Lamp strip self-adaptive adjusting method and device and storage medium

Abstract

The invention discloses a lamp strip adaptive adjustment method and device and a storage medium, and the method comprises the following steps: collecting multi-source sensing data in real time through a multi-mode sensor network, the multi-source sensing data comprising environment illumination data, spatial distribution data and temperature data; performing feature fusion and dynamic weighting on the multi-source sensing data, establishing an illumination change prediction model based on an LSTM network, and generating pre-adjustment parameters through the illumination change prediction model; and carrying out adaptive adjustment on the brightness and color temperature of the lamp strip according to the pre-adjustment parameters. The environment data is collected in real time through the multi-mode sensor network, feature fusion is performed by adopting a space-time attention mechanism, the illumination change trend is predicted based on the LSTM network, smooth adjustment is realized by applying the fuzzy PID algorithm, and the problems of single environment perception, adjustment response delay, insufficient heat management and the like in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of LED strip control technology, and in particular to an adaptive adjustment method, device and storage medium for LED strips. Background Technology

[0002] LED light strips, as a new type of lighting device, have received widespread attention and application due to their high efficiency, long lifespan, and dimmability. However, traditional LED light strips only offer a single color temperature output, which cannot meet the diverse lighting needs of users in different environments and scenarios. A suitable color temperature can provide a comfortable lighting environment, reduce visual fatigue, and have a positive impact on users' visual health. It can also enhance the sense of space and the quality of the living environment, improve work efficiency and quality, and reduce the energy consumption of the lighting system, thus achieving energy conservation and emission reduction.

[0003] LED lights typically have a constant brightness, meaning they provide the same light intensity regardless of the ambient light level. This results in weak light when strong light is needed, making it difficult for users to see, and strong light when strong light is not needed, causing eye discomfort and increasing the energy consumption of the LED lights. Summary of the Invention

[0004] The main objective of this invention is to provide a method, device, and storage medium for adaptive adjustment of LED strips, in order to solve the aforementioned technical problems.

[0005] To achieve the above objectives, the present invention provides an adaptive adjustment method for LED strips.

[0006] The adaptive adjustment method for the LED strip includes the following steps:

[0007] Multi-source sensing data is collected in real time through a multi-modal sensor network. The multi-source sensing data includes ambient light data, spatial distribution data, and temperature data.

[0008] The multi-source sensing data is fused and dynamically weighted, and an illumination change prediction model is established based on an LSTM network. Pre-adjustment parameters are generated through the illumination change prediction model.

[0009] The brightness and color temperature of the light strip are adaptively adjusted according to the pre-adjustment parameters.

[0010] In one embodiment, the step of acquiring multi-source sensing data in real time through a multimodal sensor network includes:

[0011] A three-level parallel feature extraction branch is adopted to process multi-source sensing data at different time scales;

[0012] The processed multi-source sensing data is fused with attention features in both spatial and channel dimensions.

[0013] In one embodiment, the step of fusing attention features in the spatial and channel dimensions includes:

[0014] The spatial attention submodule calculates the correlation matrix between spatial locations and outputs the weighted spatial features.

[0015] The channel weight vector is calculated through the channel attention submodule, and the channel weighted features are output.

[0016] The spatial and channel attention results are integrated through a gating mechanism.

[0017] In one embodiment, the step of calculating the correlation matrix between spatial locations and outputting the weighted spatial features through the spatial attention submodule includes:

[0018] The correlation matrix A between spatial locations is calculated using the following formula. ij :

[0019]

[0020] in:

[0021] A ij This is the spatial attention weight matrix;

[0022] Q is the query vector matrix with dimensions [N×d];

[0023] K is the key vector matrix with dimensions [N×d];

[0024] d is the dimension of the feature vector, which can be 64 or 128.

[0025] The weighted spatial feature V′ is calculated using the following formula. i :

[0026]

[0027] in:

[0028] V is a value vector matrix.

[0029] In one embodiment, the step of calculating the channel weight vector through the channel attention submodule and outputting the channel weighted features includes:

[0030] The channel weight vector W is calculated using the following formula. c :

[0031] w c =σ(W2δ(W1GAP(F)));

[0032] in:

[0033] F is the input feature map with dimensions [C×H×W];

[0034] GAP stands for Global Average Pooling.

[0035] W1 and W2 are the weight matrices of the fully connected layer;

[0036] δ is the ReLU activation function;

[0037] σ is the Sigmoid function;

[0038] The channel-weighted feature F′ is output using the following formula. c :

[0039] F′ c =w′ c ·F c .

