CN117692755A - Moving target tracking shooting method, device, equipment and storage medium - Google Patents
Moving target tracking shooting method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN117692755A CN117692755A CN202311577528.0A CN202311577528A CN117692755A CN 117692755 A CN117692755 A CN 117692755A CN 202311577528 A CN202311577528 A CN 202311577528A CN 117692755 A CN117692755 A CN 117692755A
- Authority
- CN
- China
- Prior art keywords
- target
- tracking
- polar coordinate
- target person
- angle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000000605 extraction Methods 0.000 claims abstract description 14
- 230000009471 action Effects 0.000 claims description 21
- 230000008859 change Effects 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 abstract description 15
- 230000033001 locomotion Effects 0.000 description 40
- 238000010586 diagram Methods 0.000 description 14
- 230000003287 optical effect Effects 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 7
- 238000007796 conventional method Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- QIVBCDIJIAJPQS-VIFPVBQESA-N L-tryptophane Chemical compound C1=CC=C2C(C[C@H](N)C(O)=O)=CNC2=C1 QIVBCDIJIAJPQS-VIFPVBQESA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000010361 irregular oscillation Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Closed-Circuit Television Systems (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a method, a device, equipment and a storage medium for tracking and shooting a moving target, wherein the method comprises the following steps: receiving image data of a target person shot by a camera, wherein a cradle head is arranged at the lower end of the camera; performing feature extraction processing on the image data to obtain target pedestrian features; calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on a polar coordinate angle of the target person relative to a holder; and carrying out real-time tracking shooting on the target person based on the moving track. According to the method and the device, the moving track is calculated through the polar coordinate angle of the target person, the target person is tracked and shot according to the moving track, and the tracking and shooting processes are smoothly connected.
Description
Technical Field
The present disclosure relates to the field of visual tracking technologies, and in particular, to a method, an apparatus, a device, and a storage medium for tracking and shooting a moving target.
Background
In recent years, with the rising of live broadcast, more and more remote video services are rapidly increased, and the scheme of an intelligent video system is widely applied, and the current live broadcast video service mainly shoots people by equipment comprising a host optical zoom lens, a mechanical holder, a microphone and the like, and the holder shoot people by tracking faces, so that people can be always in the center of the holder.
The conventional tracking shooting algorithm is mainly based on a character detection algorithm, and the moving distance of a current character frame is calculated, so that the movement track of a cradle head and the cradle head is controlled, the character can be always positioned in the center of the cradle head, but in the tracking process of a monocular cradle head tracking machine, the cradle head moves, pedestrians in a picture also move, at the moment, the relative coordinates of the pedestrians in the picture have uncertainty, the movement track of each target cannot be counted for tracking, and tracking shooting is not smooth.
Disclosure of Invention
The main purpose of the application is to provide a method, a device, equipment and a storage medium for tracking and shooting moving targets, which aim to solve the technical problems that in the tracking process of a monocular cradle head tracking machine in the related technology, a cradle head is moving, pedestrians in a picture also walk, at the moment, relative coordinates of the pedestrians in the picture have uncertainty, the motion trail of each target cannot be counted for tracking, and the tracking and shooting are not smooth.
In order to achieve the above object, an embodiment of the present application provides a method for tracking and shooting a moving object, where the method includes:
receiving image data of a target person shot by a camera, wherein a cradle head is arranged at the lower end of the camera;
performing feature extraction processing on the image data to obtain target pedestrian features;
calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on a polar coordinate angle of the target person relative to a holder;
and carrying out real-time tracking shooting on the target person based on the moving track.
In a possible implementation manner of the present application, the step of calculating the target pedestrian feature based on the preset track tracking model to obtain the movement track of the target person includes:
extracting action state data of the target person from the target pedestrian characteristic based on a preset track tracking model, wherein the action state data comprises center coordinates of the target person and a center coordinate change value;
converting the action state data into polar coordinate state data;
and generating a moving track based on a plurality of polar coordinate points in the polar coordinate state data.
In a possible embodiment of the present application, the step of converting the action state data into polar coordinate state data includes:
acquiring the image size of the image data, and determining a character offset value of the target character and a first polar coordinate of the target character relative to a holder based on the central coordinate in the action state data;
and determining polar coordinate state data based on the image size, the first polar coordinate, the character offset value and the current holder angle.
