A wind speed identification method, device, equipment and medium

Abstract

The application discloses a wind speed identification method and device, equipment and medium, and relates to the technical field of computers. The method comprises the following steps: acquiring an image frame sequence, wherein the image frame sequence comprises a plurality of image frames which are continuous in time; performing optical flow feature extraction on the image frame sequence to obtain optical flow features representing the motion state of a target object in the image frame sequence; processing the optical flow features from the time dimension and the space dimension respectively to obtain time dimension features and space dimension features; and determining a wind speed matched with the image frame sequence according to the time dimension features and the space dimension features. The method can improve the accuracy of wind speed identification.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a wind speed identification method, apparatus, device, and medium. Background Technology

[0002] With the rapid development of technology, 8D cockpits, as immersive interactive experience spaces, have been widely used in entertainment, simulation training, and other fields. In 8D cockpits, wind speed simulation is one of the key factors in creating a realistic environment. Accurate wind speed recognition allows users to truly feel the wind as if they were in a virtual scene, greatly enhancing the immersion and realism of the experience.

[0003] Currently, convolutional neural networks can be used to identify wind speed from a single image. However, a single image is easily affected by external environmental factors such as lighting and weather, resulting in poor robustness and generalization ability of the model, and making it prone to overfitting.

[0004] To address this, a technique for wind speed identification using optical flow algorithms on multiple images was proposed. However, traditional optical flow algorithms are typically suitable for motion analysis with simplified backgrounds. When the background is complex or the contrast between the target object and the background is low, the optical flow algorithm is easily interfered with, which affects the identification results and reduces the accuracy of wind speed identification. Summary of the Invention

[0005] This application provides a wind speed identification method, apparatus, equipment, and medium that can improve the accuracy of wind speed identification.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] In a first aspect, this application provides a wind speed identification method, the method comprising:

[0008] Acquire an image frame sequence, the image frame sequence comprising multiple temporally consecutive images;

[0009] Optical flow features are extracted from the image frame sequence to obtain optical flow features representing the motion state of the target object in the image frame sequence;

[0010] The optical flow features are processed from the time dimension and the spatial dimension respectively to obtain the time dimension features and the spatial dimension features;

[0011] Based on the time dimension features and the spatial dimension features, the wind speed matching the image frame sequence is determined.

[0012] Optionally, determining the wind speed matching the image frame sequence based on the time dimension features and the spatial dimension features includes:

[0013] The time dimension features and the spatial dimension features are input into the wind speed prediction model to obtain the wind speed matching the image frame sequence; wherein, the wind speed prediction model is trained based on the sample time dimension features, the sample spatial dimension features and the sample labels, the sample time dimension features and the sample spatial dimension features are obtained based on the sample optical flow features, and the sample optical flow features are obtained based on the sample image frame sequence.

[0014] Optionally, the temporal dimension features include: average acceleration of multiple frames, rate of change of velocity of multiple frames, and directional consistency of multiple frames; the spatial dimension features include average velocity of multiple frames, velocity gradient of multiple frames, velocity variance of multiple frames, and average percentage of motion region of multiple frames; determining the wind speed matching the image frame sequence based on the temporal dimension features and the spatial dimension features includes:

[0015] The time dimension correction coefficient is determined based on the average acceleration of the multi-frame images, the rate of change of velocity of the multi-frame images, and the orientation consistency of the multi-frame images; the spatial dimension correction coefficient is determined based on the velocity gradient of the multi-frame images, the velocity variance of the multi-frame images, and the average proportion of the motion region of the multi-frame images.

[0016] The average velocity of the multi-frame images is corrected using the time dimension correction coefficient and the spatial dimension correction coefficient to obtain the wind speed that matches the image frame sequence.

[0017] Optionally, determining the time dimension correction coefficient based on the average acceleration of the multi-frame images, the rate of change of velocity of the multi-frame images, and the orientation consistency of the multi-frame images includes:

[0018]

[0019] in, This represents the time dimension correction factor. This represents the average acceleration across multiple frames of images. Indicates the rate of change of speed across multiple frames of images. Indicates the directional consistency of multiple frames of images. Indicates the weight of the first sub-time. Indicates the weight of the second sub-time. This represents the third sub-time weight.

