Atmosphere lamp dynamic effect control method, vehicle control unit and storage medium

Abstract

The invention discloses an atmosphere lamp dynamic effect control method, a vehicle control unit and a storage medium, and relates to the technical field of automobile atmosphere lamp control. The scheme comprises: acquiring a first target video; wherein the first target video is a video uploaded by a user or a video shot by a panoramic camera; matching the first target video with the lamp bead nodes to generate first time sequence data of each lamp bead node; inputting the first time sequence data into a preset first neural network model, and outputting a first feature value; inputting the first characteristic value into a preset second neural network model, and outputting second time sequence data; and generating a first control instruction set based on the second time sequence data, so that the atmosphere lamp controller controls the lamp bead nodes according to the first control instruction set and generates an atmosphere lamp dynamic effect. The operability of the applicable video can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile atmosphere light control, and more specifically, to an atmosphere light motion effect control method, a vehicle controller and a storage medium. Background Art

[0002] With the continuous development of automobile technology and the increasing diversification of consumer needs, the application of atmosphere lights in car interiors is becoming more and more extensive. Car atmosphere lights are a kind of lighting equipment installed inside the car to create a comfortable atmosphere to enhance driving and riding comfort. It usually uses LED as a light source and achieves rich color changes through RGB color mode. Car atmosphere lights can be installed in multiple locations such as car doors, center consoles, pedals, etc., and independent lighting and color adjustment of different areas can be achieved through precise control.

[0003] At present, the control mechanism of the vehicle ambient light effect is generally based on playing music and recognizing user voice. For example, the control command corresponding to the user voice is recognized, and the brightness or color of the ambient light is controlled according to the control command. When the control source is a video, the video is generally a pre-set shape input video, and the user cannot freely and flexibly configure it, which cannot meet the preferences and needs of different users. The degree of personalization is not high, which reduces the operability of the applicable video. Summary of the invention

[0004] In view of this, the present invention provides an atmosphere light motion effect control method, a vehicle controller and a storage medium, which can improve the operability of applicable videos.

[0005] In order to achieve the above object, the present invention adopts the following technical scheme:

[0006] In a first aspect, the present invention provides an atmosphere light motion effect control method, the method comprising:

[0007] Obtain a first target video; wherein the first target video is a video uploaded by a user or a video shot by a panoramic camera;

[0008] Match the first target video with the lamp bead node to generate the first time series data of each lamp bead node;

[0009] Inputting the first time series data into a preset first neural network model and outputting a first eigenvalue;

[0010] Inputting the first characteristic value into a preset second neural network model, and outputting second time series data;

[0011] A first control instruction set is generated based on the second timing data, so that the atmosphere light controller controls the lamp bead nodes according to the first control instruction set and generates an atmosphere light motion effect.

[0012] With the method provided by the present invention, users can customize and upload DIY videos, or use the panoramic camera outside the vehicle to record videos of external scenes, perform video timing conversion and output atmosphere light effects, thereby providing users with a freer atmosphere light effect DIY function, improving the freedom of shaping atmosphere light effects, and satisfying users' personal atmosphere light experience needs.

[0013] In a possible implementation, the first time series data includes a first color value parameter and a first brightness value; matching the first target video with the lamp bead node, and generating the first time series data of each lamp bead node includes:

[0014] Divide the first target video into at least two video partitions; wherein the video partitions correspond to the lamp bead nodes one by one;

[0015] Obtain a first color histogram and a first brightness grayscale image of each video partition;

[0016] Acquire a first color value parameter based on the first color histogram;

[0017] A first brightness value is obtained based on the first brightness grayscale image.

[0018] In a possible implementation, after generating a first control instruction set based on the second time series data so that the atmosphere light controller controls the lamp bead node according to the first control instruction set and generates an atmosphere light motion effect, it also includes:

[0019] Obtain the actual atmosphere light dynamic effect video corresponding to each lamp bead node, and generate the third time series data of each lamp bead node;

[0020] Inputting the third time series data into the first neural network model and outputting the second eigenvalue;

[0021] Substitute the second eigenvalue and the first eigenvalue into the preset timing deviation function, calculate the timing deviation value between the actual atmosphere light motion effect video and the first target video, repeat the steps of obtaining the actual atmosphere light motion effect video corresponding to each lamp bead node, and generate the third timing data of each lamp bead node until the timing deviation value is lower than the preset value.

[0022] In a possible implementation, the third time series data includes a second color value parameter and a second brightness value; obtaining the actual atmosphere light dynamic effect video corresponding to each lamp bead node, and generating the third time series data for each lamp bead node includes:

[0023] Obtain the second color histogram and second brightness grayscale image of the actual atmosphere light dynamic effect video corresponding to each lamp bead node;

[0024] Acquire a second color value parameter based on the second color histogram;

[0025] A second brightness value is obtained based on the second brightness grayscale image.