In one possible embodiment of the present application, the character offset value includes a horizontal offset value and a vertical offset value, and the pan-tilt angle includes a pan-tilt horizontal angle and a pan-tilt vertical angle;
the step of determining polar coordinate state data based on the image size, the first polar coordinate, the person offset value, and the current pan/tilt angle includes:
determining second polar coordinates of an image edge relative to the pan-tilt based on the image size, the image size including a horizontal size and a vertical size;
calculating a target horizontal angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the horizontal offset value, the horizontal dimension and the current horizontal angle of the cradle head;
calculating a target vertical angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the vertical offset value, the vertical dimension and the current vertical angle of the cradle head;
determining a person image size based on the zoom magnification of the camera and the image size of the target person;
polar coordinate state data is determined based on the target horizontal angle, the target vertical angle, and the person image size.
In one possible embodiment of the present application, the step of generating a movement track based on a plurality of polar coordinate points in the polar coordinate state data includes:
and connecting a plurality of polar coordinate points in the polar coordinate state data based on a preset track strategy to obtain a moving track.
In a possible implementation manner of the present application, the step of performing feature extraction processing on the image data to obtain a target pedestrian feature includes:
based on a preset recognition model, recognizing the image data to obtain first recognition data;
and extracting character characteristic information in the first identification data to obtain the characteristics of the target pedestrians.
In a possible implementation manner of the present application, the step of performing real-time tracking shooting on the target person based on the movement track includes:
determining first angles of a plurality of polar coordinate points in the moving track;
and rotating the cradle head and the camera based on the change value of the first angle so as to track and shoot the target person in real time.
The application also provides a tracking shooting device of a moving target, the tracking shooting device of the moving target comprises:
the receiving module is used for receiving image data of a target person shot by the camera, wherein a cradle head is arranged at the lower end of the camera;
the processing module is used for carrying out feature extraction processing on the image data to obtain target pedestrian features;
the calculation module is used for calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on the polar coordinate angle of the target person relative to a holder;
and the tracking shooting module is used for carrying out real-time tracking shooting on the target person based on the moving track.
The application also provides a tracking shooting device of the moving target, wherein the tracking shooting device of the moving target is entity node equipment, and the tracking shooting device of the moving target comprises: the method comprises a memory, a processor and a program of the tracking shooting method of the moving target, wherein the program of the tracking shooting method of the moving target is stored in the memory and can be run on the processor, and the steps of the tracking shooting method of the moving target can be realized when the program of the tracking shooting method of the moving target is executed by the processor.
In order to achieve the above object, there is also provided a storage medium having stored thereon a tracking shooting program of a moving object, which when executed by a processor, implements the steps of the tracking shooting method of any one of the moving objects described above.
The application provides a tracking shooting method, a device, equipment and a storage medium of a moving target, compared with the prior art, in the tracking process of a monocular cradle head tracking machine, the cradle head is moving, pedestrians in a picture are moving, at the moment, relative coordinates of the pedestrians in the picture are uncertain, the motion trail of each target cannot be counted to track, and therefore, compared with the unsmooth tracking shooting, in the method, the moving trail is obtained through processing image data shot by a camera and calculating the polar coordinate angle of a target person relative to the cradle head, the target person is tracked and shot through the calculated moving trail, and the rotation angle of the cradle head is adjusted according to the movement angle of the person relative to the cradle head, so that the rotation angle of the cradle head is identical to the movement angle of the person, and the tracking shooting process of the camera is smooth and connected.
Drawings
Fig. 1 is a flowchart of a first embodiment of a tracking shooting method of a moving object of the present application;
fig. 2 is a schematic diagram of a conventional close-up tracking flow involved in a tracking shooting method of a moving target in the present application;
FIG. 3 is a schematic diagram of a device architecture of a hardware operating environment according to an embodiment of the present application;
FIG. 4 is a time-varying schematic diagram of the horizontal center coordinates of a target person according to the conventional method of the present application;
FIG. 5 is a schematic diagram showing the time variation of the vertical center coordinates of a target person according to the conventional method of the present application;
FIG. 6 is a schematic diagram of a target person width variation of a conventional method according to the present application;
fig. 7 is a schematic diagram of a polar coordinate system pseudo 3D modeling flow related to a tracking shooting method of a moving target in the present application;
fig. 8 is a schematic flow chart of converting relative coordinates related to the tracking shooting method of the moving object into a polar coordinate system;
fig. 9 is a schematic diagram of a horizontal movement tracking shooting flow related to a tracking shooting method of a moving target in the present application;
FIG. 10 is a schematic diagram of the relationship between the horizontal angle of a horizontal uniform motion figure and the change of time according to the tracking shooting method of a moving target;
FIG. 11 is a schematic diagram of a person's trajectory generated by a tracking shot according to a conventional method of the present application;
fig. 12 is a schematic diagram of a character track generated by the tracking shooting method of the moving object of the present application.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
An embodiment of the present application provides a method for tracking and photographing a moving object, in a first embodiment of the method for tracking and photographing a moving object of the present application, referring to fig. 1, the method includes the following steps:
step S10, receiving image data of a target person shot by a camera, wherein a cradle head is arranged at the lower end of the camera;
it should be noted that the method for tracking and shooting a moving object may be applied to a tracking and shooting device of a moving object, where the tracking and shooting device of a moving object belongs to a tracking and shooting system of a moving object, and the tracking and shooting system of a moving object belongs to a tracking and shooting device of a moving object.
