AI-Based Meteorological Scene Recognition Method and System for Media Assets

Abstract

The present application proposes an AI-based method and system for meteorological scene recognition of media assets. The meteorological scene recognition method includes: obtaining labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier; performing AI recognition on the labeled meteorological observation data to obtain a meteorological virtual scene associated with the meteorological element identifier; performing AI recognition on the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier; importing meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier. The present application aims to solve the technical problem of low meteorological scene resolution in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological services, and particularly to an AI-based method and system for meteorological scene recognition of media assets. Background Art

[0002] Meteorological data has significant guiding significance for production operations and is directly related to the efficiency of production operations. With the continuous development of technologies such as the Internet of Things, artificial intelligence (AI), and big data, the accuracy of meteorological scene recognition based on meteorological observation data as basic data has been greatly improved; it is mainly reflected in the process of dividing meteorological observation data into several subsets with similar characteristic subsets. Each subset is a clustering cluster, and the clustering center of each cluster can represent the typical characteristics of the weather conditions within the subset to a certain extent; these clustering clusters are mainly rainfall, haze amount, temperature, wind force, cloud amount, and so on. However, meteorological scene recognition still uses meteorological observation data as basic data, and the meteorological scenes obtained thereby have low discrimination. Therefore, there is an urgent need for a meteorological scene recognition method and system with high discrimination. Summary of the Invention

[0003] In view of the technical problem of low discrimination of meteorological scenes in the prior art, the present application proposes an AI-based method and system for meteorological scene recognition of media assets.

[0004] The present application proposes an AI-based method for meteorological scene recognition of media assets, including:

[0005] Obtaining labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier;

[0006] Performing AI recognition on the labeled meteorological observation data to obtain a meteorological virtual scene associated with the meteorological element identifier;

[0007] Performing AI recognition on the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier;

[0008] Importing meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier.

[0009] Optionally, before obtaining the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier, the method further includes:

[0010] Obtaining meteorological observation data and meteorological derivative data;

[0011] Identifying scene features in the meteorological derivative data, and obtaining a meteorological element identifier associated with the scene features based on the scene features;

[0012] Traverse the meteorological observation data based on the meteorological element identifier to obtain meteorological observation data containing the meteorological element identifier;

[0013] Correlate the meteorological observation data and meteorological derivative data with the same meteorological element identifier to obtain labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier.

[0014] Optionally, the AI recognition of the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier includes:

[0015] Perform AI recognition on the scene features with the same meteorological element identifier to build a meteorological virtual scene associated with the same meteorological element identifier.

[0016] Optionally, the AI recognition of the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier includes:

[0017] Based on the scene element identifier corresponding to the same meteorological element identifier, build a meteorological virtual scene associated with the meteorological element identifier with a two-dimensional scene model and / or a three-dimensional scene model.

[0018] Optionally, the meteorological observation data includes meteorological professional data and meteorological social observation data.

[0019] This application also proposes a meteorological scene recognition system for media assets based on AI, characterized in that

[0020] An acquisition module for acquiring meteorological observation data and meteorological derivative data with the same meteorological element identifier;

[0021] An AI recognition module for performing AI recognition on the meteorological observation data to obtain meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier; and performing AI recognition on the meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier;

[0022] A scene generation module that imports the meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier.

[0023] Optionally, the obtaining module is further configured to obtain meteorological observation data and meteorological derived data; the AI recognition module is further configured to recognize scene features in the meteorological derived data, and obtain meteorological element identifiers associated with the scene features based on the scene features; traverse the meteorological observation data based on the meteorological element identifiers to obtain meteorological observation data containing the meteorological element identifiers; the recognition system further includes an association module configured to associate the meteorological observation data and the meteorological derived data with the same meteorological element identifier to obtain labeled meteorological observation data and labeled meteorological derived data with the same meteorological element identifier.

[0024] Optionally, the AI recognition module is further configured to perform AI recognition on the scene features with the same meteorological element identifier to construct a meteorological virtual scene associated with the same meteorological element identifier.

[0025] Optionally, the scene generation module is configured to construct a meteorological virtual scene associated with the meteorological element identifier based on the scene element identifier corresponding to the same meteorological element identifier in a two-dimensional scene model and / or a three-dimensional scene model.