[0026] In a possible implementation, the first neural network model is a convolutional neural network model; the convolutional neural network model includes a first input layer, at least one convolutional pooling module, a fully connected layer and an activation layer; wherein,

[0027] A first input layer, used for inputting first time series data;

[0028] Each convolutional pooling module includes a double convolutional layer and a maximum pooling layer, and the input end of the maximum pooling layer is connected to the output end of the double convolutional layer;

[0029] At least one convolutional pooling module is connected in sequence; the output end of the maximum pooling layer in the previous convolutional pooling module is connected to the input end of the double convolutional layer in the next convolutional pooling module, the input end of the double convolutional layer in the first convolutional pooling module is connected to the output end of the first input layer, and the output end of the maximum pooling layer in the last convolutional pooling module is connected to the input end of the fully connected layer;

[0030] The output of the fully connected layer is connected to the input of the activation layer.

[0031] In a possible implementation, the first neural network model is a recurrent neural network model; the recurrent neural network model includes three second input layers, at least two hidden layers and at least two output layers; wherein,

[0032] The three second input layers input the first feature value, time series length, and sample number respectively;

[0033] The output of each second input layer is connected to the input of all hidden layers;

[0034] The output of each hidden layer is connected to the input of all output layers.

[0035] In a possible implementation, before acquiring the first target video, the method further includes:

[0036] Obtain a second target video; wherein the second target video is an ambient light rhythm effect diagram provided by the developer;

[0037] Matching the second target video with the lamp bead node to generate fourth time series data of each lamp bead node;

[0038] Inputting the fourth time series data into the first neural network model and outputting the third eigenvalue;

[0039] Inputting the third eigenvalue into the second neural network model and outputting fifth time series data;

[0040] generating a second control instruction set based on the fifth time series data;

[0041] Based on the second control instruction set, the atmosphere light dynamic effect simulation is performed, and the dynamic effect simulation result is output.

[0042] In a possible implementation, in response to a modification operation of a developer based on the dynamic effect simulation result, a simulation parameter corresponding to the modification operation is modified;

[0043] In response to a display adjustment operation of the developer based on the dynamic effect simulation result, the display adjustment is performed on the dynamic effect simulation result corresponding to the modification operation.

[0044] In a second aspect, the present invention provides a vehicle controller comprising a processor and a memory; the memory stores a computer program or computer instructions that can be loaded by the processor and executed by a method for designing an ambient light dynamic effect such as the first aspect and any possible implementation thereof.

[0045] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program or computer instructions that can be loaded by a processor and executed by a method for designing an atmosphere lighting motion effect such as the first aspect and any possible implementation thereof.

[0046] For the specific description of the second and third aspects of the present invention and their various implementations, reference can be made to the detailed description of the first aspect and its various implementations; and for the beneficial effects of the second and third aspects and their various implementations, reference can be made to the analysis of the beneficial effects of the first aspect and its various implementations, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0048] Figure 1 is a flow chart of a method for controlling an ambient light dynamic effect according to an embodiment of the present invention;

[0049] Figure 2 is another flow chart of a method for controlling an ambient light dynamic effect according to an embodiment of the present invention;

[0050] Figure 3 is a network structure diagram of a convolutional neural network model according to an embodiment of the present invention;

[0051] Figure 4 is a network structure diagram of a recurrent neural network model according to an embodiment of the present invention;

[0052] Figure 5 It is a block diagram of the internal structure of an atmosphere light motion effect control device according to an embodiment of the present invention;

[0053] Figure 6 It is a block diagram of the internal structure of a vehicle controller according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless otherwise specifically stated.

[0055] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0056] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered as part of the specification.

[0057] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0058] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0059] In the following, the terms "first", "second", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", etc. may explicitly or implicitly include one or more of the features.

[0060] In order to facilitate understanding of the technical solution of the present invention, the terms involved in the present invention are briefly introduced below.

[0061] 1. Vehicle Control Unit (VCU): It is the "brain-level" controller in new energy vehicles. It is responsible for coordinating and managing various systems and components of the vehicle to achieve functions such as normal driving, energy recovery, fault diagnosis and processing, and vehicle status monitoring.

[0062] 2. Ambient light controller: responsible for adjusting the brightness, color and mode of the ambient light to adapt to different driving scenarios and provide a comfortable driving experience; it integrates components such as a microcontroller (MCU), LED driver, LIN transceiver and power management, and can communicate with the vehicle's main controller, such as the vehicle controller, through the CAN bus or LIN bus to receive instructions to control the ambient light.

[0063] 3. Convolutional Neural Networks (CNN): It is a type of feedforward neural network that includes convolution calculations and has a deep structure. It simulates the human brain's visual system and uses structures such as convolution layers and pooling layers to automatically extract and learn features from data to achieve efficient and accurate classification and recognition.

