Video pushing method and system based on vehicle and storage medium

Abstract

The invention discloses a vehicle-based video pushing method and system and a storage medium, and belongs to the technical field of data processing. When the vehicle accesses the WiFi network, the server obtains an interest tag of a driver of the vehicle; the server searches for a push video matched with the interest tag in a network video library, and pushes a video clip of a predetermined duration in the push video to the vehicle through a WiFi network; the vehicle obtains vehicle state information, environment information and driver state information, and generates a safety level according to the vehicle state information, the environment information and the driver state information; the vehicle obtains visual attention information and user interaction behavior information, and an attention weight is generated according to the visual attention information and the user interaction behavior information; and the vehicle determines a playing mode according to the safety level and the attention weight, and controls playing of the video clip according to the playing mode. According to the invention, the video is pushed in advance when the WiFi is accessed, the pushing flow is saved, and the driving safety and the watching experience can also be considered.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a vehicle-based video push method, system, and storage medium. Background Art

[0002] Currently, with the increasing popularity of new energy vehicles and intelligent connected vehicles, commercial vehicles are gradually adding connected devices and applications, enabling internet-based content acquisition and a wide variety of videos. Since their launch, these services have been popular with vehicle owners due to their rapid content iteration and excellent user experience. However, due to issues related to driving safety, weak user interaction, and blind push notifications that consume high data traffic, the actual video experience is poor, resulting in low usage. Summary of the Invention

[0003] This application provides a vehicle-based video push method, system, and storage medium to address the problem that video push consumes a large amount of traffic, affects driving safety, and provides a poor viewing experience. The technical solution is as follows: According to a first aspect of the present application, a vehicle-based video push method is provided, the method comprising: When a vehicle accesses a wireless fidelity WiFi network, the server obtains an interest tag of the driver of the vehicle, where the interest tag is a tag of a video category that the driver is interested in; The server searches for a push video that matches the interest tag in a network video library, and pushes a video clip of a predetermined length in the push video to the vehicle via the WiFi network; The vehicle acquires vehicle state information, environmental information, and driver state information, and generates a safety level based on the vehicle state information, the environmental information, and the driver state information; The vehicle acquires visual attention information and user interaction behavior information, and generates an attention weight according to the visual attention information and the user interaction behavior information; The vehicle determines a playback mode according to the safety level and the attention weight, and controls playback of the video clip according to the playback mode.

[0004] In a possible implementation, the vehicle determines the playback mode according to the safety level and the attention weight, including: If the safety level is high and the attention weight is less than a weight threshold, the vehicle determines that the playback mode is full-screen playback; If the safety level is medium and the attention weight is greater than the weight threshold, the vehicle determines the playback mode to be to reduce the playback screen to a predetermined size before playback; If the security level is low, the vehicle determines that the playback mode is not to play the video.

[0005] In a possible implementation, the method further includes: After playing the video clip, the vehicle obtains the remaining video clips in the pushed video from the server; If the vehicle is connected to a mobile data network, the server determines the remaining traffic of the traffic card in the vehicle; The server determines, based on whether the remaining traffic is sufficient, push content for the remaining video clips, and pushes the push content to the vehicle via the mobile data network; The vehicle plays the pushed content in the playing manner.

[0006] In a possible implementation, the server determines, according to whether the remaining traffic is sufficient, the push content of the remaining video segments, including: If the remaining traffic is greater than a first threshold, the server determines the remaining video clips as push content; If the remaining traffic is less than a first threshold and greater than a second threshold, the server determines the audio in the remaining video clip as push content; If the remaining traffic is less than the second threshold and greater than the third threshold, the server determines the key frames in the remaining video clips as push content; If the remaining traffic is less than a third threshold and greater than a fourth threshold, the server extracts key frames from the remaining video clips, reduces the resolution of the key frames, and determines them as push content; If the remaining flow is less than a fourth threshold, the server determines that the pushed content is empty.

[0007] In a possible implementation, the vehicle plays the pushed content in the playback mode, including: When the pushed content is a key frame, the vehicle processes each key frame using a pre-trained network model to obtain a predicted intermediate frame; After the vehicle combines the intermediate frames and the key frames into a video, it plays the video in the playback mode.