[0040] In one embodiment, the step of fusing spatial and channel attention results via a gating mechanism includes:

[0041] The spatial and channel attention results are fused using the following formula:

[0042] F out =αV′ i +(1-α)W c ;

[0043] Where α is a learnable fusion weight parameter.

[0044] In addition, to achieve the above objectives, the present invention also provides a light strip adaptive adjustment method, which includes: a memory, a processor, and a light strip adaptive adjustment program stored in the memory and executable on the processor. When the light strip adaptive adjustment program is executed by the processor, it implements the steps of the light strip adaptive adjustment method as described above.

[0045] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing a light strip adaptive adjustment program, which, when executed by a processor, implements the steps of the light strip adaptive adjustment method described above.

[0046] The beneficial effects achievable by this invention are as follows: An adaptive adjustment method for LED strips proposed in this embodiment of the invention collects multi-source sensing data in real time through a multi-modal sensor network. This multi-source sensing data includes ambient light data, spatial distribution data, and temperature data. Feature fusion and dynamic weighting are performed on the multi-source sensing data, and a light change prediction model is established based on an LSTM network. Pre-adjustment parameters are generated through the light change prediction model. The brightness and color temperature of the LED strip are adaptively adjusted according to these pre-adjustment parameters.

[0047] This application acquires environmental data in real time through a multimodal sensor network, uses a spatiotemporal attention mechanism for feature fusion, predicts the trend of illumination change based on an LSTM network, and applies a fuzzy PID algorithm to achieve smooth adjustment, thus solving the problems of single environmental perception, sluggish adjustment response, and insufficient thermal management in existing technologies. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the hardware operating environment of the device involved in the embodiments of the present invention;

[0049] Figure 2 This is a flowchart illustrating the first embodiment of the adaptive adjustment method for LED strips of the present invention.

[0050] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0051] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0052] like Figure 1 As shown, Figure 1 This is a schematic diagram of the terminal structure of the hardware operating environment involved in the embodiments of the present invention.

[0053] The terminal in this invention embodiment can be a PC, or a smartphone, tablet computer, e-book reader, MP3 (Moving Picture Experts Group Audio Layer III) player, MP4 (Moving Picture Experts Group Audio Layer IV) player, portable computer, or other portable terminal devices with display functions.

[0054] like Figure 1As shown, the terminal may include: a processor 1001, such as a CPU; a communication bus 1002; a user interface 1003; a network interface 1004; and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0055] Optionally, the terminal may also include a camera, RF (Radio Frequency) circuitry, sensors, audio circuitry, a WiFi module, and so on. Sensors may include light sensors, motion sensors, and other sensors. Specifically, light sensors may include ambient light sensors and proximity sensors. The ambient light sensor can adjust the display brightness according to the ambient light level, while the proximity sensor can turn off the display and / or backlight when the mobile terminal is moved to the ear. As a type of motion sensor, a gravity accelerometer can detect the magnitude of acceleration in various directions (generally three axes). When stationary, it can detect the magnitude and direction of gravity, and can be used for applications that identify the mobile terminal's posture (such as landscape / portrait switching, related games, magnetometer posture calibration), vibration recognition functions (such as pedometers, taps), etc. Of course, the mobile terminal may also be equipped with other sensors such as gyroscopes, barometers, hygrometers, thermometers, and infrared sensors, which will not be elaborated here.

[0056] Those skilled in the art will understand that Figure 1 The terminal structure shown does not constitute a limitation on the terminal and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0057] like Figure 1 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an adaptive adjustment program for the LED strip.

[0058] exist Figure 1 In the terminal shown, network interface 1004 is mainly used to connect to the backend server and communicate with it; user interface 1003 is mainly used to connect to the client (user terminal) and communicate with it; while processor 1001 can be used to call the LED strip adaptive adjustment program stored in memory 1005 and perform the following operations:

[0059] Multi-source sensing data is collected in real time through a multi-modal sensor network. The multi-source sensing data includes ambient light data, spatial distribution data, and temperature data.

[0060] The multi-source sensing data is fused and dynamically weighted, and an illumination change prediction model is established based on an LSTM network. Pre-adjustment parameters are generated through the illumination change prediction model.

[0061] The brightness and color temperature of the light strip are adaptively adjusted according to the pre-adjustment parameters.

