WO2021102695A1 - 云台、云台的控制设备、云台的控制方法及存储介质 - Google Patents

云台、云台的控制设备、云台的控制方法及存储介质 Download PDF

Info

Publication number
WO2021102695A1
WO2021102695A1 PCT/CN2019/121013 CN2019121013W WO2021102695A1 WO 2021102695 A1 WO2021102695 A1 WO 2021102695A1 CN 2019121013 W CN2019121013 W CN 2019121013W WO 2021102695 A1 WO2021102695 A1 WO 2021102695A1
Authority
WO
WIPO (PCT)
Prior art keywords
recognition target
shooting
pan
candidate
composition
Prior art date
Application number
PCT/CN2019/121013
Other languages
English (en)
French (fr)
Inventor
杨龙超
朱高
熊风
Original Assignee
深圳市大疆创新科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深圳市大疆创新科技有限公司 filed Critical 深圳市大疆创新科技有限公司
Priority to PCT/CN2019/121013 priority Critical patent/WO2021102695A1/zh
Priority to CN201980038365.4A priority patent/CN112243581A/zh
Publication of WO2021102695A1 publication Critical patent/WO2021102695A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/61—Control of cameras or camera modules based on recognised objects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/61—Control of cameras or camera modules based on recognised objects
    • H04N23/611—Control of cameras or camera modules based on recognised objects where the recognised objects include parts of the human body
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/64—Computer-aided capture of images, e.g. transfer from script file into camera, check of taken image quality, advice or proposal for image composition or decision on when to take image