[0008] In a possible implementation, the server determines the remaining flow of the flow card in the vehicle, including: The vehicle obtains the remaining flow information of the flow card from the operator, and sends the remaining flow information to the server, and the server determines the remaining flow according to the remaining flow information; or The server obtains the remaining traffic information of the traffic card from the operator, and determines the remaining traffic according to the remaining traffic information.

[0009] In a possible implementation, the method further includes: The vehicle tracks the driver's eye focus during video playback through an in-vehicle camera, and generates the visual attention information based on the length of time the eye focus stays on the playback interface; The vehicle communication obtains the driver's touch screen interaction operation during video playback, and generates user interaction behavior information according to the touch screen interaction operation; The vehicle extracts subtitle keywords during video playback and generates semantic preference information based on the subtitle keywords; The vehicle performs weighted calculation based on the visual attention information, the user interaction behavior information, and the semantic preference information to obtain the driver's interest tag.

[0010] In a possible implementation, the method further includes: The server obtains the playback duration of the pushed video; The server determines the predetermined duration according to the playback duration.

[0011] According to a second aspect of the present application, a vehicle-based video push system is provided, the video push system comprising a vehicle and a server; When a vehicle accesses a wireless fidelity WiFi network, the server is configured to obtain an interest tag of a driver of the vehicle, where the interest tag is a tag of a video category that the driver is interested in; The server is further configured to search a network video library for a push video that matches the interest tag, and push a video clip of a predetermined length in the push video to the vehicle via the WiFi network; The vehicle is configured to obtain vehicle status information, environmental information, and driver status information, and generate a safety level based on the vehicle status information, the environmental information, and the driver status information; The vehicle is further configured to obtain visual attention information and user interaction behavior information, and generate an attention weight based on the visual attention information and the user interaction behavior information; The vehicle is further used to determine a playback mode according to the safety level and the attention weight, and control the playback of the video clip according to the playback mode.

[0012] According to a third aspect of the present application, a computer-readable storage medium is provided, in which at least one instruction is stored. The at least one instruction is loaded and executed by a processor to implement the vehicle-based video push method as described above.

[0013] The beneficial effects of the technical solution provided by this application include at least: When the vehicle is connected to the WiFi network, the server will push video clips of a predetermined length in the push video to the vehicle. In this way, the video can be pushed without consuming the traffic card's traffic. The vehicle can generate a safety level based on vehicle status information, environmental information and driver status information, and generate an attention weight based on visual attention information and user interaction behavior information, so as to determine the video playback method based on the safety level and attention weight. In this way, both driving safety and viewing experience can be taken into account. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0015] Figure 1 This is a flow chart of a vehicle-based video push method provided by one embodiment of the present application; Figure 2 This is a flowchart of a vehicle-based video push method provided by an embodiment of the present application; Figure 3 This is a structural block diagram of a vehicle-based video push system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0016] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0017] like Figure 1 As shown, it shows a method flow chart of a vehicle-based video push method provided by an embodiment of the present application. The vehicle-based video push method may include: Step 101: When a vehicle is connected to a WiFi network, the server obtains an interest tag of the driver of the vehicle, where the interest tag is a tag of a video category that the driver is interested in.

[0018] Interest tags are used to describe the video categories that drivers are interested in, such as comedy, drama, food, life skills, fashion and beauty, cute pets, film and television commentary, etc.

[0019] The vehicle may provide multiple interest tags for the driver to choose from, or the vehicle may analyze the driver's viewing habits of multiple videos to obtain interest tags. This embodiment does not limit the method for generating interest tags.

[0020] The vehicle can send the driver's interest tag to the server, and the server receives the interest tag.

[0021] In step 102 , the server searches for a push video that matches the interest tag in the online video library, and pushes a video clip of a predetermined length in the push video to the vehicle via the WiFi network.

[0022] The online video library pre-stores multiple videos, and each video corresponds to at least one interest tag. The server can filter some videos with higher popularity in the online video library according to the driver's interest tags and determine these videos as push videos.