[0020] Optionally, determining the spatial dimension correction coefficient based on the velocity gradient of the multi-frame images, the velocity variance of the multi-frame images, and the average proportion of the motion region of the multi-frame images includes:

[0021]

[0022] in, This represents the spatial dimension correction factor. Represents the velocity gradient across multiple frames of an image. This represents the velocity variance of multiple frames of images. This represents the average percentage of motion regions across multiple frames of images. Indicates the weight of the first subspace. Indicates the weight of the second subspace. This represents the weight of the third subspace.

[0023] Optionally, the average velocity of the multi-frame images is corrected using the time dimension correction coefficient and the spatial dimension correction coefficient to obtain a wind speed matching the image frame sequence, including:

[0024]

[0025]

[0026] in, This represents the time dimension correction factor. This represents the spatial dimension correction factor. This indicates the wind speed that matches the image frame sequence. This represents the average speed of multiple frames of images. S represents the time weight, and S represents the spatial weight.

[0027] Optionally, the method further includes:

[0028] The airflow of the air conditioner inside the vehicle is controlled based on the wind speed that matches the image frame sequence.

[0029] Secondly, this application provides a wind speed identification device, the device comprising:

[0030] The acquisition module is used to acquire an image frame sequence, which includes multiple temporally consecutive images;

[0031] The feature extraction module is used to extract optical flow features from the image frame sequence to obtain optical flow features representing the motion state of the target object in the image frame sequence;

[0032] The spatiotemporal processing module is used to process the optical flow features from the time dimension and the spatial dimension respectively to obtain the time dimension features and the spatial dimension features;

[0033] The identification module is used to determine the wind speed that matches the image frame sequence based on the time dimension features and the spatial dimension features.

[0034] Thirdly, this application provides a computing device, including a memory and a processor;

[0035] The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of the first aspects.

[0036] Fourthly, this application provides a computer-readable storage medium for storing a computer program for performing the method as described in any one of the first aspects.

[0037] As can be seen from the above technical solution, this application has at least the following beneficial effects:

[0038] This application provides a wind speed recognition method, which includes acquiring an image frame sequence comprising multiple temporally consecutive images; extracting optical flow features from the image frame sequence to obtain optical flow features representing the motion state of a target object in the image frame sequence; processing the optical flow features from both temporal and spatial dimensions to obtain temporal and spatial features; this application effectively separates target motion from background interference through decoupling the temporal and spatial dimensions; and determines the wind speed matching the image frame sequence based on the temporal and spatial features. The temporal features reflect the dynamic change pattern of wind speed, while the spatial features reflect the spatial consistency of wind speed. When the contrast between the target and the background decreases by 30%, the temporal and spatial fusion mechanism reduces the recognition error by 42% compared to traditional methods, verifying the effectiveness of feature complementarity. In summary, this method systematically solves the robustness problem of traditional optical flow algorithms in complex scenarios through temporal and spatial feature decoupling and multi-dimensional information fusion, providing an innovative solution for high-precision wind speed recognition.

[0039] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0040] Figure 1A flowchart illustrating a wind speed identification method provided in this application embodiment;

[0041] Figure 2 A flowchart of a wind speed identification device provided in an embodiment of this application;

[0042] Figure 3 This is a schematic diagram of a computing device provided in an embodiment of this application. Detailed Implementation

[0043] The terms "first," "second," and "third," etc., used in this application specification and accompanying drawings are used to distinguish different objects, not to limit a specific order.

[0044] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0045] Currently, commonly used wind speed identification methods can be roughly divided into two categories: single image identification based on deep learning and multi-frame image sequence identification based on optical flow algorithms.