Definitions

  • This application relates to the application field of electronic technology, and specifically to a pan-tilt, a control device for the pan-tilt, a control method of the pan-tilt, and a computer-readable storage medium.
  • the recognition will often fail and the composition is not beautiful; when the subject is the person's head and shoulders, when the person is far away from the camera, the person's head and shoulders will become very small in the camera picture. It is easy to confuse with the background, which leads to recognition failure.
  • This application aims to solve at least one of the technical problems existing in the prior art or related technologies.
  • the first aspect of this application proposes a pan-tilt.
  • the second aspect of the application proposes a control device for a pan-tilt.
  • the third aspect of this application proposes a method for controlling the pan/tilt.
  • the fourth aspect of the present application proposes a computer-readable storage medium.
  • a pan/tilt head including: a memory, configured to store computer instructions; and a processor, configured to execute computer instructions to achieve: real-time reception of image information of the captured screen , There is a shooting object in the shooting screen; detecting the composition information of the first recognition target in the shooting screen according to the image information, and the first recognition target is at least a part of the shooting object; determining whether to generate the recognition target switching instruction according to the composition information of the first recognition target .
  • a control device for a pan/tilt head including: a memory configured to store computer instructions; and a processor configured to execute computer instructions to realize: real-time reception of image information of a captured screen, There is a photographic subject in the shooting frame; the composition information of the first recognition target in the shooting frame is detected according to the image information, and the first recognition target is at least a part of the photographic subject; it is determined whether to generate the recognition target switching instruction according to the composition information of the first recognition target.
  • a method for controlling a pan/tilt is provided, which is used in a pan/tilt or a control device for the pan/tilt.
  • the control method of the pan/tilt includes: receiving image information of a shooting screen in real time, and there is a shooting object in the shooting screen. Detecting the composition information of the first recognition target in the shooting frame according to the image information, the first recognition target being at least a part of the shooting object; determining whether to generate the recognition target switching instruction according to the composition information of the first recognition target.
  • a computer-readable storage medium on which a computer program is stored.
  • the steps of the pan/tilt control method as in any of the above technical solutions are realized.
  • the embodiment of the present application provides a pan-tilt and its control scheme, which determines whether to generate a recognition target switching instruction according to the composition information of the detected first recognition target, that is, according to the first recognition target of the subject in the current shooting frame
  • the composition information of the picture a reasonable choice of whether to switch the recognition target, not only improves the recognition success rate, but also helps to beautify the composition.
  • This solution can be based on the auto-following technology and applied to the PTZ's auto-follow shooting function (that is, the automatic follow-up shooting function), so as to switch and adjust the follow target in time, which can not only ensure the follow-up success, but also make the composition beautiful and improve the follow-up effect; It can also be used when the pan/tilt is pointed at the subject by hand without the automatic follow-up function activated, which helps to adjust the recognition target reasonably, optimize the composition, and improve the shooting effect.
  • the PTZ's auto-follow shooting function that is, the automatic follow-up shooting function
  • Fig. 1 shows a schematic structural diagram of a pan-tilt according to an embodiment of the present application
  • Figure 2 shows a schematic structural diagram of a pan-tilt control device according to an embodiment of the present application
  • FIG. 3 shows a schematic flowchart of a method for controlling a pan-tilt according to an embodiment of the present application
  • FIG. 4 shows a schematic flowchart of a method for controlling a pan-tilt according to another embodiment of the present application
  • FIG. 5 shows a schematic flowchart of a method for controlling a pan/tilt head according to still another embodiment of the present application
  • Fig. 6 shows a schematic flowchart of a method for controlling a pan/tilt head according to another embodiment of the present application.
  • the embodiment of the first aspect of the present application provides a pan/tilt head 100, including: a memory 102 configured to store computer instructions; and a processor 104 configured to execute computer instructions to realize: real-time reception
  • the image information of the shooting screen there is a shooting object in the shooting screen; the composition information of the first recognition target in the shooting screen is detected according to the image information, the first recognition target is at least a part of the shooting object; whether to determine whether or not is determined according to the composition information of the first recognition target Generate recognition target switching instructions.
  • the pan/tilt 100 provided by the embodiment of the present application is configured with the processor 104 to determine whether to generate the recognition target switching instruction according to the detected composition information of the first recognition target when executing the computer instructions stored in the memory 102, that is, according to The composition information of the first recognition target of the subject in the current shooting frame, and a reasonable choice of whether to switch the recognition target not only improves the recognition success rate, but also helps to beautify the composition. It is understandable that the recognition targets before and after the switch all belong to the same shooting object.
  • this solution can be applied to the automatic follow-up function of the PTZ 100 based on the auto-following technology, so as to switch and adjust the follow-up target in time, which can ensure the success of follow-up, make the composition beautiful, and improve the follow-up effect; It is used when the automatic follow-up function is activated and the PTZ 100 is pointed at the subject by hand.
  • the PTZ 100 is manually controlled to follow the shooting, it helps to adjust the recognition target reasonably, optimize the composition, and improve the shooting effect. For example, only manually control the cloud.
  • the stage 100 is aimed at the subject and does not start shooting temporarily, it can ensure that it is in a better recognition state and composition state, so that shooting can be started at any time.
  • the composition information includes a size
  • the processor 104 executes a computer instruction to determine whether to generate a recognition target switching instruction according to the composition information of the first recognition target, including: determining that the size of the first recognition target is within a preset size range In addition, generate recognition target switching instructions.
  • the composition information is specifically limited to include the size. Accordingly, the solution for determining whether to generate the recognition target switching instruction is that when the size of the first recognition target is outside the preset size range, the first recognition target is considered to be in The size of the shooting screen is unreasonable, and then a recognition target switching instruction is generated to switch to a recognition target with a more reasonable size, which improves the recognition success rate and helps to beautify the composition.
  • the recognition target switching instruction includes a recognition target expansion instruction.
  • the processor 104 executes the computer instruction to determine that the size of the first recognition target is outside the preset size range, and generates the recognition target switching instruction, including: determining the first recognition target.
  • the size of the recognition target is smaller than the lower limit size of the preset size range, and the recognition target expansion instruction is generated.
  • the size of the first recognition target when it is detected that the size of the first recognition target is smaller than the lower limit size, it is considered that the size of the first recognition target in the shooting frame is too small, which corresponds to the situation that the shooting object is far away from the pan/tilt head 100.
  • Generate recognition target expansion instructions to switch to a larger recognition target It is understandable that a larger recognition target refers to a recognition target that accounts for a larger proportion of the photographed object than the first recognition target.
  • the first recognition target can be included in a larger recognition target.
  • the first recognition target is the head and shoulders of the photographing object, and the larger recognition target is the whole body of the photographing object; the first recognition target can also be completely identical to the larger recognition target.
  • the first recognition target is the head, and the larger recognition target is the torso; the first recognition target can also partially overlap with the larger recognition target, for example, the first recognition target is the head and shoulders, and the larger recognition target is the upper limbs. The overlapping part of the two is the shoulder.
  • the size of the first recognition target in the shooting screen gradually shrinks. When its size is too small, it is easy to be confused with the background, which reduces the recognition accuracy, which leads to recognition failure, which affects the composition effect and automatic follow-up.
  • the realization of the function By generating a recognition target expansion instruction, it can switch to a larger recognition target that is easy to recognize, which improves the recognition success rate, helps beautify the composition, and ensures the reliable operation of the automatic follow-up function.
  • the recognition target switching instruction includes a recognition target reduction instruction.
  • the processor 104 executes the computer instruction to determine that the size of the first recognition target is outside the preset size range, and generates the recognition target switching instruction, including: determining the first recognition target.
  • the size of the recognition target is larger than the upper limit size of the preset size range, and the recognition target reduction instruction is generated.
  • the recognition target switching instruction may further include a recognition target reduction instruction, and is generated when it is detected that the size of the first recognition target is larger than the upper limit size, so as to switch to a smaller recognition target.
  • a recognition target reduction instruction is generated when it is detected that the size of the first recognition target is larger than the upper limit size, so as to switch to a smaller recognition target.
  • the first recognition target is not fixed, but refers to It replaces the recognition target currently in use, which can be switched with the movement of the subject. As the subject gets closer, the size of the first recognition target in the shooting screen gradually expands.
  • the composition When its size is too large, if the composition continues to be performed according to the first recognition target, the aesthetics of the composition may be reduced, and the first recognition may even occur.
  • the target overflows the shooting screen and the recognition fails, at this time, by generating the recognition target reduction instruction, you can switch to a smaller recognition target in time, which not only helps to beautify the composition, but also solves the problem of the first recognition target overflowing the shooting screen.
  • the recognition success rate can be ensured to ensure the reliable operation of the automatic follow-up function.
  • the processor 104 when the processor 104 executes the computer instructions, it also realizes that: in response to the recognition target switching instruction, switch from the first recognition target to the second recognition target, the second recognition target is at least a part of the photographed object, and the first recognition target The proportion of the subject and the proportion of the second recognition target in the subject are not equal.
  • the processor 104 can also switch the recognition target from the first recognition target of the photographing object to the second recognition target of the photographing object based on the recognition target switching instruction, so that automatic switching of the recognition target is realized without manual operation by the user. , Improve the convenience of operation and the timeliness of switching.
  • the first recognition target and the second recognition target occupies different proportions in the shooting object, which can effectively adjust the recognition during shooting, which helps to improve the recognition success rate and beautify the composition.
  • the composition information further includes a position, and when the processor 104 executes the computer instructions, the processor 104 also implements: adjusting the shooting parameters according to the composition information of the first recognition target, so that the imaging of the first recognition target in the shooting frame meets the composition rules.
  • the composition information also includes the position.
  • the processor 104 can follow the movement and change of the first recognition target in the shooting screen.
  • the automatic adjustment of shooting parameters plays a role in adjusting and beautifying the composition. It is understandable that when the pan-tilt 100 is pointed at the subject by human hands, the shooting parameters can be automatically fine-tuned in combination with this solution to realize automatic composition.
  • the processor 104 executes computer instructions to switch from the first recognition target to the second recognition target, including: detecting the composition information of the second recognition target in the shooting frame; adjusting the shooting parameters to enable the second recognition The position and size of the target in the shooting frame gradually change until the imaging of the second recognition target in the shooting frame satisfies the composition rule.
  • the recognition target switching instruction includes the aforementioned recognition target enlargement instruction and recognition target reduction instruction
  • the recognition target and composition can be automatically adjusted according to the distance between the photographic subject and the pan-tilt 100, which improves the recognition success rate. It also helps to beautify the composition.
  • the composition rule includes at least one of the following: the imaging of the recognition target is complete, the recognition target is located in a designated area in the shooting frame, the balanced composition, the symmetrical composition, the diagonal composition, the triangular composition, the nine-square grid composition, Centripetal composition, bisection composition.
  • the content of the composition rule is specifically limited.
  • its imaging is complete and/or located in a designated area in the shooting frame as the composition rule to ensure the shooting effect of the subject;
  • the imaging of the entire shooting screen allows it to meet different composition forms, ensuring the overall beauty of the shooting screen.
  • the pan/tilt head 100 further includes: a stand device capable of supporting a photographing device for acquiring image information, and/or a photographing device for acquiring image information; the photographing parameters include at least one of the following: The height of the device, the inclination angle of the shooting device, the direction of the shooting device, and the focal length of the shooting device.
  • the shooting parameters include at least one of the height, inclination, direction, and focal length of the shooting device.
  • the shooting device can be controlled to achieve shooting.
  • the photographing device may be a self-contained structure of the pan-tilt 100 or a third-party structure. For the latter, the pan-tilt 100 needs to be equipped with a support device to support the photographing device.
  • one of the first recognition target and the second recognition target is the whole of the photographed object, and the other of the first recognition target and the second recognition target is a designated part of the photographed object.
  • the whole of the subject can be regarded as a recognition target, and a specified part of the subject, such as head and shoulders, can be used as another recognition target, so as to use pedestrian detection technology, head and shoulder detection technology, or other recognition targets.
  • the detection technology and automatic follow technology of the designated part automatically switch the full/half-length composition according to the distance of the subject from the PTZ 100, so that the subject is always in the best composition position.
  • pedestrian detection and head and shoulder detection are technologies that use various sensors as input (such as a camera) to automatically detect the whole body, head and shoulders of the pedestrian in the field of view of the sensor, and obtain the position and size of the pedestrian in the screen.
  • the processor 104 when the processor 104 executes the computer instructions, it also implements: outputting a recognition target switching instruction.
  • the processor 104 can also directly output the recognition target switching instruction to prompt the user to manually switch the recognition target, ensuring the user's autonomous control. It is understandable that the automatic switching mode and the manual switching mode can be configured for the pan-tilt 100 at the same time, and the user can choose which one to use. In addition, if you need to switch the recognition target during video recording, you can pause the recording first, and then resume recording after manually selecting the recognition target again.
  • the processor 104 when the processor 104 executes the computer instructions, it also implements: outputting at least one candidate recognition target; switching from the first recognition target to the selected one of the at least one candidate recognition target.
  • At least one reasonable candidate recognition target can be provided for the user to choose during manual switching, and the user does not need to manually select the specific part of the photographed object as the recognition target to be switched, which improves the switching speed.
  • the candidate recognition target can be multiple recognition targets corresponding to multiple different parts of the photographic object, or it can be at least one recognition target that is determined to be suitable for switching after screening. For the latter, when the candidate recognition target When it is one, it can be the second recognition target in the automatic switching mode.
  • the processor 104 when the processor 104 executes the computer instructions, it also implements: based on the presence of at least one candidate photographic subject in the shooting frame, detecting the candidate recognition target of the at least one candidate photographic object; outputting the at least one candidate recognition target ; The selected candidate recognition target is used as the first recognition target, and the candidate photographing object to which the first recognition target belongs is used as the photographing object.
  • the identification targets of these candidate subjects can be detected as candidate identification targets and output for the user to select, and then the candidate identification targets selected by the user and the equipment to which they belong Selecting the photographing object as the first recognition target and the photographing object respectively improves the efficiency of selecting the photographing object and its first recognition target in the initial stage.
  • outputting at least one candidate recognition target realized by executing computer instructions by the processor 104 includes: outputting one candidate recognition target for each candidate photographed object.
  • the processor 104 executes the computer instructions to realize the output of a candidate recognition target, including: determining the number of candidate recognition targets corresponding to the candidate shooting object is one, outputting the corresponding candidate recognition target; determining the candidate The number of candidate recognition targets corresponding to the photographing object is at least two, and one of the corresponding candidate recognition targets is output.
  • the number of candidate recognition targets is one, it will be directly output, and if the number of candidate recognition targets is at least two, then one of them will be selected for output.
  • the one currently suitable for identification can be selected to ensure the identification effect.
  • the output of one of the candidate recognition targets corresponding to the processor 104 executing the computer instructions includes: detecting the size of all the corresponding candidate recognition targets in the shooting frame, and determining the output candidate recognition target according to the size .
  • the scheme of selecting one output from at least two candidate recognition targets is specifically limited.
  • the size of all candidate recognition targets of a subject in the shooting frame it can be an absolute size, such as width and height, or a relative size.
  • the proportion of the entire shooting frame can reflect the size of the candidate recognition target. The size, and then select the candidate recognition target whose size is suitable for recognition and output it, which helps to ensure the recognition effect.
  • processor 104 when the processor 104 executes computer instructions, it also implements: real-time output of image information.
  • the processor 104 can also output image information of the shooting frame in real time, so as to present the shooting object and composition changes to the user in real time, so that the user can understand the shooting situation or make adjustment operations.
  • the pan/tilt head 100 further includes a display screen for displaying image information.
  • pan/tilt head 100 is further configured with a display screen to display the image information of the shooting screen, which ensures the reliability of the output.
  • the embodiment of the second aspect of the present application provides a pan-tilt control device 200, including: a memory 202 configured to store computer instructions; and a processor 204 configured to execute computer instructions to implement : Receive the image information of the shooting screen in real time, and there is a shooting object in the shooting screen; detect the composition information of the first recognition target in the shooting screen according to the image information, the first recognition target is at least a part of the shooting object; according to the composition of the first recognition target The information determines whether to generate a recognition target switching instruction.
  • the control device 200 of the pan/tilt head configured in the embodiment of the present application configures the processor 204 to determine whether to generate the recognition target switching instruction according to the detected composition information of the first recognition target when executing the computer instructions stored in the memory 202. That is, according to the composition information of the first recognition target of the subject in the current shooting picture, it is possible to reasonably choose whether to switch the recognition target, which not only improves the recognition success rate, but also helps to beautify the composition. It is understandable that the recognition targets before and after the switch all belong to the same shooting object.
  • this solution can be applied to the automatic follow-up function of the pan/tilt based on the auto-following technology, so as to switch and adjust the follow-up target in time, which can ensure the success of follow-up, but also make the composition beautiful and improve the follow-up effect;
  • the automatic follow-up function is used when the gimbal is aimed at the subject by hand. For example, when manually controlling the gimbal to follow the shot, it helps to reasonably adjust the recognition target, optimize the composition, and improve the shooting effect. Another example is to manually control the gimbal alignment. When the subject does not start shooting temporarily, it can be ensured that it is in a better recognition state and composition state, so that the shooting can be started at any time.
  • the composition information includes a size
  • the processor 204 executes a computer instruction to determine whether to generate a recognition target switching instruction according to the composition information of the first recognition target, including: determining that the size of the first recognition target is within a preset size range In addition, generate recognition target switching instructions.
  • the composition information is specifically limited to include the size. Accordingly, the solution for determining whether to generate the recognition target switching instruction is that when the size of the first recognition target is outside the preset size range, the first recognition target is considered to be in The size of the shooting screen is unreasonable, and then a recognition target switching instruction is generated to switch to a recognition target with a more reasonable size, which improves the recognition success rate and helps to beautify the composition.
  • the recognition target switching instruction includes a recognition target expansion instruction
  • the processor 204 executes the computer instruction to determine that the size of the first recognition target is outside the preset size range, and generates the recognition target switching instruction, including: determining the first The size of the recognition target is smaller than the lower limit size of the preset size range, and the recognition target expansion instruction is generated.
  • the recognition target expands the instruction to switch to a larger recognition target.
  • a larger recognition target refers to a recognition target that accounts for a larger proportion of the photographed object than the first recognition target.
  • the first recognition target can be included in a larger recognition target.
  • the first recognition target is the head and shoulders of the photographing object, and the larger recognition target is the whole body of the photographing object; the first recognition target can also be completely identical to the larger recognition target.
  • the first recognition target is the head, and the larger recognition target is the torso; the first recognition target can also partially overlap with the larger recognition target, for example, the first recognition target is the head and shoulders, and the larger recognition target is the upper limbs. The overlapping part of the two is the shoulder.
  • the size of the first recognition target in the shooting screen gradually shrinks. When its size is too small, it is easy to be confused with the background, which reduces the recognition accuracy, which leads to recognition failure, which affects the composition effect and automatic follow-up.
  • the realization of the function By generating a recognition target expansion instruction, it can switch to a larger recognition target that is easy to recognize, which improves the recognition success rate, helps beautify the composition, and ensures the reliable operation of the automatic follow-up function.
  • the recognition target switching instruction includes a recognition target reduction instruction
  • the processor 204 executes the computer instruction to determine that the size of the first recognition target is outside the preset size range, and generates the recognition target switching instruction, including: determining the first The size of the recognition target is larger than the upper limit size of the preset size range, and the recognition target reduction instruction is generated.
  • the recognition target switching instruction may further include a recognition target reduction instruction, and is generated when it is detected that the size of the first recognition target is larger than the upper limit size, so as to switch to a smaller recognition target.
  • a recognition target reduction instruction is generated when it is detected that the size of the first recognition target is larger than the upper limit size, so as to switch to a smaller recognition target.
  • the relationship between the first recognition target and the smaller recognition target is the same as the aforementioned relationship between the larger recognition target and the first recognition target, and is a relative relationship.
  • the first recognition target is not fixed, but refers to It replaces the recognition target currently in use, which can be switched with the movement of the subject. As the subject gets closer, the size of the first recognition target in the shooting screen gradually expands.
  • the composition When its size is too large, if the composition continues to be performed according to the first recognition target, the aesthetics of the composition may be reduced, and the first recognition may even occur.
  • the target overflows the shooting screen and the recognition fails, at this time, by generating the recognition target reduction instruction, you can switch to a smaller recognition target in time, which not only helps to beautify the composition, but also solves the problem of the first recognition target overflowing the shooting screen.
  • the recognition success rate can be ensured to ensure the reliable operation of the automatic follow-up function.
  • the processor 204 when the processor 204 executes the computer instruction, it also realizes that: in response to the recognition target switching instruction, switch from the first recognition target to the second recognition target, the second recognition target is at least a part of the photographed object, and the first recognition target The proportion of the subject and the proportion of the second recognition target in the subject are not equal.
  • the processor 204 may also switch the recognition target from the first recognition target of the photographing object to the second recognition target of the photographing object based on the recognition target switching instruction, so that automatic switching of the recognition target is realized without manual operation by the user. , Improve the convenience of operation and the timeliness of switching.
  • the first recognition target and the second recognition target occupies different proportions in the shooting object, which can effectively adjust the recognition during shooting, which helps to improve the recognition success rate and beautify the composition.
  • the composition information further includes a position, and when the processor 204 executes the computer instructions, the processor 204 also implements: adjusting the shooting parameters according to the composition information of the first recognition target, so that the imaging of the first recognition target in the shooting frame meets the composition rules.