[0023] For each pushed video, the server can determine the scheduled duration of the pushed video and push the video clip of the scheduled duration to the vehicle under the WiFi network to save data. The scheduled duration may vary for different videos. For example, if the scheduled duration of a long video is 5 minutes, the first 5 minutes of the video clip may be pushed; if the scheduled duration of a short video is 30 seconds, the first 30 seconds of the video clip may be pushed.

[0024] In step 103 , the vehicle obtains vehicle status information, environmental information, and driver status information, and generates a safety level based on the vehicle status information, environmental information, and driver status information.

[0025] The vehicle status information indicates the vehicle status and may include the frequency of sudden acceleration / sudden braking, the number of lane departure warnings, etc.

[0026] Environmental information represents the current external environment of the vehicle, which may include Advanced Driving Assistance System (ADAS) road condition score (congestion index, curve curvature), weather visibility, etc.

[0027] Driver status information indicates the driver's status and may include fatigue index (based on blinking frequency), distraction index (percentage of time the eyes are away from the road), etc.

[0028] The vehicle can use fuzzy control algorithms to calculate vehicle status information, environmental information and driver status information to obtain a safety level.

[0029] In this embodiment, the vehicle can be divided into safety levels of 1-5, and levels 1-2 represent low safety levels, indicating that the vehicle is in a high-risk state; level 3 represents medium safety levels, indicating that the vehicle is in a medium-risk state; levels 4-5 represent high safety levels, indicating that the vehicle is in a low-risk state.

[0030] In step 104 , the vehicle obtains visual attention information and user interaction behavior information, and generates attention weights based on the visual attention information and the user interaction behavior information.

[0031] Visual attention information indicates where the driver's gaze remains. In this embodiment, the vehicle can track eye focus using an in-vehicle camera and analyze the duration of the driver's attention in the area of ​​interest (e.g., the subtitle area / center of the screen).

[0032] User interaction behavior information represents the spatiotemporal distribution of touchscreen interaction operations triggered by the driver while watching a video, such as fast forward, pause, loop playback, etc. The heat map of the video can be constructed based on this information.

[0033] The vehicle processes visual attention information and user interaction behavior information based on a large model, ultimately outputting an attention weight between 0 and 1. A larger attention weight indicates more attention the driver is paying to the video playback, while a smaller attention weight indicates less attention.

[0034] In step 105 , the vehicle determines a playback mode according to the safety level and the attention weight, and controls playback of the video clip according to the playback mode.

[0035] There are many playback methods, such as full-screen playback, small-screen playback, voice playback, etc. Different playback methods will cause different degrees of distraction to the driver. Therefore, we need to select the appropriate playback method based on the current safety level and attention weight, and then use this playback method to control the playback of video clips, so as to take into account both driving safety and viewing experience.

[0036] To sum up, the vehicle-based video push method provided in the embodiment of the present application is that when the vehicle is connected to the WiFi network, the server will push the video clips of the predetermined length in the push video to the vehicle. In this way, the video push can be performed without consuming the traffic of the traffic card; the vehicle can generate a safety level based on the vehicle status information, environmental information and driver status information, and generate an attention weight based on the visual attention information and user interaction behavior information, so as to play the video according to the safety level and attention weight. In this way, driving safety and viewing experience can be taken into account.

[0037] like Figure 2As shown, it shows a flow chart of a vehicle-based video push method provided by an embodiment of the present application. The vehicle-based video push method may include: Step 201: When a vehicle accesses a WiFi network, the server obtains an interest tag of the driver of the vehicle, where the interest tag is a tag of a video category that the driver is interested in.

[0038] Interest tags are used to describe the video categories that drivers are interested in, such as comedy, drama, food, life skills, fashion and beauty, cute pets, film and television commentary, etc.

[0039] The vehicle may provide multiple interest tags for the driver to choose from, or the vehicle may analyze the driver's viewing habits of multiple videos to obtain interest tags. This embodiment does not limit the method for generating interest tags.

[0040] If the interest tag is generated by the vehicle, the vehicle tracks the driver's eye focus during video playback through the in-vehicle camera, and generates visual attention information based on the length of time the eye focus stays on the playback interface; the vehicle obtains the driver's touch screen interaction operations during video playback, and generates user interaction behavior information based on the touch screen interaction operations; the vehicle extracts subtitle keywords during video playback, and generates semantic preference information based on the subtitle keywords; the vehicle performs weighted calculation based on visual attention information, user interaction behavior information and semantic preference information to obtain the driver's interest tag.