It should be noted that, the execution subject of the method is a pan-tilt tracking shooting device, and the device predicts and calculates the image data after receiving the image data shot by the camera, so as to predict the moving track of the shot person, and further, control the pan-tilt and the camera to rotate to the corresponding angles, so as to achieve the effect of tracking and shooting the person.
It should be understood that the pan-tilt tracking shooting device comprises a single-view high-definition camera, a close-up lens, a 2-degree-of-freedom rotating pan-tilt and an optical zoom camera, fixed-point zoom shooting of a specific target can be achieved, in the shooting process, the lens/camera is used for capturing images of a user, the mechanical pan-tilt is used for supporting the camera and driving the camera to rotate for 360 degrees, and equipment such as a microphone is used for receiving and the like.
It should be understood that the number of target persons may be one or more, and that the movement trajectories of a plurality of target persons may be tracked when a plurality of target persons exist in the image data and the movement trajectories of the respective target persons are similar during photographing.
The application differs from the prior art in that:
the prior art comprises the following steps: the methods commonly used at present mainly comprise 4 types:
1. tracking the track of the suspected target by adopting a traditional tracking algorithm such as a background difference and an optical flow algorithm; the complex problems of ppt switching of different pictures, light changes, multi-person crossing, shielding and the like seriously influence the tracking effect depending on the background.
2. Based on the character detection algorithm, such as yolo series character detection algorithm+sort tracking algorithm, confirming the continuity of the track by calculating the current character frame and the IOU predicted by kalman (Kalman algorithm), wherein the method is easy to generate the phenomenon of track confusion for occlusion and aggregation scenes;
3. the character detection algorithm, the pedestrian and face features and the kalman tracking algorithm are combined, so that the problem of multi-person intersection can be effectively solved by the algorithm, but the prediction of kalman is influenced due to the fact that a holder and a target are both in motion when the pedestrian features or the face features are pulled back remotely and shielded, and the tracking effect is influenced due to the fact that the holder is swayed in disorder;
4. the binocular stereoscopic vision and depth camera scheme utilizes the double-sided stereoscopic vision technology or depth information to perform 3D modeling on the track of the target, and has the defects of high cost and incapability of being used in the original monocular machine;
the difficulties in the prior art are mainly that:
1) In the tracking process of the monocular cradle head tracking machine, the cradle head moves, pedestrians in the picture also move, and at the moment, the relative coordinates of the pedestrians in the picture have uncertainty and vibrate;
2) Due to the limitations of 1), kalman predictions become difficult, especially when occlusion or aggregation occurs;
3) The targets cannot be tracked by counting the motion trail of each target due to the limitation of 1);
4) Due to the limitation of 1), the movement of the cradle head is abnormal when the human face or the reid is pulled back to the target in a long distance, and the tracking is not smooth.
The application comprises the following steps: by establishing pseudo 3D tracking track modeling under a polar coordinate system and combining face characteristics, pedestrian characteristics and real-time modeling of the track, the accuracy of tracking a target and the smoothness of tracking are greatly guaranteed, wherein the real depth information is not obtained by means of SFM (Structure from Motion, motion structure recovery algorithm) estimation or other algorithms, but the size converted to one time by using the variable magnification of an optical zoom lens and the size of a picture pedestrian is used as the pseudo depth information for modeling.
Step S20, carrying out feature extraction processing on the image data to obtain target pedestrian features;
before image data is processed by a model, feature extraction processing needs to be performed on the image data to be input to the model, wherein the target pedestrian feature is the captured image data of the target person.
The step S20 of performing feature extraction processing on the image data to obtain the features of the target pedestrian includes:
step S21, identifying the image data based on a preset identification model to obtain first identification data;
the preset recognition model may be a trained multitask model, the multitask model belongs to a deep learning model, and the character data in the image data is determined by recognizing characters and character attributes thereof in the image data care, and operations such as rejecting, screening and the like are performed on irrelevant factors in other environments.
The first identification data is image data obtained by identifying a person, and the first identification data includes only person-related information.
