Video shooting stability control device and control system thereof

Abstract

The invention relates to the technical field of video shooting, in particular to a video shooting stability control device and a control system thereof. When the device is used, a person uses a camera to shoot animals or plants, and in the shooting process, a gyroscope sensor, an acceleration sensor, an ultrasonic wind speed and direction sensor and a visual sensor in the sensor module all work; wherein the gyroscope sensor is mainly used for measuring the rotation angle and the angular velocity of the camera; in the shooting process, the camera may rotate due to factors such as hand-held shaking and unstable mounting platform, and the gyroscope sensor can sense the rotation changes in real time and transmit data to the control algorithm module; and the control algorithm module calculates parameters needing to be adjusted based on the data, so that the execution mechanism module (such as a holder) is driven to correct the attitude of the camera, and the stability of a shot picture is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of video shooting, and specifically relates to a video shooting stability control device and its control system. Background Art

[0002] In today's video shooting field, the stability of the picture is one of the important indicators to measure the video quality. However, in the actual shooting process, there are many factors affecting the shooting stability.

[0003] Traditional shooting methods mainly rely on mechanical stabilizing devices, such as tripods, gimbals, etc. Although a tripod can provide basic stable support, it lacks flexibility and cannot meet the needs when rapid movement shooting or shooting dynamic scenes is required. Although a gimbal can control the camera angle to a certain extent, for complex jitter situations, its stabilizing effect is limited.

[0004] The emergence of electric stabilizers (handheld gimbals) has solved the stability problem of some dynamic shootings. However, current electric stabilizers on the market still have deficiencies in terms of response speed, anti-shake range, etc., and the cost is relatively high. Although electronic anti-shake technology is integrated into cameras or mobile phones and is convenient to use, it will cause picture quality loss and the anti-shake range is limited. Although hybrid anti-shake technology combines the advantages of optical anti-shake and electronic anti-shake, it is still difficult to completely eliminate the influence of jitter under certain extreme shooting conditions.

[0005] In addition, most of the existing shooting stability control systems lack intelligent learning and adaptive capabilities and cannot adjust stability parameters in real time according to different shooting environments, the motion characteristics of shooting objects, and environmental meteorological conditions. Especially when shooting dynamic animals, outdoor plants or conducting sports broadcasts, due to the uncertainty, diversity of the movement of shooting objects and the influence of environmental wind, it is very difficult to ensure the stability and clarity of the shooting picture. Therefore, a more advanced and efficient video shooting stability control system is needed to meet the requirements for video stability in different shooting scenarios.

[0006] In summary, to solve the technical problems proposed in this article, the present invention proposes a video shooting stability control device and its control system. Summary of the Invention

[0007] The present invention proposes a video shooting stability control system, including a camera; this stability control system includes:

[0008] A sensor module, which includes a gyroscope sensor, an acceleration sensor, an ultrasonic wind speed and direction sensor, and a vision sensor;

[0009] A large model database, which contains the running or walking posture information of various animals and prefabricated motion trajectories, as well as plant type data;

[0010] The control algorithm module, connected to the sensor module, automatically adjusts control parameters based on the data collected by the sensors, using an adaptive control algorithm and machine learning techniques; connected to the large model database;

[0011] The actuator module, connected to the control algorithm module, adjusts the posture of the camera according to the instructions issued by the control algorithm module;

[0012] The power management module provides a stable power supply for the entire system;

[0013] The human-computer interaction module includes a display screen and operation buttons;

[0014] The data link architecture includes a wired transmission unit and a wireless transmission unit.

[0015] As a preferred solution of the present application; the control algorithm module includes a multi-sensor fusion algorithm, and the multi-sensor fusion algorithm is;

[0016]

[0017] represents the predicted value of the state at the current time k; F represents the state transition matrix; represents the estimated value at the previous moment k-1 ; B represents the control output matrix; u k-1 represents the control input at the previous moment; W k-1 represents the process noise;

[0018] where the state transition matrix of F is the control output matrix of B is Δt is the system sampling interval time.