[0046] The first approach uses convolutional neural networks (CNNs) to extract features from single-frame images for wind speed identification. Such methods typically require collecting and labeling large amounts of image data, standardizing and augmenting the images, building a deep learning model using convolutional and fully connected layers, and selecting appropriate regression loss functions and optimizers. During model training, hyperparameters are adjusted to improve model performance and avoid overfitting. However, wind speed identification from single-frame images has significant limitations in practical applications. A single-frame image only captures static information at a specific moment, lacking dynamic data support for wind speed changes over time. This lack of spatial and temporal data limits the extraction of effective features directly related to wind speed, making it difficult for the model to accurately identify wind speed. Single-frame images are easily affected by external environmental factors such as lighting and weather, resulting in poor model robustness and generalization ability, and a tendency to overfit. This limits the reliability of the model in practical applications.

[0047] The second method utilizes traditional optical flow algorithms to identify wind speed from multiple frames of images. This method first extracts the motion vector field from the image sequence using traditional optical flow algorithms (such as Lucas-Kanade or dense optical flow), and then indirectly infers the wind speed by analyzing changes in the vector field. However, the performance of this method is highly dependent on specific environmental conditions and visible reference objects, and it has the following shortcomings: In complex backgrounds, optical flow calculations may produce large errors, and when there are few visible reference objects in the scene, the quality of the generated motion vector field will also decrease significantly, affecting the accuracy of wind speed identification. Since traditional optical flow algorithms are generally suitable for motion analysis with simplified backgrounds, they are easily interfered with when the background is complex or the contrast between the target object and the background is low, thus affecting the accuracy of the prediction.

[0048] In view of this, embodiments of this application provide a wind speed identification method, which can be executed by a processing device, which can be a terminal or a server. Terminals include, but are not limited to, smartphones, tablets, laptops, personal digital assistants, or smart wearable devices. Servers can be cloud servers, such as central servers in a central cloud computing cluster or edge servers in an edge cloud computing cluster. Of course, servers can also be servers in a local data center. A local data center refers to a data center directly controlled by the user. For example, the processing device can be a vehicle-mounted terminal. This method systematically solves the robustness problem of traditional optical flow algorithms in complex scenarios through spatiotemporal feature decoupling and multi-dimensional information fusion, providing an innovative solution for high-precision wind speed identification.

[0049] To facilitate understanding, the application scenario of this application will be introduced below. In this application scenario, taking an in-vehicle terminal as the processing device as an example, users can watch videos in the vehicle cabin through the in-vehicle terminal (also known as an in-vehicle large screen). To give users an immersive experience while watching videos in the cabin, the processing device can control the vehicle's air conditioning system, such as controlling the interior temperature and airflow, to match the scene in the video, allowing users to experience the scene in the video while inside the cabin. For example, if the video plays an animation of a gentle breeze, the processing device can control the space to make the air conditioning blow a gentle breeze on the face, allowing users to feel the gentle breeze in the cabin, greatly enhancing the immersion and realism of the experience and improving the user experience.

[0050] To make the technical solution of this application clearer and easier to understand, the technical solution of this application will be described below with reference to the accompanying drawings, such as... Figure 1 As shown in the figure, this figure is a flowchart of a wind speed identification method provided in an embodiment of this application.

[0051] The method includes:

[0052] S101, The processing device acquires the image frame sequence.

[0053] An image frame sequence consists of multiple frames of images that are consecutive in time.

[0054] In some examples, the image frame sequence can be all or part of the image frames of the video being played, which is being processed and identified. For example, if the video being played is a 2-second, 60-frame animation, the processing device can use this 2-second, 60-frame animation as the image frame sequence, or it can extract 30 frames from it, that is, use a 2-second, 30-frame animation as the image frame sequence.

[0055] After the processing device acquires the image frame sequence, it needs to be preprocessed, including standardization and size adjustment, in order to adapt to the input dimension requirements of subsequent feature extraction.

[0056] S102. The processing device extracts optical flow features from the image frame sequence to obtain the motion state and optical flow features of the target object in the image frame sequence.

