A system and method for automatic video shooting and generation by vehicle-mounted drone

Abstract

This invention discloses a vehicle-mounted drone video automatic shooting and generation system and method, belonging to the field of drones. The system includes: a vehicle-side component, comprising a main control unit, a data communication gateway, a vehicle and environmental status perception module, an interactive interface, and a cabin power supply interface; a drone-side component, comprising an airborne controller, a communication module, a drone status perception module, and a shooting and flight module; and a cloud-side component, comprising a cloud server. Both the main control unit and the cloud server have built-in AI engines. The main control unit is connected to the data communication gateway, the vehicle and environmental status perception module, the interactive interface, and the cabin power supply interface; the airborne controller is connected to the communication module, the drone status perception module, and the shooting and flight module; and the data communication gateway is connected to the communication module and the cloud server. This invention enables fully automatic generation of high-quality videos in outdoor mobile scenarios.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicles (UAVs). Specifically, this invention relates to a system and method for automatic video shooting and generation by a vehicle-mounted UAV. Background Technology

[0002] Against the backdrop of the accelerating integration of the low-altitude economy and intelligent connected pickup trucks, consumer drones, as core tools for personal mobile image creation, have become a key engine driving the upgrading of mass creative expression and the digital content industry through their intelligent and automated operation capabilities. Meanwhile, new energy vehicles, especially new energy pickup trucks with large-capacity batteries and external discharge capabilities, are rapidly gaining popularity in scenarios such as self-driving tours and outdoor adventures due to their excellent outdoor passability, spacious cargo space, and mobile energy platform attributes. Therefore, the deep-seated bottlenecks hindering the widespread application of consumer drones in the field of creative aerial photography—such as high requirements for professional operation, complex post-production, and a lack of creative planning capabilities—are expected to be systematically resolved through deep integration with intelligent connected vehicles.

[0003] In existing technologies, there are several solutions attempting to simplify the drone aerial photography process. One is a mobile device-based control solution, such as the "Target Shooting Method, Shooting Device, and Drone Based on Drone" disclosed in Chinese Patent CN111462229B. This solution provides a preset flight path through a smart terminal application, which is then automatically executed by the drone. This solution is effective in static scenes. Another solution is a vehicle-mounted mobile take-off and landing platform solution, such as the one disclosed in *Automotive Engineering*, Vol. 44, No. 8, 2022. This involves modifying a vehicle to carry the drone, achieving vehicle-mounted storage and mobile relay take-off and landing, and using Bluetooth or Wi-Fi for basic control. While these solutions attempt to simplify the aerial photography process from different angles, they still have significant limitations in achieving true "end-to-end" automation and creative micro-film generation.

[0004] Therefore, this invention proposes a system and method for automatic video shooting and generation by vehicle-mounted unmanned aerial vehicles. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies and proposes a vehicle-mounted drone video automatic shooting and generation system and method to achieve the following objective: to realize the fully automatic generation of high-quality videos in outdoor mobile scenarios.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a vehicle-mounted drone video automatic shooting and generation system, characterized in that: the system includes: a vehicle terminal 10, a drone terminal 20, and a cloud terminal 30, wherein the vehicle terminal 10 is connected to the drone terminal 20 and the cloud terminal 30 respectively. The vehicle-side 10 includes a main control unit 11, a data communication gateway 12, a vehicle and environmental status perception module 13, an interactive interface 14, and a cabin energy supply interface 15; the drone-side 20 includes an airborne controller 21, a communication module 22, a drone status perception module 23, and a shooting and flight module 24; the cloud-side 30 includes a cloud server 31; both the main control unit 11 and the cloud server 31 have built-in AI engines. The main control unit 11 is connected to the data communication gateway 12, the vehicle and environmental status perception module 13, the interactive interface 14, and the cabin energy supply interface 15, respectively; the airborne controller 21 is connected to the communication module 22, the UAV status perception module 23, and the shooting and flight module 24, respectively; and the data communication gateway 12 is connected to the communication module 22 and the cloud server 31, respectively.

[0007] Preferably, the data communication gateway 12 communicates wirelessly with the communication module 22 and the cloud server 31 respectively through wireless communication technology, including 4G, 5G, and C-V2X.

[0008] This invention also provides a method for automatic video shooting and generation by a vehicle-mounted drone, using the above-described automatic video shooting and generation system for a vehicle-mounted drone, the method comprising the following steps: Step S1: After setting the video shooting theme on the vehicle terminal 10, a shooting script is generated based on the collected vehicle and environmental status data and the cloud data obtained by accessing the cloud 30, and then sent to the drone terminal 20. Step S2: The drone terminal 20 analyzes and executes the shooting script, and sends the captured video footage back to the vehicle terminal 10; Step S3: After receiving the video material, the vehicle terminal 10 uses its built-in AI engine to perform rough editing on the video material, and then intelligently selects the rendering path according to network conditions and user settings to generate a first-edition video. Step S4: After the vehicle terminal 10 provides the initial version of the video to the user for preview and confirmation, the final version of the video is generated and saved. Step S5: After completing the shooting task, the drone returns to base.

[0009] Preferably, step S1 includes: Step S101: After the user selects a preset video theme through the interactive interface 14, the user sends it to the main control unit 11. At the same time, the vehicle and environmental status perception module 13 collects vehicle and environmental status data and sends it to the main control unit 11. Step S102: The main control unit 11 accesses the digital script library in the cloud 30 through the data communication gateway 13, and generates a shooting script based on the selected video theme, vehicle and environmental status data.