[0026] Optionally, the meteorological observation data includes meteorological professional data and meteorological social observation data.

[0027] In the technical solution of the present application, by performing AI recognition on the labeled meteorological observation data and the labeled meteorological derived data with the same meteorological identifier respectively, meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier and a meteorological virtual scene associated with the meteorological element identifier are obtained respectively, and then the meteorological monitoring data corresponding to the meteorological element identifier is imported into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier, so that the obtained meteorological virtual scene has distinctiveness and pertinence. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a schematic flowchart of a method for meteorological scene recognition of media assets based on AI proposed by the present application;

[0029] Figure 2 is a schematic structural diagram of a meteorological scene recognition system for media assets based on AI proposed by the present application. DETAILED DESCRIPTION

[0030] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] Referring to Figure 1 as shown, an AI-based meteorological scene recognition method for media assets proposed by an embodiment of the present application includes:

[0032] S100, obtaining meteorological observation data and meteorological derived data with the same meteorological element identifier;

[0033] S200, based on the meteorological observation data, obtaining meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier;

[0034] S300, performing AI recognition on the meteorological derived data to build a meteorological virtual scene associated with the meteorological element identifier;

[0035] S400, importing the meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier.

[0036] In the technical solution of the present application, by performing AI recognition on the labeled meteorological observation data and labeled meteorological derived data with the same meteorological identifier respectively, to obtain meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier and build a meteorological virtual scene associated with the meteorological element identifier respectively, and then importing the meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene, a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier is obtained, so that the obtained meteorological virtual scene has distinctiveness and pertinence.

[0037] It should be noted that in the technical solution of the present application, AI recognition can be generative adversarial network recognition, convolutional neural network recognition, etc. Perform generative adversarial training or convolutional neural training on the historical meteorological observation data and historical meteorological derived data respectively to obtain a generative adversarial network model or convolutional neural network model for meteorological observation and a generative adversarial network model or convolutional neural network model for meteorological derivation respectively. After obtaining the meteorological observation data and meteorological derived data, send them into their respective generative adversarial network models or convolutional neural network models for calculation to obtain meteorological monitoring data corresponding to the meteorological element identifier and a meteorological virtual scene.

[0038] Taking the recognition by generative adversarial network as an example: A generative adversarial network consists of two parts, a generator and a discriminator. The core idea is to establish a min-max two-player game between the generator and the discriminator. In each training stage, the parameters of the generator are updated to generate new samples, while the discriminator tries to distinguish between real historical samples and newly generated samples. Theoretically, when the game between the two reaches the Nash equilibrium, the optimal solution of the generative adversarial network will provide a generator that can accurately reflect the characteristics of the real data distribution, making it impossible for the discriminator to distinguish whether the samples come from the generator or from the historical training data. At this time, the generated scenario cannot be distinguished from the real historical data, thus ensuring the authenticity and accuracy of the simulated data or scenario.

[0039] Meteorological observation data includes meteorological professional data and meteorological social observation data. Meteorological professional data: Meteorological professional data consists of meteorological observation, meteorological products, meteorological management, meteorological business operation, and industry meteorological observation data generated by business units such as observation, forecasting, service, operation, and management affiliated to the meteorological department. Meteorological social observation data: Meteorological social observation data refers to meteorological observation data generated by socialized meteorological observations for the purpose of meteorological forecasting and service, including meteorological detection data generated by meteorological detection equipment built by research institutions, enterprises, and volunteers, meteorological element sensing data of the location where the intelligent terminal is located obtained by the intelligent terminal equipped with meteorological element sensing components, socialized meteorological data such as weather status and meteorological disaster photos taken and uploaded by Internet users, statistical analysis data of meteorological-related sensitive words by search engines, feedback information from meteorological information dissemination entities and service users, etc.

[0040] Meteorological derivative data includes data generated in society that is not directly used for meteorological-related work, but can generate certain meteorological value through analysis and mining. For example, photos of weather status in natural scenery taken and uploaded by netizens casually, the local rainfall intensity reflected by the swing frequency of the windshield wiper in front of a car during driving, images of meteorological conditions in the monitored area reflected in urban traffic surveillance cameras, humidity collected by humidity sensors in the soil, the driving speed of a car, etc.

