CN107749951A - A kind of visually-perceptible method and system for being used for unmanned photography - Google Patents
A kind of visually-perceptible method and system for being used for unmanned photography Download PDFInfo
- Publication number
- CN107749951A CN107749951A CN201711098855.2A CN201711098855A CN107749951A CN 107749951 A CN107749951 A CN 107749951A CN 201711098855 A CN201711098855 A CN 201711098855A CN 107749951 A CN107749951 A CN 107749951A
- Authority
- CN
- China
- Prior art keywords
- target
- shooting
- information
- module
- visually
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- H04N23/611—Control of cameras or camera modules based on recognised objects where the recognised objects include parts of the human body
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
A kind of visually-perceptible method and system for being used for unmanned photography, comprises the following steps：S1, camera shooting terminal is erected in support terminal, starts camera shooting terminal and target is started to shoot video, extract the foundation characteristic of target；S2, according to foundation characteristic, analysis target current behavior, posture and position, the tutorial message of next step shooting picture is exported, tutorial message is provided for the behavior of camera shooting terminal, adjusts camera terminal, according to this repeatedly, until video capture terminates.In the case that the present invention does not need people's participation, shooting is completed, human cost is reduced, saves the time, shooting effect is good.
Description
Technical field
It is specifically a kind of to be used for the unmanned visually-perceptible method photographed and be the invention belongs to unmanned shooting field
System.
Background technology
Existing camera work is required for the participation of someone.User directly holds camera terminal and shot, or borrows
Shooting tripod head etc. is controlled.Or the so-called unmanned shooting instrument of such as unmanned plane, Underwater Camera, it is substantially and by people
Manipulated, the content of shooting, time start-stop, coverage, shooting angle etc. shooting key element are determined by people.I.e. various bats
Mode is taken the photograph, is substantially all the directly or indirectly participation for needing people.Moreover, the unmanned shooting set, can only simply enter
Row shooting, can not carry out perception shooting.
The content of the invention
The technical problem to be solved in the present invention is to provide a kind of visually-perceptible method and system for being used for unmanned photography, it is not required to
In the case that very important person participates in, shooting is completed, human cost is reduced, saves the time, shooting effect is good.
In order to solve the above-mentioned technical problem, the present invention takes following technical scheme：
A kind of visually-perceptible method for being used for unmanned photography, comprises the following steps：
S1, camera shooting terminal is erected in support terminal, starts camera shooting terminal and target is started to shoot video, extract the base of target
Plinth feature；
S2, according to foundation characteristic, analysis target current behavior, posture and position, the guidance letter of output next step shooting picture
Breath, tutorial message is provided for the behavior of camera shooting terminal, adjusts camera terminal, according to this repeatedly, until video capture terminates.
The step S2 specifically includes front-end vision perception, and the front-end vision perceives and specifically includes following steps：
S201, target detection, determine particular location of the target in current shooting picture；
S202, Attitude estimation, obtain the posture feature of target, thus it is speculated that the trend of target next step；
S203, target following, according to the historical position of target and gesture distribution information, determine the position of target in subsequent frames
And attitude information.
In the step S201, during target detection, the picture frame in current shooting picture is detected every setting frame,
The target location in shooting picture is found, auxiliary information is provided for Attitude estimation and target following.
The step S2 specifically includes rear end scene and understood, the rear end scene, which understands, specifically includes following steps：
S211, behavioral value, the action to target detect, and obtain the action message of target；
S212, recognition of face, the face in shooting picture is identified, provided for identification target and object select with reference to letter
Breath；
S213, gesture identification, detect whether to represent the gesture of operational order, if so, performing corresponding operating according to gesture；
S214, focus perceive, according to the posture feature obtained in the action message and Attitude estimation obtained in behavioral value, to working as
The focus of preceding shooting video is estimated, automatically determines coverage, shooting angle and shooting path, output shooting strategy, adjusts
Whole camera terminal to relevant position is shot.
The information that the recognition of face of the step S212 obtains and the information that step S213 gesture identification obtains, are also provided
To step S203 target following, the information is analyzed and processed.
The information that the information and step S202 Attitude estimations that the target detection of the step S201 obtains obtain, is also provided to
Step S211 behavioral value, the information is analyzed and processed.
A kind of visually-perceptible system for being used for unmanned photography, the system includes foundation characteristic extraction module, for obtaining
The foundation characteristic of target in shooting picture；
Front-end vision sensing module, for receiving foundation characteristic and being analyzed and processed, export the guidance of next step shooting picture
Information；
Rear end scene Understanding Module, for receiving foundation characteristic and being analyzed and processed, export the guidance of next step shooting picture
Information.