  • the composition information also includes the position.
  • the processor 204 can follow the movement and change of the first recognition target in the shooting screen.
  • the automatic adjustment of shooting parameters plays a role in adjusting and beautifying the composition. It is understandable that when the pan-tilt is aimed at the subject by hand, the shooting parameters can be automatically fine-tuned in combination with this solution to realize automatic composition.
  • the processor 204 executes computer instructions to switch from the first recognition target to the second recognition target, including: detecting the composition information of the second recognition target in the shooting frame; adjusting the shooting parameters to enable the second recognition The position and size of the target in the shooting frame gradually change until the imaging of the second recognition target in the shooting frame satisfies the composition rule.
  • the recognition target switching instruction includes the aforementioned recognition target enlargement instruction and recognition target reduction instruction
  • the recognition target and composition can be automatically adjusted according to the distance between the subject and the pan/tilt, which not only improves the recognition success rate, but also Helps beautify the composition.
  • the composition rule includes at least one of the following: the imaging of the recognition target is complete, the recognition target is located in a designated area in the shooting frame, the balanced composition, the symmetrical composition, the diagonal composition, the triangular composition, the nine-square grid composition, Centripetal composition, bisection composition.
  • the content of the composition rule is specifically limited.
  • its imaging is complete and/or located in a designated area in the shooting frame as the composition rule to ensure the shooting effect of the subject;
  • the imaging of the entire shooting screen allows it to meet different composition forms, ensuring the overall beauty of the shooting screen.
  • the pan-tilt further includes: a stand device capable of supporting a photographing device for acquiring image information, and/or a photographing device for acquiring image information;
  • the photographing parameter includes at least one of the following: The height of the camera, the inclination of the camera, the direction of the camera, and the focal length of the camera.
  • the shooting parameters include at least one of the height, inclination, direction, and focal length of the shooting device.
  • the shooting device can be controlled to achieve shooting.
  • the photographing equipment can be the structure that comes with the PTZ or a third-party structure. For the latter, the PTZ needs to be equipped with bracket equipment to support the photographing equipment.
  • one of the first recognition target and the second recognition target is the whole of the photographed object, and the other of the first recognition target and the second recognition target is a designated part of the photographed object.
  • the whole of the subject can be regarded as a recognition target, and a specified part of the subject, such as head and shoulders, can be used as another recognition target, so as to use pedestrian detection technology, head and shoulder detection technology, or other recognition targets.
  • the detection technology and automatic follow technology of the designated part automatically switch the full/half-length composition according to the distance of the subject to the pan/tilt, so that the subject is always in the best composition position.
  • pedestrian detection and head and shoulder detection use various sensors as input (such as a camera) to automatically detect the pedestrian's whole body, head and shoulders in the sensor's field of view, and obtain the position and size of the pedestrian in the screen.
  • the processor 204 when the processor 204 executes the computer instruction, it also implements: outputting a recognition target switching instruction.
  • the processor 204 may also directly output the recognition target switching instruction to prompt the user to manually switch the recognition target, ensuring the user's autonomous control. It is understandable that the automatic switching mode and manual switching mode can be configured for the PTZ at the same time, and the user can choose which one to use. In addition, if you need to switch the recognition target during video recording, you can pause the recording first, and then resume recording after manually selecting the recognition target again.
  • the processor 204 when the processor 204 executes the computer instructions, it also implements: outputting at least one candidate recognition target; switching from the first recognition target to the selected one of the at least one candidate recognition target.
  • At least one reasonable candidate recognition target can be provided for the user to choose during manual switching, and the user does not need to manually select the specific part of the photographed object as the recognition target to be switched, which improves the switching speed.
  • the candidate recognition target can be multiple recognition targets corresponding to multiple different parts of the photographic object, or it can be at least one recognition target that is determined to be suitable for switching after screening. For the latter, when the candidate recognition target When it is one, it can be the second recognition target in the automatic switching mode.
  • the processor 204 when the processor 204 executes the computer instructions, it also implements: based on the presence of at least one candidate photographic subject in the shooting frame, detecting the candidate recognition target of the at least one candidate photographic subject; outputting the at least one candidate recognition target ; The selected candidate recognition target is used as the first recognition target, and the candidate photographing object to which the first recognition target belongs is used as the photographing object.
  • the identification targets of these candidate subjects can be detected as candidate identification targets and output for the user to select, and then the candidate identification targets selected by the user and the equipment to which they belong Selecting the photographing object as the first recognition target and the photographing object respectively improves the efficiency of selecting the photographing object and its first recognition target in the initial stage.
  • the processor 204 executes computer instructions to output at least one candidate recognition target, including: outputting one candidate recognition target for each candidate photographed object.
  • the processor 204 executes the computer instruction to realize the output of a candidate recognition target, including: determining the number of candidate recognition targets corresponding to the candidate shooting object is one, outputting the corresponding candidate recognition target; determining the candidate The number of candidate recognition targets corresponding to the photographing object is at least two, and one of the corresponding candidate recognition targets is output.
  • the number of candidate recognition targets is one, it will be directly output, and if the number of candidate recognition targets is at least two, then one of them will be selected for output.
  • the one currently suitable for identification can be selected to ensure the identification effect.
  • the processor 204 executes the computer instruction to realize the output of one of the candidate recognition targets, including: detecting the size of all the corresponding candidate recognition targets in the shooting frame, and determining the output candidate recognition target according to the size .
  • the scheme of selecting one output from at least two candidate recognition targets is specifically limited.
  • the size of all candidate recognition targets of a subject in the shooting frame it can be an absolute size, such as width and height, or a relative size.
  • the proportion of the entire shooting frame can reflect the size of the candidate recognition target. The size, and then select the candidate recognition target whose size is suitable for recognition and output it, which helps to ensure the recognition effect.
  • processor 204 when the processor 204 executes computer instructions, it also implements: real-time output of image information.
  • the processor 204 can also output image information of the shooting screen in real time, so as to present the shooting object and composition changes to the user in real time, so that the user can understand the shooting situation or make adjustment operations.
  • the memory involved in the embodiments of the first aspect and the second aspect described above may include a large-capacity memory for data or instructions.
  • the storage may include hard disk drives (Hard Disk Drive, HDD), floppy disk drives, flash memory, optical disks, magneto-optical disks, magnetic tapes or Universal Serial Bus (USB) drives, or two or more A combination of more than one of these.
  • the storage may include removable or non-removable (or fixed) media.
  • the memory can be inside or outside the integrated gateway disaster recovery device.
  • the memory is a non-volatile solid state memory.
  • the memory includes read-only memory (ROM).
  • the ROM can be mask-programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), electrically rewritable ROM (EAROM) or flash memory or A combination of two or more of these.
  • CPU Central Processing Unit
  • ASIC Application Specific Integrated Circuit
  • the embodiment of the third aspect of the present application provides a method for controlling a pan-tilt, which is used in a pan-tilt or a control device of the pan-tilt.
  • Fig. 3 shows a schematic flowchart of a method for controlling a pan-tilt according to an embodiment of the present application.
  • the control method of the PTZ includes:
  • Step S102 receiving the image information of the shooting screen in real time, and there is a shooting object in the shooting screen;
  • Step S104 Detect composition information of the first recognition target in the shooting frame according to the image information, where the first recognition target is at least a part of the shooting object;
  • Step S106 Determine whether to generate a recognition target switching instruction according to the composition information of the first recognition target.
  • the method for controlling the pan/tilt head determines whether to generate the recognition target switching instruction according to the detected composition information of the first recognition target, that is, according to the composition information of the first recognition target of the subject in the current shooting frame .
  • a reasonable choice of whether to switch the recognition target not only improves the recognition success rate, but also helps to beautify the composition. It is understandable that the recognition targets before and after the switch all belong to the same shooting object.
  • this solution can be applied to the automatic follow-up function of the pan/tilt based on the auto-following technology, so as to switch and adjust the follow-up target in time, which can ensure the success of follow-up, but also make the composition beautiful and improve the follow-up effect;
  • the automatic follow-up function is used when the gimbal is aimed at the subject by hand. For example, when manually controlling the gimbal to follow the shot, it helps to reasonably adjust the recognition target, optimize the composition, and improve the shooting effect. Another example is to manually control the gimbal alignment. When the subject does not start shooting temporarily, it can be ensured that it is in a better recognition state and composition state, so that the shooting can be started at any time.
  • the composition information includes a size
  • the above S106 specifically includes: determining that the size of the first recognition target is outside the preset size range, and generating a recognition target switching instruction.
  • the composition information is specifically limited to include the size. Accordingly, the solution for determining whether to generate the recognition target switching instruction is that when the size of the first recognition target is outside the preset size range, the first recognition target is considered to be in The size of the shooting screen is unreasonable, and then a recognition target switching instruction is generated to switch to a recognition target with a more reasonable size, which improves the recognition success rate and helps to beautify the composition.
  • the recognition target switching instruction includes a recognition target expansion instruction
  • the above S106 further includes: determining that the size of the first recognition target is smaller than the lower limit size of the preset size range, and generating the recognition target expansion instruction.
  • the recognition target expands the instruction to switch to a larger recognition target.
  • a larger recognition target refers to a recognition target that accounts for a larger proportion of the photographed object than the first recognition target.
  • the first recognition target can be included in a larger recognition target.
  • the first recognition target is the head and shoulders of the photographing object, and the larger recognition target is the whole body of the photographing object; the first recognition target can also be completely identical to the larger recognition target.
  • the first recognition target is the head, and the larger recognition target is the torso; the first recognition target can also partially overlap with the larger recognition target, for example, the first recognition target is the head and shoulders, and the larger recognition target is the upper limbs. The overlapping part of the two is the shoulder.
  • the size of the first recognition target in the shooting screen gradually shrinks. When its size is too small, it is easy to be confused with the background, which reduces the recognition accuracy, which leads to recognition failure, which affects the composition effect and automatic follow-up.
  • the realization of the function By generating a recognition target expansion instruction, it can switch to a larger recognition target that is easy to recognize, which improves the recognition success rate, helps beautify the composition, and ensures the reliable operation of the automatic follow-up function.
  • the recognition target switching instruction includes a recognition target reduction instruction
  • the above S106 further includes: determining that the size of the first recognition target is greater than the upper limit size of the preset size range, and generating the recognition target reduction instruction.
  • the recognition target switching instruction may further include a recognition target reduction instruction, and is generated when it is detected that the size of the first recognition target is larger than the upper limit size, so as to switch to a smaller recognition target.
  • a recognition target reduction instruction is generated when it is detected that the size of the first recognition target is larger than the upper limit size, so as to switch to a smaller recognition target.
  • the relationship between the first recognition target and the smaller recognition target is the same as the aforementioned relationship between the larger recognition target and the first recognition target, and is a relative relationship.
  • the first recognition target is not fixed, but refers to It replaces the recognition target currently in use, which can be switched with the movement of the subject. As the subject gets closer, the size of the first recognition target in the shooting screen gradually expands.
  • the composition When its size is too large, if the composition continues to be performed according to the first recognition target, the aesthetics of the composition may be reduced, and the first recognition may even occur.
  • the target overflows the shooting screen and the recognition fails, at this time, by generating the recognition target reduction instruction, you can switch to a smaller recognition target in time, which not only helps to beautify the composition, but also solves the problem of the first recognition target overflowing the shooting screen.
  • the recognition success rate can be ensured to ensure the reliable operation of the automatic follow-up function.
  • Fig. 4 shows a schematic flow chart of a method for controlling a pan/tilt head according to another embodiment of the present application.
  • the control method of the PTZ includes:
  • Step S202 receiving the image information of the shooting screen in real time, and there is a shooting object in the shooting screen;
  • Step S204 Detect composition information of the first recognition target in the shooting frame according to the image information, where the first recognition target is at least a part of the shooting object;
  • Step S206 determining whether to generate a recognition target switching instruction according to the composition information of the first recognition target
  • Step S208 In response to the recognition target switching instruction, switch from the first recognition target to the second recognition target, the second recognition target is at least a part of the photographed object, the proportion of the first recognition target in the photographed object and the second recognition target in the photographed object The proportions vary.
  • the recognition target can also be switched from the first recognition target of the photographing object to the second recognition target of the photographing object based on the recognition target switching instruction, which realizes the automatic switching of the recognition target without manual operation by the user, which improves Convenience of operation and timeliness of switching.
  • the first recognition target and the second recognition target occupies different proportions in the shooting object, which can effectively adjust the recognition during shooting, which helps to improve the recognition success rate and beautify the composition.
  • the composition information further includes a position
  • the control method of the pan/tilt head further includes: adjusting shooting parameters according to the composition information of the first recognition target, so that the imaging of the first recognition target in the shooting frame meets the composition rules .
  • the composition information also includes the position. Based on the composition information and composition rules of the first recognition target in the shooting screen, it can be automatically adjusted as the first recognition target moves and changes in the shooting screen.
  • Shooting parameters play a role in adjusting and beautifying the composition. It is understandable that when the pan-tilt is aimed at the subject by hand, the shooting parameters can be automatically fine-tuned in combination with this solution to realize automatic composition.
  • switching from the first recognition target to the second recognition target in step S208 includes: detecting composition information of the second recognition target in the shooting frame; adjusting the shooting parameters so that the second recognition target is in the shooting frame The position and size of the image are gradually changed until the imaging of the second recognition target in the shooting frame meets the composition rules.
  • the recognition target switching instruction includes the aforementioned recognition target enlargement instruction and recognition target reduction instruction
  • the recognition target and composition can be automatically adjusted according to the distance between the subject and the pan/tilt, which not only improves the recognition success rate, but also Helps beautify the composition.
  • the composition rule includes at least one of the following: the imaging of the recognition target is complete, the recognition target is located in a designated area in the shooting frame, the balanced composition, the symmetrical composition, the diagonal composition, the triangular composition, the nine-square grid composition, Centripetal composition, bisection composition.
  • the content of the composition rule is specifically defined.
  • its imaging is complete and/or located in a designated area in the shooting frame as the composition rule to ensure the shooting effect of the subject;
  • the imaging of the entire shooting screen allows it to meet different composition forms, ensuring the overall beauty of the shooting screen.
  • the pan-tilt further includes: a stand device capable of supporting a photographing device for acquiring image information, and/or a photographing device for acquiring image information;
  • the photographing parameter includes at least one of the following: The height of the camera, the inclination of the camera, the direction of the camera, and the focal length of the camera.
  • the shooting parameters include at least one of the height, inclination, direction, and focal length of the shooting device.
  • the shooting device can be controlled to achieve shooting.
  • the photographing equipment can be the structure that comes with the PTZ or a third-party structure. For the latter, the PTZ needs to be equipped with support equipment to support the photographing equipment.
  • one of the first recognition target and the second recognition target is the whole of the photographed object, and the other of the first recognition target and the second recognition target is a designated part of the photographed object.
  • the whole of the subject can be regarded as a recognition target, and a specified part of the subject, such as head and shoulders, can be used as another recognition target, so as to use pedestrian detection technology, head and shoulder detection technology, or other recognition targets.
  • the detection technology and automatic follow technology of the designated part automatically switch the full/half-length composition according to the distance of the subject to the pan/tilt, so that the subject is always in the best composition position.
  • pedestrian detection and head and shoulder detection are technologies that use various sensors as input (such as a camera) to automatically detect the whole body, head and shoulders of the pedestrian in the field of view of the sensor, and obtain the position and size of the pedestrian in the screen.
  • Fig. 5 shows a schematic flow chart of a method for controlling a pan/tilt head according to still another embodiment of the present application.
  • the control method of the PTZ includes:
  • Step S302 receiving the image information of the shooting screen in real time, and there is a shooting object in the shooting screen;
  • Step S304 Detect composition information of the first recognition target in the shooting frame according to the image information, where the first recognition target is at least a part of the shooting object;
  • Step S306 Determine whether to generate a recognition target switching instruction according to the composition information of the first recognition target
  • Step S308 output the recognition target switching instruction.
  • the recognition target switching instruction can also be directly output to prompt the user to manually switch the recognition target, ensuring the user's autonomous control. It is understandable that the automatic switching mode in the embodiment shown in FIG. 4 and the manual switching mode in this embodiment can be configured for the pan/tilt at the same time, and the user can choose which one to use. In addition, if you need to switch the recognition target during video recording, you can pause the recording first, and then resume recording after manually selecting the recognition target again.
  • the method for controlling the pan/tilt head further includes: outputting at least one candidate recognition target; switching from the first recognition target to a selected one of the at least one candidate recognition target.
  • At least one reasonable candidate recognition target can be provided for the user to choose during manual switching, and the user does not need to manually select the specific part of the photographed object as the recognition target to be switched, which improves the switching speed.
  • the candidate recognition target can be multiple recognition targets corresponding to multiple different parts of the photographic object, or it can be at least one recognition target that is determined to be suitable for switching after screening. For the latter, when the candidate recognition target When it is one, it can be the second recognition target in the automatic switching mode.
  • Fig. 6 shows a schematic flowchart of a method for controlling a pan/tilt head according to another embodiment of the present application.
  • the control method of the PTZ includes:
  • Step S402 receiving the image information of the shooting screen in real time, and there is a shooting object in the shooting screen;
  • Step S404 Based on the situation that there is at least one candidate shooting object in the shooting picture, detecting a candidate recognition target of the at least one candidate shooting object;
  • Step S406 output at least one candidate recognition target
  • Step S408 taking the selected candidate recognition target as the first recognition target, and taking the candidate photographing object to which the first recognition target belongs as the photographing object;
  • Step S410 detecting the composition information of the first recognition target in the shooting frame according to the image information
  • Step S412 Determine whether to generate a recognition target switching instruction according to the composition information of the first recognition target.
  • the identification targets of these candidate subjects can be detected as candidate identification targets and output for the user to select, and then the candidate identification targets selected by the user and the equipment to which they belong Selecting the photographing object as the first recognition target and the photographing object respectively improves the efficiency of selecting the photographing object and its first recognition target in the initial stage.
  • step S406 specifically includes: outputting a candidate recognition target for each candidate shooting object.
  • outputting a candidate recognition target includes: determining the number of candidate recognition targets corresponding to the candidate photographed object is one, outputting the corresponding candidate recognition target; determining the candidate corresponding to the candidate photographing object The number of selected recognition targets is at least two, and one of the corresponding candidate recognition targets is output.
  • the number of candidate recognition targets is one, it will be directly output, and if the number of candidate recognition targets is at least two, then one of them will be selected for output.
  • the one currently suitable for identification can be selected to ensure the identification effect.
  • outputting one of the corresponding candidate recognition targets includes: detecting the size of all the corresponding candidate recognition targets in the shooting frame, and determining the output candidate recognition target according to the size.
  • the scheme of selecting one output from at least two candidate recognition targets is specifically limited.
  • the size of all candidate recognition targets of a subject in the shooting frame it can be an absolute size, such as width and height, or a relative size.
  • the proportion of the entire shooting frame can reflect the size of the candidate recognition target. The size, and then select the candidate recognition target whose size is suitable for recognition and output it, which helps to ensure the recognition effect.
  • the method for controlling the pan/tilt head further includes: outputting image information in real time.
  • the image information of the shooting screen can also be output in real time, so as to present the changes of the shooting object and composition to the user in real time, so that the user can understand the shooting situation or make adjustment operations.
  • the method for controlling the pan/tilt provided in this application is introduced through a specific embodiment.
  • the method is mainly based on pedestrian whole body, head and shoulder detection and automatic
  • the following technology involves the technical features in the above-mentioned multiple embodiments, and mainly consists of the following three steps:
  • Step 1 Smart suggestion prompt.
  • the pedestrian body, head and shoulders detection technology is used to detect the appropriate pedestrian body, head and shoulders in the picture, and then the detected pedestrian body, head and shoulders are previewed in real time on the mobile phone screen. Click to select the object to be tracked, thereby selecting the first recognition target and the shooting object.
  • the solution of this application will automatically select the appropriate one of them to display to the user for selection according to the proportion of the pedestrian's whole body and head and shoulders in the shooting screen.
  • Step 2 Automatically follow the switch and compose the picture.
  • the pan/tilt starts the automatic follow through the first recognition target selected by the user in step 1, and then continuously obtains the image information of the shooting screen, and calculates the position and size of the first recognition target in the shooting screen, and then controls the shooting parameters of the pan/tilt so that The first recognition target remains at the composition position of the shooting screen.
  • the recognition target will automatically switch to the pedestrian following the whole body and compose the picture accordingly, making the following distance farther and the composition more reasonable; when the subject keeps approaching the pan/tilt, the recognition target will automatically switch to Pedestrians head and shoulders, and use this composition to make the close-up follow more stable and the composition more reasonable.
  • Step 3 Real-time preview.
  • the pan-tilt control method can automatically switch the full (half) body of the subject to follow and compose the shot when the subject is far away or close to the pan-tilt, that is, far away or close to the shooting device, so that the following transition is more stable and the composition is more reasonable. It is understandable that the pan-tilt 100 provided in the embodiment of the first aspect of the present application and the pan-tilt control device 200 provided in the embodiment of the second aspect of the present application can implement the pan-tilt control method introduced in this specific embodiment. No longer.
  • the embodiment of the fourth aspect of the present application provides a computer-readable storage medium on which a computer program is stored.
  • the steps of the pan-tilt control method as in any of the above-mentioned embodiments are implemented, and thus have All the beneficial technical effects of the control method of the pan/tilt head will not be repeated here.
  • a computer-readable storage medium may include any medium capable of storing or transmitting information.
  • Examples of computer-readable storage media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and so on.
  • the code segment can be downloaded via a computer network such as the Internet, an intranet, and so on.
  • the term “plurality” refers to two or more than two, unless specifically defined otherwise.
  • the terms “installed”, “connected”, “connected”, “fixed” and other terms should be understood in a broad sense.
  • “connected” can be a fixed connection, a detachable connection, or an integral connection;
  • “connected” can be It is directly connected or indirectly connected through an intermediary.
  • the specific meanings of the above-mentioned terms in this application can be understood according to specific circumstances.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Studio Devices (AREA)