[0010] Preferably, step S2 includes: Step S201: After receiving the shooting script through the communication module 22, the airborne controller 21 analyzes it and parses the shooting script into specific executable instructions for the flight controller, gimbal, and camera. Step S202: The airborne controller 21 sends executable commands from the flight controller, gimbal, and camera to the shooting and flight module 24 to begin the shooting task; Step S203: During the shooting mission, the video footage captured by the shooting and flight module 24 and the UAV status data collected by the UAV status perception module 23 are transmitted to the communication module 22 through the airborne controller 21. The communication module 22 then uploads it to the main control unit 11 through the data communication gateway 12. Step S204: After each shot is completed, the main control unit 11 immediately performs a multi-index lens quality check. If the technical or artistic index of the lens does not reach the corresponding preset threshold, the intelligent reshoot mechanism is triggered, and the shooting parameters are adjusted before the shooting task of the corresponding lens is re-executed.

[0011] Preferably, during the shooting mission, the main control unit 11 dynamically adjusts the executable commands of the flight controller, gimbal, and camera based on the video footage, UAV status data, and real-time environmental data collected by the vehicle and environmental status perception module 13, and sequentially sends these commands to the shooting and flight module 24 through the data communication gateway 12, communication module 22, and airborne controller 21; the dynamic adjustment includes: If the ambient light level is detected to be lower than the preset ambient light threshold, the drone camera is instructed to switch to night mode and automatically adjust the aperture and ISO parameters. If there are obstacles in the shooting path, the flight controller is instructed to replan a safe path. If the subject is lost or the composition is poor, the gimbal and flight controller parameters are dynamically adjusted, and target tracking and composition optimization are performed again.

[0012] Preferably, during the execution of the shooting task, the main control unit 11 establishes a multi-level anomaly response mechanism, including: When the wind speed exceeds the preset wind speed level, the main control unit 11 forces the drone to stop the shooting mission and return to base immediately. When the location signal strength of the drone is less than the preset strength threshold, the main control unit 11 forces the drone to stop the shooting mission and return to home immediately. When the drone's battery level is lower than the preset first battery threshold, the main control unit 11 controls the drone to only execute the key shots in the shooting script before returning to home and charging in the vehicle. When the rainfall intensity exceeds the preset rainfall intensity threshold, the main control unit 11 limits the drone's flight altitude to below the first preset altitude; When the vehicle enters a tunnel or dense forest area, the main control unit 11 limits the drone's flight altitude to above the second preset altitude and switches to aerial image acquisition mode.

[0013] Preferably, in step S3, after receiving the video footage, the vehicle-mounted device 10 performs a rough cut on the video footage using its built-in AI engine, including: The AI ​​engine of the main control unit 11 receives and classifies the shooting materials, and establishes a material index based on the timestamp and lens number; The AI ​​engine of the main control unit 11 performs time-series assembly of the shooting materials based on the shooting script; The AI ​​engine of the main control unit 11 automatically matches transition effects, background music, color filters, and sets AI subtitles based on the shooting script.

[0014] Preferably, in step S3, intelligently selecting the rendering path based on network conditions and user settings to generate the initial version of the final image includes: When the network bandwidth is less than the preset bandwidth threshold or the user selects quick rendering through the interactive interface 14, the main control unit 11 calls the vehicle-side local GPU to complete the rendering. When the network bandwidth is greater than or equal to the preset bandwidth threshold and the user selects the ultimate image quality through the interactive interface 14, the main control unit 11 uploads the corresponding project file to the cloud server 31 through the data communication gateway 12, and then uses the AI ​​engine deployed on the cloud server 31 to perform enhancement processing to complete the final rendering.

[0015] Preferably, step S5 includes: Step S501: After the drone completes the shooting mission, the onboard controller 21 executes the intelligent return-to-home procedure, and achieves precise docking and charging with the cabin energy supply interface 15 through UWB precise positioning and visual recognition guidance. Step S502: After the main control unit 11 detects the successful docking of the drone with the cabin power supply interface 15, it intelligently selects the charging mode based on the drone's current battery level, battery health status, and subsequent mission requirements, including: When the battery level is below the preset second battery threshold and there is a shooting task after a preset time, select the fast charging mode and charge with a charging parameter of 9V / 3A; otherwise, select the slow charging mode and charge with a charging parameter of 5V / 2A. Step S503: During the charging process, the main control unit 11 monitors the vehicle battery power, drone charging status and interface temperature in real time.

[0016] The technical effects of this invention are as follows: 1. Automation and ease of use: Through automatic script generation and flight control in collaboration between vehicle and machine, as well as automatic editing in collaboration between edge and cloud, the entire process from shooting to final film is automated, which greatly reduces the user's operating threshold. 2. Immediacy and efficiency: By using the onboard computing unit as an edge node for localized rough cutting and previewing, the traditionally time-consuming post-production process is greatly compressed, enabling minute-level finished products in mobile scenarios, and fundamentally improving the user experience. 3. Adaptability and Safety: Through the dynamic environment perception and safety decision-making module, the system can adaptively adjust the shooting strategy according to real-time conditions such as road conditions and weather, so as to capture the best shot while ensuring equipment safety. 4. Resource optimization: Through a hierarchical processing architecture of coarse editing on the vehicle side and fine editing in the cloud, the amount of data that needs to be uploaded to the cloud is effectively reduced, the dependence on network bandwidth and traffic costs are reduced, and the latency problem in the pure cloud processing mode is solved. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a vehicle-mounted drone video automatic shooting and generation system provided in an embodiment of the present invention. Detailed Implementation

[0018] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. This is to help those skilled in the art to have a more complete, accurate, and in-depth understanding of the inventive concept and technical solutions of the present invention, and to facilitate its implementation. It should be noted that the terms "first," "second," etc., used in this application are only for the convenience of describing the technical solutions and to distinguish components; the corresponding component configurations may be the same or different, and are not intended to limit the scope of this application. To make the technical solutions of the present invention clearer, the present invention will be explained and illustrated through the following embodiments.

