the LED fill light according to claim 4, characterized in that: Performing feature encoding on the plurality of environment area brightness histogram images to obtain the fill light environment brightness panoramic feature vector comprises: Passing the plurality of ambient area brightness histogram images through an ambient brightness convolutional neural network as a feature encoder to obtain a fill light ambient brightness feature map; The fill light environment brightness feature map is globally mean pooled to obtain the fill light environment brightness panoramic feature vector.

6. The brightness adjustment method of the LED fill light according to claim 5, characterized in that: Extracting features of the ambient light intensity information values ​​at a plurality of predetermined time points collected by the light sensor to obtain a temporal correlation feature vector of the ambient light intensity includes: Constructing the ambient light intensity information values ​​at multiple predetermined time points collected by the light sensor into multiple ambient light intensity input vectors; The multiple ambient light intensity input vectors are passed through an ambient light intensity time sequence encoder to obtain the ambient light intensity time sequence associated feature vector.

7. The brightness adjustment method of the LED fill light according to claim 6, characterized in that: Extracting features of the fill light working information collected from the database to obtain a semantic feature vector of the fill light working information includes: Segmenting the fill light working information collected from the database to obtain a plurality of fill light working information data items; The multiple fill light working information data items are passed through a fill light working information semantic feature encoder based on a converter to obtain the fill light working information semantic feature vector.

8. A brightness adjustment device for an LED fill light, characterized in that: include: A fill light brightness data acquisition module is used to acquire the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database; A fill light brightness data extraction module is used to extract a fill light environment global feature vector and a fill light working information semantic feature vector from the environment image collected by the camera, the environment light intensity information values ​​at multiple predetermined time points collected by the light sensor, and the fill light working information collected by the database; A fill light brightness feature association module is used to associate the fill light environment global feature vector with the fill light working information semantic feature vector to obtain a fill light brightness adjustment classification feature vector; A fill light brightness feature optimization module, used for performing attribute distribution supplementary description enhancement on the fill light brightness adjustment classification feature vector based on the regression parameter space to obtain an optimized fill light brightness adjustment classification feature vector; A fill light brightness adjustment judgment classification module is used to pass the optimized fill light brightness adjustment classification feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the brightness of the LED fill light needs to be adjusted; The method of performing attribute distribution supplementary description enhancement based on the regression parameter space on the fill light brightness adjustment classification feature vector to obtain an optimized fill light brightness adjustment classification feature vector includes: Multiplying the fill light brightness adjustment classification feature vector by the classification weight matrix of the classifier to obtain a first intermediate fill light brightness adjustment classification feature vector; Using a concat function, the fill light brightness adjustment classification feature vector and the first intermediate fill light brightness adjustment classification feature vector are processed to obtain a second intermediate fill light brightness adjustment classification feature vector; Add the fill light brightness adjustment classification feature vector and the first intermediate fill light brightness adjustment classification feature vector according to position to obtain a third intermediate fill light brightness adjustment classification feature vector; The first weight matrix is ​​multiplied by the second intermediate fill light brightness adjustment classification feature vector and then added with the first bias vector and then passed through a sigmoid function to obtain a first activation value; The second weight matrix is ​​multiplied by the third intermediate fill light brightness adjustment classification feature vector and then added with the second bias vector and passed through a sigmoid function to obtain a second activation value; Calculating a mean of the first activation value and the second activation value to obtain an activation mean; Calculate a difference value minus the activation mean, use the difference value as a first weighting coefficient and the activation mean value as a second weighting coefficient, and weight the second intermediate fill light brightness adjustment classification feature vector and the fill light brightness adjustment classification feature vector by position to obtain a fourth intermediate fill light brightness adjustment classification feature vector; Performing an exponential operation with a natural constant as the base on each eigenvalue of the backward anchor reference vector to obtain a fifth intermediate fill light brightness adjustment classification eigenvector; The fourth intermediate fill light brightness adjustment classification feature vector is subtracted from the fifth intermediate fill light brightness adjustment classification feature vector by position and then passed through a ReLU function to obtain an optimized fill light brightness adjustment classification feature vector.

9. An electronic device, characterized in that: include: at least one processor; And, a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the brightness adjustment method of the LED fill light as described in any one of claims 1 to 7.

Images