[0064] 4. Recurrent Neural Network (RNN): It is a neural network suitable for processing sequence data. It has memory ability and can capture the time dependency in the sequence. It processes the input sequence through internal state (hidden state) so that the output of the network depends not only on the current input, but also on the output at the previous moment.

[0065] 5. Rectified Linear Unit (ReLU): It is a nonlinear activation function that is very commonly used in deep learning, especially in convolutional neural networks (CNN); its purpose is to introduce nonlinearity into the neural network so that the network can learn and simulate more complex function mapping relationships.

[0066] Embodiment 1:

[0067] The embodiment of the present invention provides an atmosphere light dynamic effect control method, the execution subject of which is the vehicle controller. Figure 1 , which is a flow chart of the method for designing the dynamic effect of an atmosphere light provided by an embodiment of the present invention. Figure 1 As shown, the atmosphere light dynamic effect design method provided by the present invention mainly includes the following steps:

[0068] Step S101, obtaining a first target video; wherein the first target video is a video uploaded by a user or a video shot by a panoramic camera;

[0069] Step S102, matching the first target video with the lamp bead node to generate first time series data of each lamp bead node;

[0070] Step S103, inputting the first time series data into a preset first neural network model, and outputting a first eigenvalue;

[0071] Step S104, inputting the first characteristic value into a preset second neural network model, and outputting second time series data;

[0072] Step S105, generating a first control instruction set based on the second timing data, so that the atmosphere light controller controls the lamp bead nodes according to the first control instruction set and generates an atmosphere light motion effect.

[0073] In this embodiment, the user uploads a custom video through WIFI, NFC, USB flash drive, etc., or an external video captured by a panoramic camera outside the vehicle. Before matching these videos with the distribution of lamp beads, it is necessary to use the vehicle editing tool to edit and modify the style of the video. For example, the video with no obvious color changes and obvious interference from debris needs to be removed to make the video picture purer and facilitate color selection, so as to better meet the needs of the output of ambient light dynamic effects.

[0074] In some optional embodiments, for step S102, the first time series data includes a first color value parameter and a first brightness value, and the first color value parameter includes a first red value, a first green value, and a first blue value. The first target video is divided into at least two video partitions; wherein the video partitions correspond to the lamp bead nodes one by one, and each lamp bead node includes at least one lamp bead; a first color histogram and a first brightness grayscale map of each video partition are obtained; a first color value parameter is obtained based on the first color histogram; and a first brightness value is obtained based on the first brightness grayscale map.

[0075] In step S102, OpenCV can be used to perform timestamp mapping on each video partition of the second target video, obtain the first color histogram and the first brightness grayscale image frame by frame, and then obtain the second red value, second green value, second blue value and second brightness value to be input into the first neural network model based on the first color histogram and the first brightness grayscale image.

[0076] In step S105, a variable event table is first generated based on predefined variables and the second time series data; wherein the predefined variables are used to represent the state change of the second time series data, the predefined variables include type Type, time Time and value Value, the state change of the second time series data includes color change, brightness adjustment and scene switching, and scene switching refers to the scene change caused by replacing the first target video; then the above three predefined variables are used as inputs of the MCU to generate control instructions corresponding to each video partition, and the control instructions corresponding to all video partitions are collected into the first control instruction set in a time sequence.

[0077] When LED car ambient lights are working, the final light color is inconsistent with the expected or set color due to factors such as hardware (such as the brightness and color of LED lamp beads are related to their working current and junction temperature), environment (such as ambient light, material reflection) and software (such as control algorithm errors). Therefore, in order to reduce the above-mentioned ambient light color drift, the actual car ambient light dynamic effect video can be captured through the in-car camera, fed back to the deep learning model, and compared with the time series data deviation.

[0078] In some optional embodiments, such as Figure 1As shown, after generating a first control instruction set based on the second time series data so that the atmosphere light controller controls the lamp bead node according to the first control instruction set and generates an atmosphere light motion effect, it also includes:

[0079] Step S106, obtaining the actual atmosphere light dynamic effect video corresponding to each lamp bead node, and generating the third time series data of each lamp bead node;

[0080] Step S107, inputting the third time series data into the first neural network model, and outputting the second eigenvalue;

[0081] Step S108, substitute the second eigenvalue and the first eigenvalue into the preset timing deviation function, calculate the timing deviation value between the actual atmosphere light motion effect video and the first target video, repeat the steps of obtaining the actual atmosphere light motion effect video corresponding to each lamp bead node, and generate the third timing data of each lamp bead node until the timing deviation value is lower than the preset value.