[0041] The vehicle can send the driver's interest tag to the server, and the server receives the interest tag.

[0042] In step 202 , the server searches for a push video that matches the interest tag in the online video library, and pushes a video clip of a predetermined length in the push video to the vehicle via the WiFi network.

[0043] The online video library pre-stores multiple videos, and each video corresponds to at least one interest tag. The server can filter some videos with higher popularity in the online video library according to the driver's interest tags and determine these videos as push videos.

[0044] For each pushed video, the server can determine the predetermined duration corresponding to the pushed video, and push the video clip of the predetermined duration in the pushed video to the vehicle under the WiFi network to save traffic.

[0045] Specifically, the server obtains the playback duration of the pushed video and determines the scheduled duration based on the playback duration. For example, if the scheduled duration of a long video is 5 minutes, the first 5 minutes of the video can be pushed; if the scheduled duration of a short video is 30 seconds, the first 30 seconds of the video can be pushed.

[0046] In step 203 , the vehicle obtains vehicle status information, environmental information, and driver status information, and generates a safety level based on the vehicle status information, environmental information, and driver status information.

[0047] The vehicle status information indicates the vehicle status and may include the frequency of sudden acceleration / sudden braking, the number of lane departure warnings, etc.

[0048] Environmental information represents the current external environment of the vehicle, which may include ADAS road condition score (congestion index, curve curvature), weather visibility, etc.

[0049] Driver status information indicates the driver's status and may include fatigue index (based on blinking frequency), distraction index (percentage of time the eyes are away from the road), etc.

[0050] The vehicle can use fuzzy control algorithms to calculate vehicle status information, environmental information and driver status information to obtain a safety level.

[0051] In this embodiment, the vehicle can be divided into safety levels of 1-5, and levels 1-2 represent low safety levels, indicating that the vehicle is in a high-risk state; level 3 represents medium safety levels, indicating that the vehicle is in a medium-risk state; levels 4-5 represent high safety levels, indicating that the vehicle is in a low-risk state.

[0052] In step 204 , the vehicle obtains visual attention information and user interaction behavior information, and generates attention weights based on the visual attention information and the user interaction behavior information.

[0053] Visual attention information indicates where the driver's gaze remains. In this embodiment, the vehicle can track eye focus using an in-vehicle camera and analyze the duration of the driver's attention in the area of ​​interest (e.g., the subtitle area / center of the screen).

[0054] User interaction behavior information represents the spatiotemporal distribution of touchscreen interaction operations triggered by the driver while watching a video, such as fast forward, pause, loop playback, etc. The heat map of the video can be constructed based on this information.

[0055] The vehicle processes visual attention information and user interaction behavior information based on a large model, ultimately outputting an attention weight between 0 and 1. A larger attention weight indicates more attention the driver is paying to the video playback, while a smaller attention weight indicates less attention.

[0056] In step 205 , the vehicle determines a playback mode according to the safety level and the attention weight, and controls the playback of the video clip according to the playback mode.

[0057] There are many playback methods, such as full-screen playback, small-screen playback, voice playback, etc. Different playback methods will cause different degrees of distraction to the driver. Therefore, we need to select the appropriate playback method based on the current safety level and attention weight, and then use this playback method to control the playback of video clips, so as to take into account both driving safety and viewing experience.

[0058] Specifically, the vehicle determines the playback mode based on the safety level and attention weight, which may include: (1) If the safety level is high and the attention weight is less than the weight threshold, the vehicle determines that the playback mode is full-screen playback.

[0059] The weight threshold can be set according to actual needs and is not limited in this embodiment.

[0060] When the playback mode is determined to be full-screen playback, the vehicle plays the previously pushed video clip in full screen.

[0061] (2) If the safety level is medium and the attention weight is greater than the weight threshold, the vehicle determines the playback mode to be to reduce the playback screen to a predetermined size before playing.

[0062] The ratio between the predetermined size and the screen is less than 1, and the ratio may be 1 / 2, 1 / 4, and so on.