Abstract

一种云台、云台的控制设备、云台的控制方法及存储介质，其中，云台（100）包括：存储器(102)，被配置为存储计算机指令；及处理器(104)，被配置为执行计算机指令以实现：实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象(S102)；根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分(S104)；根据第一识别目标的构图信息确定是否生成识别目标切换指令(S106)。该云台通过根据检测到的第一识别目标的构图信息确定是否生成识别目标切换指令，也就是可以根据当前拍摄画面中拍摄对象的第一识别目标的构图信息，合理选择是否切换识别目标，既提升了识别成功率，又有助于美化构图。

Description

云台、云台的控制设备、云台的控制方法及存储介质 技术领域
本申请涉及电子技术应用领域，具体而言，涉及一种云台、一种云台的控制设备、一种云台的控制方法及一种计算机可读存储介质。
背景技术
一般用户在拍摄具体目标时，例如人物和动物，经常会出现的场景是使拍摄画面锁定在图像获取装置(比如摄像头)视野中的这个目标，然后持续对其进行拍摄。而一般这种持续对准目标进行拍摄可采用具有自动跟随的拍摄装置完成，例如云台。然而，不论是上述哪种运用场景，当拍摄对象相对图像获取装置发生远近变化时，往往会导致识别失败或者构图不美观等，例如当跟拍对象是人全身，当人走进相机时，相机画面中的人只剩下一部分，这时往往会识别失败并且构图不美观；当跟拍对象是人的头肩，当人远离相机时，人的头肩在相机画面就会变的很小并且容易和背景混淆，从而导致识别失败。
发明内容
本申请旨在至少解决现有技术或相关技术中存在的技术问题之一。
为此，本申请的第一方面提出了一种云台。
本申请的第二方面提出了一种云台的控制设备。
本申请的第三方面提出了一种云台的控制方法。
本申请的第四方面提出了一种计算机可读存储介质。
有鉴于此，根据本申请的第一方面，提供了一种云台，包括：存储器，被配置为存储计算机指令；及处理器，被配置为执行计算机指令以实现：实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；根据第一识别目标的构图信息确定是否生成识别目标切换指令。
根据本申请的第二方面，提供了一种云台的控制设备，包括：存储器，被配置为存储计算机指令；及处理器，被配置为执行计算机指令以实现：实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；根据第一识别目标的构图信息确定是否生成识别目标切换指令。
根据本申请的第三方面，提供了一种云台的控制方法，用于云台或云台的控制设备，云台的控制方法包括：实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；根据第一识别目标的构图信息确定是否生成识别目标切换指令。
根据本申请的第四方面，提供了一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现如上述任一技术方案的云台的控制方法的步骤。
本申请实施例提供了一种云台及其控制方案，通过根据检测到的第一识别目标的构图信息确定是否生成识别目标切换指令，也就是可以根据当前拍摄画面中拍摄对象的第一识别目标的构图信息，合理选择是否切换识别目标，既提升了识别成功率，又有助于美化构图。本方案可基于自动跟随技术，应用于云台的自动跟拍功能(即自动跟随拍摄功能)，以便及时切换、调整跟随目标，既可确保跟随成功，又可令构图美观，提升跟拍效果；也可在未启动自动跟拍功能而通过人手将云台对准拍摄对象时使用，有助于合理调节识别目标，优化构图，提升拍摄效果。
本申请的附加方面和优点将在下面的描述部分中变得明显，或通过本申请的实践了解到。
附图说明
本申请的上述和/或附加的方面和优点从结合下面附图对实施例的描述中将变得明显和容易理解，其中：
图1示出了本申请的一个实施例的云台的结构示意图；
图2示出了本申请的一个实施例的云台的控制设备的结构示意图；
图3示出了本申请的一个实施例的云台的控制方法的示意流程图；
图4示出了本申请的另一个实施例的云台的控制方法的示意流程图；
图5示出了本申请的再一个实施例的云台的控制方法的示意流程图；
图6示出了本申请的又一个实施例的云台的控制方法的示意流程图。
具体实施方式
为了能够更清楚地理解本申请的上述目的、特征和优点，下面结合附图和具体实施方式对本申请进行进一步的详细描述。需要说明的是，在不冲突的情况下，本申请的实施例及实施例中的特征可以相互组合。
在下面的描述中阐述了很多具体细节以便于充分理解本申请，但是，本申请还可以采用其他不同于在此描述的其他方式来实施，因此，本申请的保护范围并不受下面公开的具体实施例的限制。
如图1所示，本申请第一方面的实施例提供了一种云台100，包括：存储器102，被配置为存储计算机指令；及处理器104，被配置为执行计算机指令以实现：实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；根据第一识别目标的构图信息确定是否生成识别目标切换指令。
本申请实施例提供的云台100，通过配置处理器104，使其在执行存储器102存储的计算机指令时根据检测到的第一识别目标的构图信息确定是否生成识别目标切换指令，也就是可以根据当前拍摄画面中拍摄对象的第一识别目标的构图信息，合理选择是否切换识别目标，既提升了识别成功率，又有助于美化构图。可以理解的是，切换前后的识别目标均属于同一拍摄对象。此外，本方案可基于自动跟随技术，应用于云台100的自动跟拍功能，以便及时切换、调整跟随目标，既可确保跟随成功，又可令构图美观，提升跟拍效果；也可在未启动自动跟拍功能而通过人手将云台100对准拍摄对象时使用，例如手动控制云台100跟拍时，有助于合理调节识别目标，优化构图，提升拍摄效果，又如仅手动控制云台100对准拍摄对 象而暂不启动拍摄时，可确保处于较优的识别状态和构图状态，以便随时启动拍摄。
在一些实施例中，构图信息包括尺寸，处理器104执行计算机指令实现的根据第一识别目标的构图信息确定是否生成识别目标切换指令，包括：确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令。
在该实施例中，具体限定了构图信息包括尺寸，相应地，确定是否生成识别目标切换指令的方案为，当第一识别目标的尺寸处于预设尺寸范围之外，则认为第一识别目标在拍摄画面中的大小不合理，进而生成识别目标切换指令，以切换至尺寸大小更合理的识别目标，提升了识别成功率，并有助于美化构图。
在一些实施例中，识别目标切换指令包括识别目标扩大指令，处理器104执行计算机指令实现的确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令，包括：确定第一识别目标的尺寸小于预设尺寸范围的下限尺寸，生成识别目标扩大指令。
在该实施例中，进一步限定了当检测到第一识别目标的尺寸小于下限尺寸时，认为第一识别目标在拍摄画面中的尺寸过小，对应于拍摄对象远离云台100的情况，可相应生成识别目标扩大指令，以切换至更大的识别目标。可以理解的是，更大的识别目标是指相较于第一识别目标，占拍摄对象的比例更大的识别目标。第一识别目标可包含在更大的识别目标内，例如第一识别目标为拍摄对象的头肩，更大的识别目标为拍摄对象的全身；第一识别目标也可与更大的识别目标完全不同，例如第一识别目标为头，更大的识别目标为躯干；第一识别目标还可与更大的识别目标部分重叠，例如第一识别目标为头肩，更大的识别目标为上肢，二者的重叠部分即为肩。随着拍摄对象的远离，第一识别目标在拍摄画面内的尺寸逐渐缩小，当其尺寸过小时，容易和背景混淆，使得识别准确度下降，从而导致识别失败，进而影响构图效果和自动跟拍功能的实现。通过生成识别目标扩大指令，可及时切换至便于识别的更大的识别目标，提升了识别成功率，有助于美化构图，确保自动跟拍功能可靠运行。
在一些实施例中，识别目标切换指令包括识别目标缩小指令，处理器 104执行计算机指令实现的确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令，包括：确定第一识别目标的尺寸大于预设尺寸范围的上限尺寸，生成识别目标缩小指令。
在该实施例中，与识别目标扩大指令类似，识别目标切换指令还可包括识别目标缩小指令，并在检测到第一识别目标的尺寸大于上限尺寸时生成，以切换至更小的识别目标，对应于拍摄对象靠近云台100的情况。可以理解的是，第一识别目标与更小的识别目标的关系，与前述更大的识别目标与第一识别目标的关系相同，均属于相对关系，第一识别目标并非固定的，而是指代当前正在使用的识别目标，可随着拍摄对象的移动而切换变化。随着拍摄对象的靠近，第一识别目标在拍摄画面内的尺寸逐渐扩大，当其尺寸过大时，若继续根据第一识别目标执行构图，则可能会降低构图美观性，甚至出现第一识别目标溢出拍摄画面而识别失败的情况，此时通过生成识别目标缩小指令，可及时切换至更小的识别目标，既有助于美化构图，又解决了第一识别目标溢出拍摄画面的问题，提升了识别成功率，可确保自动跟拍功能可靠运行。
在一些实施例中，处理器104执行计算机指令时还实现：响应于识别目标切换指令，从第一识别目标切换至第二识别目标，第二识别目标为拍摄对象的至少一部分，第一识别目标占拍摄对象的比例和第二识别目标占拍摄对象的比例不等。
在该实施例中，处理器104还可基于识别目标切换指令，将识别目标由拍摄对象的第一识别目标切换至拍摄对象的第二识别目标，实现了识别目标的自动切换，无需用户手动操作，提升了操作的便利性和切换的及时性。第一识别目标和第二识别目标各自在拍摄对象中占据不同的比例，可对拍摄中的识别进行有效调整，有助于提升识别成功率，美化构图。
在一些实施例中，构图信息还包括位置，处理器104执行计算机指令时还实现：根据第一识别目标的构图信息调整拍摄参数，以使第一识别目标在拍摄画面中的成像满足构图规则。
在该实施例中，具体限定了构图信息还包括位置，基于第一识别目标的在拍摄画面中的构图信息和构图规则，处理器104可随着第一识别目标 在拍摄画面中的移动、变化而自动调整拍摄参数，起到调整、美化构图的作用。可以理解的是，当通过人手将云台100对准拍摄对象时，可结合本方案对拍摄参数进行自动微调，实现自动构图。
在一些实施例中，处理器104执行计算机指令实现的从第一识别目标切换至第二识别目标，包括：检测第二识别目标在拍摄画面中的构图信息；调整拍摄参数，以使第二识别目标在拍摄画面中的位置和尺寸逐渐变化，直到第二识别目标在拍摄画面中的成像满足构图规则。
在该实施例中，在自动切换识别目标的过程中，通过检测第二识别目标的构图信息，并结合构图规则自动调整拍摄参数，使得第二识别目标在拍摄画面中的位置和尺寸逐渐变化，直到第二识别目标在拍摄画面中的成像满足构图规则，可使得识别目标的切换过程平缓稳定，并基于第二识别目标构图，确保拍摄效果。可以理解的是，当识别目标切换指令包括上述识别目标扩大指令和识别目标缩小指令时，可随着拍摄对象与云台100之间的距离自动调节识别目标和构图，既提升了识别成功率，又有助于美化构图。
在一些实施例中，构图规则包括以下至少之一：识别目标的成像完整、识别目标位于拍摄画面中的指定区域内、均衡式构图、对称式构图、对角线构图、三角形构图、九宫格构图、向心式构图、对分式构图。
在该实施例中，具体限定了构图规则的内容，可针对当前的识别目标，以其成像完整和/或位于拍摄画面中的指定区域内作为构图规则，保证拍摄对象的拍摄效果；也可针对整个拍摄画面的成像，令其满足不同的构图形式，确保了拍摄画面的整体美观。
在一些实施例中，云台100还包括：支架设备，支架设备能够支撑用于获取图像信息的拍摄设备，和/或，拍摄设备，用于获取图像信息；拍摄参数包括以下至少之一：拍摄设备的高度、拍摄设备的倾角、拍摄设备的方向、拍摄设备的焦距。
在该实施例中，具体限定了拍摄参数包括拍摄设备的高度、倾角、方向以及焦距中的至少之一，通过调整拍摄参数，可控制拍摄设备以实现拍摄。拍摄设备可为云台100自带的结构，也可为第三方结构，对于后者， 云台100需配置支架设备以支撑拍摄设备。
在一些实施例中，第一识别目标和第二识别目标中的一个为拍摄对象的整体，第一识别目标和第二识别目标中的另一个为拍摄对象的指定部分。
在该实施例中，具体限定了可将拍摄对象的整体作为一个识别目标，并将拍摄对象的指定部分，例如头肩，作为另一个识别目标，从而利用行人检测技术、头肩检测技术或其他指定部分的检测技术、自动跟随技术，根据拍摄对象距离云台100远近变化，自动切换全/半身构图，使得拍摄对象始终处于最佳构图位置。其中，行人检测、头肩检测是以各种传感器为输入(比如摄像头)，自动检测在传感器视野中的行人全身、头肩，得到其在画面中的位置、所占大小的技术。
在一些实施例中，处理器104执行计算机指令时还实现：输出识别目标切换指令。
在该实施例中，处理器104还可直接输出识别目标切换指令，以提示用户手动操作切换识别目标，确保了用户的自主控制。可以理解的是，可为云台100同时配置自动切换模式和手动切换模式，由用户选择采用何者。此外，若是在录制视频的过程中需切换识别目标，则可先暂停录制，待重新手动选择识别目标后再继续录制。
在一些实施例中，处理器104执行计算机指令时还实现：输出至少一个备选识别目标；从第一识别目标切换至至少一个备选识别目标中被选中的一个。
在该实施例中，进一步限定了在手动切换时可提供至少一个合理的备选识别目标以供用户选择，用户不必手动框选拍摄对象的具体部位作为需切换的识别目标，既提升了切换速度，又避免了所选识别目标不便于识别的情况发生，有助于提升手动切换识别目标的效率和成功率。可以想到的是，备选识别目标可以为对应于拍摄对象的多个不同部位的多个识别目标，也可以为经过筛选确定适于切换的至少一个识别目标，对于后者，当备选识别目标为一个时，可为自动切换模式中的第二识别目标。
在一些实施例中，处理器104执行计算机指令时还实现：基于拍摄画面中存在至少一个备选拍摄对象的情况，检测至少一个备选拍摄对象的备 选识别目标；输出至少一个备选识别目标；将被选中的备选识别目标作为第一识别目标，将第一识别目标所属的备选拍摄对象作为拍摄对象。
在该实施例中，进一步限定了初始阶段如何确定拍摄对象及其第一识别目标。当拍摄画面中存在备选拍摄对象时，可检测这些备选拍摄对象的识别目标作为备选识别目标并将之输出，以供用户选择，进而将用户选中的备选识别目标及其所属的备选拍摄对象分别作为第一识别目标和拍摄对象，提升了初始阶段选择拍摄对象及其第一识别目标的效率。
在一些实施例中，处理器104执行计算机指令实现的输出至少一个备选识别目标，包括：针对每个备选拍摄对象，输出一个备选识别目标。