[0082] Further, for step S106, the third time series data includes a second color value parameter and a second brightness value, and the second color value parameter includes a second red value, a second green value, and a second blue value. First, the second color histogram and the second brightness grayscale map of the actual atmosphere light dynamic effect video corresponding to each lamp bead node are obtained; then, the second color value parameter is obtained based on the second color histogram, and the second brightness value is obtained based on the second brightness grayscale map.

[0083] Optionally, the time series deviation function uses the mean square error function:

[0084]

[0085] Among them, MSE is the timing deviation value, y k is the feature value of the kth actual atmosphere light dynamic effect video analysis output, is the feature value output by the first neural network model of the kth first target video, n is the number of samples, k = 1, 2...n, n is an integer. The gradient is recalculated through the back propagation algorithm, the weight and learning rate are recalculated using the gradient descent algorithm, the error caused by color drift is added to the iteration, and the final presentation effect is optimized.

[0086] In this embodiment, the current dynamic effect is recorded by the in-car camera, and the video is fed back to the tool for re-timing conversion, and the data is input into the CNN model for timing deviation calculation. After the model algorithm is iterated, new timing data is obtained, and the dynamic effect is displayed again. The cycle continues until the timing deviation value is lower than the preset value (for example, 5%), and the dynamic effect deviation adjustment is completed. After multiple iterations of the deep learning model and comparative deviation calculations, the matching degree between the timing diagram and the required dynamic effect can be improved, ensuring that the final generated ambient light effect can highly meet the design requirements and provide users with a better visual experience.

[0087] In some optional embodiments, the first neural network model may adopt a convolutional neural network model, and the second neural network model may adopt a recurrent neural network model.

[0088] CNN deep learning is used to extract feature values ​​frame by frame and generate accurate timing diagrams to ensure that the ambient light system can be precisely controlled as expected after accurate reading. Pooling and downsampling are used to efficiently compress timing data, and the RNN model is introduced for timing integration. While converting video into timing data with high precision, the consistency of timing data is increased, the main feature values ​​are retained to reduce timing data, and the shared memory usage is reduced, thereby optimizing system performance and improving system stability and response speed.

[0089] In addition, the decoupling of the brightness and color of the timing diagram and the optional realization of the ambient light area have greatly enriched the types and combinations of dynamic effects. It not only breaks the limitation of dynamic effects in traditional design that are limited to specific needs, but also provides a unique and personalized user experience for future new models or high-end models. Through flexible configuration and combination, designers can create more diverse ambient light dynamic effects to meet the needs and preferences of different users.

[0090] Embodiment 2:

[0091] Before implementing the atmosphere light motion effect control method of the first embodiment, a simulation test process needs to be performed first.

[0092] Please refer to Figure 2 , which is a flow chart of the simulation test provided by the embodiment of the present invention. Figure 2 As shown, the simulation test process mainly includes the following steps:

[0093] Step S201, obtaining a second target video; wherein the second target video is an atmosphere light rhythm effect diagram provided by a developer;

[0094] Step S202, matching the second target video with the lamp bead node to generate fourth time series data of each lamp bead node;

[0095] Step S203, inputting the fourth time series data into the first neural network model, and outputting the third eigenvalue;

[0096] Step S204, inputting the third eigenvalue into the second neural network model, and outputting fifth time series data;

[0097] Step S205, generating a second control instruction set based on the fifth time series data;

[0098] Step S206, performing an ambient light motion effect simulation based on the second control instruction set, and outputting a motion effect simulation result.

[0099] In this embodiment, after obtaining the second target video, the second target video will generally be edited and modified in style accordingly. For example, the video will be free of obvious color changes and obvious interference from debris, making the video picture purer and easier to pick colors, so as to better meet the needs of ambient light motion effect output.

[0100] In some optional embodiments, for step S202, the fourth time series data includes a third color value parameter and a third brightness value, and the third color value parameter includes a third red value, a third green value, and a third blue value. First, the second target video is divided into at least two video partitions; wherein the video partitions correspond to the lamp bead nodes one by one; a third color histogram and a third brightness grayscale map of each video partition are obtained; a third red value, a third green value, and a third blue value are obtained based on the third color histogram; and a third brightness value is obtained based on the third brightness grayscale map.

[0101] In step S202, OpenCV can be used to perform timestamp mapping on each video partition of the second target video, obtain a third color histogram and a third brightness grayscale map frame by frame, and then obtain a third red value, a third green value, a third blue value and a third brightness value to be input into the first neural network model based on the third color histogram and the third brightness grayscale map.

[0102] In step S205, a variable event table is first generated based on predefined variables and the fifth time series data; wherein the predefined variables are used to represent the state change of the fifth time series data, the predefined variables include type Type, time Time and value Value, the state change of the fifth time series data includes color change, brightness adjustment and scene switching, and scene switching refers to the scene change caused by replacing the second target video; then the above three predefined variables are used as inputs of the MCU to generate control instructions corresponding to each video partition, and the control instructions corresponding to all video partitions are collected into the second control instruction set in a time sequence.