[0063] After determining that the playback mode is to reduce the playback screen to a predetermined size and then play it, the vehicle will reduce the playback screen to 1 / 4 and play the previously pushed video clip on the 1 / 4 screen.

[0064] (3) If the security level is low, the vehicle determines that the playback mode is not to play the video.

[0065] When the safety level is low, the video will not be played regardless of the attention weight.

[0066] Step 206: After playing the video clip, the vehicle obtains the remaining video clips in the pushed video from the server.

[0067] If the driver is interested in the video, after watching the video clip, the vehicle needs to obtain the remaining video clips from the server. If the vehicle is currently connected to a WiFi network, the server directly pushes the remaining video clips to the vehicle via the WiFi network, and the vehicle plays the remaining video clips in the original playback mode. If the vehicle is currently connected to a mobile data network, step 207 is executed.

[0068] Step 207: If the vehicle is connected to the mobile data network, the server determines the remaining traffic of the traffic card in the vehicle.

[0069] When obtaining the remaining traffic, the first method is that the vehicle obtains the remaining traffic information of the traffic card from the operator, sends the remaining traffic information to the server, and the server determines the remaining traffic based on the remaining traffic information; the second method is that the server obtains the remaining traffic information of the traffic card from the operator, and determines the remaining traffic based on the remaining traffic information.

[0070] In step 208 , the server determines the push content of the remaining video clips based on whether the remaining traffic is sufficient, and pushes the push content to the vehicle via the mobile data network.

[0071] The server can generate different push content according to the amount of remaining traffic and push the push content to the vehicle through the mobile data network.

[0072] Specifically, the server determines the push content of the remaining video clips based on whether the remaining traffic is sufficient, which may include: (1) If the remaining traffic is greater than the first threshold, the server determines the remaining video clips as push content.

[0073] If the remaining traffic is greater than the first threshold, it means that the remaining traffic is sufficient. At this time, there is no need to consider how much traffic the video push consumes. The server directly determines the remaining video clips as the push content.

[0074] (2) If the remaining traffic is less than the first threshold and greater than the second threshold, the server determines the audio in the remaining video clip as the push content.

[0075] Since the traffic required to transmit audio is less than the traffic required to transmit video, the vehicle can only use audio as push content to reduce traffic consumption.

[0076] (3) If the remaining traffic is less than the second threshold and greater than the third threshold, the server determines the key frames in the remaining video clips as push content.

[0077] Since the traffic required to transmit key frames is less than the traffic required to transmit audio, the vehicle can only use key frames as push content to further reduce traffic consumption.

[0078] (4) If the remaining traffic is less than the third threshold and greater than the fourth threshold, the server extracts key frames from the remaining video clips, reduces the resolution of the key frames, and determines them as push content.

[0079] Reducing the resolution of keyframes can further reduce the bandwidth required for transmission.

[0080] (5) If the remaining traffic is less than the fourth threshold, the server determines that the pushed content is empty.

[0081] The pushed content is empty, that is, the video is not played.

[0082] It should be noted that the server can determine the push content based on the remaining traffic, or the vehicle can display options such as "download only key frames", "pause download", "switch voice mode" based on the remaining traffic for the driver to choose, and the server can determine the push content based on the driver's choice.

[0083] Step 209: The vehicle plays the pushed content in the playback mode.

[0084] When the pushed content is a video, the vehicle plays the video in the original playback mode; when the pushed content is audio, the vehicle plays the audio in the original playback mode; when the pushed content is a key frame, the vehicle needs to restore the intermediate frames based on the key frame and play the key frame and intermediate frames in the original playback mode; when the pushed content is empty, the vehicle does not play any video or audio.

[0085] Specifically, when the pushed content is a key frame, the vehicle uses a pre-trained network model to process each key frame to obtain a predicted intermediate frame; after the vehicle combines the intermediate frame and the key frame into a video, it plays the video in a playback mode.

[0086] In this embodiment, a lightweight artificial intelligence (AI) model is deployed in the vehicle to support real-time video processing and safety analysis.