在该实施例中，具体限定了针对每个备选拍摄对象，仅输出一个备选识别目标，可合理降低用户选择时面对的选择量，便于用户决策，有助于进一步提升初始阶段选择拍摄对象及其第一识别目标的效率。
在一些实施例中，处理器104执行计算机指令实现的输出一个备选识别目标，包括：确定备选拍摄对象对应的备选识别目标的数量为一个，输出对应的备选识别目标；确定备选拍摄对象对应的备选识别目标的数量为至少两个，输出对应的其中一个备选识别目标。
在该实施例中，具体限定了对于一个备选拍摄对象，若备选识别目标的数量为一个，则直接将之输出，若备选识别目标的数量为至少两个，则选择其中一个输出，具体可选择当前适于识别的一个，以确保识别效果。
在一些实施例中，处理器104执行计算机指令实现的输出对应的其中一个备选识别目标，包括：检测对应的全部备选识别目标在拍摄画面中的尺寸，根据尺寸确定输出的备选识别目标。
在该实施例中，具体限定了从至少两个备选识别目标中选择一个输出的方案。通过检测一个拍摄对象的全部备选识别目标在拍摄画面中的尺寸，可为绝对尺寸，如宽度和高度，也可为相对尺寸，例如占整个拍摄画面的比例，可以反映出备选识别目标的大小，进而选择大小适于识别的备选识别目标加以输出，有助于确保识别效果。
在一些实施例中，处理器104执行计算机指令时还实现：实时输出图像信息。
在该实施例中，进一步限定了处理器104还可实时输出拍摄画面的图像信息，以将拍摄对象和构图的变化实时地呈现给用户，便于用户了解拍摄情况或做出调整操作。
在一些实施例中，云台100还包括：显示屏，用于显示图像信息。
在该实施例中，进一步限定了云台100还配置有显示屏，以便显示拍摄画面的图像信息，保证了输出的可靠性。
如图2所示，本申请第二方面的实施例提供了一种云台的控制设备200，包括：存储器202，被配置为存储计算机指令；及处理器204，被配置为执行计算机指令以实现：实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；根据第一识别目标的构图信息确定是否生成识别目标切换指令。
本申请实施例提供的云台的控制设备200，通过配置处理器204，使其在执行存储器202存储的计算机指令时根据检测到的第一识别目标的构图信息确定是否生成识别目标切换指令，也就是可以根据当前拍摄画面中拍摄对象的第一识别目标的构图信息，合理选择是否切换识别目标，既提升了识别成功率，又有助于美化构图。可以理解的是，切换前后的识别目标均属于同一拍摄对象。此外，本方案可基于自动跟随技术，应用于云台的自动跟拍功能，以便及时切换、调整跟随目标，既可确保跟随成功，又可令构图美观，提升跟拍效果；也可在未启动自动跟拍功能而通过人手将云台对准拍摄对象时使用，例如手动控制云台跟拍时，有助于合理调节识别目标，优化构图，提升拍摄效果，又如仅手动控制云台对准拍摄对象而暂不启动拍摄时，可确保处于较优的识别状态和构图状态，以便随时启动拍摄。
在一些实施例中，构图信息包括尺寸，处理器204执行计算机指令实现的根据第一识别目标的构图信息确定是否生成识别目标切换指令，包括：确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令。
在该实施例中，具体限定了构图信息包括尺寸，相应地，确定是否生成识别目标切换指令的方案为，当第一识别目标的尺寸处于预设尺寸范围 之外，则认为第一识别目标在拍摄画面中的大小不合理，进而生成识别目标切换指令，以切换至尺寸大小更合理的识别目标，提升了识别成功率，并有助于美化构图。
在一些实施例中，识别目标切换指令包括识别目标扩大指令，处理器204执行计算机指令实现的确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令，包括：确定第一识别目标的尺寸小于预设尺寸范围的下限尺寸，生成识别目标扩大指令。
在该实施例中，进一步限定了当检测到第一识别目标的尺寸小于下限尺寸时，认为第一识别目标在拍摄画面中的尺寸过小，对应于拍摄对象远离云台的情况，可相应生成识别目标扩大指令，以切换至更大的识别目标。可以理解的是，更大的识别目标是指相较于第一识别目标，占拍摄对象的比例更大的识别目标。第一识别目标可包含在更大的识别目标内，例如第一识别目标为拍摄对象的头肩，更大的识别目标为拍摄对象的全身；第一识别目标也可与更大的识别目标完全不同，例如第一识别目标为头，更大的识别目标为躯干；第一识别目标还可与更大的识别目标部分重叠，例如第一识别目标为头肩，更大的识别目标为上肢，二者的重叠部分即为肩。随着拍摄对象的远离，第一识别目标在拍摄画面内的尺寸逐渐缩小，当其尺寸过小时，容易和背景混淆，使得识别准确度下降，从而导致识别失败，进而影响构图效果和自动跟拍功能的实现。通过生成识别目标扩大指令，可及时切换至便于识别的更大的识别目标，提升了识别成功率，有助于美化构图，确保自动跟拍功能可靠运行。
在一些实施例中，识别目标切换指令包括识别目标缩小指令，处理器204执行计算机指令实现的确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令，包括：确定第一识别目标的尺寸大于预设尺寸范围的上限尺寸，生成识别目标缩小指令。
在该实施例中，与识别目标扩大指令类似，识别目标切换指令还可包括识别目标缩小指令，并在检测到第一识别目标的尺寸大于上限尺寸时生成，以切换至更小的识别目标，对应于拍摄对象靠近云台的情况。可以理解的是，第一识别目标与更小的识别目标的关系，与前述更大的识别目标 与第一识别目标的关系相同，均属于相对关系，第一识别目标并非固定的，而是指代当前正在使用的识别目标，可随着拍摄对象的移动而切换变化。随着拍摄对象的靠近，第一识别目标在拍摄画面内的尺寸逐渐扩大，当其尺寸过大时，若继续根据第一识别目标执行构图，则可能会降低构图美观性，甚至出现第一识别目标溢出拍摄画面而识别失败的情况，此时通过生成识别目标缩小指令，可及时切换至更小的识别目标，既有助于美化构图，又解决了第一识别目标溢出拍摄画面的问题，提升了识别成功率，可确保自动跟拍功能可靠运行。
在一些实施例中，处理器204执行计算机指令时还实现：响应于识别目标切换指令，从第一识别目标切换至第二识别目标，第二识别目标为拍摄对象的至少一部分，第一识别目标占拍摄对象的比例和第二识别目标占拍摄对象的比例不等。
在该实施例中，处理器204还可基于识别目标切换指令，将识别目标由拍摄对象的第一识别目标切换至拍摄对象的第二识别目标，实现了识别目标的自动切换，无需用户手动操作，提升了操作的便利性和切换的及时性。第一识别目标和第二识别目标各自在拍摄对象中占据不同的比例，可对拍摄中的识别进行有效调整，有助于提升识别成功率，美化构图。
在一些实施例中，构图信息还包括位置，处理器204执行计算机指令时还实现：根据第一识别目标的构图信息调整拍摄参数，以使第一识别目标在拍摄画面中的成像满足构图规则。
在该实施例中，具体限定了构图信息还包括位置，基于第一识别目标的在拍摄画面中的构图信息和构图规则，处理器204可随着第一识别目标在拍摄画面中的移动、变化而自动调整拍摄参数，起到调整、美化构图的作用。可以理解的是，当通过人手将云台对准拍摄对象时，可结合本方案对拍摄参数进行自动微调，实现自动构图。
在一些实施例中，处理器204执行计算机指令实现的从第一识别目标切换至第二识别目标，包括：检测第二识别目标在拍摄画面中的构图信息；调整拍摄参数，以使第二识别目标在拍摄画面中的位置和尺寸逐渐变化，直到第二识别目标在拍摄画面中的成像满足构图规则。
在该实施例中，在自动切换识别目标的过程中，通过检测第二识别目标的构图信息，并结合构图规则自动调整拍摄参数，使得第二识别目标在拍摄画面中的位置和尺寸逐渐变化，直到第二识别目标在拍摄画面中的成像满足构图规则，可使得识别目标的切换过程平缓稳定，并基于第二识别目标构图，确保拍摄效果。可以理解的是，当识别目标切换指令包括上述识别目标扩大指令和识别目标缩小指令时，可随着拍摄对象与云台之间的距离自动调节识别目标和构图，既提升了识别成功率，又有助于美化构图。
在一些实施例中，构图规则包括以下至少之一：识别目标的成像完整、识别目标位于拍摄画面中的指定区域内、均衡式构图、对称式构图、对角线构图、三角形构图、九宫格构图、向心式构图、对分式构图。
在该实施例中，具体限定了构图规则的内容，可针对当前的识别目标，以其成像完整和/或位于拍摄画面中的指定区域内作为构图规则，保证拍摄对象的拍摄效果；也可针对整个拍摄画面的成像，令其满足不同的构图形式，确保了拍摄画面的整体美观。
在一些实施例中，云台还包括：支架设备，支架设备能够支撑用于获取图像信息的拍摄设备，和/或，拍摄设备，用于获取图像信息；拍摄参数包括以下至少之一：拍摄设备的高度、拍摄设备的倾角、拍摄设备的方向、拍摄设备的焦距。
在该实施例中，具体限定了拍摄参数包括拍摄设备的高度、倾角、方向以及焦距中的至少之一，通过调整拍摄参数，可控制拍摄设备以实现拍摄。拍摄设备可为云台自带的结构，也可为第三方结构，对于后者，云台需配置支架设备以支撑拍摄设备。
在一些实施例中，第一识别目标和第二识别目标中的一个为拍摄对象的整体，第一识别目标和第二识别目标中的另一个为拍摄对象的指定部分。
在该实施例中，具体限定了可将拍摄对象的整体作为一个识别目标，并将拍摄对象的指定部分，例如头肩，作为另一个识别目标，从而利用行人检测技术、头肩检测技术或其他指定部分的检测技术、自动跟随技术，根据拍摄对象距离云台远近变化，自动切换全/半身构图，使得拍摄对象始终处于最佳构图位置。其中，行人检测、头肩检测是以各种传感器为输入 (比如摄像头)，自动检测在传感器视野中的行人全身、头肩，得到其在画面中的位置、所占大小的技术。
在一些实施例中，处理器204执行计算机指令时还实现：输出识别目标切换指令。
在该实施例中，处理器204还可直接输出识别目标切换指令，以提示用户手动操作切换识别目标，确保了用户的自主控制。可以理解的是，可为云台同时配置自动切换模式和手动切换模式，由用户选择采用何者。此外，若是在录制视频的过程中需切换识别目标，则可先暂停录制，待重新手动选择识别目标后再继续录制。
在一些实施例中，处理器204执行计算机指令时还实现：输出至少一个备选识别目标；从第一识别目标切换至至少一个备选识别目标中被选中的一个。
在该实施例中，进一步限定了在手动切换时可提供至少一个合理的备选识别目标以供用户选择，用户不必手动框选拍摄对象的具体部位作为需切换的识别目标，既提升了切换速度，又避免了所选识别目标不便于识别的情况发生，有助于提升手动切换识别目标的效率和成功率。可以想到的是，备选识别目标可以为对应于拍摄对象的多个不同部位的多个识别目标，也可以为经过筛选确定适于切换的至少一个识别目标，对于后者，当备选识别目标为一个时，可为自动切换模式中的第二识别目标。
在一些实施例中，处理器204执行计算机指令时还实现：基于拍摄画面中存在至少一个备选拍摄对象的情况，检测至少一个备选拍摄对象的备选识别目标；输出至少一个备选识别目标；将被选中的备选识别目标作为第一识别目标，将第一识别目标所属的备选拍摄对象作为拍摄对象。
在该实施例中，进一步限定了初始阶段如何确定拍摄对象及其第一识别目标。当拍摄画面中存在备选拍摄对象时，可检测这些备选拍摄对象的识别目标作为备选识别目标并将之输出，以供用户选择，进而将用户选中的备选识别目标及其所属的备选拍摄对象分别作为第一识别目标和拍摄对象，提升了初始阶段选择拍摄对象及其第一识别目标的效率。
在一些实施例中，处理器204执行计算机指令实现的输出至少一个备 选识别目标，包括：针对每个备选拍摄对象，输出一个备选识别目标。
在该实施例中，具体限定了针对每个备选拍摄对象，仅输出一个备选识别目标，可合理降低用户选择时面对的选择量，便于用户决策，有助于进一步提升初始阶段选择拍摄对象及其第一识别目标的效率。
在一些实施例中，处理器204执行计算机指令实现的输出一个备选识别目标，包括：确定备选拍摄对象对应的备选识别目标的数量为一个，输出对应的备选识别目标；确定备选拍摄对象对应的备选识别目标的数量为至少两个，输出对应的其中一个备选识别目标。
在该实施例中，具体限定了对于一个备选拍摄对象，若备选识别目标的数量为一个，则直接将之输出，若备选识别目标的数量为至少两个，则选择其中一个输出，具体可选择当前适于识别的一个，以确保识别效果。
在一些实施例中，处理器204执行计算机指令实现的输出对应的其中一个备选识别目标，包括：检测对应的全部备选识别目标在拍摄画面中的尺寸，根据尺寸确定输出的备选识别目标。
在该实施例中，具体限定了从至少两个备选识别目标中选择一个输出的方案。通过检测一个拍摄对象的全部备选识别目标在拍摄画面中的尺寸，可为绝对尺寸，如宽度和高度，也可为相对尺寸，例如占整个拍摄画面的比例，可以反映出备选识别目标的大小，进而选择大小适于识别的备选识别目标加以输出，有助于确保识别效果。
在一些实施例中，处理器204执行计算机指令时还实现：实时输出图像信息。
在该实施例中，进一步限定了处理器204还可实时输出拍摄画面的图像信息，以将拍摄对象和构图的变化实时地呈现给用户，便于用户了解拍摄情况或做出调整操作。
具体地，上述第一方面和第二方面的实施例涉及的存储器可以包括用于数据或指令的大容量存储器。举例来说而非限制，存储器可包括硬盘驱动器(Hard Disk Drive，HDD)、软盘驱动器、闪存、光盘、磁光盘、磁带或通用串行总线(Universal Serial Bus，USB)驱动器或者两个或更多个以上这些的组合。在合适的情况下，存储器可包括可移除或不可移除(或 固定)的介质。在合适的情况下，存储器可在综合网关容灾设备的内部或外部。在特定实施例中，存储器是非易失性固态存储器。在特定实施例中，存储器包括只读存储器(ROM)。在合适的情况下，该ROM可以是掩模编程的ROM、可编程ROM(PROM)、可擦除PROM(EPROM)、电可擦除PROM(EEPROM)、电可改写ROM(EAROM)或闪存或者两个或更多个以上这些的组合。
上述第一方面和第二方面的实施例涉及的处理器(CPU)，或者特定集成电路(Application Specific Integrated Circuit，ASIC)，或者可以被配置成实施本申请实施例的一个或多个集成电路。
本申请第三方面的实施例提供了一种云台的控制方法，用于云台或云台的控制设备。
图3示出了本申请的一个实施例的云台的控制方法的示意流程图。如图3所示，该云台的控制方法包括：
步骤S102，实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；
步骤S104，根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；
步骤S106，根据第一识别目标的构图信息确定是否生成识别目标切换指令。