[0057] In some examples, the processing device can extract optical flow features from image frame sequences using a Recurrent All Pairs Field Transforms (RAFT) model. The RAFT model can calculate motion features between two consecutive frames. After preprocessing the image frame sequence, the processing device can generate batch data using a dataloader (a key component for efficient batch data processing in deep learning tasks). The data dimensions are [B, F, C, H, W], where B is the batch size, representing how much data is sent to the RAFT model at a time (e.g., F=30 represents 30 frames per data set), C=2 is the number of channels, H=224 represents the image height, and W=224 represents the image width.

[0058] In some embodiments, the processing device may sequentially select multiple pairs of consecutive two-frame images from the image frame sequence as image pairs, and then input them sequentially into the RAFT model to obtain the optical flow features of each image pair. The optical flow features of each image pair represent the motion vectors of each pixel in the image, and are used to represent the motion information (e.g., horizontal motion information and vertical motion information) between frames.

[0059] S103 The processing device processes the optical flow features from the time dimension and the spatial dimension respectively to obtain the time dimension features and the spatial dimension features.

[0060] In this application, after obtaining the optical flow features, the processing device does not directly rely on the optical flow features for wind speed identification, but performs multi-dimensional processing on the optical flow features, which can enhance the feature expression capability and improve the computational efficiency.

[0061] In some embodiments, the processing device can use the TimeSformer model to process the optical flow features described above. The TimeSformer model can capture spatiotemporal information in the video through a self-attention mechanism, and can effectively process the optical flow features in the image frame sequence, extracting features in the temporal dimension and spatial dimension respectively.

[0062] When processing optical flow features, the TimeSformer model treats the input optical flow feature sequence as a series of tokens and then uses a multi-head self-attention mechanism to calculate the correlation between these tokens. Spatially, the TimeSformer model focuses on the relationship between optical flow features at different locations within the same frame; temporally, it focuses on the relationship between optical flow features at the same or different locations across different frames.

[0063] For example, the TimeSformer model includes a spatial attention module and a temporal attention module. The spatial attention module learns spatial features within each frame through a self-attention mechanism, while the temporal attention module captures feature dependencies in the temporal dimension between frames through a self-attention mechanism.

[0064] The optical flow feature has dimensions [B, T, C, H, W], where B is the batch size, T represents the number of image pairs (e.g., T=29), F=30 represents 30 frames per data set, C=2 is the number of channels, H=224 represents the image height, and W=224 represents the image width. The feature embedding layer divides each frame of the input sequence into several small patches (usually fixed-size local regions) and flattens them to form an input format suitable for the Transformer model. The temporal attention module and spatial attention module use two Transformer submodules for temporal and spatial dimension processing, respectively, to obtain temporal and spatial features, ensuring the extraction of effective temporal and spatial dependency information within the sequence and frames.

[0065] S104. The processing device determines the wind speed that matches the image frame sequence based on the time dimension characteristics and spatial dimension characteristics.

[0066] After obtaining the temporal and spatial features, the processing device can determine the wind speed that matches the image frame sequence based on these features.

[0067] In some examples, the processing device can input temporal and spatial features into the wind speed prediction model to obtain wind speeds that match the image frame sequence. The training method for the wind speed prediction model is described below:

[0068] For example, the processing device can acquire sample image frame sequences and corresponding sample labels, then extract optical flow features from the sample image frame sequences to obtain sample optical flow features. These features are then processed in multiple dimensions, including time and space, to obtain sample time-dimensional features and sample space-dimensional features. These features, along with the sample labels, constitute a set of training data. The processing device can obtain multiple sets of training data, such as 1000 sets, in a similar manner. These training data are then used to train the model, resulting in a wind speed prediction model. Specifically, the processing device can calculate the loss based on the prediction results of the wind speed prediction model and the sample labels, for example, using a mean squared error loss function. The calculated loss is used for backpropagation to update the gradients of each layer of the model, thereby adjusting the weight parameters. After multiple rounds of iterative training, when the model gradually stabilizes and the prediction error converges within a reasonable range, the final model weights are saved. During the inference phase, the model parameters remain unchanged. At this time, the output result obtained by processing the input short video (a continuous sequence of image frames) through the wind speed prediction model is the wind speed that matches the sequence of image frames, for example, [0.3, 0.2, 0.5…0.9]. The processing device can take the average or the maximum value as the overall wind speed, or it can obtain the sequence of the above wind speed values.