[0041] For example, the meteorological observation data is the weather conditions of a certain place, including rainfall, temperature, humidity, and wind force, and the meteorological derivative data is the wiper frequency of the cars running in a certain place; the wiper frequency can reflect the amount of rainfall to a certain extent; thus, the same meteorological identifier for both is rain; performing AI recognition on the labeled meteorological observation data to obtain meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier, that is, obtaining the rainfall; performing AI recognition on the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier, that is, a road with cars running; importing the rainfall into this road to obtain how much rainfall there is on this road, rather than how much rainfall there is in this place, so that the obtained meteorological virtual scene has a higher discrimination degree and can specifically guide the traffic on this road. In the presented meteorological virtual scene, both the characteristic data of the meteorological element identifier and the characteristic data of the scene characteristics are presented. For example, the rainfall is presented while the wiper frequency is also presented, so that the rainfall and the wiper frequency can be compared with each other.

[0042] Also for example, the meteorological observation data is the weather conditions of a certain place, including rainfall; the meteorological derivative data is the humidity data of the soil humidity sensor near the long-distance oil and gas pipeline in a certain place; the humidity data of the soil humidity sensor can reflect the amount of rainfall to a certain extent; thus, the same meteorological identifier for both is rain; performing AI recognition on the labeled meteorological observation data to obtain meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier, that is, obtaining the rainfall; performing AI recognition on the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier, that is, the soil environment around the long-distance oil and gas pipeline; importing the rainfall into the soil environment around the long-distance oil and gas pipeline to obtain how much rainfall there is in the soil environment around the long-distance oil and gas pipeline, rather than how much rainfall there is in this place, so that the obtained meteorological virtual scene has a higher discrimination degree and can specifically monitor the soil conditions near the long-distance oil and gas pipeline. In the presented meteorological virtual scene, both the characteristic data of the meteorological element identifier and the characteristic data of the scene characteristics are presented. For example, the rainfall is presented while the soil humidity is also presented, so that the rainfall and the soil humidity can be compared with each other.

[0043] That is to say, in the technical solution of the embodiment of the present application, the meteorological derivative data usually comes from the operation characteristics possessed by the observed object itself and can reflect the meteorological conditions to a certain extent; while the meteorological observation data usually comes from the sensors arranged in the environment and can collect the meteorological data in the environment in real time.

[0044] Meteorological observation data and meteorological derivative data are continuously generated. In order to effectively perform scene recognition, the meteorological observation data and meteorological derivative data are preliminarily screened so that the associated meteorological observation data and meteorological derivative data can be directly read at the key stage of meteorological recognition, improving the recognition efficiency. Specifically, as an alternative implementation manner of the above embodiment, before obtaining the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier, the method further includes:

[0045] Obtain meteorological observation data and meteorological derivative data;

[0046] Identify the scene features in the meteorological derivative data, and obtain the meteorological element identifier associated with the scene features based on the scene features;

[0047] Traverse the meteorological observation data based on the meteorological element identifier to obtain the meteorological observation data containing the meteorological element identifier;

[0048] Correlate the meteorological observation data and meteorological derivative data with the same meteorological element identifier to obtain the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier.

[0049] In this embodiment, the meteorological element identifier associated with the scene feature is obtained through the scene features in the identified meteorological derivative data; the meteorological observation data containing the meteorological element identifier is obtained by traversing the meteorological observation data through the meteorological element identifier, and then the meteorological observation data and meteorological derivative data with the same meteorological element identifier are correlated with each other, facilitating the direct and rapid construction of the scene in subsequent scene recognition.

[0050] For example, in the embodiment, the scene feature identified by the wiper frequency of a car driving on the road is the wiper frequency, and then the meteorological element identifier matching the scene feature is further obtained as rain; then the meteorological observation data containing the rain on the road is traversed and screened from the meteorological observation data through "rain", and then the two are correlated and labeled to build the scene for the follow-up.

[0051] As an alternative implementation manner of the above embodiment, the AI recognition of the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier includes: performing AI recognition on the scene features of the same meteorological element identifier to build a meteorological virtual scene associated with the same meteorological element identifier.

[0052] In an embodiment, an AI recognition is performed on the scene features with the same meteorological element identifier to build a meteorological virtual scene. For example, in the embodiment, the scene element identifier corresponding to rain is a car driving on a road. Then, the scene element identifier corresponding to the built meteorological virtual scene is a model of a car driving on a road, which can be a static picture or a dynamic picture, can be a three-dimensional model, or can be a planar model.