And S22, extracting character characteristic information in the first identification data to obtain the characteristics of the target pedestrians.
The feature extraction is performed on the character information in the first identification data to obtain the pedestrian feature of the target character.
It should be noted that the target pedestrian features include a pedestrian feature and a pedestrian face feature, and both the face attribute and the pedestrian attribute are taken into consideration when performing tracking shooting.
Step S30, calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on the polar coordinate angle of the target person relative to a holder;
the preset track tracking model includes a deep learning model and a kalman filtering algorithm, and performs state prediction on the target person according to each coordinate of the target pedestrian feature to determine the motion state of the target person, thereby obtaining the motion track of the target person.
It should be noted that, in the conventional practice, the relative coordinates x of the target person in the camera frame are mainly used as follows, and the state x of kalma motion estimation is as formula 1:
where x_center, y_center, w, h are the center coordinates and width and height of the human target, the units are pixels,the speed of the change of the center and the width and the height of the human object target is the coordinate under 640x360 resolution in the formula;
because the target person and the tripod head are simultaneously performing irregular motions, the relative coordinates of the person in the motion of the tripod head camera are used as motion states to perform modeling, and specifically referring to fig. 2, the left person starts to move the tripod head to accelerate slowly, the motion speed of the middle person is greater than the rotation speed of the tripod head, and the rotation speed of the right tripod head is greater than the motion speed of the person. The irregular motion state in fig. 2 causes the coordinates of the target person to show an irregular oscillation state, wherein the change schematic diagram of the horizontal center coordinates (X coordinates) of the target person is shown in fig. 4, the change schematic diagram of the horizontal center coordinates (Y coordinates) of the target person is shown in fig. 5, and the change schematic diagram of the width of the target person is shown in fig. 6, so that the kalman prediction effect is poor, the x_center and y_center/width of the target person are influenced with time, and the camera performs zoom magnification according to the size of the pedestrian under a certain condition, so that the size of the target pedestrian is also oscillated; the method can clearly know that the motion trail of each target based on the center and the size is disordered and can be circulated, so that the effect of linear filtering kalman is affected, and meanwhile, the continuity of the target trail cannot be optimized by using a trail strategy, so that the effect of long-range tracking cannot be achieved in use.
It should be noted that, in view of the disadvantage that the relative coordinates of the close-up tracking camera are taken as kalman states, according to the fact that the soc shares the current pan_angle (zoom_ratio) state to the tracking program, the invention constructs a pseudo 3D tracking track model based on the polar coordinate system, as shown in fig. 7, and the angles and zoom values of the pan-tilt are taken as the state quantities of the targets, that is, the target pedestrian features.
And step S40, carrying out real-time tracking shooting on the target person based on the moving track.
After accurately predicting the moving track of the target person, the target person is tracked and shot in real time by rotating the pan-tilt.
The application provides a tracking shooting method, a device, equipment and a storage medium of a moving target, compared with the prior art, in the tracking process of a monocular cradle head tracking machine, the cradle head is moving, pedestrians in a picture are moving, at the moment, relative coordinates of the pedestrians in the picture are uncertain, the motion trail of each target cannot be counted to track, and therefore, compared with the unsmooth tracking shooting, in the method, the moving trail is obtained through processing image data shot by a camera and calculating the polar coordinate angle of a target person relative to the cradle head, the target person is tracked and shot through the calculated moving trail, and the rotation angle of the cradle head is adjusted according to the movement angle of the person relative to the cradle head, so that the rotation angle of the cradle head is identical to the movement angle of the person, and the tracking shooting process of the camera is smooth and connected.
Further, based on the first embodiment of the present application, another embodiment of the present application is provided, in this embodiment, the step S30 of performing calculation processing on the target pedestrian feature based on the preset trajectory tracking model to obtain the movement trajectory of the target person includes:
step S31, extracting action state data of the target person from the target pedestrian characteristic based on a preset track tracking model, wherein the action state data comprises center coordinates of the target person and a center coordinate change value;
it should be noted that, the action state data is shown in formula 1:
wherein x_center, y_center, w, h are the center coordinates of the person object and the width and height of the person, the units are pixels,the center coordinates of the human object target and the width and height changing speed, namely the center coordinate changing value.
Step S32, converting the action state data into polar coordinate state data;
it should be noted that, the present application converts the original action state data into polar coordinate state data, so that the angle and the size of the target are not affected by the pan-tilt during the movement process, and the position and the size are not changed along with the change of the pan-tilt, thereby constructing a track tracking method of 3DOF (degree of freedom) based on the camera as an origin, and the pixel size under the horizontal angle, the vertical angle and the object zoom as one time, and making the track tracking process smoother.