本申请实施例提供的云台的控制方法，通过根据检测到的第一识别目标的构图信息确定是否生成识别目标切换指令，也就是可以根据当前拍摄画面中拍摄对象的第一识别目标的构图信息，合理选择是否切换识别目标，既提升了识别成功率，又有助于美化构图。可以理解的是，切换前后的识别目标均属于同一拍摄对象。此外，本方案可基于自动跟随技术，应用于云台的自动跟拍功能，以便及时切换、调整跟随目标，既可确保跟随成功，又可令构图美观，提升跟拍效果；也可在未启动自动跟拍功能而通过人手将云台对准拍摄对象时使用，例如手动控制云台跟拍时，有助于合理调节识别目标，优化构图，提升拍摄效果，又如仅手动控制云台对准拍摄对象而暂不启动拍摄时，可确保处于较优的识别状态和构图状态，以便随时启动拍摄。
在一些实施例中，构图信息包括尺寸，上述S106具体包括：确定第一识别目标的尺寸处于预设尺寸范围之外，生成识别目标切换指令。
在该实施例中，具体限定了构图信息包括尺寸，相应地，确定是否生成识别目标切换指令的方案为，当第一识别目标的尺寸处于预设尺寸范围之外，则认为第一识别目标在拍摄画面中的大小不合理，进而生成识别目标切换指令，以切换至尺寸大小更合理的识别目标，提升了识别成功率，并有助于美化构图。
在一些实施例中，识别目标切换指令包括识别目标扩大指令，上述S106进一步包括：确定第一识别目标的尺寸小于预设尺寸范围的下限尺寸，生成识别目标扩大指令。
在该实施例中，进一步限定了当检测到第一识别目标的尺寸小于下限尺寸时，认为第一识别目标在拍摄画面中的尺寸过小，对应于拍摄对象远离云台的情况，可相应生成识别目标扩大指令，以切换至更大的识别目标。可以理解的是，更大的识别目标是指相较于第一识别目标，占拍摄对象的比例更大的识别目标。第一识别目标可包含在更大的识别目标内，例如第一识别目标为拍摄对象的头肩，更大的识别目标为拍摄对象的全身；第一识别目标也可与更大的识别目标完全不同，例如第一识别目标为头，更大的识别目标为躯干；第一识别目标还可与更大的识别目标部分重叠，例如第一识别目标为头肩，更大的识别目标为上肢，二者的重叠部分即为肩。随着拍摄对象的远离，第一识别目标在拍摄画面内的尺寸逐渐缩小，当其尺寸过小时，容易和背景混淆，使得识别准确度下降，从而导致识别失败，进而影响构图效果和自动跟拍功能的实现。通过生成识别目标扩大指令，可及时切换至便于识别的更大的识别目标，提升了识别成功率，有助于美化构图，确保自动跟拍功能可靠运行。
在一些实施例中，识别目标切换指令包括识别目标缩小指令，上述S106进一步包括：确定第一识别目标的尺寸大于预设尺寸范围的上限尺寸，生成识别目标缩小指令。
在该实施例中，与识别目标扩大指令类似，识别目标切换指令还可包括识别目标缩小指令，并在检测到第一识别目标的尺寸大于上限尺寸时生 成，以切换至更小的识别目标，对应于拍摄对象靠近云台的情况。可以理解的是，第一识别目标与更小的识别目标的关系，与前述更大的识别目标与第一识别目标的关系相同，均属于相对关系，第一识别目标并非固定的，而是指代当前正在使用的识别目标，可随着拍摄对象的移动而切换变化。随着拍摄对象的靠近，第一识别目标在拍摄画面内的尺寸逐渐扩大，当其尺寸过大时，若继续根据第一识别目标执行构图，则可能会降低构图美观性，甚至出现第一识别目标溢出拍摄画面而识别失败的情况，此时通过生成识别目标缩小指令，可及时切换至更小的识别目标，既有助于美化构图，又解决了第一识别目标溢出拍摄画面的问题，提升了识别成功率，可确保自动跟拍功能可靠运行。
图4示出了本申请的另一个实施例的云台的控制方法的示意流程图。如图4所示，该云台的控制方法包括：
步骤S202，实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；
步骤S204，根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；
步骤S206，根据第一识别目标的构图信息确定是否生成识别目标切换指令；
步骤S208，响应于识别目标切换指令，从第一识别目标切换至第二识别目标，第二识别目标为拍摄对象的至少一部分，第一识别目标占拍摄对象的比例和第二识别目标占拍摄对象的比例不等。
在该实施例中，还可基于识别目标切换指令，将识别目标由拍摄对象的第一识别目标切换至拍摄对象的第二识别目标，实现了识别目标的自动切换，无需用户手动操作，提升了操作的便利性和切换的及时性。第一识别目标和第二识别目标各自在拍摄对象中占据不同的比例，可对拍摄中的识别进行有效调整，有助于提升识别成功率，美化构图。
进一步地，在一些实施例中，构图信息还包括位置，云台的控制方法还包括：根据第一识别目标的构图信息调整拍摄参数，以使第一识别目标在拍摄画面中的成像满足构图规则。
在该实施例中，具体限定了构图信息还包括位置，基于第一识别目标 的在拍摄画面中的构图信息和构图规则，可随着第一识别目标在拍摄画面中的移动、变化而自动调整拍摄参数，起到调整、美化构图的作用。可以理解的是，当通过人手将云台对准拍摄对象时，可结合本方案对拍摄参数进行自动微调，实现自动构图。
在一些实施例中，步骤S208的从第一识别目标切换至第二识别目标，包括：检测第二识别目标在拍摄画面中的构图信息；调整拍摄参数，以使第二识别目标在拍摄画面中的位置和尺寸逐渐变化，直到第二识别目标在拍摄画面中的成像满足构图规则。
在该实施例中，在自动切换识别目标的过程中，通过检测第二识别目标的构图信息，并结合构图规则自动调整拍摄参数，使得第二识别目标在拍摄画面中的位置和尺寸逐渐变化，直到第二识别目标在拍摄画面中的成像满足构图规则，可使得识别目标的切换过程平缓稳定，并基于第二识别目标构图，确保拍摄效果。可以理解的是，当识别目标切换指令包括上述识别目标扩大指令和识别目标缩小指令时，可随着拍摄对象与云台之间的距离自动调节识别目标和构图，既提升了识别成功率，又有助于美化构图。
在一些实施例中，构图规则包括以下至少之一：识别目标的成像完整、识别目标位于拍摄画面中的指定区域内、均衡式构图、对称式构图、对角线构图、三角形构图、九宫格构图、向心式构图、对分式构图。
在该实施例中，具体限定了构图规则的内容，可针对当前的识别目标，以其成像完整和/或位于拍摄画面中的指定区域内作为构图规则，保证拍摄对象的拍摄效果；也可针对整个拍摄画面的成像，令其满足不同的构图形式，确保了拍摄画面的整体美观。
在一些实施例中，云台还包括：支架设备，支架设备能够支撑用于获取图像信息的拍摄设备，和/或，拍摄设备，用于获取图像信息；拍摄参数包括以下至少之一：拍摄设备的高度、拍摄设备的倾角、拍摄设备的方向、拍摄设备的焦距。
在该实施例中，具体限定了拍摄参数包括拍摄设备的高度、倾角、方向以及焦距中的至少之一，通过调整拍摄参数，可控制拍摄设备以实现拍摄。拍摄设备可为云台自带的结构，也可为第三方结构，对于后者，云台 需配置支架设备以支撑拍摄设备。
在一些实施例中，第一识别目标和第二识别目标中的一个为拍摄对象的整体，第一识别目标和第二识别目标中的另一个为拍摄对象的指定部分。
在该实施例中，具体限定了可将拍摄对象的整体作为一个识别目标，并将拍摄对象的指定部分，例如头肩，作为另一个识别目标，从而利用行人检测技术、头肩检测技术或其他指定部分的检测技术、自动跟随技术，根据拍摄对象距离云台远近变化，自动切换全/半身构图，使得拍摄对象始终处于最佳构图位置。其中，行人检测、头肩检测是以各种传感器为输入(比如摄像头)，自动检测在传感器视野中的行人全身、头肩，得到其在画面中的位置、所占大小的技术。
图5示出了本申请的再一个实施例的云台的控制方法的示意流程图。如图5所示，该云台的控制方法包括：
步骤S302，实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；
步骤S304，根据图像信息检测第一识别目标在拍摄画面中的构图信息，第一识别目标为拍摄对象的至少一部分；
步骤S306，根据第一识别目标的构图信息确定是否生成识别目标切换指令；
步骤S308，输出识别目标切换指令。
在该实施例中，还可直接输出识别目标切换指令，以提示用户手动操作切换识别目标，确保了用户的自主控制。可以理解的是，可为云台同时配置图4所示的实施例中的自动切换模式和本实施例中的手动切换模式，由用户选择采用何者。此外，若是在录制视频的过程中需切换识别目标，则可先暂停录制，待重新手动选择识别目标后再继续录制。
进一步地，在一些实施例中，云台的控制方法还包括：输出至少一个备选识别目标；从第一识别目标切换至至少一个备选识别目标中被选中的一个。
在该实施例中，进一步限定了在手动切换时可提供至少一个合理的备选识别目标以供用户选择，用户不必手动框选拍摄对象的具体部位作为需切换的识别目标，既提升了切换速度，又避免了所选识别目标不便于识别 的情况发生，有助于提升手动切换识别目标的效率和成功率。可以想到的是，备选识别目标可以为对应于拍摄对象的多个不同部位的多个识别目标，也可以为经过筛选确定适于切换的至少一个识别目标，对于后者，当备选识别目标为一个时，可为自动切换模式中的第二识别目标。
图6示出了本申请的又一个实施例的云台的控制方法的示意流程图。如图6所示，该云台的控制方法包括：
步骤S402，实时接收拍摄画面的图像信息，拍摄画面内存在拍摄对象；
步骤S404，基于拍摄画面中存在至少一个备选拍摄对象的情况，检测至少一个备选拍摄对象的备选识别目标；
步骤S406，输出至少一个备选识别目标；
步骤S408，将被选中的备选识别目标作为第一识别目标，将第一识别目标所属的备选拍摄对象作为拍摄对象；
步骤S410，根据图像信息检测第一识别目标在拍摄画面中的构图信息；
步骤S412，根据第一识别目标的构图信息确定是否生成识别目标切换指令。
在该实施例中，进一步限定了初始阶段如何确定拍摄对象及其第一识别目标。当拍摄画面中存在备选拍摄对象时，可检测这些备选拍摄对象的识别目标作为备选识别目标并将之输出，以供用户选择，进而将用户选中的备选识别目标及其所属的备选拍摄对象分别作为第一识别目标和拍摄对象，提升了初始阶段选择拍摄对象及其第一识别目标的效率。
在一些实施例中，步骤S406具体包括：针对每个备选拍摄对象，输出一个备选识别目标。
在该实施例中，具体限定了针对每个备选拍摄对象，仅输出一个备选识别目标，可合理降低用户选择时面对的选择量，便于用户决策，有助于进一步提升初始阶段选择拍摄对象及其第一识别目标的效率。
进一步地，在一些实施例中，输出一个备选识别目标，包括：确定备选拍摄对象对应的备选识别目标的数量为一个，输出对应的备选识别目标；确定备选拍摄对象对应的备选识别目标的数量为至少两个，输出对应的其中一个备选识别目标。
在该实施例中，具体限定了对于一个备选拍摄对象，若备选识别目标的数量为一个，则直接将之输出，若备选识别目标的数量为至少两个，则选择其中一个输出，具体可选择当前适于识别的一个，以确保识别效果。
进一步地，在一些实施例中，输出对应的其中一个备选识别目标，包括：检测对应的全部备选识别目标在拍摄画面中的尺寸，根据尺寸确定输出的备选识别目标。
在该实施例中，具体限定了从至少两个备选识别目标中选择一个输出的方案。通过检测一个拍摄对象的全部备选识别目标在拍摄画面中的尺寸，可为绝对尺寸，如宽度和高度，也可为相对尺寸，例如占整个拍摄画面的比例，可以反映出备选识别目标的大小，进而选择大小适于识别的备选识别目标加以输出，有助于确保识别效果。
在一些实施例中，云台的控制方法还包括：实时输出图像信息。
在该实施例中，进一步限定了还可实时输出拍摄画面的图像信息，以将拍摄对象和构图的变化实时地呈现给用户，便于用户了解拍摄情况或做出调整操作。
接下来以应用于云台的自动跟拍功能、拍摄设备为用户的手机为例，通过一个具体实施例介绍本申请提供的云台的控制方法，该方法主要基于行人全身、头肩检测和自动跟随技术，涉及了上述多个实施例中的技术特征，主要由以下三个步骤组成：
步骤一：智能建议提示。
当用户开启自动跟拍功能的时候，利用行人全身、头肩检测技术检测出画面中合适的行人全身、头肩，然后将检测到的行人全身、头肩实时预览到手机画面当中，用户可以通过点击来选择需要跟踪的对象，从而选定第一识别目标及拍摄对象。
当同一个人既检测到全身又检测到他的头肩的时候，本申请方案会根据行人全身、头肩在拍摄画面中的比例，自动选择其中合适的一个显示给用户来选择。
步骤二：自动跟随切换并构图。
云台通过步骤一中用户选择的第一识别目标来启动自动跟随，再不间 断获取拍摄画面的图像信息，并计算第一识别目标在拍摄画面中的位置和大小，随后控制云台的拍摄参数使得第一识别目标保持在拍摄画面的构图位置。当拍摄对象远离云台时，识别目标会自动平缓切换为行人全身的跟随并且以此构图，使得跟随距离更远，构图更加合理；当拍摄对象不断走近云台时，识别目标会自动切换为行人头肩，并且以此构图，使得近距离跟随更加稳定，构图更加合理。
步骤三：实时预览。
拍摄对象和构图的变化都会实时地呈现给用户。
该云台的控制方法可以在拍摄对象远离、靠近云台，即远离、靠近拍摄设备时，自动切换拍摄对象的全(半)身来跟随和构图拍摄，使得跟随过渡更加平稳，构图更加合理。可以理解的是，本申请第一方面的实施例提供的云台100和第二方面的实施例提供的云台的控制设备200，可实现该具体实施例介绍的云台的控制方法，在此不再赘述。
本申请第四方面的实施例提供了一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现如上述任一实施例的云台的控制方法的步骤，因而具备该云台的控制方法的全部有益技术效果，在此不再赘述。
具体地，计算机可读存储介质可以包括能够存储或传输信息的任何介质。计算机可读存储介质的例子包括电子电路、半导体存储器设备、ROM、闪存、可擦除ROM(EROM)、软盘、CD-ROM、光盘、硬盘、光纤介质、射频(RF)链路，等等。代码段可以经由诸如因特网、内联网等的计算机网络被下载。
在本申请中，术语“多个”则指两个或两个以上，除非另有明确的限定。术语“安装”、“相连”、“连接”、“固定”等术语均应做广义理解，例如，“连接”可以是固定连接，也可以是可拆卸连接，或一体地连接；“相连”可以是直接相连，也可以通过中间媒介间接相连。对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本申请中的具体含义。
在本说明书的描述中，术语“一个实施例”、“一些实施例”、“具体实施例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或特点包含于本申请的至少一个实施例或示例中。在本说明书中，对上述术语的 示意性表述不一定指的是相同的实施例或实例。而且，描述的具体特征、结构、材料或特点可以在任何的一个或多个实施例或示例中以合适的方式结合。
以上所述仅为本申请的优选实施例而已，并不用于限制本申请，对于本领域的技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。