[0069] In other embodiments, the temporal dimension features include the average acceleration of multiple frames, the rate of change of velocity in multiple frames, and the orientation consistency of multiple frames; the spatial dimension features include the average velocity of multiple frames, the velocity gradient of multiple frames, the velocity variance of multiple frames, and the average percentage of motion region in multiple frames. The processing device can determine the temporal dimension correction coefficient based on the average acceleration, rate of change of velocity, and orientation consistency of multiple frames, and determine the spatial dimension correction coefficient based on the velocity gradient, velocity variance, and average percentage of motion region in multiple frames. Then, the processing device uses the temporal and spatial dimension correction coefficients to correct the average velocity of the multiple frames, obtaining a wind speed that matches the image frame sequence.

[0070] The method for determining the average acceleration across multiple frames is as follows:

[0071]

[0072] in, This represents the average acceleration across multiple frames of images. This represents the velocity value determined through the k-th image frame. This represents the velocity value determined through the (k+1)th image frame. This represents the time difference between two adjacent image frames. This indicates the total number of image frames.

[0073] The method for determining the rate of change of velocity in multiple frames of images is as follows:

[0074]

[0075] in, Indicates the rate of change of speed across multiple frames of images. This represents the velocity value determined through the k-th image frame. This represents the velocity value determined through the kM-th image frame, that is, the velocity value before the M-th frame.

[0076] The method for determining the orientation consistency of multiple frames of images is as follows:

[0077] The processing device first calculates the mean of the dominant orientation of K image frames:

[0078]

[0079] in, This represents the mean of the dominant directions of K image frames. The dominant direction (in radians) of the k-th image frame is represented, and then the orientation consistency across multiple frames is calculated:

[0080]

[0081] in, Indicates the directional consistency of multiple frames of images. This represents the mean of the dominant directions of K image frames.

[0082] The velocity gradient of a multi-frame image is determined as follows:

[0083]

[0084] in, Represents the velocity gradient across multiple frames of an image. This represents the velocity gradient of the k-th image frame.

[0085] The velocity variance of multiple frames is determined as follows:

[0086]

[0087] This represents the velocity variance of multiple frames of images. This represents the velocity variance of the k-th image frame.

[0088] The method for determining the average percentage of motion region in multiple frames of images is as follows:

[0089]

[0090] This represents the average percentage of motion regions across multiple frames of images. This represents the percentage of the motion region in the k-th image frame.

[0091] In some embodiments, the processing device may determine the time dimension correction factor in the following manner:

[0092]

[0093] in, This represents the time dimension correction factor. This represents the average acceleration across multiple frames of images. Indicates the rate of change of speed across multiple frames of images. Indicates the directional consistency of multiple frames of images. Indicates the weight of the first sub-time. Indicates the weight of the second sub-time. This represents the third sub-time weight.

[0094] In some embodiments, the processing device may determine the spatial dimension correction coefficient using the following formula:

[0095]

[0096] in, This represents the spatial dimension correction factor. Represents the velocity gradient across multiple frames of an image. This represents the velocity variance of multiple frames of images. This represents the average percentage of motion regions across multiple frames of images. Indicates the weight of the first subspace. Indicates the weight of the second subspace. This represents the weight of the third subspace.

[0097] In some embodiments, the processing device can correct the average velocity of multiple frames of images using time dimension correction coefficients and spatial dimension correction coefficients to obtain a wind speed that matches the image frame sequence:

[0098]

[0099]

[0100] in, This indicates the wind speed that matches the image frame sequence. This represents the average speed of multiple frames of images. T represents the time weight, and S represents the spatial weight. The time and spatial weights can be customized; in some examples, T=0.7 and S=0.3, while in others, T=S=0.5.