It should be noted that, the polar coordinate state data may be:
w 1 =w/zoom_ratio equation 3
h 1 =h/zoom_ratio equation 4
Wherein pan_angle, tilt_angle, w 1 ,h 1 The horizontal angle, vertical angle (in degrees) and width and height (in pixels at 4K) of the target person relative to the camera at one magnification,and the zoom_ratio is the current multiplying power of the optical zoom lens and is the moving speed corresponding to the target person.
And step S33, generating a moving track based on a plurality of polar coordinate points in the polar coordinate state data.
In the acquired image data, each frame of image corresponds to a polar coordinate point, and after a plurality of points of the multi-frame image are determined, a corresponding movement track can be generated by a point connection mode.
Wherein the step S32 of converting the action state data into polar coordinate state data includes:
step S320, obtaining the image size of the image data, and determining the character offset value of the target character and the first polar coordinate of the target character relative to the holder based on the central coordinate in the action state data;
it should be noted that the image size may be obtained from a fixed setting of the camera, wherein the image size is associated with the pixels of the camera.
After the center coordinates are determined, the offset distance between the target person and the center corresponding to the camera may also be determined, and the offset distance is the person offset value.
The first polar coordinate is the angle α in fig. 8, that is, the angle of the target person with respect to the center point of the camera.
Step S321, determining polar coordinate state data based on the image size, the first polar coordinate, the character offset value and the current pan/tilt angle.
The image size, the first polar coordinate, the person offset value and the current pan/tilt angle are calculated to obtain polar coordinate state data.
Wherein the character offset value comprises a horizontal offset value and a vertical offset value, and the pan-tilt angle comprises a pan-tilt horizontal angle and a pan-tilt vertical angle;
the step S321 of determining polar coordinate state data based on the image size, the first polar coordinate, the person offset value, and the current pan/tilt angle includes:
step S3210, determining second polar coordinates of an image edge relative to the pan-tilt based on the image size, wherein the image size comprises a horizontal size and a vertical size;
the horizontal offset value is an offset value of the target person in the horizontal direction and the center point corresponding to the camera, the vertical offset value is an offset value of the target person in the vertical direction and the center point corresponding to the camera, and similarly, the horizontal angle of the pan-tilt is an angle value of rotation of the pan-tilt in the horizontal direction, and the vertical angle of the pan-tilt is an angle value of rotation of the pan-tilt in the vertical direction.
It should be noted that, the second polar coordinate is an offset angle of the image edge with respect to the corresponding center point of the pan/tilt head, as shown in the angle β of fig. 8.
Step S3211, calculating a target horizontal angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the horizontal offset value, the horizontal dimension, and the current horizontal angle of the pan-tilt;
it should be noted that, through the first polar coordinate, the second polar coordinate, the horizontal offset value, the horizontal dimension and the current horizontal angle of the pan-tilt, the calculation formula of the target horizontal angle of the target person relative to the camera is as follows:
pan_angle＝arctan(x_shift*tan(β)/(img_width/2))+pan_now
wherein pan_angle is a target horizontal angle, pan_now is a current pan-tilt horizontal angle, α is a corresponding angle of a first polar coordinate, β is a corresponding angle of a second polar coordinate, x_shift is a horizontal offset value, and img_width is a horizontal dimension.
Step S3212, calculating a target vertical angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the vertical offset value, the vertical dimension and the current vertical angle of the cradle head;
it should be noted that, according to the first polar coordinate, the second polar coordinate, the vertical offset value, the vertical dimension, and the current vertical angle of the pan-tilt, the calculation formula for calculating the target vertical angle may be:
tilt_angle＝arctan(y_shift*tan(β)/(img_height/2))+tilt_now
wherein pan_angle is a target vertical angle, tilt_now is a current pan-tilt horizontal angle, α is a first polar coordinate corresponding angle, β is a second polar coordinate corresponding angle, y_shift is a horizontal offset value, and img_height is a horizontal dimension.
Step S3213, determining the size of the image of the person based on the zoom magnification of the camera and the size of the image of the target person;
the person image size is the person image size of the target person in the image data, and the person width converted to one time is:
the current pan/tilt state amounts of width_now=w/zoom_ratio are pan_now (horizontal angle), tilt_now (vertical angle), zoom_ratio (current magnification of the optical zoom lens).
Step S3214, determining polar coordinate state data based on the target horizontal angle, the target vertical angle, and the person image size.