[0053] As an alternative implementation of the above embodiment, the AI recognition of the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier includes: based on the scene element identifier corresponding to the same meteorological element identifier, building a meteorological virtual scene associated with the meteorological element identifier with a two-dimensional scene model and / or a three-dimensional scene model. In the embodiment, the corresponding meteorological element identifier and scene element identifier are determined according to the scene features, and the two-dimensional scene model and / or the three-dimensional scene model are built with the scene element identifier. For example, both the road and the car are scene element identifiers, that is, a two-dimensional scene model and / or a three-dimensional scene model of a car driving on the road are built; for another example, the oil and gas pipeline and the soil are also scene elements, that is, a two-dimensional scene model and / or a three-dimensional scene model of the oil and gas pipeline buried in the soil are built; then, in the built model, the meteorological element identifier is labeled, such as rainfall, etc., to achieve visual display and integrated detection, effectively improving the immediacy of meteorological monitoring.

[0054] As an alternative implementation of the above embodiment, the scene element identifiers of the meteorological virtual scenes corresponding to different meteorological derivative data may be the same. For example, the windshield wiper frequency and the fog lamp gear both correspond to a car driving on a road; at this time, the corresponding meteorological observation data is rain and / or fog, that is:

[0055] By recognizing different meteorological derivative data, different meteorological element identifiers associated with the same scene element identifier are recognized, and the meteorological characteristic data of the meteorological observation data respectively corresponding to the different meteorological element identifiers are fused into the meteorological virtual scene corresponding to the scene element identifier. Furthermore, a meteorological virtual scene containing different meteorological elements can be obtained, so that the meteorological virtual scene contains different meteorological characteristic data under the same scene element identifier, so as to fuse the meteorological element identifiers with relevance, enrich the fusion effect of meteorological scene recognition, and have a more professional prevention effect.

[0056] As Figure 2 shown, the present application also proposes a meteorological scene recognition system based on AI for media assets, including:

[0057] An acquisition module 100, configured to acquire meteorological observation data and meteorological derivative data with the same meteorological element identifier;

[0058] The AI recognition module 200 is used to perform AI recognition on the meteorological observation data to obtain meteorological monitoring data corresponding to the meteorological element identifiers associated with the meteorological element identifiers; and perform AI recognition on the meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifiers.

[0059] The scene generation module 300 imports the meteorological monitoring data corresponding to the meteorological element identifiers into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifiers.

[0060] In the technical solution of this application, by performing AI recognition on the labeled meteorological observation data and the labeled meteorological derivative data with the same meteorological identifier respectively, to obtain the meteorological monitoring data corresponding to the meteorological element identifiers associated with the meteorological element identifiers and build a meteorological virtual scene associated with the meteorological element identifiers respectively, and then import the meteorological monitoring data corresponding to the meteorological element identifiers into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifiers, so that the obtained meteorological virtual scene has distinctiveness and pertinence.

[0061] It should be noted that in the technical solution of this application, AI recognition can be generative adversarial network recognition, convolutional neural network recognition, etc. Perform generative adversarial training or convolutional neural training on the historical meteorological observation data and the historical meteorological derivative data respectively to obtain the generative adversarial network model or convolutional neural network model of meteorological observation and the generative adversarial network model or convolutional neural network model of meteorological derivative respectively. After obtaining the meteorological observation data and the meteorological derivative data, send them into their respective generative adversarial network models or convolutional neural network models for calculation to obtain the meteorological monitoring data corresponding to the meteorological element identifiers and the meteorological virtual scene.

[0062] As an alternative implementation of the above embodiment, the acquisition module is further configured to acquire meteorological observation data and meteorological derivative data.

[0063] The AI recognition module is further configured to identify the scene features in the meteorological derivative data, and obtain the meteorological element identifiers associated with the scene features based on the scene features; traverse the meteorological observation data based on the meteorological element identifiers to obtain the meteorological observation data containing the meteorological element identifiers.

[0064] The recognition system further includes an association module, configured to associate the meteorological observation data and the meteorological derivative data with the same meteorological element identifier to obtain the labeled meteorological observation data and the labeled meteorological derivative data with the same meteorological element identifier.