[0062] Furthermore, the processor 1001 can call the LED strip adaptive adjustment program stored in the memory 1005 and also perform the following operations:

[0063] A three-level parallel feature extraction branch is adopted to process multi-source sensing data at different time scales;

[0064] The processed multi-source sensing data is fused with attention features in both spatial and channel dimensions.

[0065] Furthermore, the processor 1001 can call the LED strip adaptive adjustment program stored in the memory 1005 and also perform the following operations:

[0066] The spatial attention submodule calculates the correlation matrix between spatial locations and outputs the weighted spatial features.

[0067] The channel weight vector is calculated through the channel attention submodule, and the channel weighted features are output.

[0068] The spatial and channel attention results are integrated through a gating mechanism.

[0069] Furthermore, the processor 1001 can call the LED strip adaptive adjustment program stored in the memory 1005 and also perform the following operations:

[0070] The correlation matrix A between spatial locations is calculated using the following formula. ij :

[0071]

[0072] in:

[0073] A ij This is the spatial attention weight matrix;

[0074] Q is the query vector matrix with dimensions [N×d];

[0075] K is the key vector matrix with dimensions [N×d];

[0076] d is the dimension of the feature vector, which can be 64 or 128.

[0077] The weighted spatial feature V′ is calculated using the following formula. i :

[0078]

[0079] in:

[0080] V is a value vector matrix.

[0081] Furthermore, the processor 1001 can call the LED strip adaptive adjustment program stored in the memory 1005 and also perform the following operations:

[0082] The channel weight vector W is calculated using the following formula. c :

[0083] w c =σ(W2δ(W1GAP(F)));

[0084] in:

[0085] F is the input feature map with dimensions [C×H×W];

[0086] GAP stands for Global Average Pooling.

[0087] W1 and W2 are the weight matrices of the fully connected layer;

[0088] δ is the ReLU activation function;

[0089] σ is the Sigmoid function;

[0090] The channel-weighted feature F′ is output using the following formula. c :

[0091] F′ c =w c ·F c .

[0092] Furthermore, the processor 1001 can call the LED strip adaptive adjustment program stored in the memory 1005 and also perform the following operations:

[0093] The spatial and channel attention results are fused using the following formula:

[0094] F out =αV′ i +(1-α)W c ;

[0095] Where α is a learnable fusion weight parameter.

[0096] The specific embodiments of the present invention using data storage devices are basically the same as the embodiments of the adaptive adjustment method for light strips described below, and will not be repeated here.

[0097] Reference Figure 2 The first embodiment of the present invention provides a method for adaptive adjustment of a light strip, the method comprising:

[0098] Step S10: Collect multi-source sensing data in real time through a multi-modal sensor network. The multi-source sensing data includes ambient light data, spatial distribution data, and temperature data.

[0099] In this embodiment, a flexible COB light strip with a width of 8mm can be used, with 60 LEDs per meter and a sensing node deployed every 3 meters. The node includes a TSL25911 spectral sensor, a VL53L8CX spatial sensor, and an MLX90640 thermal imager. The control terminal uses an STM32H743VI as the main controller and supports CAN bus networking.

[0100] Specifically, data can be collected by collecting environmental data at preset intervals. For example, environmental data can be collected every 100ms to form a multi-dimensional time series input.

[0101] Step S20: Perform feature fusion and dynamic weighting on the multi-source sensing data, and establish an illumination change prediction model based on the LSTM network. Generate pre-adjustment parameters through the illumination change prediction model.

[0102] Specifically, a three-level parallel feature extraction branch can be adopted to process multi-source sensing data at different time scales. For example, the illumination data can be divided into three parallel input streams at different time scales (100ms, 1s, 10s), and each input stream can extract local features through an independent convolutional layer.

[0103] The processed multi-source sensing data is then fused with attention features in both spatial and channel dimensions.

[0104] Specifically, for illumination data of different resolutions, a three-level parallel feature extraction branch is designed to process features at different time scales:

[0105] 1. Fast Response Branch (100ms Scale):

[0106] 1D convolution kernel size: 5

[0107] Step size: 1

[0108] Number of channels: 32

[0109] Activation function: LeakyReLU (α=0.1)

[0110] Output dimension: [time step × 32]

[0111] 2. Medium-speed response branch (1s scale):

[0112] 1D convolution kernel size: 15

[0113] Step size: 5

[0114] Number of channels: 64

[0115] Activation function: Swish

[0116] Output dimension: [time step / 5 × 64]

[0117] 3. Slow response branch (10s scale):

[0118] 1D convolution kernel size: 30

[0119] Step size: 20

[0120] Number of channels: 128

[0121] Activation function: GELU

[0122] Output dimension: [time step / 20 × 128]

[0123] Each branch is followed by an LSTM layer to extract temporal features, with the number of hidden units being 16, 32, and 64 respectively. Feature maps of different scales are spliced ​​together in the channel dimension through skip connections to form multi-scale fused features.