[0103] In step S206, the atmosphere light motion simulation model can be constructed through tools such as OpenCV or Python Imaging library, and the second control instruction set can be read through the interface to perform motion simulation of the lamp end modeling. The second control instruction set can be read in a time sequence to change the brightness and color of the model lamp end, and the simulation parameters can be recorded through logs.

[0104] After outputting the dynamic effect simulation result in step S106, the vehicle controller also responds to the developer's modification operation based on the dynamic effect simulation result and modifies the simulation parameters corresponding to the modification operation.

[0105] Specifically, the following debugging and parameter changes can be performed on the simulation results:

[0106] In response to the developer's modification operation based on the dynamic effect simulation result, the simulation parameters corresponding to the modification operation are modified, wherein the simulation parameters include the lamp bead node rhythm speed, the lamp bead node color parameter, the lamp bead node brightness parameter and the lamp bead node partition parameter;

[0107] In response to the developer's display adjustment operation based on the motion effect simulation result, the motion effect simulation result corresponding to the modification operation is displayed and adjusted, wherein the display adjustment operation includes fast forward, rewind, pause and resume, so as to find unnecessary jump scenes in the motion effect and delete or modify them.

[0108] The real-time adjustment function of the atmosphere light parameters allows designers to instantly observe and adjust the motion effects during the motion design process, thereby reducing the waste of manpower and time caused by subjective judgment differences. This feature enhances the flexibility and efficiency of the design, ensuring that the final motion effects can accurately meet the design expectations. In addition, it provides developers with efficient and convenient means of testing, comparing and checking timing and motion effects. This function allows developers to more intuitively observe the details of the development process and quickly locate the problem, thereby improving development efficiency and quality.

[0109] Through the CNN model deep learning algorithm, the required feature data of each frame is extracted through convolution, pooling, and regression values, and the timestamps of consecutive frames are mapped to input the feature data into the RNN neural network model. The smoothness of the time series changes is increased by predicting the time series, and the fluency of the time series data with time series changes is increased. Time series graph data is generated to make the output time series data more accurate and the connection changes of the feature values ​​smoother.

[0110] Embodiment three:

[0111] Since the first time series data, the third time series data, and the fourth time series data all include four types of data, namely, red value, green value, blue value, and brightness value, the convolutional neural network model is designed as a network structure with four input channels.

[0112] Based on the first and second embodiments, the present invention provides a convolutional neural network model. The convolutional neural network model includes:

[0113] The convolutional neural network model includes a first input layer, at least one convolutional pooling module, a fully connected layer, and an activation layer;

[0114] The first input layer has four input terminals, and the four input terminals sequentially input red value, green value, blue value and brightness value;

[0115] Each convolutional pooling module includes a double convolutional layer and a maximum pooling layer, and the input end of the maximum pooling layer is connected to the output end of the double convolutional layer;

[0116] At least one convolutional pooling module is connected in sequence; the output end of the maximum pooling layer in the previous convolutional pooling module is connected to the input end of the double convolutional layer in the next convolutional pooling module, the input end of the double convolutional layer in the first convolutional pooling module is connected to the output end of the first input layer, and the output end of the maximum pooling layer in the last convolutional pooling module is connected to the input end of the fully connected layer;

[0117] The output of the fully connected layer is connected to the input of the activation layer.

[0118] It should be noted that a double convolutional layer refers to two convolutional layers stacked continuously, each of which contains a set of learnable weights and bias items, and the output of the previous convolutional layer is connected to the input of the next convolutional layer.

[0119] Figure 3 FIG. 4 shows a network structure of a convolutional neural network model using four convolutional pooling modules. Figure 3 As shown in the figure, the dimension of the input data of the first input layer is 256x256x4; each convolution layer in the double convolution layer of the first convolution pooling module connected to the output of the first input layer has 32 convolution kernels, each convolution layer in the double convolution layer of the second convolution pooling module has 64 convolution kernels, each convolution layer in the double convolution layer of the third convolution pooling module has 128 convolution kernels, and each convolution layer in the double convolution layer of the last convolution pooling module has 256 convolution kernels. The size of all convolution kernels is 3*3, that is, a matrix of 3 rows and 3 columns. The pooling window used by all maximum pooling layers is 2*2, that is, the local area size is 2 rows and 2 columns, and the step size is also 2. The fully connected layer has 512 neurons. The weights are initialized to a Gaussian distribution of (0,0.01).

[0120] In addition, in order to make the data adapt to the input requirements of the fully connected layer while retaining all the extracted feature information, the feature map output by the last convolutional pooling module needs to be one-dimensionalized, that is, the multi-dimensional feature map (two-dimensional matrix) is converted into a one-dimensional feature vector.

[0121] Embodiment 4:

[0122] Based on the first and second embodiments, the present invention provides a recurrent neural network model. Figure 4 , which is a network structure diagram of the recurrent neural network model provided by an embodiment of the present invention. Figure 4 As shown, the recurrent neural network model includes three second input layers, at least two hidden layers and at least two output layers; wherein,

[0123] The second input layer 1 inputs the time series length (seq_len); the second input layer 2 inputs the number of samples (batch_size); the second input layer 3 inputs the first eigenvalue and the third eigenvalue (input_size);

[0124] The output end of each second input layer is connected to the input ends of all hidden layers. Taking the second input layer 1 as an example, the output end of the second input layer 1 is respectively connected to the input ends of hidden layer 1, hidden layer 2, ... hidden layer M. Wherein, M ≥ 2, and M is an integer.

[0125] The output of each hidden layer is connected to the input of all output layers. Taking hidden layer 1 as an example, the output of hidden layer 1 is connected to the input of output layer 1, output layer 2, ... output layer N. N ≥ 2, and N is an integer.

[0126] It should be noted that the values ​​of M and N are related to factors such as the complexity of solving the problem, the amount of data, the dimension of data features, the length of the sequence, and hardware limitations, and need to be set according to actual conditions.

[0127] Embodiment five:

[0128] Based on the same inventive concept, the embodiment of the present invention also provides an ambient light motion effect control device for implementing the above-mentioned ambient light motion effect control method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more embodiments of the ambient light motion effect control device provided below can refer to the limitations of the ambient light motion effect control method above, and will not be repeated here.

[0129] Figure 5 FIG. 5 shows a structural block diagram of an ambient light motion effect control device 500. Figure 5 As shown, the atmosphere light motion effect control device 500 includes: a first video acquisition module 501, a first matching module 502, a first feature value output module 503, a first time series data output module 504 and a first instruction set generation module 505;

[0130] The first video acquisition module 501 is used to acquire a first target video; wherein the first target video is a video uploaded by a user or a video shot by a panoramic camera;

[0131] A first matching module 502 is used to match the first target video with the lamp bead node to generate first time series data for each of the lamp bead nodes;

[0132] A first eigenvalue output module 503, used to input the first time series data into a preset first neural network model and output a first eigenvalue;

[0133] A first time series data output module 504, used to input the first feature value into a preset second neural network model and output second time series data;

[0134] The first instruction set generation module 505 is used to generate a first control instruction set based on the second timing data, so that the atmosphere light controller controls the lamp bead node according to the first control instruction set and generates an atmosphere light motion effect.

[0135] In some optional embodiments, the first timing data includes a first color value parameter and a first brightness value; the first matching module 502 is specifically used to divide the first target video into at least two video partitions; wherein the video partitions correspond one-to-one to the lamp bead nodes; obtain the first color histogram and the first brightness grayscale map of each video partition; obtain the first color value parameter based on the first color histogram; and obtain the first brightness value based on the first brightness grayscale map.

[0136] Some optional embodiments further include:

[0137] A second video acquisition module is used to generate a first control instruction set based on the second timing data, so that the atmosphere light controller controls the lamp bead node according to the first control instruction set and generates an atmosphere light motion effect, obtain the actual atmosphere light motion effect video corresponding to each lamp bead node, and generate third timing data for each lamp bead node;

[0138] A second eigenvalue output module, used for inputting the third time series data into the first neural network model and outputting a second eigenvalue;

[0139] The timing deviation calculation module is used to substitute the second eigenvalue and the first eigenvalue into a preset timing deviation function, calculate the timing deviation value between the actual atmosphere light motion effect video and the first target video, repeat the steps of obtaining the actual atmosphere light motion effect video corresponding to each lamp bead node, and generate the third timing data of each lamp bead node until the timing deviation value is lower than the preset value.

[0140] Furthermore, the third timing data includes a second color value parameter and a second brightness value; a second video acquisition module is used to obtain a second color histogram and a second brightness grayscale map of the actual atmosphere light motion effect video corresponding to each lamp bead node; obtain the second color value parameter based on the second color histogram; and obtain the second brightness value based on the second brightness grayscale map.

[0141] In some optional embodiments, the first neural network model is a convolutional neural network model, which includes a first input layer, at least one convolutional pooling module, a fully connected layer and an activation layer; wherein,

[0142] A first input layer, used for inputting first time series data;

[0143] Each convolutional pooling module includes a double convolutional layer and a maximum pooling layer, and the input end of the maximum pooling layer is connected to the output end of the double convolutional layer;

[0144] At least one convolutional pooling module is connected in sequence; the output end of the maximum pooling layer in the previous convolutional pooling module is connected to the input end of the double convolutional layer in the next convolutional pooling module, the input end of the double convolutional layer in the first convolutional pooling module is connected to the output end of the first input layer, and the output end of the maximum pooling layer in the last convolutional pooling module is connected to the input end of the fully connected layer;

[0145] The output of the fully connected layer is connected to the input of the activation layer.

[0146] In some optional embodiments, the second neural network model is a recurrent neural network model, and the recurrent neural network model includes three second input layers, at least two hidden layers, and at least two output layers;

[0147] in,

[0148] The three second input layers input the first feature value, time series length, and sample number respectively;

[0149] The output of each second input layer is connected to the input of all hidden layers;

[0150] The output of each hidden layer is connected to the input of all output layers.

[0151] In some optional embodiments, the device further comprises:

[0152] The third video acquisition module is used to acquire the second target video; wherein the second target video is an atmosphere light rhythm effect diagram provided by the developer;

[0153] A second matching module is used to match the second target video with the lamp bead node to generate fourth time series data of each lamp bead node;

[0154] A third eigenvalue output module, used for inputting the fourth time series data into the first neural network model and outputting a third eigenvalue;

[0155] A second time series data output module, used for inputting the third eigenvalue into the second neural network model and outputting fifth time series data;

[0156] A first instruction set generating module, used for generating a second control instruction set based on the fifth time series data;

[0157] The simulation module is used to perform the ambient light motion effect simulation based on the second control instruction set and output the motion effect simulation result.

[0158] Furthermore, the device also includes:

[0159] A simulation parameter modification module, which is used to modify the simulation parameters corresponding to the modification operation in response to the developer's modification operation based on the dynamic effect simulation result, wherein the simulation parameters include the rhythm speed of the lamp bead node, the color parameter of the lamp bead node, the brightness parameter of the lamp bead node and the partition parameter of the lamp bead node;

[0160] The simulation display adjustment module is used to respond to the developer's display adjustment operation based on the dynamic effect simulation result, and to adjust the display of the dynamic effect simulation result corresponding to the modification operation, wherein the display adjustment operation includes fast forward, rewind, pause and resume.

[0161] It should be noted that the functional modules in the fifth embodiment can be integrated together to form an independent unit, for example, integrated into a processing unit, or each module can exist physically separately, or two or more modules can be integrated to form an independent unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0162] Embodiment six:

[0163] Based on the same inventive concept, the embodiment of the present invention further provides a vehicle controller 600, the internal structure of the vehicle controller 600 is as follows: Figure 6 The vehicle controller 600 includes a memory 601 , a processor 602 and a communication bus 603 ; the memory 601 and the processor 602 are connected via the communication bus 603 .

[0164] The memory 601 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 601 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing the method for designing the dynamic effect of an atmosphere light provided in the above embodiment, etc.; the data storage area may store data involved in the method for designing the dynamic effect of an atmosphere light provided in the above embodiment, etc.

[0165] The processor 602 may include one or more processing cores. The processor 602 executes various functions and processes data of the present invention by running or executing instructions, programs, code sets or instruction sets stored in the memory 601, calling the data stored in the memory 601. The processor 602 may be at least one of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller and a microprocessor. It is understandable that for different devices, the electronic device used to implement the above-mentioned processor function may also be other, and the embodiment of the present invention is not specifically limited.

[0166] The communication bus 603 may include a path to transmit information between the above components. The communication bus 603 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The communication bus 603 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one double arrow is used in the diagram, but it does not mean that there is only one bus or one type of bus. Figure 6 The vehicle controller shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0167] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a partial structure related to the scheme of the present invention, and does not constitute a limitation on the vehicle controller to which the scheme of the present invention is applied. The specific vehicle controller may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0168] Embodiment nine:

[0169] In particular, according to an embodiment of the present invention, the process described in the above flowchart may be implemented as a computer software program.

[0170] The technical solution of the present invention, in essence, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, that is, the embodiment of the present invention also provides a computer-readable storage medium, which includes a number of computer programs or computer instructions, which are used to enable an electronic device (which can be a personal computer, server or network device, etc.) or a processor to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory, random access memory, disk or optical disk, etc., various media that can store program codes.

[0171] Although some specific embodiments of the present invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It will be appreciated by those skilled in the art that the above embodiments may be modified without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A method for controlling an ambient light dynamic effect, characterized in that: include: Obtain a first target video; wherein the first target video is a video uploaded by a user or a video shot by a panoramic camera; Matching the first target video with the lamp bead node to generate first time series data for each of the lamp bead nodes; Inputting the first time series data into a preset first neural network model, and outputting a first eigenvalue; Inputting the first characteristic value into a preset second neural network model, and outputting second time series data; A first control instruction set is generated based on the second timing data, so that the atmosphere light controller controls the lamp bead node according to the first control instruction set and generates an atmosphere light motion effect.

2. The method for controlling the dynamic effect of an ambient light according to claim 1, characterized in that: The first time series data includes a first color value parameter and a first brightness value; The matching of the first target video with the lamp bead node to generate the first time series data of each lamp bead node includes: Divide the first target video into at least two video partitions; wherein the video partitions correspond to the lamp bead nodes one by one; Obtaining a first color histogram and a first brightness grayscale image of each of the video partitions; Acquire the first color value parameter based on the first color histogram; The first brightness value is acquired based on the first brightness grayscale image.

3. The method for controlling the dynamic effect of an ambient light according to claim 1, characterized in that: After generating the first control instruction set based on the second time series data so that the atmosphere light controller controls the lamp bead node according to the first control instruction set and generates the atmosphere light dynamic effect, the method further includes: Obtaining the actual atmosphere light dynamic effect video corresponding to each of the lamp bead nodes, and generating the third time series data of each of the lamp bead nodes; Inputting the third time series data into the first neural network model, and outputting a second eigenvalue; Substitute the second eigenvalue and the first eigenvalue into the preset timing deviation function, calculate the timing deviation value between the actual atmosphere light motion effect video and the first target video, repeat the steps of obtaining the actual atmosphere light motion effect video corresponding to each of the lamp bead nodes, and generate the third timing data of each of the lamp bead nodes until the timing deviation value is lower than the preset value.

4. The method for controlling the dynamic effect of an ambient light according to claim 3, characterized in that: The third time series data includes a second color value parameter and a second brightness value; The obtaining of the actual atmosphere light dynamic effect video corresponding to each of the lamp bead nodes and generating the third time series data of each of the lamp bead nodes includes: Obtain a second color histogram and a second brightness grayscale image of the actual atmosphere light dynamic effect video corresponding to each of the lamp bead nodes; Acquire the second color value parameter based on the second color histogram; The second brightness value is acquired based on the second brightness grayscale image.

5. The method for controlling the dynamic effect of an ambient light according to any one of claims 1 to 4, characterized in that: The first neural network model is a convolutional neural network model; the convolutional neural network model includes a first input layer, at least one convolutional pooling module, a fully connected layer and an activation layer; wherein, The first input layer is used to input the first time series data; Each of the convolutional pooling modules includes a double convolutional layer and a maximum pooling layer, and the input end of the maximum pooling layer is connected to the output end of the double convolutional layer; The at least one convolutional pooling module is connected in sequence; the output end of the maximum pooling layer in the previous convolutional pooling module is connected to the input end of the double convolutional layer in the next convolutional pooling module, the input end of the double convolutional layer in the first convolutional pooling module is connected to the output end of the first input layer, and the output end of the maximum pooling layer in the last convolutional pooling module is connected to the input end of the fully connected layer; The output end of the fully connected layer is connected to the input end of the activation layer.

6. The method for controlling the dynamic effect of an ambient light according to any one of claims 1 to 4, characterized in that: The second neural network model is a recurrent neural network model; the recurrent neural network model includes three second input layers, at least two hidden layers and at least two output layers; wherein, The three second input layers respectively input the first feature value, time series length and sample quantity; The output end of each of the second input layers is connected to the input ends of all the hidden layers; The output end of each hidden layer is connected to the input ends of all the output layers.

7. The method for controlling the dynamic effect of an ambient light according to any one of claims 1 to 4, characterized in that: Before acquiring the first target video, the method further includes: Obtain a second target video; wherein the second target video is an atmosphere light rhythm effect diagram provided by the developer; Matching the second target video with the lamp bead node to generate fourth time series data of each lamp bead node; Inputting the fourth time series data into the first neural network model, and outputting a third eigenvalue; Inputting the third characteristic value into the second neural network model, and outputting fifth time series data; Generate a second control instruction set based on the fifth time series data; Based on the second control instruction set, an atmosphere light motion effect simulation is performed and a motion effect simulation result is output.

8. The method for controlling the dynamic effect of an ambient light according to claim 7, characterized in that: After performing the ambient light dynamic effect simulation based on the second control instruction set and outputting the dynamic effect simulation result, the method further includes: In response to a modification operation performed by the developer based on the dynamic effect simulation result, modifying a simulation parameter corresponding to the modification operation; In response to the developer's display adjustment operation based on the motion effect simulation result, the display adjustment is performed on the motion effect simulation result corresponding to the modification operation.

9. A vehicle controller, characterized in that: including a processor and a memory; The memory stores the method for controlling the ambient light motion effect as described in any one of claims 1 to 8 which can be loaded and executed by the processor.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program or computer instructions that can be loaded by a processor and execute the method for controlling the ambient light motion effect as described in any one of claims 1 to 8.

Images