[0065] As an alternative implementation of the above embodiment, the AI recognition module is further configured to perform AI recognition on the scene features of the same meteorological element identifier, and construct a meteorological virtual scene associated with the same meteorological element identifier. As an alternative implementation of the above embodiment, the scene generation module is configured to construct a meteorological virtual scene associated with the meteorological element identifier based on the scene element identifier corresponding to the same meteorological element identifier, in a two-dimensional scene model and / or a three-dimensional scene model.

[0066] As an alternative implementation of the above embodiment, the meteorological observation data includes meteorological professional data and meteorological social observation data.

[0067] An embodiment of the present invention further provides an electronic device, which includes a central processing unit (CPU), a system memory including a random access memory (RAM) and a read-only memory (ROM), and a system bus connecting the system memory and the central processing unit. The control device further includes a basic input / output system (I / O system) for facilitating the transfer of information between various devices within the computer, and a mass storage device for storing an operating system, application programs, and other program modules.

[0068] The basic input / output system includes a display for displaying information and input devices such as a mouse, keyboard, etc. for user input of information. Among them, the display and the input devices are both connected to the central processing unit through an input / output controller connected to the system bus. The basic input / output system may further include an input / output controller for receiving and processing inputs from multiple other devices such as a keyboard, mouse, or electronic stylus. Similarly, the input / output controller also provides output to a display screen, printer, or other types of output devices.

[0069] The mass storage device is connected to the central processing unit through a mass storage controller (not shown) connected to the system bus. The mass storage device and its associated computer-readable medium provide non-volatile storage for the control device. That is to say, the mass storage device may include a computer-readable medium (not shown) such as a hard disk or a CD-ROM drive.

[0070] Without loss of generality, the computer-readable medium may include computer storage media and communication media. Computer storage media includes volatile and non-volatile, removable and non-removable media implemented by any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes RAM, ROM, EPROM, EEPROM, flash memory or other solid-state storage technologies, CD-ROM, DVD or other optical storage, magnetic tape cartridges, tapes, disk storage or other magnetic storage devices. Of course, those skilled in the art will know that the computer storage media is not limited to the above several types. The above system memory and mass storage devices may be collectively referred to as memory.

[0071] According to various embodiments of the present invention, the control device may also be connected to a remote computer on the network through a network such as the Internet. That is, the control device may be connected to the network through a network interface unit connected to the system bus, or rather, a network interface unit may also be used to connect to other types of networks or remote computer systems (not shown).

[0072] The memory further includes one or more programs, and the one or more programs are stored in the memory, and the one or more programs are used to execute the methods provided in the above embodiments:

[0073] Obtain the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier;

[0074] Perform AI recognition on the labeled meteorological observation data to obtain a meteorological virtual scene associated with the meteorological element identifier;

[0075] Perform AI recognition on the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier;

[0076] Import the meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier.

[0077] In the technical solution of the present application, by performing AI recognition on the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological identifier respectively, to obtain the meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier and build a meteorological virtual scene associated with the meteorological element identifier respectively, and then import the meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier, so that the obtained meteorological virtual scene has distinctiveness and pertinence.

[0078] It should be noted that in the technical solution of this application, AI recognition can be generative adversarial network recognition, convolutional neural network recognition, etc. The meteorological observation historical data and meteorological derivative historical data are respectively subjected to generative adversarial training or convolutional neural training to obtain a generative adversarial network model or a convolutional neural network model for meteorological observation and a generative adversarial network model or a convolutional neural network model for meteorological derivatives. After obtaining the meteorological observation data and meteorological derivative data, they are respectively sent into their respective generative adversarial network models or convolutional neural network models for calculation to obtain meteorological monitoring data and meteorological virtual scenes corresponding to meteorological element identifiers.

[0079] Further, before obtaining the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier, the method further includes:

[0080] Identify the scene features in the meteorological derivative data, and obtain a meteorological element identifier associated with the scene features based on the scene features;

[0081] Traverse the meteorological observation data based on the meteorological element identifier to obtain meteorological observation data containing the meteorological element identifier;

[0082] Correlate the meteorological observation data and meteorological derivative data with the same meteorological element identifier to obtain the labeled meteorological observation data and labeled meteorological derivative data with the same meteorological element identifier.

[0083] Further, the method further includes performing AI recognition on the scene features with the same meteorological element identifier to build a meteorological virtual scene associated with the same meteorological element identifier.