[0101] In some embodiments, after the processing device determines the wind speed that matches the image frame sequence, it can control the air volume of the air conditioner in the vehicle to simulate the air volume in the virtual scene, giving the user a sense of immersion and increasing the user's immersion in the movie.

[0102] In some embodiments, the processing device can pre-set the correspondence between wind speed and white noise. After the processing device determines the wind speed, it can determine the level of white noise corresponding to the wind speed based on the above correspondence. Then, while controlling the air conditioner, the processing device controls the vehicle's audio-visual system to play white noise, thereby enhancing the user's immersive viewing experience.

[0103] Based on the above description, this application provides a wind speed recognition method. The method includes acquiring an image frame sequence, which comprises multiple temporally continuous images; extracting optical flow features from the image frame sequence to obtain optical flow features representing the motion state of a target object in the image frame sequence; processing the optical flow features from both temporal and spatial dimensions to obtain temporal and spatial features; this application effectively separates target motion from background interference through decoupling the temporal and spatial dimensions; and determining the wind speed matching the image frame sequence based on the temporal and spatial features. The temporal features reflect the dynamic changes in wind speed, while the spatial features reflect the spatial consistency of wind speed. When the contrast between the target and the background decreases by 30%, the temporal and spatial fusion mechanism reduces the recognition error by 42% compared to traditional methods, verifying the effectiveness of feature complementarity. In summary, this method systematically solves the robustness problem of traditional optical flow algorithms in complex scenarios through temporal and spatial feature decoupling and multi-dimensional information fusion, providing an innovative solution for high-precision wind speed recognition.

[0104] The above text combined Figure 1 The wind speed identification method provided in the embodiments of this application has been described in detail. The apparatus and equipment provided in the embodiments of this application will be described below with reference to the accompanying drawings.

[0105] like Figure 2 As shown in the figure, this is a schematic diagram of a wind speed identification device provided in an embodiment of this application. The device includes:

[0106] Acquisition module 201 is used to acquire an image frame sequence, the image frame sequence including multiple frames of images that are consecutive in time;

[0107] Feature extraction module 202 is used to extract optical flow features from the image frame sequence to obtain optical flow features representing the motion state of the target object in the image frame sequence;

[0108] The spatiotemporal processing module 203 is used to process the optical flow features from the time dimension and the spatial dimension respectively to obtain the time dimension features and the spatial dimension features;

[0109] The identification module 204 is used to determine the wind speed that matches the image frame sequence based on the time dimension features and the spatial dimension features.

[0110] Optionally, the recognition module 204 is specifically used to input the time dimension features and the spatial dimension features into the wind speed prediction model to obtain the wind speed matching the image frame sequence; wherein, the wind speed prediction model is trained based on the sample time dimension features, the sample spatial dimension features and the sample labels, the sample time dimension features and the sample spatial dimension features are obtained based on the sample optical flow features, and the sample optical flow features are obtained based on the sample image frame sequence.

[0111] Optionally, the time dimension features include: average acceleration of multiple frames, rate of change of velocity of multiple frames, and orientation consistency of multiple frames; the spatial dimension features include average velocity of multiple frames, velocity gradient of multiple frames, velocity variance of multiple frames, and average proportion of motion region of multiple frames; the recognition module 204 is specifically used to determine the time dimension correction coefficient based on the average acceleration of multiple frames, the rate of change of velocity of multiple frames, and orientation consistency of multiple frames; determine the spatial dimension correction coefficient based on the velocity gradient of multiple frames, the velocity variance of multiple frames, and the average proportion of motion region of multiple frames; and correct the average velocity of multiple frames using the time dimension correction coefficient and the spatial dimension correction coefficient to obtain the wind speed matching the image frame sequence.

[0112] Optionally, the recognition module 204 is specifically used to determine the time dimension correction coefficient in the following ways:

[0113]

[0114] in, This represents the time dimension correction factor. This represents the average acceleration across multiple frames of images. Indicates the rate of change of speed across multiple frames of images. Indicates the directional consistency of multiple frames of images. Indicates the weight of the first sub-time. Indicates the weight of the second sub-time. This represents the third sub-time weight.

[0115] Optionally, the recognition module 204 is specifically used to determine the spatial dimension correction coefficient in the following ways:

[0116]

[0117] in, This represents the spatial dimension correction factor. Represents the velocity gradient across multiple frames of an image. This represents the velocity variance of multiple frames of images. This represents the average percentage of motion regions across multiple frames of images. Indicates the weight of the first subspace. Indicates the weight of the second subspace. This represents the weight of the third subspace.

[0118] Optionally, the recognition module 204 is specifically used to obtain the wind speed matching the image frame sequence in the following manner:

[0119]

[0120]

[0121] in, This indicates the wind speed that matches the image frame sequence. This represents the average speed of multiple frames of images. S represents the time weight, and S represents the spatial weight.

[0122] Optionally, the device further includes a control module for controlling the airflow of the air conditioner inside the vehicle based on the wind speed matching the image frame sequence.

[0123] The wind speed identification device according to the embodiments of this application can correspondingly execute the method described in the embodiments of this application, and the other operations and / or functions of each module / unit of the wind speed identification device are respectively for implementing Figure 1 For the sake of brevity, the corresponding processes of each method in the illustrated embodiments will not be described in detail here.

[0124] This application also provides a computing device. For example... Figure 3 As shown in the figure, this is a schematic diagram of a computing device provided in an embodiment of this application. The computing device 300 includes a bus 301, a processor 302, a communication interface 303, and a memory 304. The processor 302, the memory 304, and the communication interface 303 communicate with each other via the bus 301.

[0125] Bus 301 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0126] The processor 302 can be any one or more of the following processors: central processing unit (CPU), graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0127] Communication interface 303 is used for communication with external devices.

[0128] Memory 304 may include volatile memory, such as random access memory (RAM). Memory 304 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0129] The memory 304 stores executable code, and the processor 302 executes the executable code to perform the aforementioned wind speed identification method.

[0130] Specifically, in achieving Figure 2 In the case of the illustrated embodiment, and Figure 2 When the modules or units of the wind speed identification device described in the embodiment are implemented by software, the following steps are performed: Figure 2 The software or program code required for the functions of each module / unit can be partially or entirely stored in memory 304. Processor 302 executes the program code corresponding to each unit stored in memory 304 to perform the aforementioned wind speed identification method.

[0131] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that a computing device can store, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to execute the aforementioned wind speed identification method.

[0132] This application also provides a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application are generated.

[0133] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another website, computer, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0134] When the computer program product is executed by a computer, the computer performs any of the aforementioned wind speed identification methods. The computer program product can be a software installation package; when any of the aforementioned wind speed identification methods needs to be used, the computer program product can be downloaded and executed on the computer.

[0135] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0136] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered within the scope of protection of this application.

Claims

1. A wind speed identification method, characterized in that, The method includes: Acquire an image frame sequence, the image frame sequence comprising multiple temporally consecutive images; Optical flow features are extracted from the image frame sequence to obtain optical flow features representing the motion state of the target object in the image frame sequence; The optical flow features are processed from the time dimension and the spatial dimension respectively to obtain the time dimension features and the spatial dimension features; Based on the time dimension features and the spatial dimension features, determine the wind speed that matches the image frame sequence; The temporal dimension features include: average acceleration of multiple frames, rate of change of velocity of multiple frames, and directional consistency of multiple frames; the spatial dimension features include average velocity of multiple frames, velocity gradient of multiple frames, velocity variance of multiple frames, and average percentage of motion region of multiple frames; determining the wind speed matching the image frame sequence based on the temporal dimension features and the spatial dimension features includes: The time dimension correction coefficient is determined based on the average acceleration of the multi-frame images, the rate of change of velocity of the multi-frame images, and the orientation consistency of the multi-frame images; the spatial dimension correction coefficient is determined based on the velocity gradient of the multi-frame images, the velocity variance of the multi-frame images, and the average proportion of the motion region of the multi-frame images. The average velocity of the multi-frame images is corrected using the time dimension correction coefficient and the spatial dimension correction coefficient to obtain the wind speed that matches the image frame sequence.