[0019] In existing technologies, simple control schemes based on mobile devices cannot adapt their preset static flight paths to the real-time changes in the vehicle's movement and the dynamic landscape. Essentially, this confines creative thinking to limited, fixed templates, resulting in a severe disconnect between the captured footage and the actual movement of the vehicle and the scenery outside the window, failing to generate an immersive, integrated film. While vehicle-mounted mobile take-off and landing platforms improve take-off and landing convenience, they generally treat drones merely as independent shooting tools, with shooting and editing processes separated: the drone collects raw footage, while complex editing still requires manual completion by the user on their personal computer after the trip. This process is not only cumbersome and time-consuming but also misses valuable opportunities for real-time reshoots and creative adjustments during the journey due to the inability to preview the final product immediately. Although cloud-based material management architectures attempt to automate editing, their "shoot first, upload later, wait for cloud processing" model generates significant network latency and data costs due to the need to transmit large amounts of high-bitrate raw material, severely disrupting the immediacy and continuity from shooting to final product, making the "one-click micro-movie" experience a misnomer. Crucially, existing technology systems have failed to build an end-to-end solution with the vehicle's infotainment system as the intelligent processing hub. They cannot perform real-time pre-editing based on the vehicle's computing power during shooting, or immediately use the vehicle's large screen for AI-powered fine editing and generate the final video in seconds after shooting. This results in a significant lack of autonomy and immediacy in the system, making it difficult to meet users' ultimate demand for "what you see is what you get" creative expression in mobile scenarios.

[0020] Therefore, embodiments of the present invention provide a system and method for automatic video shooting and generation by a vehicle-mounted drone, aiming to solve the following problems: (1) Solve the technical problems of fully automated take-off, landing, recovery and charging of UAVs on mobile vehicle platforms; (2) Construct an intelligent creation system that integrates vehicle, machine, and cloud technologies to solve the problems of camera movement control and creative planning for ordinary users; (3) Based on real-time vehicle location, environmental landscape and user theme, generate and execute dynamic shooting scripts autonomously; (4) To solve the problems of fragmented shooting and processing in the existing solution, as well as high latency and high traffic consumption in cloud processing mode, end-to-end automated video generation is achieved through vehicle edge computing and cloud collaboration.

[0021] Specifically, such as Figure 1 As shown, the system in this embodiment includes: vehicle terminal 10, drone terminal 20, and cloud terminal 30, wherein the vehicle terminal 10 is connected to the drone terminal 20 and the cloud terminal 30 respectively; The vehicle-side 10 includes a main control unit 11, a data communication gateway 12, a vehicle and environmental status perception module 13, an interactive interface 14, and a cabin energy supply interface 15; the drone-side 20 includes an airborne controller 21, a communication module 22, a drone status perception module 23, and a shooting and flight module 24; the cloud-side 30 includes a cloud server 31; both the main control unit 11 and the cloud server 31 have built-in AI engines (artificial intelligence engines, such as chatgpt, deepseek, etc.). The main control unit 11 is connected to the data communication gateway 12, the vehicle and environmental status perception module 13, the interactive interface 14, and the cabin energy supply interface 15, respectively; the airborne controller 21 is connected to the communication module 22, the UAV status perception module 23, and the shooting and flight module 24, respectively; and the data communication gateway 12 is connected to the communication module 22 and the cloud server 31, respectively.

[0022] The main control unit 11 is the data processing and control core of the vehicle-side 10, and is equipped with an AI engine. It can typically use a high-performance integrated controller such as the vehicle control unit (VCU). Based on this, the main control unit 11 can integrate vehicle GPS trajectory, vehicle speed, surrounding environmental visual information and user-selected video themes in real time, and simultaneously access the cloud-based digital script library to dynamically generate a personalized shooting script that includes multiple camera angles, camera movement trajectories and emotional tone.

[0023] The data communication gateway 12 communicates wirelessly with the communication module 22 and the cloud server 31 via wireless communication technologies, including 4G, 5G, and C-V2X. Thus, the system in this embodiment establishes remote wireless communication between the vehicle, the drone, and the cloud via the data communication gateway 12, which forms the hardware foundation for subsequent cloud-based collaboration.

[0024] The vehicle and environment status perception module 13 is used for data acquisition of the vehicle and the environment in which the vehicle is located. It includes devices such as GPS navigation and positioning system, vehicle IMU sensor, ambient light sensor, rainfall intensity sensor, and wind speed sensor, so as to realize the acquisition of data such as vehicle GPS position, speed, heading angle, ambient light data, rainfall intensity, and wind speed. It provides multi-dimensional sensor data to the main control unit to facilitate subsequent drone shooting control.

[0025] The interactive interface 14 is the human-machine interface provided by the vehicle for the user. Based on this, on the one hand, the user can input user commands through the interactive interface 14, and on the other hand, the user can play videos captured by the drone through the interactive interface. Typically, the interactive interface 14 is an in-vehicle smart display screen. In a preferred embodiment of this application, the interactive interface 14 can also be a mobile device (such as a smart tablet, mobile phone, etc.). Correspondingly, in this case, it is also necessary to use devices such as Bluetooth and WIFI to establish a communication connection between the mobile device and the main control unit 11.