The front-end vision sensing module includes：
Module of target detection, for determining particular location of the target in current shooting picture；
Attitude estimation module, obtain the posture feature of target, thus it is speculated that the trend of target next step；
Target tracking module, according to the historical position of target and gesture distribution information, determine the position of target in subsequent frames
And attitude information；
Rear end scene Understanding Module includes：Behavioral value module, the action to target detect, and obtain the action letter of target
Breath；
Face recognition module, the face in shooting picture is identified, reference information is provided for identification target and object select；
Gesture recognition module, detect whether to represent the gesture of operational order, if so, performing corresponding operating according to gesture；
Focus sensing module, according to the posture feature obtained in the action message and Attitude estimation obtained in behavioral value, to working as
The focus of preceding shooting video is estimated, automatically determines coverage, shooting angle and shooting path, output shooting strategy, adjusts
Whole camera terminal to relevant position is shot.
In the case that the present invention does not need people's participation, shooting is completed, human cost is reduced, is extracted, had by foundation characteristic
The time that effect reduces the amount of calculation of shooting and expended, shooting effect are good.
Brief description of the drawings
Accompanying drawing 1 is the catenation principle schematic diagram of system in the present invention；
The overall flow schematic diagram that accompanying drawing 2 is shot for the present invention with reference to camera terminal.
Embodiment
For the feature of the present invention, technological means and the specific purposes reached, function can be further appreciated that, with reference to
Accompanying drawing is described in further detail with embodiment to the present invention.
A kind of visually-perceptible method for being used for unmanned photography, comprises the following steps：
S1, camera shooting terminal is erected in support terminal, starts camera shooting terminal and target is started to shoot video, extract the base of target
Plinth feature.It can be head either other movable support meanss to support terminal, and camera shooting terminal is such as camera, mobile phone has
The electronic equipment of camera function.Obtain a variety of foundation characteristics of target, the foundation characteristic is as the relative of target and surrounding enviroment
Relation, action, posture etc..
S2, according to foundation characteristic, analysis target current behavior, posture and position, the finger of output next step shooting picture
Information is led, tutorial message is provided for the behavior of camera shooting terminal, adjusts camera terminal, according to this repeatedly, until video capture terminates.
The step S2 specifically includes front-end vision perception, and the front-end vision perceives and specifically includes following steps：
S201, target detection, determine particular location of the target in current shooting picture.Every setting frame to current shooting picture
In picture frame detected, the target location in shooting picture is found, so as to acquire more accurate target location.
S202, Attitude estimation, the posture feature of target is obtained, according to the inner link between posture and behavior, thus it is speculated that mesh
The trend of next step is marked, the target for obtaining a supposition tends to.
S203, target following, according to the historical position of target and gesture distribution information, determine target in subsequent frames
Position and attitude information, export the target location in subsequent frame and attitude information.Mainly the action to target is tracked, can
To be more smoothly rapidly performed by tracking, according to the position of target and attitude information, tracking effect can be effectively improved.And
And among the scene of complexity, the information that is obtained by target detection and Attitude estimation can also find the people in video
Body position, so as to aid in the work of human body tracking, or target following is helped to recover from failure, so as to continue to track.
The step S2 specifically includes rear end scene and understood, the rear end scene, which understands, specifically includes following steps：
S211, behavioral value, the action to target detect, and obtain the action message of target, while can be used for estimation and work as
The behavior focus of preceding object.
S212, recognition of face, the face in shooting picture is identified, reference is provided for identification target and object select
Information；
S213, gesture identification, detect whether to represent the gesture of operational order, if so, performing corresponding operating according to gesture.Such as
Target is people, then people can be sent to video camera by gesture before camera and start to image, stop the instruction such as shooting.Pass through knowledge
These other gestures, it is possible to achieve corresponding operation.
S214, focus perceives, according to the posture feature obtained in the action message and Attitude estimation obtained in behavioral value,
The focus of current shooting video is estimated, automatically determines coverage, shooting angle and shooting path, output shooting plan
Slightly, adjustment camera terminal to relevant position is shot.So as to according to obtained shooting strategy, adjust camera terminal, Huo Zhetong
Adjustment support terminal is crossed to reach the purpose of adjustment camera terminal, is once shot under realizing more preferably.With control camera shooting terminal or
Person supports the control system of terminal to carry out the communication of director data, realizes specific control.
In order to preferably be shot and be perceived, the information that information and gesture identification that recognition of face obtains obtain also provides
To target following, target following step is conveyed to, is advantageous to preferably be tracked.