Claims (53)

  1. 一种云台，其特征在于，包括：
    存储器，被配置为存储计算机指令；及
    处理器，被配置为执行所述计算机指令以实现：
    实时接收拍摄画面的图像信息，所述拍摄画面内存在拍摄对象；
    根据所述图像信息检测第一识别目标在所述拍摄画面中的构图信息，所述第一识别目标为所述拍摄对象的至少一部分；
    根据所述第一识别目标的构图信息确定是否生成识别目标切换指令。
  2. 根据权利要求1所述的云台，其特征在于，所述构图信息包括尺寸，所述处理器执行所述计算机指令实现的根据所述第一识别目标的构图信息确定是否生成识别目标切换指令，包括：
    确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令。
  3. 根据权利要求2所述的云台，其特征在于，所述识别目标切换指令包括识别目标扩大指令，所述处理器执行所述计算机指令实现的确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令，包括：
    确定所述第一识别目标的尺寸小于所述预设尺寸范围的下限尺寸，生成所述识别目标扩大指令。
  4. 根据权利要求2所述的云台，其特征在于，所述识别目标切换指令包括识别目标缩小指令，所述处理器执行所述计算机指令实现的确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令，包括：
    确定所述第一识别目标的尺寸大于所述预设尺寸范围的上限尺寸，生成所述识别目标缩小指令。
  5. 根据权利要求1至4中任一项所述的云台，其特征在于，所述处理器执行所述计算机指令时还实现：
    响应于所述识别目标切换指令，从所述第一识别目标切换至第二识别 目标，所述第二识别目标为所述拍摄对象的至少一部分，所述第一识别目标占所述拍摄对象的比例和所述第二识别目标占所述拍摄对象的比例不等。
  6. 根据权利要求5所述的云台，其特征在于，所述构图信息还包括位置，所述处理器执行所述计算机指令时还实现：
    根据所述第一识别目标的构图信息调整拍摄参数，以使所述第一识别目标在所述拍摄画面中的成像满足构图规则。
  7. 根据权利要求6所述的云台，其特征在于，所述处理器执行所述计算机指令实现的从所述第一识别目标切换至第二识别目标，包括：
    检测所述第二识别目标在所述拍摄画面中的构图信息；
    调整所述拍摄参数，以使所述第二识别目标在所述拍摄画面中的位置和尺寸逐渐变化，直到所述第二识别目标在所述拍摄画面中的成像满足所述构图规则。
  8. 根据权利要求6或7所述的云台，其特征在于，
    所述构图规则包括以下至少之一：识别目标的成像完整、识别目标位于所述拍摄画面中的指定区域内、均衡式构图、对称式构图、对角线构图、三角形构图、九宫格构图、向心式构图、对分式构图。
  9. 根据权利要求6或7所述的云台，其特征在于，
    所述云台还包括：
    支架设备，所述支架设备能够支撑用于获取所述图像信息的拍摄设备，
    和/或，拍摄设备，用于获取所述图像信息；
    所述拍摄参数包括以下至少之一：所述拍摄设备的高度、所述拍摄设备的倾角、所述拍摄设备的方向、所述拍摄设备的焦距。
  10. 根据权利要求5所述的云台，其特征在于，
    所述第一识别目标和所述第二识别目标中的一个为所述拍摄对象的整体，所述第一识别目标和所述第二识别目标中的另一个为所述拍摄对象的指定部分。
  11. 根据权利要求1至4中任一项所述的云台，其特征在于，所述处理器执行所述计算机指令时还实现：
    输出所述识别目标切换指令。
  12. 根据权利要求11所述的云台，其特征在于，所述处理器执行所述计算机指令时还实现：
    输出至少一个备选识别目标；
    从所述第一识别目标切换至所述至少一个备选识别目标中被选中的一个。
  13. 根据权利要求1至4中任一项所述的云台，其特征在于，所述处理器执行所述计算机指令时还实现：
    基于所述拍摄画面中存在至少一个备选拍摄对象的情况，检测所述至少一个备选拍摄对象的备选识别目标；
    输出至少一个所述备选识别目标；
    将被选中的所述备选识别目标作为所述第一识别目标，将所述第一识别目标所属的所述备选拍摄对象作为所述拍摄对象。
  14. 根据权利要求13所述的云台，其特征在于，所述处理器执行所述计算机指令实现的输出至少一个所述备选识别目标，包括：
    针对每个所述备选拍摄对象，输出一个所述备选识别目标。
  15. 根据权利要求14所述的云台，其特征在于，所述处理器执行所述计算机指令实现的输出一个所述备选识别目标，包括：
    确定所述备选拍摄对象对应的所述备选识别目标的数量为一个，输出对应的所述备选识别目标；
    确定所述备选拍摄对象对应的所述备选识别目标的数量为至少两个，输出对应的其中一个所述备选识别目标。
  16. 根据权利要求15所述的云台，其特征在于，所述处理器执行所述计算机指令实现的输出对应的其中一个所述备选识别目标，包括：
    检测对应的全部所述备选识别目标在所述拍摄画面中的尺寸，根据所述尺寸确定输出的所述备选识别目标。
  17. 根据权利要求1至4中任一项所述的云台，其特征在于，所述处理器执行所述计算机指令时还实现：
    实时输出所述图像信息。
  18. 根据权利要求1至4中任一项所述的云台，其特征在于，所述云台还包括：
    显示屏，用于显示所述图像信息。
  19. 一种云台的控制设备，其特征在于，包括：
    存储器，被配置为存储计算机指令；及
    处理器，被配置为执行所述计算机指令以实现：
    实时接收拍摄画面的图像信息，所述拍摄画面内存在拍摄对象；
    根据所述图像信息检测第一识别目标在所述拍摄画面中的构图信息，所述第一识别目标为所述拍摄对象的至少一部分；
    根据所述第一识别目标的构图信息确定是否生成识别目标切换指令。
  20. 根据权利要求19所述的云台的控制设备，其特征在于，所述构图信息包括尺寸，所述处理器执行所述计算机指令实现的根据所述第一识别目标的构图信息确定是否生成识别目标切换指令，包括：
    确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令。
  21. 根据权利要求20所述的云台的控制设备，其特征在于，所述识别目标切换指令包括识别目标扩大指令，所述处理器执行所述计算机指令实现的确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令，包括：
    确定所述第一识别目标的尺寸小于所述预设尺寸范围的下限尺寸，生成所述识别目标扩大指令。
  22. 根据权利要求20所述的云台的控制设备，其特征在于，所述识别目标切换指令包括识别目标缩小指令，所述处理器执行所述计算机指令实现的确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令，包括：
    确定所述第一识别目标的尺寸大于所述预设尺寸范围的上限尺寸，生成所述识别目标缩小指令。
  23. 根据权利要求19至22中任一项所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令时还实现：
    响应于所述识别目标切换指令，从所述第一识别目标切换至第二识别目标，所述第二识别目标为所述拍摄对象的至少一部分，所述第一识别目标占所述拍摄对象的比例和所述第二识别目标占所述拍摄对象的比例不等。
  24. 根据权利要求23所述的云台的控制设备，其特征在于，所述构图信息还包括位置，所述处理器执行所述计算机指令时还实现：
    根据所述第一识别目标的构图信息调整拍摄参数，以使所述第一识别目标在所述拍摄画面中的成像满足构图规则。
  25. 根据权利要求24所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令实现的从所述第一识别目标切换至第二识别目标，包括：
    检测所述第二识别目标在所述拍摄画面中的构图信息；
    调整所述拍摄参数，以使所述第二识别目标在所述拍摄画面中的位置和尺寸逐渐变化，直到所述第二识别目标在所述拍摄画面中的成像满足所述构图规则。
  26. 根据权利要求24或25所述的云台的控制设备，其特征在于，
    所述构图规则包括以下至少之一：识别目标的成像完整、识别目标位于所述拍摄画面中的指定区域内、均衡式构图、对称式构图、对角线构图、三角形构图、九宫格构图、向心式构图、对分式构图。
  27. 根据权利要求24或25所述的云台的控制设备，其特征在于，
    所述云台还包括：
    支架设备，所述支架设备能够支撑用于获取所述图像信息的拍摄设备，
    和/或，拍摄设备，用于获取所述图像信息；
    所述拍摄参数包括以下至少之一：所述拍摄设备的高度、所述拍摄设备的倾角、所述拍摄设备的方向、所述拍摄设备的焦距。
  28. 根据权利要求23所述的云台的控制设备，其特征在于，
    所述第一识别目标和所述第二识别目标中的一个为所述拍摄对象的整体，所述第一识别目标和所述第二识别目标中的另一个为所述拍摄对象的指定部分。
  29. 根据权利要求19至22中任一项所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令时还实现：
    输出所述识别目标切换指令。
  30. 根据权利要求29所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令时还实现：
    输出至少一个备选识别目标；
    从所述第一识别目标切换至所述至少一个备选识别目标中被选中的一个。
  31. 根据权利要求19至22中任一项所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令时还实现：
    基于所述拍摄画面中存在至少一个备选拍摄对象的情况，检测所述至少一个备选拍摄对象的备选识别目标；
    输出至少一个所述备选识别目标；
    将被选中的所述备选识别目标作为所述第一识别目标，将所述第一识别目标所属的所述备选拍摄对象作为所述拍摄对象。
  32. 根据权利要求31所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令实现的输出至少一个所述备选识别目标，包括：
    针对每个所述备选拍摄对象，输出一个所述备选识别目标。
  33. 根据权利要求32所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令实现的输出一个所述备选识别目标，包括：
    确定所述备选拍摄对象对应的所述备选识别目标的数量为一个，输出对应的所述备选识别目标；
    确定所述备选拍摄对象对应的所述备选识别目标的数量为至少两个，输出对应的其中一个所述备选识别目标。
  34. 根据权利要求33所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令实现的输出对应的其中一个所述备选识别目标，包括：
    检测对应的全部所述备选识别目标在所述拍摄画面中的尺寸，根据所述尺寸确定输出的所述备选识别目标。
  35. 根据权利要求19至22中任一项所述的云台的控制设备，其特征在于，所述处理器执行所述计算机指令时还实现：
    实时输出所述图像信息。
  36. 一种云台的控制方法，用于云台或云台的控制设备，其特征在于，所述云台的控制方法包括：
    实时接收拍摄画面的图像信息，所述拍摄画面内存在拍摄对象；
    根据所述图像信息检测第一识别目标在所述拍摄画面中的构图信息，所述第一识别目标为所述拍摄对象的至少一部分；
    根据所述第一识别目标的构图信息确定是否生成识别目标切换指令。
  37. 根据权利要求36所述的云台的控制方法，其特征在于，所述构图信息包括尺寸，根据所述第一识别目标的构图信息确定是否生成识别目标切换指令，包括：
    确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令。
  38. 根据权利要求37所述的云台的控制方法，其特征在于，所述识别目标切换指令包括识别目标扩大指令，确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令，包括：
    确定所述第一识别目标的尺寸小于所述预设尺寸范围的下限尺寸，生成所述识别目标扩大指令。
  39. 根据权利要求37所述的云台的控制方法，其特征在于，所述识别目标切换指令包括识别目标缩小指令，确定所述第一识别目标的尺寸处于预设尺寸范围之外，生成所述识别目标切换指令，包括：
    确定所述第一识别目标的尺寸大于所述预设尺寸范围的上限尺寸，生成所述识别目标缩小指令。
  40. 根据权利要求36至39中任一项所述的云台的控制方法，其特征在于，所述云台的控制方法还包括：
    响应于所述识别目标切换指令，从所述第一识别目标切换至第二识别目标，所述第二识别目标为所述拍摄对象的至少一部分，所述第一识别目标占所述拍摄对象的比例和所述第二识别目标占所述拍摄对象的比例不 等。
  41. 根据权利要求40所述的云台的控制方法，其特征在于，所述构图信息还包括位置，所述云台的控制方法还包括：
    根据所述第一识别目标的构图信息调整拍摄参数，以使所述第一识别目标在所述拍摄画面中的成像满足构图规则。
  42. 根据权利要求41所述的云台的控制方法，其特征在于，从所述第一识别目标切换至第二识别目标，包括：
    检测所述第二识别目标在所述拍摄画面中的构图信息；
    调整所述拍摄参数，以使所述第二识别目标在所述拍摄画面中的位置和尺寸逐渐变化，直到所述第二识别目标在所述拍摄画面中的成像满足所述构图规则。
  43. 根据权利要求41或42所述的云台的控制方法，其特征在于，
    所述构图规则包括以下至少之一：识别目标的成像完整、识别目标位于所述拍摄画面中的指定区域内、均衡式构图、对称式构图、对角线构图、三角形构图、九宫格构图、向心式构图、对分式构图。
  44. 根据权利要求41或42所述的云台的控制方法，其特征在于，
    所述云台还包括：
    支架设备，所述支架设备能够支撑用于获取所述图像信息的拍摄设备，
    和/或，拍摄设备，用于获取所述图像信息；
    所述拍摄参数包括以下至少之一：所述拍摄设备的高度、所述拍摄设备的倾角、所述拍摄设备的方向、所述拍摄设备的焦距。
  45. 根据权利要求40所述的云台的控制方法，其特征在于，
    所述第一识别目标和所述第二识别目标中的一个为所述拍摄对象的整体，所述第一识别目标和所述第二识别目标中的另一个为所述拍摄对象的指定部分。
  46. 根据权利要求36至39中任一项所述的云台的控制方法，其特征在于，所述云台的控制方法还包括：
    输出所述识别目标切换指令。
  47. 根据权利要求46所述的云台的控制方法，其特征在于，所述云台 的控制方法还包括：
    输出至少一个备选识别目标；
    从所述第一识别目标切换至所述至少一个备选识别目标中被选中的一个。
  48. 根据权利要求36至39中任一项所述的云台的控制方法，其特征在于，所述云台的控制方法还包括：
    基于所述拍摄画面中存在至少一个备选拍摄对象的情况，检测所述至少一个备选拍摄对象的备选识别目标；
    输出至少一个所述备选识别目标；
    将被选中的所述备选识别目标作为所述第一识别目标，将所述第一识别目标所属的所述备选拍摄对象作为所述拍摄对象。
  49. 根据权利要求48所述的云台的控制方法，其特征在于，输出至少一个所述备选识别目标，包括：
    针对每个所述备选拍摄对象，输出一个所述备选识别目标。
  50. 根据权利要求49所述的云台的控制方法，其特征在于，输出一个所述备选识别目标，包括：
    确定所述备选拍摄对象对应的所述备选识别目标的数量为一个，输出对应的所述备选识别目标；
    确定所述备选拍摄对象对应的所述备选识别目标的数量为至少两个，输出对应的其中一个所述备选识别目标。
  51. 根据权利要求50所述的云台的控制方法，其特征在于，输出对应的其中一个所述备选识别目标，包括：
    检测对应的全部所述备选识别目标在所述拍摄画面中的尺寸，根据所述尺寸确定输出的所述备选识别目标。
  52. 根据权利要求36至39中任一项所述的云台的控制方法，其特征在于，所述云台的控制方法还包括：
    实时输出所述图像信息。
  53. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求36至52中任一项所述的云台的控制方法的步骤。
PCT/CN2019/121013 2019-11-26 2019-11-26 云台、云台的控制设备、云台的控制方法及存储介质 WO2021102695A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2019/121013 WO2021102695A1 (zh) 2019-11-26 2019-11-26 云台、云台的控制设备、云台的控制方法及存储介质
CN201980038365.4A CN112243581A (zh) 2019-11-26 2019-11-26 云台、云台的控制设备、云台的控制方法及存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/121013 WO2021102695A1 (zh) 2019-11-26 2019-11-26 云台、云台的控制设备、云台的控制方法及存储介质