[0087] In this embodiment, the vehicle can also upload the attention weights and images captured during eye tracking to a server for storage. To protect user privacy, the images can be desensitized to convert them into vectors, and the attention weights and vectors can then be uploaded to the server for storage.

[0088] In summary, the vehicle-based video push method provided by the embodiments of the present application is that when a vehicle is connected to a WiFi network, the server will push a video clip of a predetermined length in the push video to the vehicle. In this way, video push can be performed without consuming the traffic of the data card; the vehicle can generate a safety level based on vehicle status information, environmental information, and driver status information, and generate an attention weight based on visual attention information and user interaction behavior information, so as to adjust the video playback mode according to the safety level and attention weight. In this way, driving safety and viewing experience can be taken into account.

[0089] In the urban commuting scenario, video push includes the following stages: Pre-caching stage: The vehicle is connected to a normal WiFi network and automatically downloads the first 5 minutes of the "Morning News" subscribed by the driver (including keyframe metadata) based on the driver's interest tags.

[0090] Driving phase: During the morning rush hour, the vehicle detected that the driver frequently checked the rearview mirror and determined the overall safety level to be Level 3. Trigger recommendation strategy: The news video automatically shrinks to a floating window and a voice summary is broadcast simultaneously; When the remaining traffic reaches a certain threshold, the system triggers the push of optional options (such as "download only key frames", "pause download", and "switch audio mode"), which are immediately executed after the driver selects them; When the remaining traffic reaches 50MB, it switches to keyframe mode (bitrate drops from 4Mbps to 1Mbps), and the vehicle uses Generative Adversarial Networks (GAN) to complete the picture and maintain a smooth 45fps.

[0091] In long-distance self-driving scenarios, video push includes the following stages: Pre-caching stage: The vehicle is connected to a normal WiFi network and automatically downloads the first 10 minutes of high-definition video (including multi-camera perspective data) of the "Self-Driving Route Guide" based on the driver's interest tags.

[0092] Driving phase: On highway curves (radius of curvature <100m), the vehicle is determined to have a safety level of 2, video playback is paused, and the head-up display (HUD) navigation guidance is magnified. On straight sections, after the safety level is determined to have returned to level 5, based on the driver's repeated viewing behavior of "scenery clips" (e.g., an attention weight of 0.9), documentaries about similar scenic spots are recommended and subsequent content is pre-cached.

[0093] When the remaining traffic reaches a certain threshold, the system triggers the push of optional options (such as "download only key frames", "pause download", and "switch audio mode"), which are immediately executed after the driver selects them; When the remaining traffic reaches 50MB, it switches to keyframe mode (bitrate drops from 4Mbps to 1Mbps), and the vehicle uses GAN to complete the picture and maintain a smooth 45fps.

[0094] like Figure 3 , which shows a structural block diagram of a vehicle-based video push system provided by an embodiment of the present application, the vehicle-based video push system includes a vehicle 310 and a server 320; When the vehicle 310 accesses the WiFi network, the server 320 is configured to obtain interest tags of the driver of the vehicle 310 , where the interest tags are tags of video categories that the driver is interested in; The server 320 is further configured to search for a push video matching the interest tag in an online video library, and push a video segment of a predetermined length in the push video to the vehicle 310 via a WiFi network; Vehicle 310, for obtaining vehicle state information, environmental information, and driver state information, and generating a safety level based on the vehicle state information, environmental information, and driver state information; The vehicle 310 is further configured to obtain visual attention information and user interaction behavior information, and generate an attention weight based on the visual attention information and the user interaction behavior information; The vehicle 310 is further configured to determine a playback mode according to the safety level and the attention weight, and control playback of the video clip according to the playback mode.

[0095] In an optional embodiment, the vehicle 310 is further configured to: If the security level is high and the attention weight is less than the weight threshold, the playback mode is determined to be full-screen playback; If the security level is medium and the attention weight is greater than the weight threshold, the playback mode is determined to be to reduce the playback image to a predetermined size before playback; If the security level is low, the playback mode is determined to be not playing the video.