[0019] As a preferred solution of the present application; the control algorithm module includes an animal motion prediction algorithm, and the animal motion prediction algorithm is;

[0020] State definition: x = [θ, ω]

[0021] The angle changes with the angular velocity: θ k = θ k-1 + ω k-1 Δt

[0022] Control input: u = a

[0023] where θ is the angle; ω is the angular velocity; a is the angular acceleration.

[0024] As a preferred solution of the present application; the control algorithm module further includes a plant swing compensation algorithm, and the plant swing compensation algorithm is;

[0025] State definition:

[0026] State transition matrix:

[0027] Control input: u = F d

[0028] State update at the next moment: x k = F k-1 + Bu k-1 ;

[0029] where is the blade angle; is the angular velocity; ω n is the natural frequency; is the damping ratio; F d is the air resistance; B is the control input matrix.

[0030] As a preferred solution of the present application; the control algorithm module further includes an animal movement wind force algorithm and a grass leaf swing amplitude algorithm; the animal movement wind force algorithm is; F d = 0.5ρAv 2 C d

[0031] where F d is the air resistance; ρ is the air density; A is the windward area during animal movement; v is the animal movement speed; C d is the air resistance coefficient;

[0032] The grass leaf swing amplitude algorithm is;

[0033] where θ(t) is the angle of grass leaf swing at time t; θ0 is the initial angle of the grass leaf; k is the stiffness of the grass leaf.

[0034] As a preferred solution of the present application; the data link architecture includes;

[0035] Metadata embedding unit, which writes the animal movement wind force data F d and the grass leaf swing angle θ(t) into the MXF file metadata through a recorder;

[0036] Post-processing unit, which supports the DaVinci Resolve script to align the CSV file with the video timeline, and the synchronization accuracy error < 1ms.

[0037] As a preferred solution of the present application; the stability control system further includes;

[0038] Learning module, which is connected to the mobile terminal through the Internet of Things and shares and exchanges data with the AI software inside the mobile terminal.

[0039] As a preferred solution of the present application; the learning module includes:

[0040] A reinforcement learning optimization unit that iteratively updates the control algorithm parameters through user feedback data to optimize the animal motion prediction model and the plant swing compensation model;

[0041] A knowledge transfer unit that transfers the animal motion patterns in the large model database to the identification of new species to expand the scope of application of the system.

[0042] A video shooting stability control device, which is used in the above-mentioned video shooting stability control system. The control device includes a camera mounting pan-tilt head, and the pan-tilt head is an electric pan-tilt head for mounting and supporting the camera.

[0043] The beneficial effects of the present invention are as follows:

[0044] When in use, a person uses the camera to shoot animals or plants. During the shooting process, the gyroscope sensor, acceleration sensor, ultrasonic wind speed and direction sensor, and vision sensor in the sensor module all work; among them, the gyroscope sensor is mainly used to measure the rotation angle and angular velocity of the camera itself; during the shooting process, the camera may generate rotational motion due to factors such as hand jitter and unstable mounting platform. The gyroscope sensor can sense these rotational changes in real time and transmit the data to the control algorithm module; based on these data, the control algorithm module calculates the parameters that need to be adjusted, so as to drive the actuator module (such as the pan-tilt head) to correct the camera posture and ensure the stability of the shooting image. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is the overall architecture diagram of the present invention;

[0046] Figure 2 is the view of the data link architecture of the present invention; DETAILED DESCRIPTION OF THE INVENTION

[0047] In order to make the technical means, creative features, achieved purposes and functions of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0048] Example 1:

[0049] A video shooting stability control system includes a camera; the stability control system includes:

[0050] A sensor module, and the sensor module includes a gyroscope sensor, an acceleration sensor, an ultrasonic wind speed and direction sensor, and a vision sensor;