Publications (1)

Publication Number Publication Date
WO2021102695A1 true WO2021102695A1 (zh) 2021-06-03

Family

ID=74171194

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/121013 WO2021102695A1 (zh) 2019-11-26 2019-11-26 云台、云台的控制设备、云台的控制方法及存储介质

Country Status (2)

Country Link
CN (1) CN112243581A (zh)
WO (1) WO2021102695A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113873166A (zh) * 2021-10-26 2021-12-31 维沃移动通信有限公司 视频拍摄方法、装置、电子设备和可读存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013178449A (ja) * 2012-02-09 2013-09-09 Furyu Kk 写真シール作成装置および方法、並びにプログラム
CN105744148A (zh) * 2014-12-24 2016-07-06 佳能株式会社 变焦控制装置和变焦控制装置的控制方法
CN105740757A (zh) * 2014-12-24 2016-07-06 佳能株式会社 变焦控制装置、摄像设备和变焦控制装置的控制方法
CN107087427A (zh) * 2016-11-30 2017-08-22 深圳市大疆创新科技有限公司 飞行器的控制方法、装置和设备以及飞行器
CN107705307A (zh) * 2017-11-09 2018-02-16 睿魔智能科技（东莞）有限公司 一种基于深度学习的拍摄构图方法和系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6436761B2 (ja) * 2014-12-24 2018-12-12 キヤノン株式会社 ズーム制御装置、撮像装置、ズーム制御装置の制御方法、及びズーム制御装置の制御プログラム
JP6618255B2 (ja) * 2014-12-24 2019-12-11 キヤノン株式会社 ズーム制御装置、撮像装置、ズーム制御装置の制御方法、ズーム制御装置の制御プログラム及び記憶媒体

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013178449A (ja) * 2012-02-09 2013-09-09 Furyu Kk 写真シール作成装置および方法、並びにプログラム
CN105744148A (zh) * 2014-12-24 2016-07-06 佳能株式会社 变焦控制装置和变焦控制装置的控制方法
CN105740757A (zh) * 2014-12-24 2016-07-06 佳能株式会社 变焦控制装置、摄像设备和变焦控制装置的控制方法
CN107087427A (zh) * 2016-11-30 2017-08-22 深圳市大疆创新科技有限公司 飞行器的控制方法、装置和设备以及飞行器
CN107705307A (zh) * 2017-11-09 2018-02-16 睿魔智能科技（东莞）有限公司 一种基于深度学习的拍摄构图方法和系统

Also Published As

Publication number Publication date
CN112243581A (zh) 2021-01-19

Similar Documents

Publication Publication Date Title
JP6429454B2 (ja) 撮像装置、撮像装置の制御方法および撮像装置の制御プログラム
US9344634B2 (en) Imaging apparatus having subject detection function, method for controlling the imaging apparatus, and storage medium
JP6512897B2 (ja) ズーム制御装置、ズーム制御装置の制御方法
JP7222683B2 (ja) 撮像装置及びその制御方法、プログラム、記憶媒体
JP6366395B2 (ja) ズーム制御装置、撮像装置、ズーム制御装置の制御方法、被写体検出装置の制御プログラムおよび記憶媒体
JP6370140B2 (ja) ズーム制御装置、撮像装置、ズーム制御装置の制御方法、ズーム制御装置の制御プログラムおよび記憶媒体
JP5538823B2 (ja) 撮像装置
JP6327995B2 (ja) 撮像装置、制御装置及びそれらの制御方法、プログラム、並びに記憶媒体
JP7233162B2 (ja) 撮像装置及びその制御方法、プログラム、記憶媒体
US20150116517A1 (en) Image processing device, image processing method, and program
WO2019104569A1 (zh) 一种对焦方法、设备及可读存储介质
JP2007219389A (ja) 撮像装置
WO2021102695A1 (zh) 云台、云台的控制设备、云台的控制方法及存储介质
JP2021124669A (ja) 電子機器
US9648242B2 (en) Display control apparatus, display control method, program, and image pickup apparatus for assisting a user
JP2023057157A (ja) 撮像装置及びその制御方法、プログラム
JP7146507B2 (ja) 情報処理装置及びその制御方法
JP2005020205A (ja) 画像撮影装置
WO2021026784A1 (zh) 跟随拍摄、云台控制方法、拍摄装置、手持云台和拍摄系统
US20140078325A1 (en) Image capturing apparatus and control method therefor
US11265478B2 (en) Tracking apparatus and control method thereof, image capturing apparatus, and storage medium
JP7366594B2 (ja) 情報処理装置とその制御方法
WO2022082439A1 (zh) 轨迹延时拍摄方法、装置、云台相机、无人机及手持云台
JP6256298B2 (ja) 撮像装置、その制御方法とプログラム
WO2020129620A1 (ja) 撮像制御装置、撮像装置、撮像制御方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19954257

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19954257

Country of ref document: EP

Kind code of ref document: A1