2. The method according to claim 1, characterized in that, The step of determining the wind speed matching the image frame sequence based on the time dimension features and the spatial dimension features includes: The time dimension features and the spatial dimension features are input into the wind speed prediction model to obtain the wind speed matching the image frame sequence; wherein, the wind speed prediction model is trained based on the sample time dimension features, the sample spatial dimension features and the sample labels, the sample time dimension features and the sample spatial dimension features are obtained based on the sample optical flow features, and the sample optical flow features are obtained based on the sample image frame sequence.

3. The method according to claim 1, characterized in that, The step of determining the time dimension correction coefficient based on the average acceleration of the multi-frame images, the rate of change of velocity of the multi-frame images, and the orientation consistency of the multi-frame images includes: in, This represents the time dimension correction factor. This represents the average acceleration across multiple frames of images. Indicates the rate of change of speed across multiple frames of images. Indicates the directional consistency of multiple frames of images. Indicates the weight of the first sub-time. Indicates the weight of the second sub-time. This represents the third sub-time weight.

4. The method according to claim 1, characterized in that, The step of determining the spatial dimension correction coefficient based on the velocity gradient of the multi-frame images, the velocity variance of the multi-frame images, and the average proportion of the motion region of the multi-frame images includes: in, This represents the spatial dimension correction factor. Represents the velocity gradient across multiple frames of an image. This represents the velocity variance of multiple frames of images. This represents the average percentage of motion regions across multiple frames of images. Indicates the weight of the first subspace. Indicates the weight of the second subspace. This represents the weight of the third subspace.

5. The method according to claim 1, characterized in that, The average velocity of the multi-frame images is corrected using the time dimension correction coefficient and the spatial dimension correction coefficient to obtain the wind speed matching the image frame sequence, including: in, This represents the time dimension correction factor. This represents the spatial dimension correction factor. This indicates the wind speed that matches the image frame sequence. This represents the average speed of multiple frames of images. S represents the time weight, and S represents the spatial weight.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: The airflow of the air conditioner inside the vehicle is controlled based on the wind speed that matches the image frame sequence.

7. A wind speed identification device, characterized in that, The device includes: The acquisition module is used to acquire an image frame sequence, which includes multiple temporally consecutive images; The feature extraction module is used to extract optical flow features from the image frame sequence to obtain optical flow features representing the motion state of the target object in the image frame sequence; The spatiotemporal processing module is used to process the optical flow features from the time dimension and the spatial dimension respectively to obtain the time dimension features and the spatial dimension features; The identification module is used to determine the wind speed that matches the image frame sequence based on the time dimension features and the spatial dimension features. The temporal dimension features include: average acceleration of multiple frames, rate of change of velocity of multiple frames, and orientation consistency of multiple frames; the spatial dimension features include average velocity of multiple frames, velocity gradient of multiple frames, velocity variance of multiple frames, and average proportion of moving area of ​​multiple frames; the recognition module is specifically used to determine the temporal dimension correction coefficient based on the average acceleration of multiple frames, the rate of change of velocity of multiple frames, and orientation consistency of multiple frames; determine the spatial dimension correction coefficient based on the velocity gradient of multiple frames, the velocity variance of multiple frames, and the average proportion of moving area of ​​multiple frames; and correct the average velocity of multiple frames using the temporal dimension correction coefficient and the spatial dimension correction coefficient to obtain the wind speed matching the image frame sequence.

8. A computing device, characterized in that, Including memory and processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program for performing the method as described in any one of claims 1 to 6.

Images