[0051] The large model database contains information on the running or walking postures of various animals, prefabricated motion trajectories, and plant type data;

[0052] The control algorithm module is connected to the sensor module. Based on the data collected by the sensor, it uses an adaptive control algorithm and machine learning technology to automatically adjust the control parameters; it is connected to the large model database;

[0053] The actuator module is connected to the control algorithm module and adjusts the posture of the camera according to the instructions issued by the control algorithm module;

[0054] The power management module provides a stable power supply for the entire system;

[0055] The human-computer interaction module includes a display screen and operation buttons;

[0056] The data link architecture includes a wired transmission unit and a wireless transmission unit;

[0057] The control algorithm module includes a multi-sensor fusion algorithm, and the multi-sensor fusion algorithm is;

[0058]

[0059] represents the predicted value of the state at the current time k; F represents the state transition matrix; represents the estimated value at the previous moment k-1 ; B represents the control output matrix; u k-1 represents the control input at the previous moment; W k-1 represents the process noise;

[0060] Among them, the state transition matrix of F is The control output matrix of B is Δt is the system sampling interval time;

[0061] The control algorithm module includes an animal motion prediction algorithm, and the animal motion prediction algorithm is;

[0062] State definition: x = [θ, ω]

[0063] The angle changes with the angular velocity: θ k = θ k-1 + ω k-1 Δt

[0064] Control input: u = a

[0065] Where θ is the angle; ω is the angular velocity; a is the angular acceleration.

[0066] The control algorithm module also includes a plant swing compensation algorithm, and the plant swing compensation algorithm is;

[0067] State definition:

[0068] State transition matrix:

[0069] Control input: u = F d

[0070] Next moment state update: x k = F k-1 + Bu k-1 ;

[0071] where is the blade angle; is the angular velocity; ω n is the natural frequency; is the damping ratio; F d is the air resistance; B is the input matrix.

[0072] The specific principle is as follows;

[0073] When in use, the user uses a camera to take pictures of animals or plants. During the shooting process, the gyroscope sensor, acceleration sensor, ultrasonic wind speed and direction sensor, and vision sensor in the sensor module all work; among them, the gyroscope sensor is mainly used to measure the rotation angle and angular velocity of the camera itself; during the shooting process, the camera may generate rotational motion due to factors such as hand jitter and unstable mounting platform. The gyroscope sensor can sense these rotational changes in real time and transmit the data to the control algorithm module; based on this data, the control algorithm module calculates the parameters that need to be adjusted, so as to drive the actuator module (such as a pan-tilt head) to correct the camera attitude and ensure the stability of the shooting image;

[0074] Among them, the acceleration sensor is used to detect the acceleration of the camera in each direction; sense the acceleration, deceleration, and impact force of the camera, etc.; when shooting a dynamic scene, actions such as the rapid movement or sudden stop of the camera will be detected by the acceleration sensor and converted into electrical signals and transmitted to the control algorithm module; the control algorithm module comprehensively analyzes the acceleration data, judges the motion state of the camera, and then adjusts the parameters of the camera or controls the pan-tilt head to make corresponding actions to offset the image instability caused by the movement of the camera itself;

[0075] Among them, the vision sensor can detect the movement speed and position information of animals. By analyzing the position changes and movement trajectories of animals in the picture, the vision sensor can obtain the movement speed of animals in real time, providing key data for the animal movement prediction algorithm, enabling the system to predict the movement trajectory of animals. For example, by detecting the angular velocity of the animal's head, the pan-tilt can be driven to make compensatory actions in advance. During the monitoring of animals, due to the running postures of various animals recorded in the large model database and the data on plant species, when photographing animals, such as photographing a running antelope, the large model database transmits the recorded posture information and prefabricated movement trajectory information of the antelope during running to the control algorithm module, facilitating the camera to obtain the prediction information of the antelope during running. Subsequently, the animal movement prediction algorithm in the control algorithm module obtains the information obtained, as well as the speed data of the antelope, the head angular velocity ω, angular acceleration a, and Δt sampling interval obtained by the vision sensor during the monitoring of the running antelope;