[0096] In an optional embodiment, after playing the video clip, the vehicle 310 is further configured to obtain the remaining video clips in the pushed video from the server 320; If the vehicle 310 is connected to the mobile data network, the server 320 is further used to determine the remaining flow of the flow card in the vehicle 310; The server 320 is further configured to determine the push content of the remaining video clips based on whether the remaining traffic is sufficient, and push the push content to the vehicle 310 via the mobile data network; The vehicle 310 is also used to play the pushed content in a playback mode.

[0097] In an optional embodiment, the server 320 is further configured to: If the remaining traffic is greater than the first threshold, the remaining video clips are determined as push content; If the remaining traffic is less than the first threshold and greater than the second threshold, the audio in the remaining video clip is determined as the pushed content; If the remaining traffic is less than the second threshold and greater than the third threshold, determining the key frames in the remaining video clips as push content; If the remaining traffic is less than the third threshold and greater than the fourth threshold, extracting key frames from the remaining video clips, reducing the resolution of the key frames, and determining them as push content; If the remaining flow is less than the fourth threshold, it is determined that the pushed content is empty.

[0098] In an optional embodiment, the vehicle 310 is further configured to: When the pushed content is a key frame, the pre-trained network model is used to process each key frame to obtain the predicted intermediate frame; After combining the intermediate frames and key frames into a video, play the video in playback mode.

[0099] In an optional embodiment, the vehicle 310 is further configured to obtain the remaining flow information of the flow card from the operator, and send the remaining flow information to the server 320, and the server 320 is further configured to determine the remaining flow according to the remaining flow information; or The server 320 is further configured to obtain the remaining traffic information of the traffic card from the operator, and determine the remaining traffic according to the remaining traffic information.

[0100] In an optional embodiment, the vehicle 310 is further configured to: The driver's eye focus during video playback is tracked using the in-car camera, and visual attention information is generated based on the length of time the eye focus remains on the playback interface. By acquiring the driver's touch screen interaction operations during video playback, user interaction behavior information is generated based on the touch screen interaction operations; Extract subtitle keywords during video playback and generate semantic preference information based on the subtitle keywords; The driver's interest tag is obtained by performing weighted calculation based on visual attention information, user interaction behavior information and semantic preference information.

[0101] In an optional embodiment, the server 320 is further configured to: Get the playback duration of the pushed video; The scheduled duration is determined based on the playback duration.

[0102] To sum up, the vehicle-based video push system provided by the embodiment of the present application is such that when the vehicle is connected to a WiFi network, the server will push video clips of a predetermined length in the push video to the vehicle. In this way, video push can be performed without consuming the traffic of the traffic card; the vehicle can generate a safety level based on vehicle status information, environmental information and driver status information, and generate an attention weight based on visual attention information and user interaction behavior information, thereby determining the video playback method based on the safety level and attention weight. In this way, both driving safety and viewing experience can be taken into account.

[0103] One embodiment of the present application provides a computer-readable storage medium, wherein the storage medium stores at least one instruction, and the at least one instruction is loaded and executed by a processor to implement the vehicle-based video push method as described above.

[0104] It should be noted that the vehicle-based video push system provided in the above embodiment only uses the division of the above functional modules as an example to illustrate when performing vehicle-based video push. In actual applications, the above functional distribution can be completed by different functional modules as needed, that is, the internal structure of the vehicle-based video push system can be divided into different functional modules to complete all or part of the functions described above. In addition, the vehicle-based video push system provided in the above embodiment and the vehicle-based video push method embodiment are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0105] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, which may be a read-only memory, a disk, or an optical disk, etc.

[0106] The above description is not intended to limit the embodiments of the present application. Any adjustments, equivalent replacements, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the embodiments of the present application.

Claims

1. A vehicle-based video push method, characterized in that: The method comprises: When a vehicle accesses a wireless fidelity WiFi network, the server obtains an interest tag of the driver of the vehicle, where the interest tag is a tag of a video category that the driver is interested in; The server searches for a push video that matches the interest tag in a network video library, and pushes a video clip of a predetermined length in the push video to the vehicle via the WiFi network; The vehicle acquires vehicle state information, environmental information, and driver state information, and generates a safety level based on the vehicle state information, the environmental information, and the driver state information; The vehicle acquires visual attention information and user interaction behavior information, and generates an attention weight according to the visual attention information and the user interaction behavior information; The vehicle determines a playback mode according to the safety level and the attention weight, and controls playback of the video clip according to the playback mode.