[0084] Further, the performing AI recognition on the labeled meteorological derivative data to build a meteorological virtual scene associated with the meteorological element identifier includes:

[0085] Based on the scene element identifier corresponding to the same meteorological element identifier, build a meteorological virtual scene associated with the meteorological element identifier with a two-dimensional scene model and / or a three-dimensional scene model.

[0086] Further, the meteorological observation data includes meteorological professional data and meteorological social observation data.

[0087] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An AI-based meteorological scene recognition method for media assets, characterized in that: include: Obtain annotated meteorological observation data and annotated meteorological derivative data with the same meteorological element identifier; Performing AI recognition on the annotated meteorological observation data to obtain a meteorological virtual scene associated with the meteorological element identifier; Performing AI recognition on the annotated meteorological derived data to build a meteorological virtual scene associated with the meteorological element identifier; the meteorological derived data includes weather status photos, windshield wiper swing frequency, meteorological condition images, soil moisture or vehicle driving speed; Importing the meteorological monitoring data corresponding to the meteorological element identifier into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identifier; Before obtaining the annotated meteorological observation data and the annotated meteorological derived data having the same meteorological element identifier, the method further includes: Obtain meteorological observation data and meteorological derived data; Identifying scene features in the meteorological derived data, and obtaining meteorological element identifiers associated with the scene features based on the scene features; Traversing the meteorological observation data based on the meteorological element identifier to obtain the meteorological observation data containing the meteorological element identifier; The meteorological observation data and the meteorological derived data having the same meteorological element identification are correlated with each other to obtain the annotated meteorological observation data and the annotated meteorological derived data having the same meteorological element identification; The step of performing AI recognition on the annotated meteorological derived data to build a meteorological virtual scene associated with the meteorological element identifier includes: Perform AI recognition on the scene features of the same meteorological element identifier and build a meteorological virtual scene associated with the same meteorological element identifier.

2. The meteorological scene recognition method according to claim 1, characterized in that: The step of performing AI recognition on the annotated meteorological derived data to build a meteorological virtual scene associated with the meteorological element identifier includes: Based on the scene element identifier corresponding to the same meteorological element identifier, a meteorological virtual scene associated with the meteorological element identifier is constructed using a two-dimensional scene model and / or a three-dimensional scene model.

3. The meteorological scene recognition method according to claim 1, characterized in that: The meteorological observation data include meteorological professional data and meteorological social observation data.

4. An AI-based meteorological scene recognition system for media assets, characterized in that: An acquisition module, used for acquiring meteorological observation data and meteorological derived data having the same meteorological element identifier; An AI recognition module is used to perform AI recognition on the meteorological observation data to obtain meteorological monitoring data corresponding to the meteorological element identifier associated with the meteorological element identifier; and perform AI recognition on the meteorological derived data to build a meteorological virtual scene associated with the meteorological element identifier; The scene generation module imports the meteorological monitoring data corresponding to the meteorological element identification into the meteorological virtual scene to obtain a meteorological virtual scene containing the meteorological monitoring data corresponding to the meteorological element identification; the acquisition module is also used to acquire meteorological observation data and meteorological derived data; the meteorological derived data includes weather status photos, windshield wiper swing frequency, meteorological condition images, soil moisture or vehicle driving speed; The AI ​​recognition module is also used to identify scene features in the meteorological derived data, and obtain a meteorological element identifier associated with the scene feature based on the scene feature; traverse the meteorological observation data based on the meteorological element identifier to obtain the meteorological observation data containing the meteorological element identifier; The identification system also includes an association module, which is used to associate meteorological observation data and meteorological derived data with the same meteorological element identification to obtain labeled meteorological observation data and labeled meteorological derived data with the same meteorological element identification; the AI ​​identification module is also used to perform AI identification on scene features of the same meteorological element identification to build a meteorological virtual scene associated with the same meteorological element identification.

5. The weather scene recognition system according to claim 4, characterized in that: The scene generation module is configured to build a meteorological virtual scene associated with the meteorological element identifier using a two-dimensional scene model and / or a three-dimensional scene model based on the scene element identifier corresponding to the same meteorological element identifier.

6. The weather scene recognition system according to claim 4, characterized in that: The meteorological observation data include meteorological professional data and meteorological social observation data.

Images