In addition, the information that information and step S202 Attitude estimations that the target detection of the step S201 obtains obtain, also
Step S211 behavioral value is supplied to, the information is analyzed and processed.Understand so as to which front-end vision is perceived as rear end scene
Substantial amounts of Back ground Information, such as posture feature, position are provided.And rear end scene understanding simultaneously can also feed back to front-end vision
Perceive, perceived by front-end vision and rear end scene understands that the next step of dual output shoots tutorial message, it is comprehensive obtain one compared with
The shooting strategy of extension.
Present invention further teaches a kind of visually-perceptible system for being used for unmanned photography, as shown in Figure 1, the system includes
Foundation characteristic extraction module, for obtaining the foundation characteristic of target in shooting picture, by foundation characteristic extraction module, as list
Only extraction module, save and subsequently clarification of objective is extracted, a unified characteristic extracting module is formed, so as to save several times
Time and computer memory space.The deep neural network parameter of foundation characteristic extraction module, first comes from what is trained
ImageNet disaggregated model, then using transfer learning, trained successively by different modules, obtain shared foundation characteristic
Extract network.
Front-end vision sensing module, for receiving foundation characteristic and being analyzed and processed, output next step shooting picture
Tutorial message；Rear end scene Understanding Module, for receiving foundation characteristic and being analyzed and processed, output next step shooting picture
Tutorial message.
The front-end vision sensing module includes：Module of target detection, for determining target in current shooting picture
Particular location；Attitude estimation module, obtain the posture feature of target, thus it is speculated that the trend of target next step；Target tracking module, root
According to the historical position and gesture distribution information of target, target position in subsequent frames and attitude information are determined, after output
Target location and attitude information in continuous frame.
Rear end scene Understanding Module includes：Behavioral value module, the action to target detect, and obtain the action of target
Information；Face recognition module, the face in shooting picture is identified, provided for identification target and object select with reference to letter
Breath；Gesture recognition module, detect whether to represent the gesture of operational order, if so, performing corresponding operating according to gesture；Focus sense
Module is known, according to the posture feature obtained in the action message and Attitude estimation obtained in behavioral value, to current shooting video
Focus estimated, automatically determine coverage, shooting angle and shooting path, output shooting strategy, adjust camera terminal
Shot to relevant position.
Target detection can find target location, determine skeleton location for Attitude estimation, there is provided auxiliary information.Target
Detection module provides auxiliary information for behavioral value module.
Attitude estimation module provides auxiliary information for target tracking module：Attitude estimation module can obtain each portion of target
Position where position, more auxiliary informations are provided for the reset of target following.Attitude estimation module provides for behavioral value module
Auxiliary information.Attitude estimation module can obtain the posture feature of target, exist between posture and behavior and contact, therefore can pass through
Posture feature, thus it is speculated that possible goal behavior.
Attitude estimation module provides auxiliary information for focus sensing module：The definition of focus depends on target to a certain extent
Posture.
Face recognition module provides auxiliary information for target tracking module：Carried with module of target detection for target tracking module
It is similar for auxiliary information.
Behavioral value module provides auxiliary information for focus sensing module：The definition of focus depends on target to a certain extent
Behavior.
Finally, the output information of comprehensive front-end visually-perceptible module and rear end scene analysis module, for support terminal and take the photograph
As the behavior offer tutorial message of terminal, so as to complete to shoot.
As shown in Figure 2, goal-setting is behaved, and support terminal uses head.Current picture is shot, then into we
In visually-perceptible system in method, the information such as position, posture to people obtains, thus it is speculated that and next posture of people tends to, and
And it is tracked.The action also to people detects in addition, is perceived by various focuses, obtains a preferably shooting strategy,
Then by the control system of head control head turn particularly to and angle, realize steering and the angular adjustment of camera shooting terminal.
According to this repeatedly, until shooting terminates.Button can be either pressed by people's gesture identification or other signals are controlled.
It should be noted that these are only the preferred embodiments of the present invention, it is not intended to limit the invention, although ginseng
The present invention is described in detail according to embodiment, for those skilled in the art, it still can be to foregoing reality
Apply the technical scheme described in example to modify, or equivalent substitution is carried out to which part technical characteristic, but it is all in this hair
Within bright spirit and principle, any modification, equivalent substitution and improvements made etc., protection scope of the present invention should be included in
Within.
Claims (8)
1. a kind of visually-perceptible method for being used for unmanned photography, comprises the following steps：
S1, camera shooting terminal is erected in support terminal, starts camera shooting terminal and target is started to shoot video, extract the base of target
Plinth feature；
S2, according to foundation characteristic, analysis target current behavior, posture and position, the guidance letter of output next step shooting picture
Breath, tutorial message is provided for the behavior of camera shooting terminal, adjusts camera terminal, according to this repeatedly, until video capture terminates.
2. the visually-perceptible method according to claim 1 for being used for unmanned photography, it is characterised in that the step S2 is specific
Perceived including front-end vision, the front-end vision perceives and specifically includes following steps：
S201, target detection, determine particular location of the target in current shooting picture；
S202, Attitude estimation, obtain the posture feature of target, thus it is speculated that the trend of target next step；
S203, target following, according to the historical position of target and gesture distribution information, determine the position of target in subsequent frames
And attitude information.
3. the visually-perceptible method according to claim 2 for being used for unmanned photography, it is characterised in that the step S201
In, during target detection, the picture frame in current shooting picture is detected every setting frame, finds the target in shooting picture
Position, auxiliary information is provided for Attitude estimation and target following.
4. the visually-perceptible method according to claim 3 for being used for unmanned photography, it is characterised in that the step S2 is specific
Understand including rear end scene, the rear end scene, which understands, specifically includes following steps：
S211, behavioral value, the action to target detect, and obtain the action message of target；
S212, recognition of face, the face in shooting picture is identified, provided for identification target and object select with reference to letter
Breath；
S213, gesture identification, detect whether to represent the gesture of operational order, if so, performing corresponding operating according to gesture；
S214, focus perceive, according to the posture feature obtained in the action message and Attitude estimation obtained in behavioral value, to working as
The focus of preceding shooting video is estimated, automatically determines coverage, shooting angle and shooting path, output shooting strategy, adjusts
Whole camera terminal to relevant position is shot.
5. the visually-perceptible method according to claim 4 for being used for unmanned photography, it is characterised in that the step S212's
The information that recognition of face obtains and the information that step S213 gesture identification obtains, are also provided to step S203 target following,
The information is analyzed and processed.
6. the visually-perceptible method according to claim 5 for being used for unmanned photography, it is characterised in that the step S201's
The information that the information and step S202 Attitude estimations that target detection obtains obtain, is also provided to step S211 behavioral value, right
The information is analyzed and processed.
A kind of 7. visually-perceptible system for being used for unmanned photography, it is characterised in that the system includes foundation characteristic extraction module,
For obtaining the foundation characteristic of target in shooting picture；
Front-end vision sensing module, for receiving foundation characteristic and being analyzed and processed, export the guidance of next step shooting picture
Information；
Rear end scene Understanding Module, for receiving foundation characteristic and being analyzed and processed, export the guidance of next step shooting picture
Information.
8. the visually-perceptible system according to claim 7 for being used for unmanned photography, it is characterised in that the front-end vision sense
Know that module includes：
Module of target detection, for determining particular location of the target in current shooting picture；
Attitude estimation module, obtain the posture feature of target, thus it is speculated that the trend of target next step；
Target tracking module, according to the historical position of target and gesture distribution information, determine the position of target in subsequent frames
And attitude information；
Rear end scene Understanding Module includes：Behavioral value module, the action to target detect, and obtain the action letter of target
Breath；
Face recognition module, the face in shooting picture is identified, reference information is provided for identification target and object select；
Gesture recognition module, detect whether to represent the gesture of operational order, if so, performing corresponding operating according to gesture；
Focus sensing module, according to the posture feature obtained in the action message and Attitude estimation obtained in behavioral value, to working as
The focus of preceding shooting video is estimated, automatically determines coverage, shooting angle and shooting path, output shooting strategy, adjusts