[0026] The cabin power supply interface 15 is an interface provided by the vehicle end 10 for charging the drone. It is electrically connected to the vehicle's power battery, thereby using the vehicle's power battery to charge the drone. This configuration ensures the drone's endurance and facilitates the shooting of more videos.

[0027] The airborne controller 21 is the core of data reception and analysis for the UAV terminal 20. It is mainly used to receive and parse the scripts or instructions issued by the main control unit 11, and generate corresponding control instructions to be sent to the shooting and flight module 24; and to control the UAV status perception module 23 to collect UAV status data and send it to the main control unit 11.

[0028] The communication module 22 is a communication device provided on the UAV terminal 20, used for communication between at least two UAVs and communication between the UAV and the vehicle terminal 10. In the system of this embodiment, the communication module 22 adopts an intelligent dual-mode communication mechanism, using direct connection mode (bandwidth 100Mbps) within a 500-meter range with the vehicle terminal 10, and automatically switching to relay mode when the signal strength between the UAV and the vehicle terminal 10 is <-85dBm for 3 seconds.

[0029] The UAV status perception module 23 is used for detecting and collecting the UAV's own status, including RTK-GNSS (real-time dynamic carrier phase differential, used for high-precision positioning), UAV battery management system, etc., thereby collecting data such as UAV position, battery level, and temperature, and providing data support for UAV control.

[0030] The shooting and flight module 24 is used to execute control commands and serves as the mission execution terminal for the drone. It includes a flight controller, gimbal, and camera. The flight controller executes flight control commands to enable the drone to navigate, take off, and land. The gimbal controls the camera's attitude to prevent camera shake caused by drone turbulence. The camera is used to capture video footage.

[0031] The cloud server 31 is a terminal device of the cloud 30, and typically has data processing and data storage capabilities far exceeding those of conventional controllers or processors. In this embodiment, the cloud server 31 is also equipped with an AI engine. Compared to the AI ​​engine deployed in the main control unit 11, the AI ​​engine of the cloud server 31 has stronger computing power and supports enhanced processing operations such as super-resolution rendering and advanced color correction of video materials to obtain high-quality videos.

[0032] Based on the aforementioned system, this embodiment achieves a fully automated video production process, from dynamic script generation and collaborative shooting to real-time editing, through the coordinated efforts of vehicle-side decision-making, drone-side execution, and cloud-based enhancement. It primarily addresses the issues of poor shooting results, cumbersome production processes, and limited creative expression faced by ordinary users in scenarios such as travel, outdoor adventures, and daily recording, due to a lack of professional aerial photography operation skills, video post-production capabilities, and creative content planning abilities.

[0033] In short, this embodiment constructs a "vehicle-machine-cloud" collaborative decision-making system based on the above system, including: (1) Intelligent planning and camera movement planning mechanism: The main control unit 11 integrates vehicle GPS trajectory, vehicle speed, visual information of the surrounding environment and video theme selected by the user in real time, and accesses the cloud digital script library simultaneously. Based on this, a set of personalized shooting scripts containing multiple shot sizes, camera movement trajectories and emotional tone are dynamically generated; (2) Automated collaborative shooting execution mechanism: Once the script is planned, the system automatically triggers the collaborative shooting process; Vehicle-side control: The main control unit 11 acts as the control center, sending the shooting script to the airborne controller on the UAV via a low-latency communication link. The airborne controller then converts the shooting script into precise UAV flight control commands (such as flight path, speed, and gimbal angle). Vehicle coordination: Depending on the shooting needs, the main control unit 11 can send instructions to the vehicle's power system to request that it maintain a constant speed or make a short stop on a specific road section in order to cooperate with the drone to capture the best shot.

[0034] (3) AI editing and finalization mechanism with edge-cloud collaboration: Constructing a hierarchical processing architecture to balance efficiency and quality: Vehicle-side real-time rough cut and preview (device-side): Utilizing the computing power of the vehicle, the footage stream transmitted back from the drone is received in real time during the shooting process, and preliminary screening and splicing are performed. Users can preview the prototype of the video in real time on the interactive interface 14. Cloud-based AI editing and rendering (cloud side): After shooting, the system automatically uploads the rough cut and the high bitrate original to the cloud. The AI ​​algorithm performs fine editing, beat-matching music, intelligent color grading and subtitle generation based on the digital script, and finally renders and outputs the finished product and sends it back to the vehicle. (4) Multi-level anomaly response mechanism: When environmental anomalies or insufficient drone power occur, the drone shooting mission will be forcibly terminated to ensure equipment safety and data integrity.

[0035] Specifically, this embodiment provides a method for automatic video shooting and generation by a vehicle-mounted drone. Using the aforementioned automatic video shooting and generation system for a vehicle-mounted drone, the method includes the following steps: Step S1: After setting the video shooting theme on the vehicle terminal 10, a shooting script is generated based on the collected vehicle and environmental status data and the cloud data obtained by accessing the cloud 30, and then sent to the drone terminal 20. Step S2: The drone terminal 20 analyzes and executes the shooting script, and sends the captured video footage back to the vehicle terminal 10; Step S3: After receiving the video material, the vehicle terminal 10 uses its built-in AI engine to perform rough editing on the video material, and then intelligently selects the rendering path according to network conditions and user settings to generate a first-edition video. Step S4: After the vehicle terminal 10 provides the initial version of the video to the user for preview and confirmation, the final version of the video is generated and saved. Step S5: After completing the shooting task, the drone returns to base.