2. The vehicle-based video push method according to claim 1, characterized in that: The vehicle determines a playback mode according to the safety level and the attention weight, including: If the safety level is high and the attention weight is less than a weight threshold, the vehicle determines that the playback mode is full-screen playback; If the safety level is medium and the attention weight is greater than the weight threshold, the vehicle determines the playback mode to be to reduce the playback screen to a predetermined size before playback; If the security level is low, the vehicle determines that the playback mode is not to play the video.

3. The vehicle-based video push method according to claim 1, characterized in that: The method further comprises: After playing the video clip, the vehicle obtains the remaining video clips in the pushed video from the server; If the vehicle is connected to a mobile data network, the server determines the remaining traffic of the traffic card in the vehicle; The server determines, based on whether the remaining traffic is sufficient, push content for the remaining video clips, and pushes the push content to the vehicle via the mobile data network; The vehicle plays the pushed content in the playing manner.

4. The vehicle-based video push method according to claim 3, characterized in that: The server determines, according to whether the remaining traffic is sufficient, the push content of the remaining video clips, including: If the remaining traffic is greater than a first threshold, the server determines the remaining video clips as push content; If the remaining traffic is less than a first threshold and greater than a second threshold, the server determines the audio in the remaining video clip as push content; If the remaining traffic is less than the second threshold and greater than the third threshold, the server determines the key frames in the remaining video clips as push content; If the remaining traffic is less than a third threshold and greater than a fourth threshold, the server extracts key frames from the remaining video clips, reduces the resolution of the key frames, and determines them as push content; If the remaining flow is less than a fourth threshold, the server determines that the pushed content is empty.

5. The vehicle-based video push method according to claim 4, characterized in that: The vehicle plays the pushed content in the playing mode, including: When the pushed content is a key frame, the vehicle processes each key frame using a pre-trained network model to obtain a predicted intermediate frame; After the vehicle combines the intermediate frames and the key frames into a video, it plays the video in the playback mode.

6. The vehicle-based video push method according to claim 1, characterized in that: The server determines the remaining flow of the flow card in the vehicle, including: The vehicle obtains the remaining flow information of the flow card from the operator, and sends the remaining flow information to the server, and the server determines the remaining flow according to the remaining flow information; or The server obtains the remaining traffic information of the traffic card from the operator, and determines the remaining traffic according to the remaining traffic information.

7. The vehicle-based video push method according to claim 1, characterized in that: The method further comprises: The vehicle tracks the driver's eye focus during video playback through an in-vehicle camera, and generates the visual attention information based on the length of time the eye focus stays on the playback interface; The vehicle communication obtains the driver's touch screen interaction operation during video playback, and generates user interaction behavior information according to the touch screen interaction operation; The vehicle extracts subtitle keywords during video playback and generates semantic preference information based on the subtitle keywords; The vehicle performs weighted calculation based on the visual attention information, the user interaction behavior information, and the semantic preference information to obtain the driver's interest tag.

8. The vehicle-based video push method according to any one of claims 1 to 7, characterized in that: The method further comprises: The server obtains the playback duration of the pushed video; The server determines the predetermined duration according to the playback duration.

9. A vehicle-based video push system, characterized in that: The video push system includes a vehicle and a server; When a vehicle accesses a wireless fidelity WiFi network, the server is configured to obtain an interest tag of a driver of the vehicle, where the interest tag is a tag of a video category that the driver is interested in; The server is further configured to search a network video library for a push video that matches the interest tag, and push a video clip of a predetermined length in the push video to the vehicle via the WiFi network; The vehicle is configured to obtain vehicle status information, environmental information, and driver status information, and generate a safety level based on the vehicle status information, the environmental information, and the driver status information; The vehicle is further configured to obtain visual attention information and user interaction behavior information, and generate an attention weight based on the visual attention information and the user interaction behavior information; The vehicle is further configured to determine a playback mode according to the safety level and the attention weight, and control playback of the video clip according to the playback mode.

10. A computer-readable storage medium, characterized in that The storage medium stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the vehicle-based video push method according to any one of claims 1 to 8.

Images