It should be noted that, the polar state data is expressed as:
w 1 ＝w/zoom_ratio
h 1 ＝h/zoom_ratio
wherein pan_angle, tilt_angle, w 1 ,h 1 The horizontal angle, vertical angle (in degrees) and width and height (in pixels at 4K) of the person relative to the camera at one magnification,for the corresponding speed, zoom_ratio is the current magnification of the optical zoom lens;
in this application, the range of pan_angle is [0 °,340 ° ], the range of tilt_angle is [0 °,120 ° ], and width is the pixel value at 4k, and the range of tilt_angle is [0,3840].
The step S33 of generating a moving track based on the plurality of polar coordinate points in the polar coordinate state data includes:
step S330, based on a preset track policy, connects the plurality of polar coordinate points in the polar coordinate state data, so as to obtain a moving track.
The movement trajectory is determined by connecting a plurality of polar coordinates in the polar coordinate state data.
It should be noted that, the predicted trajectory graph of the conventional model and the predicted trajectory graph in the present application are shown in fig. 11 and 12, and it can be seen that the obtained coordinates are converted into polar coordinates, so that the generated trajectory is clear, and a more accurate movement trajectory can be obtained.
In this embodiment, the relative coordinates acquired by the camera are converted into polar coordinates, and the trajectory tracking is performed according to the changing angle of the movement of the person, so that the trajectory tracking of the target person is smoothed.
Further, based on the first embodiment and the second embodiment in the present application, there is provided another embodiment in the present application, in this embodiment, the step S40 of performing real-time tracking shooting on the target person based on the movement track includes:
step S41, determining first angles of a plurality of polar coordinate points in the moving track;
in the image data, each frame of image corresponds to a polar coordinate point, and each polar coordinate point corresponds to a plurality of angles with respect to the camera.
It should be noted that, the first angle may be an angle value formed by the polar coordinate point in each frame of image and the camera/cradle head, and the value of the first angle will also change correspondingly as the task moves continuously.
And step S42, rotating the cradle head and the camera based on the change value of the first angle so as to track and shoot the target person in real time.
It should be noted that, according to the real-time variation value of the first angle, the cradle head is rotated to track the target person or the target face, and because the camera is arranged above the cradle head, when rotating, only the cradle head is rotated to drive the camera to synchronously rotate, thereby realizing the effect of real-time tracking and shooting of the target person.
In this embodiment, according to a plurality of polar coordinate points in the movement track, the rotation angle of the target person is determined, and the real-time tracking shooting is performed on the target person by rotating the pan-tilt in real time.
Referring to fig. 3, fig. 3 is a schematic device structure diagram of a hardware running environment according to an embodiment of the present application.
As shown in fig. 3, the tracking photographing apparatus of the moving object may include: a processor 1001, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to enable connected communication between the processor 1001 and the memory 1005.
Optionally, the tracking shooting device of the moving object may further include a user interface, a network interface, a cradle head, an RF (Radio Frequency) circuit, a sensor, a WiFi module, and so on. The user interface may include a Display, an input sub-module such as a Keyboard (Keyboard), and the optional user interface may also include a standard wired interface, a wireless interface. The network interface may include a standard wired interface, a wireless interface (e.g., WI-FI interface).
It will be appreciated by those skilled in the art that the tracking camera configuration of the moving object shown in fig. 3 does not constitute a limitation of the tracking camera of the moving object, and may include more or less components than illustrated, or may combine certain components, or may be arranged in different components.
As shown in fig. 3, an operating system, a network communication module, and a tracking shooting program of a moving object may be included in a memory 1005 as one type of storage medium. The operating system is a program that manages and controls the tracking camera hardware and software resources of the moving object, supporting the tracking camera of the moving object and the execution of other software and/or programs. The network communication module is used for realizing communication among components in the memory 1005 and communication with other hardware and software in the tracking shooting system of the moving object.
In the moving object tracking shooting apparatus shown in fig. 3, a processor 1001 is configured to execute a moving object tracking shooting program stored in a memory 1005, and to implement the steps of the moving object tracking shooting method described in any one of the above.
The specific implementation manner of the tracking shooting device of the moving target is basically the same as the above embodiments of the tracking shooting method of the moving target, and will not be described herein.
The application also provides a tracking shooting device of a moving target, the tracking shooting device of the moving target comprises:
the receiving module is used for receiving image data of a target person shot by the camera, wherein a cradle head is arranged at the lower end of the camera;
the processing module is used for carrying out feature extraction processing on the image data to obtain target pedestrian features;
the calculation module is used for calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on the polar coordinate angle of the target person relative to a holder;
and the tracking shooting module is used for carrying out real-time tracking shooting on the target person based on the moving track.