Whole camera terminal to relevant position is shot.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711098855.2A CN107749951A (en) | 2017-11-09 | 2017-11-09 | A kind of visually-perceptible method and system for being used for unmanned photography |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711098855.2A CN107749951A (en) | 2017-11-09 | 2017-11-09 | A kind of visually-perceptible method and system for being used for unmanned photography |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107749951A true CN107749951A (en) | 2018-03-02 |
Family
ID=61250890
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711098855.2A Pending CN107749951A (en) | 2017-11-09 | 2017-11-09 | A kind of visually-perceptible method and system for being used for unmanned photography |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107749951A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111801932A (en) * | 2018-08-31 | 2020-10-20 | 华为技术有限公司 | Image shooting method and device |
| CN112639652A (en) * | 2020-05-07 | 2021-04-09 | 深圳市大疆创新科技有限公司 | Target tracking method and device, movable platform and imaging platform |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573617A (en) * | 2013-10-28 | 2015-04-29 | 季春宏 | Video shooting control method |
| CN105718862A (en) * | 2016-01-15 | 2016-06-29 | 北京市博汇科技股份有限公司 | Method, device and recording-broadcasting system for automatically tracking teacher via single camera |
| CN105847681A (en) * | 2016-03-30 | 2016-08-10 | 乐视控股（北京）有限公司 | Shooting control method, device and system |
| CN105979147A (en) * | 2016-06-22 | 2016-09-28 | 上海顺砾智能科技有限公司 | Intelligent shooting method of unmanned aerial vehicle |
| CN106454108A (en) * | 2016-11-04 | 2017-02-22 | 北京百度网讯科技有限公司 | Tracking shooting method and apparatus based on artificial intelligence, and electronic device |
-
2017
- 2017-11-09 CN CN201711098855.2A patent/CN107749951A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573617A (en) * | 2013-10-28 | 2015-04-29 | 季春宏 | Video shooting control method |
| CN105718862A (en) * | 2016-01-15 | 2016-06-29 | 北京市博汇科技股份有限公司 | Method, device and recording-broadcasting system for automatically tracking teacher via single camera |
| CN105847681A (en) * | 2016-03-30 | 2016-08-10 | 乐视控股（北京）有限公司 | Shooting control method, device and system |
| CN105979147A (en) * | 2016-06-22 | 2016-09-28 | 上海顺砾智能科技有限公司 | Intelligent shooting method of unmanned aerial vehicle |
| CN106454108A (en) * | 2016-11-04 | 2017-02-22 | 北京百度网讯科技有限公司 | Tracking shooting method and apparatus based on artificial intelligence, and electronic device |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111801932A (en) * | 2018-08-31 | 2020-10-20 | 华为技术有限公司 | Image shooting method and device |
| CN112639652A (en) * | 2020-05-07 | 2021-04-09 | 深圳市大疆创新科技有限公司 | Target tracking method and device, movable platform and imaging platform |
| WO2021223171A1 (en) * | 2020-05-07 | 2021-11-11 | 深圳市大疆创新科技有限公司 | Target tracking method and apparatus, movable platform, and imaging platform |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101783019B (en) | Subject tracking apparatus and control method therefor, image capturing apparatus, and display apparatus | |
| CN103310186B (en) | The method and portable terminal of the direction of gaze of user in correction chart picture | |
| US20170223261A1 (en) | Image pickup device and method of tracking subject thereof | |
| WO2019137131A1 (en) | Image processing method, apparatus, storage medium, and electronic device | |
| CN107707871B (en) | Image processing apparatus, image capturing apparatus, image processing method, and storage medium | |
| CN109453517B (en) | Virtual character control method and device, storage medium and mobile terminal | |
| US11036966B2 (en) | Subject area detection apparatus that extracts subject area from image, control method therefor, and storage medium, as well as image pickup apparatus and display apparatus | |
| KR101783999B1 (en) | Gesture recognition using chroma-keying | |
| CN112287880A (en) | Cloud deck attitude adjusting method, device and system and electronic equipment | |
| CN111654624B (en) | Shooting prompting method and device and electronic equipment | |
| CN114697539A (en) | Photographing recommendation method and device, electronic equipment and storage medium | |
| CN107749951A (en) | A kind of visually-perceptible method and system for being used for unmanned photography | |
| CN110378183B (en) | Image analysis device, image analysis method, and recording medium | |
| CN112733773B (en) | Object detection method, device, computer equipment and storage medium | |
| WO2017113674A1 (en) | Method and system for realizing motion-sensing control based on intelligent device, and intelligent device | |
| JP6406044B2 (en) | Camera calibration unit, camera calibration method, and camera calibration program | |
| KR20140137564A (en) | 3 Dimensional Image Sensing System for Human Tracking and Its Method | |
| CN113891002B (en) | Shooting method and device | |
| JP5539565B2 (en) | Imaging apparatus and subject tracking method | |
| CN111726531B (en) | Image shooting method, processing method, device, electronic equipment and storage medium | |
| CN113837033A (en) | Face recognition method carrying TOF module | |
| CN115037867B (en) | Shooting method, shooting device, computer readable storage medium and electronic equipment | |
| CN112533071B (en) | Image processing method and device and electronic equipment | |
| CN115278060B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN113055602B (en) | Remote group photo method, system, intelligent terminal and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180302 |
|
| RJ01 | Rejection of invention patent application after publication |