[0076] For example: the angular velocity ω of the antelope is 30° / s, the angular acceleration is 5° / s², and Δt is 0.01s;

[0077] At this time, by substituting Calculate the attitude adjustment amount of the control output matrix;

[0078] That is:

[0079] Subsequently, substitute θ k = θ k-1 + ω k-1 Δt; Calculate the predicted value of the antelope's head swing; Assume the initial angle ω k-1 is 5°;

[0080] That is: θ k = 5° + 30° / s × 0.01s = 5.3°, that is, the predicted value of the antelope's head swing angle at the next moment is 5.3°;

[0081] At this time, the control algorithm module feeds back the calculation result to the actuator module, and the actuator module drives the pan-tilt to adjust 5.3° in the opposite direction of the antelope's head swing 0.05s in advance, offsetting the picture offset caused by the movement, ensuring that the antelope is always in the center of the picture and achieving stable shooting;

[0082] Secondly, the vision sensor can also be used to identify some feature information in the captured scene. For example, it can assist in identifying the types of grass leaves. By combining the parameters of different grass leaves in the large model database, it can provide more accurate input for the plant swing compensation algorithm, improving the accuracy of predicting the swing of grass leaves. The vision sensor, in cooperation with the ultrasonic wind speed and direction sensor, can measure the wind speed and direction information in the captured environment, as well as the impact of wind speed and direction information on outdoor plants. For example, when photographing plants, the control algorithm module calculates by combining the wind direction and speed information and the types of plants in the large model database, and uses the plant swing compensation algorithm to calculate the possible impact of wind speed and direction on the grass leaves. When photographing reeds, through the vision sensor and by combining the reed parameters in the large model database, such as obtaining the reed parameters as; The damping ratio is 0.3; F d The air resistance is 1N; ω n The natural frequency is 2.24 rad / s; the Δt interval is 0.1; and in the initial state The blade angle is 0°, The angular velocity is 0 rad / s; perform the calculation; according to Substitute the parameters;

[0083] Subsequently, calculate the state of the plant at the next moment, that is, substitute x k = F k-1 + Bu k-1 , u k-1 = Fd = 1;

[0084] That is, the blade angle is

[0085] That is, the angular velocity is

[0086] Therefore, the state of the reed at the next moment is x1 = [0, 1], the blade angle remains 0°, and the angular velocity becomes 1 rad / s. This result can be used to further predict the subsequent swing amplitude of the reed, assist the shooting system in adjusting the pan-tilt, and compensate for the impact of the reed swing on the picture, further improving the stability of photographing the reed and the stability of the picture;

[0087] Embodiment 2:

[0088] The control algorithm module further includes an animal movement wind force algorithm and a grass leaf swing amplitude algorithm; the animal movement wind force algorithm is; F d = 0.5ρAv 2 C d ;

[0089] Where F dis the air resistance; ρ is the air density; A is the windward area during animal movement; v is the animal movement speed; C d is the air resistance coefficient;

[0090] The grass leaf swing amplitude algorithm is;

[0091] where θ(t) is the angle of grass leaf swing at time t; θ0 is the initial angle of the grass leaf; k is the stiffness of the grass leaf;

[0092] The data link architecture includes;

[0093] The metadata embedding unit writes the animal movement wind force data F d and the grass leaf swing angle θ(t) into the MXF file metadata through a recorder;

[0094] The post-processing unit supports the DaVinci Resolve script to align the CSV file with the video timeline, and the synchronization accuracy error < 1ms;

[0095] The specific principle is as follows;

[0096] During the process of photographing the animal above, since the animal's running will drive the surrounding air to disturb, that is, generate air currents, and the air currents will drive the surrounding plants, causing the plants to swing; provide the animal movement wind force algorithm and the grass leaf swing amplitude algorithm to calculate the wind speed generated by the animal, and then calculate the influence of the wind speed on the swing of the grass leaf, and transfer the data to the actuator module to adjust the pan-tilt head; for example, obtain the movement parameters of the antelope through the visual sensor and the large model database; that is, ρ is 1.29 kg / m 3 ; A is 1.2 m 2 ; v is 15 m / s; C d is 0.6; and obtain the parameters of the reed, that is, k is 0.8 N / m; θ0 is 0°; substitute the above parameters into F d = 0.5ρAv 2 C d and ;

[0097] That is: Fd = 0.5 × 1.29 × 1.2 × 152 × 0.6 = 104.49 N

[0098] That is:

[0099] According to the above content, it can be obtained that if the reed swings 89.56° to the right due to the wind force generated by the antelope running, the actuator module controls the camera to turn 89.56° to the left through the pan-tilt head to offset the influence of the reed swing on the captured image and ensure the stability of the image; thereby improving the shooting effect;

[0100] And when shooting animals and plants, the recorder in the data link architecture writes the animal movement wind data Fd and the grass blade swing angle θ(t) into the MXF file metadata (following the SMPTE ST291 standard), giving the data standardized labels; this allows for direct and quick reading of key environmental data in post-production without the need for additional data extraction or calibration, thus avoiding data loss or confusion; and supports Da Vinci Resolve scripts to align CSV files with the video timeline, with a synchronization accuracy error of less than 1ms; this ensures the net gain of data such as wind force and grass blade swing with the video screen in the time dimension, providing a basis for refined production and thus improving the stability of camera videos.

[0101] Embodiment three:

[0102] The stability control system further comprises:

[0103] The learning module is connected to the mobile terminal through the Internet of Things, and shares and interacts with the AI ​​software inside the mobile terminal.

[0104] The learning module includes:

[0105] Reinforcement learning optimization unit, which iteratively updates control algorithm parameters through user feedback data, and optimizes animal motion prediction models and plant swing compensation models;

[0106] Knowledge transfer unit, which transfers animal movement patterns in a large model database to new species identification, expanding the scope of application of the system;

[0107] The specific principles are as follows;

[0108] Reinforcement Learning Optimization Unit

[0109] Operation mechanism: Through the network, the learning module can collect feedback data from users when using the system (such as shooting picture stability evaluation, algorithm prediction error, etc.), and use reinforcement learning technology to iteratively update the control algorithm parameters; then optimize the animal motion prediction model in the system to more accurately predict the animal's running trajectory, angle, speed change, etc., for example, improve the prediction accuracy of animal movement posture in a new environment; optimize the plant swing compensation model: more accurately calculate the swing direction and amplitude of grass leaves caused by wind or animal movement, assist the camera in early adjustment, and enhance the stability of the shooting picture;

[0110] Extract existing animal motion patterns (such as the running posture data of common animals) from the large model database through the knowledge migration unit, and migrate and apply them to the recognition and analysis of new species; thereby avoiding the need to build a motion model for each new species separately and expanding the scope of application of the system; for example, when photographing an animal that has never been recorded, the system can quickly analyze its motion characteristics based on the motion patterns of similar animals in the database, achieve shooting stability control, and improve the versatility and scalability of the system.

[0111] Embodiment 4:

[0112] A video shooting stability control device, which is used in the above-mentioned video shooting stability control system. The control device includes a camera mounting pan-tilt head. The pan-tilt head is an electric pan-tilt head and is used for mounting and supporting the camera.

[0113] The specific principle is as follows;

[0114] The pan-tilt head selects an electric pan-tilt head, which is built-in with a motor and an intelligent control system, and can accurately control the movement of the pan-tilt head through a preset program or an external control device (such as a mobile phone, a remote control); when photographing an antelope running, it can quickly and accurately follow the movement trajectory of the antelope, and can also achieve some complex motion patterns, such as automatic tracking, fixed-point shooting, etc., improving the shooting efficiency and picture quality.

Claims

1. A video shooting stability control system, comprising a camera; characterized in that ； The stability control system includes: Sensor module, the sensor module includes a gyroscope sensor, an acceleration sensor, an ultrasonic wind speed and direction sensor, and a visual sensor; A large model database, including information on the running or walking postures of various animals and pre-made motion trajectories, as well as plant type data; A control algorithm module is connected to the sensor module, and automatically adjusts control parameters based on data collected by the sensor using an adaptive control algorithm and machine learning technology; and is connected to a large model database; An actuator module is connected to the control algorithm module and adjusts the posture of the camera according to the instructions issued by the control algorithm module; Power management module, providing stable power supply for the entire system; Human-computer interaction module, including display screen and operation buttons; The data link architecture includes a wired transmission unit and a wireless transmission unit.

2. A video shooting stability control system as claimed in claim 1, characterized in that: The control algorithm module includes a multi-sensor fusion algorithm, and the multi-sensor fusion algorithm is: represents the state prediction value of the current time k; F represents the state transfer matrix; Indicates the last moment k-1 The estimated value of; B represents the control output matrix; u k-1 Represents the control input at the previous moment; W k-1 represents process noise; The state transfer matrix of F is The control output matrix of B is Δt is the system sampling interval.

3. A video shooting stability control system as claimed in claim 2, characterized in that: The control algorithm module includes an animal motion prediction algorithm, and the animal motion prediction algorithm is; State definition: x = θ, ω Angle changes with angular velocity: θ k =θ k-1 +ω k-1 Δt Control input: u = a Where θ is the angle; ω is the angular velocity; and a is the angular acceleration.

4. A video shooting stability control system as claimed in claim 3, characterized in that: The control algorithm module also includes a plant swing compensation algorithm, and the plant swing supplement algorithm is; Status definition: State transition matrix: Control input: u=F d Next moment status update: x k =F k-1 +Bu k-1 ; in is the blade angle; is the angular velocity; ω n is the natural frequency; is the damping ratio; F d is the air resistance; B is the control input matrix.

5. A video shooting stability control system as claimed in claim 4, characterized in that: The control algorithm module also includes an animal movement wind force algorithm and a grass leaf swing amplitude algorithm; the animal movement wind force algorithm is: d =0.5ρAv 2 C d where F d is air resistance; ρ is air density; A is the windward area of ​​the animal when it moves; v is the speed of the animal; C d is the air resistance coefficient; The algorithm for the grass blade swing amplitude is: Where θ(t) is the swing angle of the grass blade at time t; θ0 is the initial angle of the grass blade; and k is the stiffness of the grass blade.

6. A video shooting stability control system as claimed in claim 1, characterized in that: The data link architecture includes: Metadata embedding unit, through the recorder to record animal movement wind data F d and the grass blade swing angle θ(t) are written into the MXF file metadata; The post-processing unit supports DaVinci Resolve scripts to align CSV files with the video timeline, with a synchronization accuracy error of less than 1ms.

7. A video shooting stability control system as claimed in claim 6, characterized in that: The stability control system further comprises: The learning module is connected to the mobile terminal through the Internet of Things, and shares and interacts with the AI ​​software inside the mobile terminal.

8. A video shooting stability control system as claimed in claim 7, characterized in that: The learning module includes: Reinforcement learning optimization unit, which iteratively updates control algorithm parameters through user feedback data, and optimizes animal motion prediction models and plant swing compensation models; The knowledge transfer unit transfers the animal movement patterns in the large model database to new species identification, expanding the scope of application of the system.

9. A video shooting stability control device, which is used in a video shooting stability control system in claims 1-8, characterized in that: The control device comprises a camera mounting platform, which is an electrically controlled platform for mounting and supporting the camera.

Images