In one possible embodiment of the present application, the calculation module includes:
a first extraction unit, configured to extract, based on a preset trajectory tracking model, action state data of the target person from the target pedestrian feature, where the action state data includes a center coordinate of the target person and a center coordinate variation value;
a conversion unit for converting the action state data into polar coordinate state data;
and the generation unit is used for generating a moving track based on a plurality of polar coordinate points in the polar coordinate state data.
In a possible embodiment of the present application, the conversion unit includes:
an acquisition subunit, configured to acquire an image size of the image data, and determine a person offset value of the target person and a first polar coordinate of the target person relative to a pan-tilt based on a center coordinate in the action state data;
and the determining subunit is used for determining polar coordinate state data based on the image size, the first polar coordinate, the character offset value and the current holder angle.
In one possible embodiment of the present application, the determining subunit includes:
a first determining component configured to determine a second polar coordinate of an image edge relative to a pan-tilt based on the image size, the image size including a horizontal size and a vertical size;
the first calculating component is used for calculating a target horizontal angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the horizontal offset value, the horizontal dimension and the current horizontal angle of the holder;
the second calculating component is used for calculating a target vertical angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the vertical offset value, the vertical dimension and the current vertical angle of the cradle head;
a second determining component for determining a person image size based on a zoom magnification of the camera and the image size of the target person;
and a third determining component for determining polar coordinate state data based on the target horizontal angle, the target vertical angle, and the person image size.
In a possible embodiment of the present application, the generating unit includes:
and the connection subunit is used for connecting a plurality of polar coordinate points in the polar coordinate state data based on a preset track strategy to obtain a moving track.
In one possible embodiment of the present application, the processing module includes:
the identification unit is used for identifying the image data based on a preset identification model to obtain first identification data;
and the second extraction unit is used for extracting the character characteristic information in the first identification data to obtain the characteristics of the target pedestrians.
In one possible embodiment of the present application, the tracking shooting module includes:
a determining unit, configured to determine first angles of a plurality of polar coordinate points in the movement track;
and the rotation unit is used for rotating the cradle head and the camera based on the change value of the first angle so as to track and shoot the target person in real time.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present application are merely for describing, and do not represent advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) as described above, including several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method described in the embodiments of the present application.
The foregoing description is only of the preferred embodiments of the present application, and is not intended to limit the scope of the claims, and all equivalent structures or equivalent processes using the descriptions and drawings of the present application, or direct or indirect application in other related technical fields are included in the scope of the claims of the present application.
Claims (10)
1. A tracking shooting method of a moving object, characterized in that the method comprises the following steps:
receiving image data of a target person shot by a camera, wherein a cradle head is arranged at the lower end of the camera;
performing feature extraction processing on the image data to obtain target pedestrian features;
calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on a polar coordinate angle of the target person relative to a holder;
and carrying out real-time tracking shooting on the target person based on the moving track.
2. The method for tracking and photographing a moving object according to claim 1, wherein the step of calculating the characteristics of the target pedestrian based on a preset trajectory tracking model to obtain the moving trajectory of the target person comprises:
extracting action state data of the target person from the target pedestrian characteristic based on a preset track tracking model, wherein the action state data comprises center coordinates of the target person and a center coordinate change value;
converting the action state data into polar coordinate state data;
and generating a moving track based on a plurality of polar coordinate points in the polar coordinate state data.
3. The tracking shooting method of a moving object according to claim 2, wherein the step of converting the action state data into polar state data comprises:
acquiring the image size of the image data, and determining a character offset value of the target character and a first polar coordinate of the target character relative to a holder based on the central coordinate in the action state data;
and determining polar coordinate state data based on the image size, the first polar coordinate, the character offset value and the current holder angle.
4. The method for tracking and photographing a moving object according to claim 3, wherein the character offset value includes a horizontal offset value and a vertical offset value, and the pan-tilt angle includes a pan-tilt horizontal angle and a pan-tilt vertical angle;
the step of determining polar coordinate state data based on the image size, the first polar coordinate, the person offset value, and the current pan/tilt angle includes:
determining second polar coordinates of an image edge relative to the pan-tilt based on the image size, the image size including a horizontal size and a vertical size;
calculating a target horizontal angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the horizontal offset value, the horizontal dimension and the current horizontal angle of the cradle head;
calculating a target vertical angle of the target person relative to the camera based on the first polar coordinate, the second polar coordinate, the vertical offset value, the vertical dimension and the current vertical angle of the cradle head;
determining a person image size based on the zoom magnification of the camera and the image size of the target person;
polar coordinate state data is determined based on the target horizontal angle, the target vertical angle, and the person image size.