[0124] The correlation matrix A between spatial locations can be calculated using the following formula. ij :

[0125]

[0126] in:

[0127] A ij This is the spatial attention weight matrix;

[0128] Q is the query vector matrix with dimensions [N×d];

[0129] K is the key vector matrix with dimensions [N×d];

[0130] d is the dimension of the feature vector, which can be 64 or 128.

[0131] The weighted spatial feature V′ is calculated using the following formula. i :

[0132]

[0133] in:

[0134] V is a value vector matrix.

[0135] The channel weight vector W is calculated using the following formula. c :

[0136] w c =σ(W2δ(W1GAP(F)));

[0137] in:

[0138] F is the input feature map with dimensions [C×H×W];

[0139] GAP stands for Global Average Pooling.

[0140] W1 and W2 are the weight matrices of the fully connected layer;

[0141] δ is the ReLU activation function;

[0142] σ is the Sigmoid function;

[0143] The channel-weighted feature F′ is output using the following formula. c :

[0144] F' c =w c ·F c .

[0145] Finally, the spatial and channel attention results are fused using the following formula:

[0146] F out =αV′ i +(1-α)W c ;

[0147] Here, α is a learnable fusion weight parameter, typically set to 0.5, which can be adjusted during model training. For example, the weights can be increased on cloudy days.

[0148] Step S30: Adaptively adjust the brightness and color temperature of the light strip according to the pre-adjustment parameters.

[0149] In this embodiment, the hidden state of the LSTM is mapped to the predicted light intensity (in lx) and color temperature (in K) within the next 500ms through a fully connected layer, and the output dimension is 2×5 (predicting the values ​​at the next 5 time points).

[0150] Finally, the predicted illuminance value E (lx) is converted into the target luminance B (%):

[0151]

[0152] Among them, E maxand E min Presets can be made based on the environment, such as 500, 1000, etc.

[0153] The predicted color temperature T(K) is mapped to RGB values ​​using the Robertson algorithm:

[0154] Calculate chromaticity coordinates:

[0155]

[0156] After converting to RGB space, PWM values ​​for each channel are generated, allowing for adjustment of the color temperature.

[0157] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the above-described apparatus and modules can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0158] Furthermore, embodiments of the present invention also propose a computer-readable storage medium storing a light strip adaptive adjustment program, which, when executed by a processor, performs the following operations:

[0159] Multi-source sensing data is collected in real time through a multi-modal sensor network. The multi-source sensing data includes ambient light data, spatial distribution data, and temperature data.

[0160] The multi-source sensing data is fused and dynamically weighted, and an illumination change prediction model is established based on an LSTM network. Pre-adjustment parameters are generated through the illumination change prediction model.

[0161] The brightness and color temperature of the light strip are adaptively adjusted according to the pre-adjustment parameters.

[0162] Furthermore, when the LED strip adaptive adjustment program is executed by the processor, it also performs the following operations:

[0163] A three-level parallel feature extraction branch is adopted to process multi-source sensing data at different time scales;

[0164] The processed multi-source sensing data is fused using attention features in both spatial and channel dimensions. Furthermore, when the LED strip adaptive adjustment program is executed by the processor, it also performs the following operations:

[0165] The spatial attention submodule calculates the correlation matrix between spatial locations and outputs the weighted spatial features.

[0166] The channel weight vector is calculated through the channel attention submodule, and the channel weighted features are output.

[0167] The spatial and channel attention results are integrated through a gating mechanism.

[0168] Furthermore, when the LED strip adaptive adjustment program is executed by the processor, it also performs the following operations:

[0169] The correlation matrix A between spatial locations is calculated using the following formula. ij :

[0170]

[0171] in:

[0172] A ij This is the spatial attention weight matrix;

[0173] Q is the query vector matrix with dimensions [N×d];

[0174] K is the key vector matrix with dimensions [N×d];

[0175] d is the dimension of the feature vector, which can be 64 or 128.

[0176] The weighted spatial feature V′ is calculated using the following formula. i :

[0177]

[0178] in:

[0179] V is a value vector matrix.

[0180] Furthermore, when the LED strip adaptive adjustment program is executed by the processor, it also performs the following operations:

[0181] The channel weight vector W is calculated using the following formula. c :

[0182] w c =σ(W2δ(W1GAP(F)));

[0183] in:

[0184] F is the input feature map with dimensions [C×H×W];

[0185] GAP stands for Global Average Pooling.

[0186] W1 and W2 are the weight matrices of the fully connected layer;

[0187] δ is the ReLU activation function;

[0188] σ is the Sigmoid function;

[0189] The channel-weighted feature F′ is output using the following formula.c :

[0190] F′ c =w c ·F c .

[0191] Furthermore, when the LED strip adaptive adjustment program is executed by the processor, it also performs the following operations:

[0192] The spatial and channel attention results are fused using the following formula:

[0193] F out =αV′ i +(1-α)W c ;

[0194] Where α is a learnable fusion weight parameter.

[0195] The specific embodiments of the computer-readable storage medium of the present invention are basically the same as the embodiments of the above-described adaptive adjustment method for LED strips, and will not be described in detail here.

[0196] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0197] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0198] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0199] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for adaptive adjustment of LED strip lights, characterized in that, The adaptive adjustment method for the LED strip includes the following steps: Multi-source sensing data is collected in real time through a multi-modal sensor network. The multi-source sensing data includes ambient light data, spatial distribution data, and temperature data. The multi-source sensing data is fused and dynamically weighted, and an illumination change prediction model is established based on an LSTM network. Pre-adjustment parameters are generated through the illumination change prediction model. The brightness and color temperature of the light strip are adaptively adjusted according to the pre-adjustment parameters.

2. The adaptive adjustment method for LED strips as described in claim 1, characterized in that, The step of acquiring multi-source sensing data in real time through a multimodal sensor network includes: A three-level parallel feature extraction branch is adopted to process multi-source sensing data at different time scales; The processed multi-source sensing data is fused with attention features in both spatial and channel dimensions.

3. The adaptive adjustment method for LED strips as described in claim 2, characterized in that, The steps for fusing attention features in the spatial and channel dimensions include: The spatial attention submodule calculates the correlation matrix between spatial locations and outputs the weighted spatial features. The channel weight vector is calculated through the channel attention submodule, and the channel weighted features are output. The spatial and channel attention results are integrated through a gating mechanism.

4. The adaptive adjustment method for LED strips according to claim 3, characterized in that, The step of calculating the correlation matrix between spatial locations and outputting the weighted spatial features through the spatial attention submodule includes: The correlation matrix A between spatial locations is calculated using the following formula. ij : in: A ij This is the spatial attention weight matrix; Q is the query vector matrix with dimensions [N×d]; K is the key vector matrix with dimensions [N×d]; d is the dimension of the feature vector, which can be 64 or 128. The weighted spatial feature V is calculated using the following formula. ' i : in: V is a value vector matrix.

5. The adaptive adjustment method for LED strips according to claim 4, characterized in that, The steps of calculating the channel weight vector through the channel attention submodule and outputting the channel weighted features include: The channel weight vector W is calculated using the following formula. c : w c =σ(W2δ(W1GAP(F))); in: F is the input feature map with dimensions [C×H×W]; GAP stands for Global Average Pooling. W1 and W2 are the weight matrices of the fully connected layer; δ is the ReLU activation function; σ is the Sigmoid function; The channel-weighted feature F is output using the following formula. ' c : F′ c =w c ·F c 。 6. The adaptive adjustment method for LED strips according to claim 4, characterized in that, The step of fusing spatial and channel attention results through a gating mechanism includes: The spatial and channel attention results are fused using the following formula: F out =αV' i +(1-α)W c ; Where α is a learnable fusion weight parameter.

7. A light strip adaptive adjustment device, characterized in that, The LED strip adaptive adjustment device includes: a memory, a processor, and an LED strip adaptive adjustment program stored in the memory and executable on the processor. When the LED strip adaptive adjustment program is executed by the processor, it implements the steps of the LED strip adaptive adjustment method as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an adaptive adjustment program for the light strip, which, when executed by a processor, implements the steps of the adaptive adjustment method for the light strip as described in any one of claims 1 to 6.

Images