[0036] Referring to step S1, the first step is to perform system initialization and creative planning, which specifically includes: Step S101: After the user selects a preset video theme (such as "Road Narrative", "Forest Exploration" etc.) through the interactive interface 14, the user sends it to the main control unit 11. At the same time, the vehicle and environmental status perception module 13 collects vehicle and environmental status data and sends it to the main control unit 11. The vehicle and environmental status data includes vehicle GPS position, speed, heading angle and ambient light data. Step S102: The main control unit 11 accesses the digital script library in the cloud 30 through the data communication gateway 13, and finally generates a personalized shooting script containing shot sequence, camera movement type and emotional parameters based on the selected video theme, vehicle and environmental status data.

[0037] For example, the user selects the theme: highway narrative. Real-time status data: GPS shows the vehicle is traveling on an open coastal highway at a speed of 80 km / h, with a stable heading angle, and the light sensor indicates it is dusk. Correspondingly, the main control unit 11, based on its AI engine, matches templates and successful case data related to the theme and real-time status data from the digital script library in the cloud 30. For example, it suggests dynamic tracking shots based on high-speed driving; recommends emotional parameters for warm tones and backlit silhouettes based on dusk light; and suggests grand compositions including the sea and skyline based on GPS information from the coastal highway. Finally, a personalized shooting script is generated, for example: Opening (Grand Presentation): Using equipment from the passenger seat or back seat, shoot a time-lapse video shot that shows the changing colors of the skyline and the trajectory of the vehicle moving at a constant speed. Set the emotional parameters to be open and free.

[0038] Process (Dynamic Capture): Based on a stable vehicle speed, the main control unit suggests a series of low-angle follow shots and side panning shots to capture the dynamic movement of the wheels turning and the roadside scenery rapidly receding.

[0039] Climax (Emotional Elevation): Taking advantage of the twilight lighting, the script will recommend backlit silhouette shots to capture the outlines of the driver or passengers, emphasizing warmth and contemplation in the emotional parameters. Simultaneously, using heading angle data, a smooth panning shot from inside the vehicle will be automatically triggered when the vehicle makes a slight turn, panning from the driver's perspective to the sunset seascape outside the window.

[0040] Ending (Conclusion): Generates a wide-angle shot script that fades into the distance, suggesting the journey continues, with emotional parameters of tranquility and longing.

[0041] Referring to step S2, after the shooting script is generated, it is sent to the airborne controller 21 on the drone. Specifically, step S2 includes: Step S201: After receiving the shooting script through the communication module 22, the airborne controller 21 analyzes it and parses the shooting script into specific executable instructions for the flight controller, gimbal, and camera. Step S202: The airborne controller 21 sends executable commands from the flight controller, gimbal, and camera to the shooting and flight module 24 to begin the shooting task; Step S203: During the shooting mission, the video footage captured by the shooting and flight module 24 and the UAV status data collected by the UAV status perception module 23 are transmitted to the communication module 22 through the airborne controller 21. The communication module 22 then uploads it to the main control unit 11 through the data communication gateway 12. Step S204: After each shot is completed, the main control unit 11 immediately performs a multi-index lens quality check. If the technical or artistic index of the lens does not reach the corresponding preset threshold, the intelligent reshoot mechanism is triggered, and the shooting parameters are adjusted before the shooting task of the corresponding lens is re-executed.

[0042] In step S201, the airborne controller 21 parses the shooting script into specific executable commands for the flight controller, gimbal, and camera through three core steps: command decoding, parameter mapping, and path pre-simulation.

[0043] Instruction decoding is used to interpret the artistic descriptions in the script and convert them into quantifiable technical parameters, including: Analyzing the shot sequence: The system identified that the shooting script contained an ordered list of shots, such as: "Opening time-lapse photography → Low-angle tracking shot in the process → Climax backlit silhouette → Ending wide-angle shot"; Extract camera movement type: Capture keywords from the natural language descriptions in the script. For example, "low-angle follow shot" will be parsed as a combined camera movement that includes the drone's flight altitude (low angle), flight direction (in the same direction as the vehicle), and speed (matching the vehicle's speed); Interpreting Emotional Parameters: The emotional parameters of "openness and freedom" will be decoded into the visual feeling that the lens should present, which will then affect the subsequent specific parameter mapping. For example, the drone's flight path can be more stable and smooth, and the gimbal's turning can be more gentle.

[0044] Parameter mapping is used to precisely map the quantifiable technical parameters obtained after decoding the command to the executable parameters of the drone's flight and shooting units. For example, the quantifiable technical parameters corresponding to "low-angle follow shooting" are: the drone's flight altitude, flight direction, and speed. The corresponding mapped executable parameters are: flight altitude: 5-10 meters above the ground; flight speed: locked to the vehicle's speed at 80 km / h; flight direction (heading angle): consistent with the vehicle's heading angle.

[0045] Path pre-simulation is used by the airborne controller 21 to perform a virtual pre-simulation before generating the final command, ensuring the feasibility of the plan. For example, based on the mapped flight path, the airborne controller 21 will perform dynamic simulation to check for potential collision risks with terrain and obstacles (such as streetlights and trees). At the same time, it will also verify the coordination between flight control commands and gimbal commands to avoid the gimbal turning to extreme angles or the image shaking violently due to excessively fast drone movements.

[0046] Referring to step S203, during the execution of the shooting task, the main control unit 11 in this embodiment also dynamically adjusts the executable commands of the flight controller, gimbal, and camera based on the video material, UAV status data, and real-time environmental data collected by the vehicle and environmental status perception module 13, and sends them sequentially to the shooting and flight module 24 through the data communication gateway 12, communication module 22, and airborne controller 21; the dynamic adjustment includes: If the ambient illuminance is detected to be lower than the preset ambient illuminance threshold (set to 1000 lux in this embodiment, but can be flexibly set according to needs in specific implementation), the drone camera is instructed to switch to night scene mode and automatically adjust the aperture and ISO parameters. If there are obstacles in the shooting path, the flight controller is instructed to replan a safe path based on path planning algorithms such as the A* algorithm. If the subject is lost or the composition is poor, the gimbal and flight controller parameters are dynamically adjusted, and target tracking and composition optimization are performed again.

[0047] Meanwhile, to ensure the safety and data integrity of the drone, the main control unit 11 also establishes a multi-level anomaly response mechanism during the execution of the shooting mission, including: When the wind speed exceeds the preset wind speed level (set to level 8 in this embodiment), the main control unit 11 forces the drone to stop the shooting mission and return immediately. When the location signal strength of the drone is less than the preset strength threshold (set to -80dBm in this embodiment), the main control unit 11 forces the drone to stop the shooting mission and return to home immediately. When the drone's battery level is less than the preset first battery threshold (set to 25% in this embodiment), the main control unit 11 controls the drone to only execute the key shots in the shooting script before returning to home and charging in the vehicle. When the rainfall intensity is greater than the preset rainfall intensity threshold (set to 20 mm / h in this embodiment), the main control unit 11 restricts the drone's flight altitude to below the first preset altitude (set to 50 m in this embodiment). When the vehicle enters a tunnel or dense forest area, the main control unit 11 restricts the drone's flight altitude to above the second preset altitude (set to 80m in this embodiment) and switches to aerial shot acquisition mode. All the preset values ​​mentioned above are reference values ​​provided in this embodiment and can be flexibly adjusted according to actual conditions during specific implementation.

[0048] Referring to step S3, after receiving the video footage, the vehicle-mounted unit 10 implements an edge-cloud collaborative AI editing and finalization mechanism. Based on this mechanism, the main control unit 11 first uses its built-in AI engine to perform rough editing on the video footage, including: The AI ​​engine of the main control unit 11 receives and classifies the shooting materials, and establishes a material index based on the timestamp and lens number; The AI ​​engine of the main control unit 11 performs time-series assembly of the shooting materials based on the shooting script; The AI ​​engine of the main control unit 11 automatically matches transition effects, background music, color filters, and sets AI subtitles based on the shooting script.

[0049] For example, the first step is for the AI ​​engine to read the metadata embedded in each video clip, including timestamps accurate to milliseconds, GPS location (e.g., "coastal highway at 36° North latitude"), and device parameters (e.g., focal length, shutter speed). This data forms the basis for subsequent classification and retrieval. Next, the AI ​​uses its built-in computer vision model to perform deep analysis of the video frames. For instance, it might identify that shot A primarily depicts "an open sea and sky," shot B includes "fast-moving roads and vehicles," shot C shows "the outline of a person's profile," and shot D is a "wide-angle view of vehicles receding into the distance." Simultaneously, it analyzes the emotional characteristics of the scene, such as "openness," "freedom," and "warmth." Then, combining the metadata and AI analysis results, the AI ​​engine generates a digital ID for each clip and stores it in the database. This digital ID includes: shot number, timestamp, GPS location, content tags (e.g., coastal highway, dusk, time-lapse), and emotional parameters. In this way, based on the shooting script, when an "opening time-lapse shot" is needed, the AI ​​can instantly and accurately locate all matching segments from hours of footage.

[0050] The second step: With a clearly categorized media library, the AI ​​engine begins editing strictly according to the narrative logic of the script. For example, the AI ​​engine reads and deconstructs the shooting script for a "road trip narrative," understanding it as an ordered sequence of shots: [Opening Time-lapse] - [Follow-up Shot] - [Climax Silhouette] - [Closing Wide-Angle]. Based on the index built by the AI ​​engine, it finds the most matching physical media for each shot description in the script. It prioritizes clips with stable footage, accurate exposure, and content tags highly consistent with the script description. Finally, the AI ​​precisely arranges the selected footage clips on a timeline according to the script's order. For example, it ensures that the opening shot of the time-lapse is long enough to showcase the color changes of the sky and precisely places the backlit silhouette shot at the anticipated climax of the music.

[0051] Step 3: The AI ​​engine performs intelligent transitions and rhythm optimization. Based on the emotional parameters of the script, it automatically adds transitions, music, color tones, and subtitles to the rough cut video. This completes the rough cut.

[0052] Based on an edge-cloud collaborative AI editing and finalization mechanism, after the rough cut is completed, the system intelligently selects the rendering path according to network conditions and user settings to generate the initial final version, including: When the network bandwidth is less than the preset bandwidth threshold (set to 10Mbps in this embodiment) or when the user selects quick rendering through the interactive interface 14, the main control unit 11 calls the vehicle-side local GPU to complete the rendering. When the network bandwidth is greater than or equal to the preset bandwidth threshold (set to 50Mbps in this embodiment) and the user selects the ultimate image quality through the interactive interface 14, the main control unit 11 uploads the corresponding project file to the cloud server 31 through the data communication gateway 12, and completes the final rendering by using the AI ​​engine deployed on the cloud server 31 for enhancement processing.

[0053] Referring to step S4, the vehicle terminal 10 provides the initial version of the video to the user for preview and confirmation through the interactive interface 14. If the user is satisfied, it is saved and exported directly; if the user wants to make modifications, it enters the manual fine-tuning mode, which supports personalized adjustments to editing points, background music, filter intensity, etc., and finally generates and saves the final version of the video. Referring to step S5, after completing the shooting mission, the drone returns to base to recharge, preparing for the next shooting mission. Specifically, step S5 includes: Step S501: After the drone completes the shooting mission, the onboard controller 21 executes the intelligent return-to-home procedure, and achieves precise docking and charging with the cabin energy supply interface 15 through UWB precise positioning and visual recognition guidance. Step S502: After the main control unit 11 detects the successful docking of the drone with the cabin power supply interface 15, it intelligently selects the charging mode based on the drone's current battery level, battery health status, and subsequent mission requirements, including: When the battery level is less than the preset second battery threshold (set to 30% in this embodiment) and there is a shooting task after a preset time (set to 30 minutes in this embodiment), the fast charging mode is selected to charge the drone with a charging parameter of 9V / 3A, thereby quickly increasing the drone's battery level and ensuring the timely and complete execution of the next shooting task; otherwise, the slow charging mode is selected to charge the drone with a charging parameter of 5V / 2A, thereby optimizing the drone's battery life. Step S503: During the charging process, the main control unit 11 monitors the vehicle battery power, drone charging status and interface temperature in real time. Based on the pre-established multiple safety protection mechanisms such as overvoltage, overcurrent, short circuit and temperature rise, once an abnormality is detected (for example, the interface temperature is detected to exceed the set temperature threshold), the main control unit 11 will promptly remind the user through the interactive interface 14.

[0054] Compared with the prior art, the beneficial effects of the embodiments of the present invention include: 1. Automation and ease of use: Through automatic script generation and flight control via vehicle-machine collaboration, and automatic editing via end-to-cloud collaboration, the entire process from shooting to final film is automated, greatly reducing the user's operating threshold.

[0055] 2. Immediacy and efficiency: By using the onboard computing unit as an edge node for localized rough cutting and previewing, the traditionally time-consuming post-production process is greatly compressed, enabling minute-level finished products in mobile scenarios, and fundamentally improving the user experience.

[0056] 3. Adaptability and Safety: Through the dynamic environment perception and safety decision-making module, the system can adaptively adjust the shooting strategy according to real-time conditions such as road conditions and weather, so as to capture the best shots while ensuring equipment safety.

[0057] 4. Resource optimization: By adopting a hierarchical processing architecture of coarse editing on the vehicle side and fine editing in the cloud, the amount of data that needs to be uploaded to the cloud is effectively reduced, the dependence on network bandwidth and traffic costs are reduced, and the latency problem in the pure cloud processing mode is solved.

[0058] In summary, the vehicle-mounted drone video automatic shooting and generation system and method provided by this invention constructs a collaborative architecture based on "vehicle-side decision-making, aerial execution, and cloud enhancement" across the vehicle (10), drone (20), and cloud (30), forming a complete closed loop for automatic video generation. Through intelligent creative planning, dynamic shooting control, and real-time rendering technology, it achieves fully automatic generation of high-quality videos in outdoor mobile scenarios. Furthermore, this embodiment, through the above implementation methods, realizes intelligent management of the entire process from creative inspiration to final output, encapsulating professional aerial photography techniques and complex post-production processes within an automated system, establishing a complete vehicle-mounted drone video creation ecosystem, enabling users to automatically generate high-quality video content in mobile travel scenarios.

[0059] The present invention has been described above by way of example with reference to the accompanying drawings. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any non-substantial improvements made using the inventive concept and technical solution; or the direct application of the inventive concept and technical solution to other situations without modification, are all within the protection scope of the present invention.

Claims

1. A vehicle-mounted unmanned aerial vehicle (UAV) video automatic shooting and generation system, characterized in that: The system includes: a vehicle terminal (10), a drone terminal (20), and a cloud terminal (30), wherein the vehicle terminal (10) is connected to the drone terminal (20) and the cloud terminal (30) respectively; The vehicle terminal (10) includes a main control unit (11), a data communication gateway (12), a vehicle and environmental status perception module (13), an interactive interface (14), and a cabin energy supply interface (15); the drone terminal (20) includes an airborne controller (21), a communication module (22), a drone status perception module (23), and a shooting and flight module (24); the cloud terminal (30) includes a cloud server (31); both the main control unit (11) and the cloud server (31) have built-in AI engines; The main control unit (11) is connected to the data communication gateway (12), the vehicle and environmental status perception module (13), the interactive interface (14), and the cabin energy supply interface (15), respectively; the airborne controller (21) is connected to the communication module (22), the UAV status perception module (23), and the shooting and flight module (24), respectively; the data communication gateway (12) is connected to the communication module (22) and the cloud server (31), respectively.

2. The vehicle-mounted unmanned aerial vehicle (UAV) video automatic shooting and generation system according to claim 1, characterized in that: The data communication gateway (12) communicates wirelessly with the communication module (22) and the cloud server (31) respectively through wireless communication technology, including 4G, 5G and C-V2X.

3. A method for automatic video shooting and generation by a vehicle-mounted drone, using the automatic video shooting and generation system for a vehicle-mounted drone according to any one of claims 1-2, characterized in that: The method includes the following steps: Step S1: After setting the video shooting theme on the vehicle end (10), the shooting script is generated based on the collected vehicle and environmental status data and the cloud data obtained by accessing the cloud (30) and sent to the drone end (20). Step S2: The drone terminal (20) analyzes and executes the shooting script and sends the captured video footage back to the vehicle terminal (10). Step S3: After receiving the video material, the vehicle (10) uses its built-in AI engine to perform rough editing on the video material, and then intelligently selects the rendering path according to network conditions and user settings to generate the initial version of the finished film. Step S4: The vehicle (10) provides the initial version of the video to the user for preview and confirmation, and finally generates and saves the final version of the video. Step S5: The drone (20) returns to base after completing the shooting task.

4. The method for automatic video shooting and generation by a vehicle-mounted drone according to claim 3, characterized in that: Step S1 includes: Step S101: The user selects a preset video theme through the interactive interface (14) and sends it to the main control unit (11). At the same time, the vehicle and environmental status perception module (13) collects vehicle and environmental status data and sends it to the main control unit (11). Step S102: The main control unit (11) accesses the digital script library in the cloud (30) through the data communication gateway (13) and generates a shooting script based on the selected video theme, vehicle and environmental status data.

5. The method for automatic video shooting and generation by a vehicle-mounted drone according to claim 3, characterized in that: Step S2 includes: Step S201: The airborne controller (21) receives the shooting script through the communication module (22), analyzes it, and parses the shooting script into specific executable instructions for the flight controller, gimbal and camera; Step S202: The airborne controller (21) sends the executable commands of the flight controller, gimbal and camera to the shooting and flight module (24) to start the shooting task; Step S203: During the shooting mission, the video footage captured by the shooting and flight module (24) and the UAV status data collected by the UAV status perception module (23) are transmitted to the communication module (22) through the airborne controller (21), and the communication module (22) then uploads it to the main control unit (11) through the data communication gateway (12). Step S204: After each shot is completed, the main control unit (11) immediately performs a multi-index lens quality check. If the technical or artistic index of the lens does not reach the corresponding preset threshold, the intelligent reshoot mechanism is triggered, and the shooting parameters are adjusted before the shooting task of the corresponding lens is re-executed.

6. The method for automatic video shooting and generation by a vehicle-mounted drone according to claim 5, characterized in that: During the shooting mission, the main control unit (11) dynamically adjusts the executable commands of the flight controller, gimbal, and camera based on the video material, UAV status data, and real-time environmental data collected by the vehicle and environmental status perception module (13), and sends them sequentially to the shooting and flight module (24) through the data communication gateway (12), communication module (22), and airborne controller (21); the dynamic adjustment includes: If the ambient light level is detected to be lower than the preset ambient light threshold, the drone camera is instructed to switch to night mode and automatically adjust the aperture and ISO parameters. If there are obstacles in the shooting path, the flight controller is instructed to replan a safe path. If the subject is lost or the composition is poor, the gimbal and flight controller parameters are dynamically adjusted, and target tracking and composition optimization are performed again.

7. The method for automatic video shooting and generation by a vehicle-mounted drone according to claim 5, characterized in that: During the execution of the shooting task, the main control unit (11) establishes a multi-level abnormal response mechanism, including: When the wind speed exceeds the preset wind speed level, the main control unit (11) forces the drone to stop shooting and return immediately; When the position signal strength of the drone is less than the preset strength threshold, the main control unit (11) forces the drone to stop the shooting mission and return immediately; When the drone's battery level is less than the preset first battery threshold, the main control unit (11) controls the drone to only execute the key shots in the shooting script and then return to the vehicle for charging. When the rainfall intensity is greater than the preset rainfall intensity threshold, the main control unit (11) limits the flight altitude of the UAV to below the first preset altitude; When the vehicle enters a tunnel or dense forest area, the main control unit (11) limits the drone's flight altitude to above the second preset altitude and switches to the aerial camera acquisition mode.

8. The method for automatic video shooting and generation by a vehicle-mounted drone according to claim 3, characterized in that: In step S3, after receiving the video material, the vehicle (10) performs a rough cut on the video material using its built-in AI engine, including: The AI ​​engine of the main control unit (11) receives and classifies the shooting materials, and establishes a material index based on the timestamp and lens number; The AI ​​engine of the main control unit (11) performs time-series assembly of the shooting materials based on the shooting script; The AI ​​engine of the main control unit (11) automatically matches transition effects, background music, color filters, and sets AI subtitles based on the shooting script.

9. The method for automatic video shooting and generation by a vehicle-mounted unmanned aerial vehicle according to claim 3, characterized in that: In step S3, intelligently selecting the rendering path based on network conditions and user settings to generate the initial version of the final image includes: When the network bandwidth is less than the preset bandwidth threshold or the user selects fast rendering through the interactive interface (14), the main control unit (11) calls the vehicle-side local GPU to complete the rendering. When the network bandwidth is greater than or equal to the preset bandwidth threshold and the user selects the ultimate picture quality through the interactive interface (14), the main control unit (11) uploads the corresponding project file to the cloud server (31) through the data communication gateway (12), and completes the rendering of the finished film by using the AI ​​engine deployed on the cloud server (31) for enhancement processing.

10. A method for automatic video shooting and generation by a vehicle-mounted unmanned aerial vehicle according to claim 3, characterized in that: Step S5 includes: Step S501: After the UAV completes the shooting task, the onboard controller (21) executes the intelligent return-to-home procedure, and achieves precise docking and charging with the cabin energy supply interface (15) through UWB precise positioning and visual recognition guidance; Step S502: After the main control unit (11) detects the successful docking of the UAV with the cabin energy supply interface (15), it intelligently selects the charging mode based on the UAV's current power level, battery health status, and subsequent mission requirements, including: When the battery level is below the preset second battery threshold and there is a shooting task after a preset time, select the fast charging mode and charge with a charging parameter of 9V / 3A; otherwise, select the slow charging mode and charge with a charging parameter of 5V / 2A. Step S503: During the charging process, the main control unit (11) monitors the vehicle battery power, the drone charging status and the interface temperature in real time.

Images