5. The method of tracking a moving object according to claim 2, wherein the step of generating a moving trajectory based on a plurality of polar coordinate points in the polar coordinate state data includes:
and connecting a plurality of polar coordinate points in the polar coordinate state data based on a preset track strategy to obtain a moving track.
6. The method for tracking and photographing a moving object according to claim 1, wherein the step of performing feature extraction processing on the image data to obtain the features of the pedestrian as the object comprises:
based on a preset recognition model, recognizing the image data to obtain first recognition data;
and extracting character characteristic information in the first identification data to obtain the characteristics of the target pedestrians.
7. The method of tracking and photographing a moving object according to claim 1, wherein the step of tracking and photographing the object person in real time based on the moving trajectory comprises:
determining first angles of a plurality of polar coordinate points in the moving track;
and rotating the cradle head and the camera based on the change value of the first angle so as to track and shoot the target person in real time.
8. A tracking camera of a moving object, characterized in that the tracking camera of a moving object comprises:
the receiving module is used for receiving image data of a target person shot by the camera, wherein a cradle head is arranged at the lower end of the camera;
the processing module is used for carrying out feature extraction processing on the image data to obtain target pedestrian features;
the calculation module is used for calculating the characteristics of the target pedestrians based on a preset track tracking model to obtain a moving track of the target person, wherein the moving track is calculated based on the polar coordinate angle of the target person relative to a holder;
and the tracking shooting module is used for carrying out real-time tracking shooting on the target person based on the moving track.
9. A tracking shooting apparatus of a moving object, characterized by comprising: a memory, a processor, and a tracking shooting program of a moving object stored on the memory and executable on the processor, the tracking shooting program of a moving object configured to implement the steps of the tracking shooting method of a moving object according to any one of claims 1 to 7.
10. A computer storage medium, wherein a tracking shooting program of a moving object is stored on the computer storage medium, and the tracking shooting program of the moving object, when executed by a processor, realizes the steps of the tracking shooting method of the moving object according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311577528.0A CN117692755A (en) | 2023-11-22 | 2023-11-22 | Moving target tracking shooting method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311577528.0A CN117692755A (en) | 2023-11-22 | 2023-11-22 | Moving target tracking shooting method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117692755A true CN117692755A (en) | 2024-03-12 |
Family
ID=90136248
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311577528.0A Pending CN117692755A (en) | 2023-11-22 | 2023-11-22 | Moving target tracking shooting method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117692755A (en) |
-
2023
- 2023-11-22 CN CN202311577528.0A patent/CN117692755A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110866480B (en) | Object tracking method and device, storage medium and electronic device | |
| CN104685513B (en) | According to the high-resolution estimation of the feature based of the low-resolution image caught using array source | |
| US9159169B2 (en) | Image display apparatus, imaging apparatus, image display method, control method for imaging apparatus, and program | |
| CN111935393A (en) | Shooting method, shooting device, electronic equipment and storage medium | |
| CN105678809A (en) | Handheld automatic follow shot device and target tracking method thereof | |
| US8903130B1 (en) | Virtual camera operator | |
| CN112207821B (en) | Target searching method of visual robot and robot | |
| CN111382613B (en) | Image processing method, device, equipment and medium | |
| JP2015522959A (en) | Systems, methods, and media for providing interactive refocusing in images | |
| WO2007032819A2 (en) | Camera placement and virtual-scene construction | |
| JP6624877B2 (en) | Information processing apparatus, information processing method and program | |
| JP6610535B2 (en) | Image processing apparatus and image processing method | |
| CN112640419B (en) | Following method, movable platform, device and storage medium | |
| JP2004062393A (en) | Method and device for determining attention | |
| CN112001224A (en) | Video acquisition method and video acquisition system based on convolutional neural network | |
| Stancil et al. | Active multicamera networks: From rendering to surveillance | |
| CN117014716A (en) | Target tracking method and electronic equipment | |
| CN111385481A (en) | Image processing method and device, electronic device and storage medium | |
| CN117692755A (en) | Moving target tracking shooting method, device, equipment and storage medium | |
| CN115589532A (en) | Anti-shake processing method and device, electronic equipment and readable storage medium | |
| CN113409358A (en) | Image tracking method, image tracking device, storage medium and electronic equipment | |
| EP3624050B1 (en) | Method and module for refocusing at least one plenoptic video | |
| CN113950705A (en) | Image processing method and device and movable platform | |
| Ambadkar et al. | Deep reinforcement learning approach to predict head movement in 360 videos | |
| Lee | Novel video stabilization for real-time optical character recognition applications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |