CN106251334B - A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system - Google Patents
A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system Download PDFInfo
- Publication number
- CN106251334B CN106251334B CN201610562671.6A CN201610562671A CN106251334B CN 106251334 B CN106251334 B CN 106251334B CN 201610562671 A CN201610562671 A CN 201610562671A CN 106251334 B CN106251334 B CN 106251334B
- Authority
- CN
- China
- Prior art keywords
- camera
- instructor
- broadcasting
- video
- dimensional coordinate
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
Abstract
The embodiment of the invention discloses a kind of camera parameters method of adjustment, instructor in broadcasting's video camera and systems, wherein this method comprises: determining the target video object for needing to shoot；The target video camera for shooting the target video object is filtered out from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera according to preset instructor in broadcasting's strategy；The first three-dimensional coordinate of the target video object is obtained, first three-dimensional coordinate is three-dimensional coordinate of the target video object under corresponding first coordinate system of the target video camera；The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and the video image after output adjustment camera parameter.Using this programme, the efficiency of camera parameters adjustment can be improved, and promote video camera shooting effect.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of camera parameters methods of adjustment, instructor in broadcasting's video camera
And system.
Background technique
With the continuous development of image processing techniques and internet, need to use the scene of video conference more and more,
Remote communication of the video conference between user brings great convenience.Currently, generally requiring to dispose when carrying out video conference
Multiple video cameras are shot, to obtain the direct picture of participant.For example, referring to Figure 1, Fig. 1 is a kind of video conference
Schematic diagram of a scenario.As shown in Figure 1, in this scenario, meeting room uses strip ellipse conference table, and the seat of participant surrounds
Conference table, participant include A and B, and participant A and B sit facing each other, and the projection screen two sides in the front of the A and B are disposed with camera shooting
Machine C0 and C1.Then for A, the direct picture of A only can be just photographed by video camera C0, and video camera C1 can not take A's
Front.And for B, the direct picture of B only can be just photographed by video camera C1, and video camera C0 can not take the front of B.
It can be seen that multiple video cameras need to be used to be shot to realize video conference.
When disposing multiple video cameras progress meeting shootings, camera shooting is manually adjusted generally by remote controler or other modes
Machine parameter is to obtain a preferable shooting effect.However, the mode manually adjusted needs operator to have certain camera shooting
Machine professional knowledge, and operating process is cumbersome, it is lower that this allows for adjusting efficiency, can not guarantee preferable shooting effect in time.This
Outside, the video camera of shooting can be also determined by way of auditory localization and adjusts video camera shooting effect.The auditory localization side
Formula is to shoot the spokesman by positioning and tracking the participant to make a speech (i.e. " spokesman "), while using a video camera
Feature, when feature the face location of tracking spokesman and carry out camera lens PTZ (Pan Tilt Zoom, i.e., " translation, inclination,
Zoom ") adjustment, so that the face of spokesman is located at the intermediate region of image.However, the auditory localization mode only accounts for send out
Speech people front is adjusted to image center, and does not consider to shoot the effect problem of image, also not can guarantee a preferably shooting
Effect.
Summary of the invention
The embodiment of the present invention provides a kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system, can be improved camera shooting
The efficiency of machine parameter adjustment, and promote video camera shooting effect.
In a first aspect, the method is applied to instructor in broadcasting the embodiment of the invention provides a kind of camera parameters method of adjustment
In video camera, comprising:
Determine the target video object for needing to shoot；
It is screened from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera according to preset instructor in broadcasting's strategy
Out for shooting the target video camera of the target video object；
The first three-dimensional coordinate of the target video object is obtained, first three-dimensional coordinate is the target video object
Three-dimensional coordinate under corresponding first coordinate system of the target video camera；
The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and defeated
The video image after camera parameter is adjusted out.
Wherein, which can be the target video object under corresponding first coordinate system of target video camera
Three-dimensional coordinate.The target video camera can be instructor in broadcasting's video camera or be common Pan/Tilt/Zoom camera, then the target video camera is corresponding
First coordinate system can refer to the three-dimensional system of coordinate established using target camera optical center as origin, or with any other object of reference
For the three-dimensional system of coordinate that origin is established, the embodiment of the present invention is without limitation.
Wherein, which can be the corresponding photographed scene of instructor in broadcasting's camera system where instructor in broadcasting's video camera
In any one or more the video objects.
In some embodiments, first three-dimensional coordinate for obtaining the target video object, comprising:
Obtain the second three-dimensional coordinate of the binocular camera transmission connecting with instructor in broadcasting's video camera, the described second three-dimensional seat
It is designated as three-dimensional coordinate of the target video object under corresponding second coordinate system of the binocular camera；
It is three-dimensional by described second according to the positional relationship of the binocular camera and the target video camera demarcated in advance
Coordinate is converted to the first three-dimensional coordinate.
In some embodiments, second three-dimensional coordinate can be the binocular camera and pass through the video that obtains respectively
Object is in the binocular camera of two-dimensional coordinate and acquisition in the left view and right view of the binocular camera
What outer parameter evidence was calculated.
Wherein, which is the target video object in corresponding second coordinate system of the binocular camera
Under three-dimensional coordinate, corresponding second coordinate system of the binocular camera can refer to using binocular camera optical center as origin establish
Three-dimensional system of coordinate, or the three-dimensional system of coordinate established using any other object of reference as origin.The two-dimensional coordinate can be specially should
Target video object corresponding pixel coordinate in the left view and right view of the binocular camera.
In some embodiments, the determining target video object for needing to shoot, comprising:
The shooting image of binocular camera transmission is obtained, includes at least one the video object in the shooting image；
The video object model of at least one the video object described in including is established, and from least one described the video object
Determine target video object；
It is described according to preset instructor in broadcasting's strategy from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera
Filter out the target video camera for shooting the target video object, comprising:
The target video is determined from the shooting image that each video camera in instructor in broadcasting's camera system obtains respectively
Object, and the target video object is obtained in the shooting effect parameter of each video camera；
It is determined as the video camera that shooting effect parameter meets default instructor in broadcasting's strategy to be used to shoot the target video object
Target video camera.
In some embodiments, the target video object is determined from the shooting image that video camera obtains, comprising:
According to the positional relationship of the binocular camera and current camera demarcated in advance, by second three-dimensional coordinate
Be converted to third three-dimensional coordinate；
Judge that region and the current camera of the target video object under the third three-dimensional coordinate detect
The overlapping area in region of the video object under the three-dimensional coordinate of the video object whether be more than preset area threshold；
If being more than, which is determined as the target video object.
Wherein, which is to remove in instructor in broadcasting's camera system with instructor in broadcasting's camera calibration good position relationship
Any video camera other than the binocular camera, the third three-dimensional coordinate are target video object video camera in this prior
Three-dimensional coordinate under corresponding third coordinate system.
In some embodiments, the shooting effect parameter includes that the target video object is corresponding in current camera
Eye under coordinate system is to any one of scenario objects parameter of eye efficacy parameter, hiding relation parameter and shooting area or more
, the current camera is any video camera in instructor in broadcasting's camera system in addition to the binocular camera.
Wherein, which may include the target video object relative to the corresponding coordinate system of current camera to eye efficacy parameter
Rotation angle, the rotation angle be according to target video object the rotation angle of second coordinate system and in advance calibration
The binocular camera and the positional relationship of the current camera determine.The rotation angle is smaller, and eye is better to eye effect.
Wherein, the hiding relation parameter and the scenario objects parameter can be according to the binocular camera demarcated in advance and
The current camera is thrown into the region for the scenario objects that the current camera detects by the positional relationship of current camera again
What imaging plane was determined.Output image effect is better when not having hiding relation (hiding relation parameter is smaller).The scenario objects
Area is smaller, number is smaller, then it is better to export image effect；Conversely, it is poorer then to export image effect.
Second aspect, the embodiment of the invention also provides a kind of instructor in broadcasting's video cameras, comprising: memory and processor, it is described
Processor is connect with the memory；Wherein,
The memory is used for storage driving software；
The processor reads the drive software from the memory and executes under the action of the drive software
State some or all of the camera parameters method of adjustment of first aspect step.
The third aspect, the embodiment of the invention also provides a kind of parameter adjustment controls, including object determination unit, selection list
Member, acquiring unit and parameter adjustment unit, the parameter adjustment controls realize that the video camera of first aspect is joined by said units
Some or all of number adjusting method step.
Fourth aspect, the embodiment of the invention also provides a kind of computer storage medium, the computer storage medium is deposited
Program is contained, described program includes the step of some or all of camera parameters method of adjustment of above-mentioned first aspect when executing
Suddenly.
5th aspect, the embodiment of the invention also provides a kind of instructor in broadcasting's camera systems, including the first video camera and at least one
A second video camera, first video camera include instructor in broadcasting's video camera and binocular camera, instructor in broadcasting's video camera and described double
It is connected between lens camera, between first video camera and second video camera by wireline interface or wireless interface；Its
In,
Instructor in broadcasting's video camera, for determining the target video object that shoots of needs, and according to preset instructor in broadcasting's strategy from
The target video camera for shooting the target video object is filtered out in the video camera of instructor in broadcasting's camera system；
The binocular camera, for obtaining the second three-dimensional coordinate of the target video object, and by the described 2nd 3
Dimension coordinate is transferred to instructor in broadcasting's video camera；Wherein, second three-dimensional coordinate is the target video object in the binocular
Three-dimensional coordinate under corresponding second coordinate system of video camera；
Instructor in broadcasting's video camera, for receiving second three-dimensional coordinate of the binocular camera transmission；According to preparatory
The binocular camera of calibration and the positional relationship of the target video camera, are converted to first for second three-dimensional coordinate
Three-dimensional coordinate；The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and
Video image after output adjustment camera parameter；Wherein, first three-dimensional coordinate is the target video object in the mesh
Mark the three-dimensional coordinate under corresponding first coordinate system of video camera.
In some embodiments, which may include instructor in broadcasting's video camera and binocular camera, then the target is taken the photograph
Camera can be any instructor in broadcasting's video camera in instructor in broadcasting's camera system；Alternatively, second video camera can also be taken the photograph for common PTZ
Camera, then the target video camera can be instructor in broadcasting's video camera or common Pan/Tilt/Zoom camera.
In some embodiments, which may be disposed on preset instructor in broadcasting's bracket, and pass through instructor in broadcasting's bracket
It is connect with instructor in broadcasting's video camera.
The implementation of the embodiments of the present invention has the following beneficial effects:
It in embodiments of the present invention, can be after the target video object for determining to need to shoot, according to preset instructor in broadcasting
Strategy is filtered out from each video camera of instructor in broadcasting's camera system shoots the optimal target video camera of target video object effect, and
Three-dimensional coordinate of the target video object under the corresponding coordinate system of target video camera is acquired, to control target camera shooting
Machine carries out camera parameters adjustment, and the video after output adjustment camera parameter according to the three-dimensional coordinate of the target video object
Image enables instructor in broadcasting's camera system to be based on three-dimensional coordinate detection and preset instructor in broadcasting strategy, to improve the video object detection
With the precision of tracking, while the efficiency of camera parameters adjustment is improved, and effectively improves the shooting effect of video camera.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of schematic diagram of a scenario of video conference；
Fig. 2 is a kind of flow diagram of camera parameters method of adjustment provided in an embodiment of the present invention；
Fig. 3 a is a kind of camera imaging model schematic diagram provided in an embodiment of the present invention；
Fig. 3 b is a kind of calibration schematic diagram of a scenario of multiple-camera provided in an embodiment of the present invention；
Fig. 3 c is a kind of three-dimensional localization schematic diagram of binocular camera provided in an embodiment of the present invention；
Fig. 3 d is a kind of Pan/Tilt/Zoom camera rotating model schematic diagram provided in an embodiment of the present invention；
Fig. 4 a is a kind of the video object matching schematic diagram of a scenario provided in an embodiment of the present invention；
Fig. 4 b is one group of the video object image in Fig. 4 a；
Fig. 5 is a kind of structural schematic diagram of parameter adjustment controls provided in an embodiment of the present invention；
Fig. 6 is a kind of structural schematic diagram of instructor in broadcasting's camera system provided in an embodiment of the present invention；
Fig. 7 is a kind of structural schematic diagram of first video camera provided in an embodiment of the present invention；
Fig. 8 is a kind of networking schematic diagram of instructor in broadcasting's camera system provided in an embodiment of the present invention；
Fig. 9 is a kind of structural schematic diagram of instructor in broadcasting's video camera provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
It should be understood that the present embodiments relate to " first ", " second " and " third " etc. be for distinguishing different objects, and
It is non-to be used to describe a particular order.In addition, term " includes " and their any deformations, it is intended that cover and non-exclusive include.
Such as contain a series of steps or units process, method, system, product or equipment be not limited to listed step or
Unit, but optionally further comprising the step of not listing or unit, or optionally further comprising for these process, methods, produce
The other step or units of product or equipment inherently.
It should be understood that the present embodiments relate to instructor in broadcasting's video camera can particularly for execute the embodiment of the present invention technology
The Pan/Tilt/Zoom camera of scheme can be connect with binocular camera, which can be applied to the scenes such as meeting, training, and can
The deployment of instructor in broadcasting's camera position and number is carried out according to different scenes.
In some embodiments, the binocular camera is mountable on instructor in broadcasting's bracket, i.e., instructor in broadcasting's video camera can be by leading
Bracket (referred to as " bracket ") is broadcast to connect with binocular camera.Wherein, instructor in broadcasting's video camera is for carrying out instructor in broadcasting's shooting and tracking.This
Outside, also mountable microphone, the microphone can be used for realizing the functions such as auditory localization, identification of sound source on the bracket.The instructor in broadcasting takes the photograph
Camera and bracket can be separation, be also possible to integrate, and control can be used between instructor in broadcasting's video camera and bracket
Interface processed such as serial line interface is communicated.
In some embodiments, binocular camera can be used for video acquisition, video pre-filtering, motion detection, Face datection,
Humanoid detection, scenario objects detection, feature detection/matching, binocular camera calibration, multiple cameras calibration etc., microphone can
For audio collection, audio pretreatment, video acquisition, sound source Activity recognition etc., instructor in broadcasting's video camera can be used for audio-video
(Audio Video, referred to as " AV ") object 3D positioning, AV object modeling, AV control image tracing, movement/gesture recognition, instructor in broadcasting
System and video switching/synthesis etc..Wherein, video acquisition includes the video of synchronous acquisition binocular camera and instructor in broadcasting's video camera
Stream；Video pre-filtering includes pre-processing to the binocular image of input, such as carry out noise reduction, change resolution ratio and frame per second behaviour
Make；Motion detection includes the Moving Objects detected in scene, and Moving Objects and static background are separated, and is moved
The region of object；Face datection includes the human face target object detected in scene, the detection information of face is exported, such as face position
It sets, region, the information such as direction；Humanoid detection includes the humanoid head-and-shoulder area region in detection scene, exports detection information；Scene
Object detection includes the other objects detected in scene other than people, such as fluorescent tube, window, conference table；Feature detection/matching
Moving Objects region including obtaining to detection carries out feature detection and matching, detects the properties object in an image (such as spy
Sign point) and matched in another image, export matched feature object information；Binocular camera calibration includes to binocular
Video camera is demarcated, and the inside and outside ginseng of binocular camera is obtained, for calculating the three-dimensional coordinate of the video object in video image；It is more
Camera calibration includes demarcating to the relative positional relationship of multiple instructor in broadcasting's video cameras, obtains the opposite of multiple instructor in broadcasting's video cameras
Outer ginseng information, for positioning of the video object in multiple camera coordinate systems.Further, audio collection includes synchronous acquisition
The multi-path audio-frequency data of microphone；Audio pretreatment includes carrying out 3A processing to the multi-path audio-frequency data of input, and wherein 3A is handled
Including auto-exposure control (AE), auto focus control (AF), automatic white balance control (AWB)；Auditory localization includes to input
Multi-path audio-frequency data detected, find the two-dimensional position information of sounding object；Sound source Activity recognition includes detection and statistics
The speech act of the video object in scene.Further, AV object 3D positioning includes according to the inside and outside ginseng of binocular camera and spy
The parallax information that sign detection/matching obtains, obtains the depth information of objects in images feature, in conjunction with audio positioning as a result, obtaining
To three dimensional local information of the characteristics of objects under single instructor in broadcasting's camera coordinate system, according to feature under single instructor in broadcasting's coordinate system
The relative positional relationship of position and multiple instructor in broadcasting's video cameras, position of the available feature in other instructor in broadcasting's camera coordinate systems
Confidence breath；AV object modeling includes combining the information architectures AV objects such as source of sound positioning, face information, feature object and scenario objects
Model；The tracking of AV object includes tracking to multiple AV objects in scene, the status information of upgating object；Movement/appearance
State identification includes that the movement to AV object, posture etc. identify, such as identify stance, the gesture motion etc. of object；It leads
Broadcast Control system includes that movement/gesture recognition and the result of source of sound Activity recognition is combined to determine that instructor in broadcasting's strategy, instructor in broadcasting's camera control are defeated
The corresponding control instruction of instructor in broadcasting's strategy, the video object and scene characteristic information and video output policy etc. out.Wherein, video camera
Control instruction can be used for controlling Pan/Tilt/Zoom camera and carry out PTZ operation, i.e. translation, inclination, zoom operation etc., the video object and field
The information sharing that scape characteristic information can be used between multiple instructor in broadcasting's video cameras, video output policy can be used for controlling single or more
The output policy of a instructor in broadcasting's camera video stream.
The embodiment of the invention provides a kind of camera parameters method of adjustment, instructor in broadcasting's video camera and systems, can be improved and take the photograph
The efficiency of camera parameter adjustment, and promote video camera shooting effect.It is described in detail individually below.
Further, Fig. 2 is referred to, Fig. 2 is a kind of stream of camera parameters method of adjustment provided in an embodiment of the present invention
Journey schematic diagram.Specifically, the method for the embodiment of the present invention can be applied particularly in above-mentioned instructor in broadcasting's video camera.Such as Fig. 2 institute
Show, the camera parameters method of adjustment of the embodiment of the present invention may comprise steps of:
101, the target video object for needing to shoot is determined.
102, according to preset instructor in broadcasting's strategy from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera
Filter out the target video camera for shooting the target video object.
Optionally, the determining target video object for needing to shoot, can be with specifically: obtains binocular camera transmission
Image is shot, includes at least one the video object in the shooting image；Establish the view of at least one the video object described in including
Frequency object model, and target video object is determined from least one described the video object.Further, described according to default
Instructor in broadcasting's strategy filter out from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera for shooting the mesh
The target video camera for marking the video object, can be with specifically: the bat obtained respectively from each video camera in instructor in broadcasting's camera system
It takes the photograph in image and determines the target video object, and obtain the target video object and join in the shooting effect of each video camera
Number；It is determined as the video camera that shooting effect parameter meets default instructor in broadcasting's strategy to be used to shoot the target of the target video object
Video camera.Wherein, one or more instructor in broadcasting's video cameras can be disposed in instructor in broadcasting's camera system, i.e., instructor in broadcasting's camera system can be
Instructor in broadcasting's video camera+instructor in broadcasting's video camera deployment way can also be (such as common PTZ camera shooting of instructor in broadcasting's video camera+common camera
Machine) deployment way.Specifically, may include instructor in broadcasting's camera system correspondence where instructor in broadcasting's video camera in the video object model
Photographed scene in all the video objects.If other video cameras in instructor in broadcasting's camera system also include instructor in broadcasting's video camera,
It also can receive the shooting image that the binocular camera connecting with other instructor in broadcasting's video cameras is sent, carry out the video object model modification,
To acquire the video object model of all the video objects in photographed scene.Wherein, which can be the bat
Take the photograph any one or more the video objects in scene.
Optionally, the shooting effect parameter may include the target video object in the corresponding coordinate system of current camera
Under eye to any one of scenario objects parameter of eye efficacy parameter, hiding relation parameter and shooting area or multinomial.Its
In, the current camera is any video camera in instructor in broadcasting's camera system in addition to the binocular camera, that is,
The current camera can be any instructor in broadcasting's video camera or common Pan/Tilt/Zoom camera in instructor in broadcasting's camera system.
Wherein, the eye may include the target video object relative to the corresponding seat of current camera to eye efficacy parameter
The rotation angle of system is marked, the rotation angle can be the rotation angle according to the target video object in second coordinate system
What the positional relationship of degree and the binocular camera and the current camera demarcated in advance was determined.Specifically, the mesh
Mark the video object can be relative to the rotation angle of the corresponding coordinate system of current camera refers to that the target video object is corresponding
The optical axis angle of face or humanoid subject relative to the current camera (instructor in broadcasting's video camera or common Pan/Tilt/Zoom camera).The angle
It is smaller, it can indicate that face is more presented in a manner of positive face namely eye is better to eye effect, output image effect is better.
Wherein, the hiding relation parameter and the scenario objects parameter can be takes the photograph according to the binocular demarcated in advance
Institute is thrown into the region for the scenario objects that the current camera detects by the positional relationship of camera and the current camera again
State what the imaging plane of current camera was determined.Specifically, then believing using depth such as the region overlapping of two the video objects
The hiding relation determined between two objects is ceased, this can block farther away video pair apart from the closer the video object of binocular camera
As.Output image effect is better when not having hiding relation (hiding relation parameter is smaller).The scene of scenario objects parameter instruction
Object may include fluorescent tube, window, desk etc., and the area of the scenario objects is smaller, number is smaller, then it is better to export image effect；
Conversely, it is poorer then to export image effect.
103, the first three-dimensional coordinate of the target video object is obtained.
Wherein, which can be the target video object under corresponding first coordinate system of target video camera
Three-dimensional coordinate.The target video camera can be configured to above-mentioned instructor in broadcasting's video camera or for common Pan/Tilt/Zoom camera, then the target is taken the photograph
Corresponding first coordinate system of camera can refer to the three-dimensional system of coordinate established using target camera optical center as origin, or with other
Any object of reference is the three-dimensional system of coordinate that origin is established, and the embodiment of the present invention is without limitation.
Optionally, instructor in broadcasting's video camera can be connected with preset binocular camera.It is then described to obtain the target view
First three-dimensional coordinate of frequency object, can be with specifically: obtain that the binocular camera connecting with instructor in broadcasting's video camera transmits the
Two three-dimensional coordinates；According to the positional relationship of the binocular camera and the target video camera demarcated in advance, by described second
Three-dimensional coordinate is converted to the first three-dimensional coordinate.Further alternative, second three-dimensional coordinate can be the binocular camera
By two-dimensional coordinate in the left view and right view of the binocular camera of the target video object that obtains respectively and
What the inside and outside parameter evidence of the binocular camera obtained was calculated.Wherein, which is the target video
Three-dimensional coordinate of the object under corresponding second coordinate system of the binocular camera, corresponding second coordinate of the binocular camera
System can refer to the three-dimensional system of coordinate established using binocular camera optical center as origin, or build using any other object of reference as origin
Vertical three-dimensional system of coordinate.The two-dimensional coordinate can be specially left view and right view of the target video object in the binocular camera
Corresponding pixel coordinate in figure.
It, can be in advance to positional relationship, instructor in broadcasting's video camera and the binocular camera between binocular camera in specific embodiment
Between positional relationship and instructor in broadcasting's camera system in multimachine position video camera between positional relationship demarcate.Wherein,
The parameter that binocular camera system calibrating obtains can be used for calculating three of the video object under the corresponding coordinate system of binocular camera
Tie up coordinate；Positional relationship calibration between instructor in broadcasting's video camera and binocular camera can be used for calculating the video object in instructor in broadcasting's video camera
Three-dimensional coordinate under coordinate system；And the parameter of the positional relationship calibration between the video camera of multimachine position can be used for calculating multimachine position portion
When affixing one's name to scene, three-dimensional coordinate of the video object under the camera coordinate system of each seat in the plane, in order to carry out coordinate conversion.The multimachine
The deployment way of the video camera of position can be above-mentioned instructor in broadcasting's video camera+instructor in broadcasting's video camera deployment way, or take the photograph for instructor in broadcasting
The deployment way of the Pan/Tilt/Zoom camera of camera+common.Wherein, each instructor in broadcasting's video camera can be described as a seat in the plane, lead when using multiple
Broadcast video camera carry out cooperation shooting when, can therefrom determine a host bit, then remaining be from seat in the plane, as the instructor in broadcasting from seat in the plane
Video camera can be by the information registerings to host bit such as the IP of itself, so that host bit, which can be realized, carries out pipe from seat in the plane to multiple
Reason.Specifically, calibration process is briefly described below.Wherein, binocular camera includes left video camera and right video camera, a left side
The image that video camera obtains can be described as left view, and the image that right video camera obtains can be described as right view.Then wherein single camera
Imaging (projection) model can be described by following formula:
X=PX=K [R | t] X
As shown in Figure 3a, x is that some point in scene (i.e. the video object, concretely the corresponding characteristic point of the video object) exists
Pixel coordinate under image coordinate system is two-dimensional coordinate；X is that some in scene puts the position coordinates under world coordinate system；P
For 3 × 4 projection matrix.PX refers to P × X.Wherein, the video camera internal reference matrix that K is 3 × 3, can indicate are as follows:
Wherein, fx,fyFor the equivalent focal length in the direction x and y, cx,cyFor the image coordinate of optical center, s is skew deformation coefficient
(sensor and optical axis out of plumb cause, and usual very little can be ignored during the calibration process).
In addition, R and t is to join outside video camera, it is expressed as 3 × 3 spin matrix and 3 × 1 translation vector, following institute
Show:
R=[r1 r2 r3]
T=[t1 t2 t3]T
Wherein, r1,r2,r3For in spin matrix 3 × 1 column vector.
Due to factors such as the optical characteristics of camera lens, the manufacture of image sensitive device and installations, video camera is practical to be clapped
The image taken the photograph is not preferably, can have distortion, therefore can model to pattern distortion, to obtain ideal image.Tool
Body, the model of camera review distortion can describe according to the following formula:
Wherein, xp,ypFor location of pixels after correction, xd,ydTo correct preceding pixel position, k1,k2,k3For radial distortion system
Number, p1,p2For tangential distortion coefficient.
Based on the imaging model of above-mentioned monocular-camera, when the known world coordinate system transformation to left camera coordinate system and the right side
The spin matrix R of camera coordinate system1And R2And translation vector t1And t2When, then it is opposite between available binocular camera
Outer ginseng, including spin matrix R and translation vector T:
It should be understood that in embodiments of the present invention, positional relationship and instructor in broadcasting's video camera such as PTZ between binocular camera are taken the photograph
Positional relationship between camera and binocular camera is fixed and invariable, both calibration can be completed before factory, i.e., this two
Kind data obtained by calibrating such as inside and outside parameter evidence is fixed and invariable.Optionally, in embodiments of the present invention, the calibration of video camera
Kinds of schemes can be used, such as the plane reference method (also known as " Zhang Shi standardization ") of Zhang, distortion parameter, which calculates, to be used
The method of Brown, does not repeat herein.
Further, by above-mentioned binocular camera calibration principle it is found that for example more instructor in broadcasting positions for video camera of the video camera of multimachine position
The essence for setting the calibration of relationship is the relatively outer ginseng asked between adjacent two-by-two instructor in broadcasting's video camera, according to adjacent instructor in broadcasting's video camera it
Between relatively outer ginseng calculate the outer ginseng between any two instructor in broadcasting's video camera, to obtain between any two instructor in broadcasting's video camera
Positional relationship.Biggish shooting overlapping region is needed between instructor in broadcasting's video camera two-by-two when more instructor in broadcasting's video cameras are disposed, it is multiple
Seat in the plane, which is constituted, to be similar to around multi-camera system, spin matrix and translation vector of i-th of video camera with respect to j-th of video camera
Are as follows:
Ri,i-1Ri-1,i-2...Rj+1,j
Ri,i-1Ri-1,i-2...Rj+2,j+1Tj+Ri,i-1Ri-1,i-2...Rj+3,j+2Tj+1+...+Ri,i-1Ti-2+Ti-1
Wherein, Ri,i-1Ri-1,i-2...Rj+1,jIndicate Ri,i-1×Ri-1,i-2×...×Rj+1,j.Due to instructor in broadcasting's camera unit
When administration, the position of camera used for positioning is changed according to actual deployment scene, therefore more on different instructor in broadcasting's brackets
Positional relationship between instructor in broadcasting's video camera can not be demarcated in advance before equipment is dispatched from the factory, then can be carried out when instructor in broadcasting's video camera is disposed
Field calibration.
Further, it is assumed that the video camera in instructor in broadcasting's camera system is instructor in broadcasting's video camera, and each instructor in broadcasting images
Machine is connected with binocular camera.Fig. 3 b is referred to, Fig. 3 b is a kind of calibration of more instructor in broadcasting's video cameras provided in an embodiment of the present invention
Schematic diagram of a scenario.As shown in Figure 3b, when being demarcated, between two adjacent instructor in broadcasting's video cameras or binocular camera and instructor in broadcasting
Local area network (Local Area Network, referred to as " LAN ") or Wireless Fidelity (Wireless can be passed through between video camera
Fidelity, referred to as " Wi-Fi ") network is communicated, including transmission calibrating template image and calibrating parameters etc., transmission association
Multiple network agreement can be used by discussing, for example, hypertext transfer protocol (HyperText Transfer Protocol, referred to as
"HTTP").Specifically, can be in advance each instructor in broadcasting's video camera and the video camera (binocular camera) being connect with instructor in broadcasting's video camera point
With global ID number.Such as it is initial position that some video camera, which may be selected, for example the video camera work of Far Left or rightmost may be selected
For initial position, the ID number of other cameras is by counterclockwise or incremental clockwise.One group is chosen in all video cameras to take the photograph
Camera participates in calibration, and the principle of selection can be to ensure that the maximum of the overlapping region between adjacent camera.As shown in Figure 3b, it is assumed that
Be deployed with 3 seats in the plane D1, D2, D3 in present filming scene, each seat in the plane respectively include instructor in broadcasting's video camera (be denoted as respectively PTZ0,
PTZ1 and PTZ2) and binocular camera (binocular camera includes left and right two video cameras, be denoted as respectively C0, C1, C2, C3,
C4,C5).Assuming that ID number is selected to be demarcated for the video camera of C0, C2 and C4, and select one of instructor in broadcasting's video camera as mark
Devise a stratagem calculates equipment, such as instructor in broadcasting's video camera of above-mentioned host bit.Further, can by direction from left to right or from right to left into
Outer ginseng calibration between capable video camera two-by-two obtains the relatively outer ginseng between two video cameras.Optionally, in each instructor in broadcasting's video camera
A video camera relative positional relationship table can be safeguarded, as shown in following table one.Wherein, calibration can increase or update therein one newly every time
A list item, each list item are uniquely determined by two video camera ID number.
Table one
| List item | Video camera ID1 | Video camera ID2 | Relatively outer ginseng |
| 1 | C0 | C2 | R02,T02 |
| 2 | C2 | C4 | R24,T24 |
| ... | ...... | ... | ...... |
After the completion of calibration, instructor in broadcasting's video camera as calibrated and calculated equipment can be sent out the positional relationship table by network
Other all instructor in broadcasting's video cameras are given to save.Further, according to the outer of the positional relationship table and binocular camera
Join (the outer ginseng can demarcate completion before factory), arbitrarily video camera (including binocular two-by-two can be calculated in the calibration scene
Between video camera, between binocular camera and Pan/Tilt/Zoom camera and between Pan/Tilt/Zoom camera) positional relationship.
As an example it is assumed that instructor in broadcasting's video camera D3 in Fig. 3 b is calibrated and calculated equipment, instructor in broadcasting video camera D1 and D2 are to want
Instructor in broadcasting's video camera D3, can be set as being used for the host bit demarcated by the video camera for demarcating positional relationship, other video cameras be set as from
Seat in the plane, and calibration is initiated by instructor in broadcasting's video camera D3.It before calibration, need to ensure to be interconnected by network between video camera, need
Overlapping region can be taken between the video camera to be demarcated, and there is calibrating template (such as checkerboard pattern) in overlapping region
Deng.When needing to demarcate, instructor in broadcasting's video camera D3 starts calibration process, and sends image acquisition commands and give instructor in broadcasting video camera D1, acquires
The ID number (C4 or C5) of ID number (i.e. D1) and the binocular camera for needing to acquire in order comprising instructor in broadcasting's video camera D1.Instructor in broadcasting takes the photograph
Camera D1 receives the Image Acquisition of the laggard rower solid plate of the acquisition, and the image data of acquisition is transferred to instructor in broadcasting
Video camera D3.Similar, instructor in broadcasting's video camera D3 acquires the calibrating template of the shooting of the binocular camera on instructor in broadcasting's video camera D2
Image.If the binocular camera that need to be demarcated is located on instructor in broadcasting's video camera D3, instructor in broadcasting's video camera D3 can directly acquire the binocular and take the photograph
The calibrating template image of camera.After obtaining the calibrating template image for needing the video camera demarcated, instructor in broadcasting's video camera D3 can be right
The calibrating template image of two video cameras carries out X-comers detection, if two images are capable of detecting when all gridiron patterns
Angle point then can be shown that and acquire successfully；Otherwise the two images be can drop, the image is resurveyed.Further, change can be passed through
The position of calibrating template cyclically obtains and several calibrating template images for two video cameras demarcated is needed to be saved in instructor in broadcasting's camera shooting
In machine D3, after meeting the quantitative requirement of calibrating template image, instructor in broadcasting's video camera D3 can carry out camera calibration, each camera shooting
The internal reference of machine has been demarcated before factory, therefore can be used as the input initial value of calibration.Two camera shootings are obtained after the completion of calibration
Relatively outer ginseng R and T between machine, and calculate and throw whether (shadow) error is less than preset threshold value again, if throwing error again is greater than threshold value,
It then can be shown that calibration failure；Otherwise, then it can be shown that and demarcate successfully.Instructor in broadcasting's video camera D3 can be based on being calculated after completing calibration
Video camera relative positional relationship update and positional relationship table and be sent to other instructor in broadcasting's video cameras.
Further, it is realizing between positional relationship, instructor in broadcasting's video camera and the binocular camera between binocular camera
Positional relationship and more instructor in broadcasting's video cameras between positional relationship calibration after, can be to the coverage of instructor in broadcasting's video camera
Interior the video object is positioned, its three dimensional local information is obtained, and is determined suitably with the three dimensional local information according to acquisition
Instructor in broadcasting's video camera position, and parameter adjustment is carried out to instructor in broadcasting's video camera according to the corresponding instructor in broadcasting's strategy of the three dimensional local information,
It controls instructor in broadcasting's Camera Positioning and carries out the video object shooting to suitable position.It wherein, include binocular to the positioning of the video object
Three-dimensional localization between the video camera of video camera three-dimensional localization, single instructor in broadcasting's video camera such as Pan/Tilt/Zoom camera positioning and multimachine position.
Specifically, during the three-dimensional localization of binocular camera, using the stereo-picture that binocular camera is shot, meter
Calculation obtains depth location information of some observation point in camera coordinate system in scene, so that it is determined that the three-dimensional position of the observation point
Confidence breath.Which is identical with the principle of human eye perceived depth distance, referred to as binocular camera ranging.As shown in Figure 3c, it mentions
A kind of three-dimensional localization schematic diagram of binocular camera has been supplied, has briefly been situated between to the range measurement principle of the binocular camera system below
It continues.Wherein, P is the observation point under world coordinate system, by the video camera shooting imaging of left and right two.Wherein, the P point is in left video camera
Position in physical coordinates system is XL, YL, ZL, it is x in the imaging point location of pixels coordinate of left viewl, yl；In right video camera physics
Position in coordinate system is XR, YR, ZR, it is x in the imaging point location of pixels coordinate of right viewr, yr, it is assumed that the phase of left and right cameras
Externally ginseng is R, T；The focal length of left and right cameras is respectively as follows: fl, fr.According to binocular camera model, it is known that left and right cameras at
As the physical coordinates positional relationship of model and left and right cameras are as follows:
It can be derived by according to above-mentioned formula:
Wherein, xl, yl, xr, yrValue can be obtained by images match, fl, fr, R, T can pass through binocular camera mark
Surely it obtains, therefore X can be calculatedL, YL, ZLAnd XR, YR, ZRValue, so that it is determined that observation point is in binocular camera pair in scene
Three-dimensional coordinate under the coordinate system answered.
Further, during instructor in broadcasting's video camera such as Pan/Tilt/Zoom camera three-dimensional localization, the basic mesh of Pan/Tilt/Zoom camera positioning
Be physical coordinates of some the known target in Pan/Tilt/Zoom camera coordinate system, how by rotate Pan/Tilt/Zoom camera make the mesh
Specific pixel coordinate position of some point location of target into image.Physical coordinates of the target in Pan/Tilt/Zoom camera coordinate system
It can be taken the photograph by three-dimensional position of the target in binocular camera coordinate system and binocular camera obtained by calibrating and PTZ
Positional relationship between camera is calculated.Fig. 3 d is referred to, is a kind of Pan/Tilt/Zoom camera rotating mould provided in an embodiment of the present invention
Type schematic diagram.As shown in Figure 3d, it is assumed that it is desirable that the position coordinates that target point P occurs are x0,y0, the physical coordinates position of target point P
For X, Y, Z, the pixel coordinate position on imaging plane is xc,yc, then can be rotated rotating around X-axis and Y-axis, make the pixel position of point P
It sets and is overlapped with target position, then the rotation angle of Pan (translation) and Tilt (inclination) operation can be modeled by following equation:
Since Pan/Tilt/Zoom camera is zoom camera, it is therefore desirable to obtain the internal references such as zoom magnification Z and focal length, distortion factor
Functional relation.For example, using fitting of a polynomial zoom magnification Z and focal length fx, fyRelationship, obtain following relationship:
fx=a0+a1Z+a2Z2+...anZn
fy=b0+b1Z+b2Z2+...bnZn
Specifically, calibration obtains video camera internal reference under different Z values, corresponding f is calculatedx, fyAnd distortion factor,
And coefficient is fitted using least square method.The internal references such as other distortion factors can also be handled according to similar approach.It obtains not
After internal reference with video camera under Z value, Δ p, the value of Δ t can be calculated according to Pan/Tilt model formation.
It is further alternative, under more instructor in broadcasting's camera scenes, it can also obtain the binocular connecting with other instructor in broadcasting's video cameras
The shooting image that video camera is sent is updated the video object model after carrying out the video object matching.Then obtained from video camera
Shooting image in determine the target video object, can be with specifically: according to the binocular camera demarcated in advance and
Second three-dimensional coordinate is converted to third three-dimensional coordinate by the positional relationship of current camera；Judge the target video pair
As under the third three-dimensional coordinate region and the video object that detects of the current camera the three of the video object
Whether the overlapping area for tieing up the region under coordinate is more than preset area threshold；If being more than, which is determined as institute
State target video object.Namely the video object successful match.Wherein, the third three-dimensional coordinate is that the target video object exists
Three-dimensional coordinate under the corresponding third coordinate system of the current camera, the current camera are in instructor in broadcasting's camera system
Any video camera in addition to the binocular camera.For example, under more instructor in broadcasting's camera scenes, which can be with
For other binocular cameras in addition to the binocular camera of host bit.
In specific embodiment, the purpose of multimachine digital video object three-dimensional localization is to be imaged according to the video object in some instructor in broadcasting
It is calculated in other instructor in broadcasting's video camera binocular cameras or some PZT video camera in three-dimensional coordinate in machine binocular coordinate system
Three-dimensional coordinate in coordinate system.Some known observation point (i.e. the video object specifically can be a certain characteristic point of the video object)
Coordinate vector in video camera D1 is X1And outer ginseng R of the video camera D2 relative to video camera D121, T21(pass through binocular camera shooting
Machine is demarcated to obtain), coordinate vector X of the observation point in video camera D2 can be calculated2:
X2=R21X1+t21
It is a kind of object matching schematic diagram of a scenario provided in an embodiment of the present invention specifically, referring to Fig. 4 a.Multimachine position view
Frequency object three-dimensional localization is determined for the corresponding relationship of multiple the video objects.As shown in fig. 4 a, three machines are deployed in scene
Video camera D1, D2 and D3 of position have O in scene1、O2And O3Totally 3 participants.It further, as shown in Figure 4 b, is in Fig. 4 a
One group of the video object image, be participant O1Imaging in instructor in broadcasting's video camera D1, D2 and D3 of different perspectives.For
Participant O1, the binocular camera in the D1 of seat in the plane detects to obtain video object using Face datection scheduling algorithm11, then using double
Lens camera three-dimensional location obtains three-dimensional position of the object in D1 binocular coordinate system.Similarly D2 and D3 can also be detected
Video object12And VO13, and the three-dimensional coordinate of the object under D2, D3 binocular coordinate system is calculated.In multimachine digital video
When object three-dimensional localization, using D1, the positional relationship between D2 and D3 is demarcated, by video object under D1 coordinate system11Three
Dimension position is transformed under D2 and D3 coordinate system.And detect its overlapping region.If detecting VO after conversion11Three-dimensional position and
VO12、VO13Position overlapping region be more than certain area threshold, i.e., it is believed that VO11、VO12And VO13For same the video object, depending on
The success of frequency object matching.Further, if there are the closer multiple the video objects of neighbor distance in image, position weight is utilized merely
It closes region and determines that the corresponding relationship of the video object may result in matching error.It as a result, can be further combined with the figure of the video object
The accuracy of determining corresponding relationship is promoted as information and by matching algorithm.Wherein, which may include template matching
Algorithm etc..Such as using template matching algorithm, the video object that the binocular camera of some seat in the plane such as host bit is detected
Two dimensional image as known template, and the video object that the binocular camera of other seats in the plane is detected carries out one by one with it
Match, the most matched object of the video object is found by difference of two squares matching, relevant matches scheduling algorithm, to set up pair of object
It should be related to.
It further, can be to view after the three-dimensional localization and Pan/Tilt/Zoom camera three-dimensional localization for determining binocular camera
The tracking of frequency object detection and scene modeling.Wherein, the video object detection/tracking purpose is present in building and description scene
The video object, and these objects are tracked and identified.The video object includes participant's object and scenario objects, such as lamp
Pipe, window, conference table etc..System needs cyclically to handle the image data of the binocular camera of input, including carries out
Face datection and matching, humanoid detection and matching, Moving Objects detection and matching, scenario objects detection and matching etc., to video
Object establishes model and updates model parameter, to carry out the modeling of entire photographed scene according to the object model that detection obtains.
Modeling obtained model of place can be used for subsequent Object identifying and the processing of instructor in broadcasting's strategy.Wherein, Face datection can be used for examining
Ranging is from closer the video object, such as the closer participant of detecting distance, for farther away region, due to the smaller nothing of face area
Method detects well, then the method for humanoid or Moving Objects detection can be used.The Face datection can acquire face video
The parameters of object, two-dimensional coordinate, center point coordinate, rectangular area, face including face circumscribed rectangular region are around coordinate
The position of the organs such as eyes, nose and mouth in the rotation angle (representing face deflection, pitching and the degree of rotation) of axis, face
The parameters such as set.
Further, after detecting the video object in each frame image, it is also necessary to the video object in sequence of frames of video
It is tracked, to set up the corresponding relationship of the video object in the time domain.The widely used Video object tracking of tradition at present
Algorithm includes the template matching based on gray scale, MeanShift, CamShift, Kalman filter algorithm etc..Wherein, the video object
Matching can be applied in binocular camera, utilize the video object detected in a camera review in the binocular camera
Corresponding video image region is found, in another camera review so as in the matching area of the video object in region
Carry out the calculating of characteristic matching and three-dimensional coordinate.The matching algorithm of the video object is similar with track algorithm, can be used based on gray scale
Template matching and MeanShift scheduling algorithm.
In embodiments of the present invention, the video object can be indicated by its feature, and usually used feature includes spy
Levy point, image texture, histogram information etc..This feature detection and matching can carry out in the video object region detected, from
And the three dimensional local information of the video object can be calculated according to characteristic point information, i.e. three-dimensional coordinate, and can according to texture information and
The tracking of histogram information progress the video object.Wherein, this feature point is main characteristic type, and feature point detection algorithm includes
Harris Corner Detection, SIFT feature detect scheduling algorithm.Further, characteristic matching is for establishing the same view of binocular camera
The corresponding relationship of frequency characteristics of objects, characteristic point can be matched using matching algorithms such as FLANN algorithm, KLT optical flow methods, image
Texture can be matched using gray scale template matching scheduling algorithm, and histogram can be using the progress of Histogram Matching scheduling algorithm
Match.To sum up, the characteristic information and above-mentioned binocular camera three-dimensional location obtained according to matching, then can be calculated
The three-dimensional coordinate of the video object feature under single instructor in broadcasting's video camera three-dimensional system of coordinate, so as to position in three dimensions and with
Some the video object of track.
Further, it detects and matches according to the video object, feature detection and matching algorithm obtain data and video pair
As three-dimensional position calculate as a result, the model of multiple the video objects can be set up in single instructor in broadcasting's camera coordinate system, and can
Model data is updated by face, humanoid and motion detection track algorithm.Specifically, can be each the video object model
A unique ID number is distributed, the data in model represent the attribute of the video object.For example, for Moving Objects model,
Data in model may include object ID, boundary rectangle two-dimensional coordinate, the three-dimensional coordinate of characteristics of objects point, moving region texture number
According to attributes such as, histogram datas.When Moving Objects position changes, attribute can be according to above-mentioned detection and matching algorithm
Output is refreshed, but the ID of object is remained unchanged.The foundation of face and humanoid subject are similar with Moving Objects model, herein not
It repeats.
It should be understood that in the application scenarios of multimachine position video can be exchanged by network communication between multiple instructor in broadcasting's video cameras
Object model data, after single instructor in broadcasting's video camera has obtained the video object model data of other instructor in broadcasting's video cameras, using upper
The more instructor in broadcasting's video camera three-dimensional localizations and the matched algorithm of the video object stated establish the corresponding relationship of the video object model, thus
To instructor in broadcasting's strategy to entire scene.Network communication protocol when communication can use standard agreement such as http protocol, can also be with
Using custom protocol, the data of the video object model are by according to certain format such as extensible markup language (eXtensible
Markup Language, referred to as " XML ") format is formatted, is packaged and transmits.By being regarded to multiple instructor in broadcasting's video cameras
The matching and merging of frequency object model, single instructor in broadcasting's video camera can set up the model of entire photographed scene.It is wrapped in model of place
The model for having contained multiple the video objects reflects the feature and distribution situation in three dimensions of the video object.Instructor in broadcasting's camera shooting
Machine needs to safeguard model of place, including newly-increased, deletion object model and object model attributes.Such as it is new when having in scene
Participant when occurring object mould is added after establishing object model when binocular camera detects new face or humanoid subject
In type set；After thering is participant to leave in scene, the model of the object is deleted；After participant position changes, update pair
Answer the parameter of object model.To formulate instructor in broadcasting's strategy according to newest the video object model, the video camera of best seat in the plane is selected
It is shot.
104, the camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate,
And the video image after output adjustment camera parameter.
In specific embodiment, obtain including the video object model of all the video objects and determining mesh in foundation (update)
It, can be according to instructor in broadcasting's video camera machine of the one or more best shooting effects of preset instructor in broadcasting's policy selection after marking the video object
Position, such as determine have to the scenario objects parameter of eye efficacy parameter, hiding relation parameter and shooting area according to eye
The video camera of better shooting effect.Specifically, the eye needs according to face/humanoid subject relative to Pan/Tilt/Zoom camera eye effect
Optical axis angle determine that angle is smaller, face is more presented in a manner of positive face, and eye is better to eye effect.Specifically, can
What is obtained by face/humanoid detection algorithm is that 3-D walls and floor centered on face/humanoid is sat relative to binocular camera
The rotation angle (deflection, pitching and rotation angle) of system is marked, and utilizes the formula of coordinate system conversion between aforementioned video camera,
By face/humanoid rotation angle relative to binocular camera, the rotation angle relative to Pan/Tilt/Zoom camera is converted to.In conversion process
Need using the outer ginseng between the aforementioned binocular camera and Pan/Tilt/Zoom camera demarcate, and difference seats in the plane binocular camera it
Between outer ginseng, determine the eye to eye efficacy parameter.And an eye can be set up further directed to each the video object, eye is imitated
The Pan/Tilt/Zoom camera priority query of fruit, eye have higher priority to the better video camera of eye effect.
It further, can be according to video camera projection equation, it is known that some instructor in broadcasting when obtaining the hiding relation of the video object
The region (such as boundary rectangle) for the video object that video camera detects, and taken the photograph using the single instructor in broadcasting's video camera binocular demarcated
Outer ginseng between outer ginseng between camera and Pan/Tilt/Zoom camera, and the binocular camera of different seats in the plane, which is thrown into respectively again
On the imaging plane of seat in the plane Pan/Tilt/Zoom camera.If the region overlapping of two the video objects, can determine that the two are right using depth information
As hiding relation, i.e., can block farther away the video object apart from the closer the video object of binocular camera.As a result, for every
A the video object can establish the Pan/Tilt/Zoom camera priority query an of hiding relation, and the video camera not blocked has more
High priority.
Further, other than the video object detection based on people, system is also to other videos pair interested in scene
As (scenario objects) are detected, such as fluorescent tube, window and conference table etc..The detection of these objects can be using based on image face
Color and the algorithm of edge feature etc..For example, being detected for fluorescent tube, the edge of fluorescent tube can be first gone out with Canny operator extraction, is obtained
To its long linear feature, adjacent area is then detected again with the presence or absence of overexposure pixel region (light emitting region), according to the two spies
Sign can detecte out fluorescent tube object, obtain the coordinate of its boundary rectangle.The detection of window is similar with fluorescent tube detection, can pass through side
Edge detects to obtain quadrangle feature, then according in quadrangle whether there is certain area overexposure pixel region to determine whether
For window.The edge feature that conference table also can use in image is detected to obtain.Then when obtaining the scenario objects parameter,
It can be according to video camera projection equation, it is known that the region for the scenario objects that some instructor in broadcasting's video camera detects utilizes the list demarcated
It is outer between outer ginseng between a instructor in broadcasting's video camera binocular camera and Pan/Tilt/Zoom camera, and the binocular camera of different seats in the plane
Ginseng, which is thrown into again on the imaging plane of each seat in the plane Pan/Tilt/Zoom camera.For objects such as fluorescent tube and windows, it will usually exist
The overexposure region of large area causes the automatic exposure effect of video camera to be deteriorated, and scene is dimmed, and the scenario objects such as desk can exist
The color regions such as the red or yellow of large area lead to the automatic white balance colour cast of video camera, these scenario objects should be as far as possible
It avoids the occurrence of in the picture.The PTZ camera shooting that whether can take the scenario objects for being unfavorable for image effect can be set up as a result,
Machine priority query takes the lesser video camera of scenario objects probability with higher priority.
Further, instructor in broadcasting's video camera of instructor in broadcasting's video camera such as host bit is according to the image effect parameter of acquisition, in conjunction with pre-
If instructor in broadcasting strategy, priority query is established to the Pan/Tilt/Zoom camera of each seat in the plane, that is, can determine that the video camera for needing selection.Tool
Body, the one or more the video objects for needing to shoot, i.e. target video object can be predefined out, such as according to auditory localization knot
The video object spoken that fruit determines, to shoot the feature of the video object；Or using AutoFrame strategy, need by
When all the video objects are as the target video object, then it can adjust Pan/Tilt for all the video objects in scene and be included in bat
Range is taken the photograph, and adjusting Zoom makes object have suitable size, etc..For target reference object, eye effect is joined using eye
The Pan/Tilt/Zoom camera priority query of number, hiding relation parameter and scenario objects parameter can determine according to certain instructor in broadcasting's strategy
A comprehensive Pan/Tilt/Zoom camera priority query out.Instructor in broadcasting's strategy can be calculated automatically by system, or preparatory by user
Setting, the embodiment of the present invention is without limitation.
As an example it is assumed that eye Pan/Tilt/Zoom camera best to eye effect, unobstructed is paid the utmost attention to, if there is meeting this
Multiple video cameras of part, reselection scenario objects parameter are most preferably the best Pan/Tilt/Zoom camera of image effect as target video camera
It is shot.After completing Pan/Tilt/Zoom camera selection, host bit can be according to the three-dimensional coordinate of the target video object to selected
The PTZ parameter of Pan/Tilt/Zoom camera is adjusted, to obtain optimal image effect as far as possible.Such as in tone tracking, shooting
Avoid taking the object of the influence brightness of image effect such as fluorescent tube, window when participant's feature；In AutoFrame, adjust
Avoid taking the white balance effect, etc. of large area desk object influences image when Zoom size.
Further, the exportable selected Pan/Tilt/Zoom camera video image of host bit (instructor in broadcasting's video camera of host bit) or
ID.Optionally, for more instructor in broadcasting's camera chains of support video cascade, host bit can directly export the figure of selected video camera
Picture；For the more instructor in broadcasting's camera chains exported by video matrix, communication interface (such as serial ports or net can be passed through by host bit
Mouthful) export the ID of selected Pan/Tilt/Zoom camera to video matrix, by the switching of video matrix completion camera review.
It in embodiments of the present invention, can be after the target video object for determining to need to shoot, according to preset instructor in broadcasting
Strategy is filtered out from each video camera of instructor in broadcasting's camera system shoots the optimal target video camera of target video object effect, and
Three-dimensional coordinate of the target video object under the corresponding coordinate system of target video camera is acquired, to control target camera shooting
Machine carries out camera parameters adjustment, and the video after output adjustment camera parameter according to the three-dimensional coordinate of the target video object
Image enables instructor in broadcasting's camera system to be based on three-dimensional coordinate detection and preset instructor in broadcasting strategy, to improve the video object detection
With the precision of tracking, while the efficiency of camera parameters adjustment is improved, and effectively improves the shooting effect of video camera.
Fig. 5 is referred to, Fig. 5 is a kind of structural schematic diagram of parameter adjustment controls provided in an embodiment of the present invention.Specifically,
The described device of the embodiment of the present invention can be specifically set in above-mentioned instructor in broadcasting's video camera, as shown in fig. 6, the embodiment of the present invention
The parameter adjustment controls may include object determination unit 10, selecting unit 20, acquiring unit 30 and parameter adjustment unit
40.Wherein,
The object determination unit 10, the target video object shot for determining needs.
The selecting unit 20, for imaging system according to the instructor in broadcasting where preset instructor in broadcasting's strategy from instructor in broadcasting's video camera
The target video camera for shooting the target video object is filtered out in each video camera of system.
Optionally, the shooting effect parameter may include the target video object in the corresponding coordinate system of current camera
Under eye to any one of scenario objects parameter of eye efficacy parameter, hiding relation parameter and shooting area or multinomial.Its
In, the current camera is any video camera in instructor in broadcasting's camera system in addition to the binocular camera.
Wherein, the eye may include the target video object relative to the corresponding seat of current camera to eye efficacy parameter
The rotation angle of system is marked, the rotation angle can be the rotation angle according to the target video object in second coordinate system
What the positional relationship of degree and the binocular camera and the current camera demarcated in advance was determined.
Wherein, the hiding relation parameter and the scenario objects parameter can be takes the photograph according to the binocular demarcated in advance
Institute is thrown into the region for the scenario objects that the current camera detects by the positional relationship of camera and the current camera again
State what the imaging plane of current camera was determined.
The acquiring unit 30, for obtaining the first three-dimensional coordinate of the target video object.
Wherein, which can be the target video object under corresponding first coordinate system of target video camera
Three-dimensional coordinate.The target video camera can be above-mentioned instructor in broadcasting's video camera or be common Pan/Tilt/Zoom camera, then the target video camera
Corresponding first coordinate system can refer to the three-dimensional system of coordinate established using target camera optical center as origin, or with any other
Object of reference is the three-dimensional system of coordinate that origin is established, and the embodiment of the present invention is without limitation.
The parameter adjustment unit 40, for the camera parameter of the target video camera to be adjusted to and first three-dimensional
The corresponding camera parameter of coordinate, and the video image after output adjustment camera parameter.
Optionally, the acquiring unit 30 can be specifically used for:
Obtain the second three-dimensional coordinate of the binocular camera transmission connecting with instructor in broadcasting's video camera, the described second three-dimensional seat
It is designated as three-dimensional coordinate of the target video object under corresponding second coordinate system of the binocular camera；
It is three-dimensional by described second according to the positional relationship of the binocular camera and the target video camera demarcated in advance
Coordinate is converted to the first three-dimensional coordinate.
Further alternative, second three-dimensional coordinate can be the binocular camera and pass through the target obtained respectively
The binocular camera of two-dimensional coordinate and acquisition of the video object in the left view and right view of the binocular camera
Inside and outside parameter according to being calculated.Wherein, which is the target video object in the binocular camera
Three-dimensional coordinate under corresponding second coordinate system, corresponding second coordinate system of the binocular camera can refer to binocular camera
Optical center is the three-dimensional system of coordinate that origin is established, or the three-dimensional system of coordinate established using any other object of reference as origin.The two dimension
Coordinate can be specially the target video object corresponding pixel coordinate in the left view and right view of the binocular camera.
Optionally, the object determination unit 10 can be specifically used for:
The shooting image of binocular camera transmission is obtained, includes at least one the video object in the shooting image；
The video object model of at least one the video object described in including is established, and from least one described the video object
Determine target video object；
The selecting unit 20 can be specifically used for:
The target video is determined from the shooting image that each video camera in instructor in broadcasting's camera system obtains respectively
Object, and the target video object is obtained in the shooting effect parameter of each video camera；
It is determined as the video camera that shooting effect parameter meets default instructor in broadcasting's strategy to be used to shoot the target video object
Target video camera.
Further alternative, the execution of selecting unit 20 determines the target from the shooting image that video camera obtains
The concrete mode of the video object can be with are as follows:
According to the positional relationship of the binocular camera and current camera demarcated in advance, by second three-dimensional coordinate
Be converted to third three-dimensional coordinate, wherein the current camera be instructor in broadcasting's camera system in except the binocular camera with
Outer any video camera, the third three-dimensional coordinate are that the target video object is sat in the corresponding third of the current camera
Three-dimensional coordinate under mark system；
Judge that region and the current camera of the target video object under the third three-dimensional coordinate detect
The overlapping area in region of the video object under the three-dimensional coordinate of the video object whether be more than preset area threshold；
If being more than, which is determined as the target video object.
It in embodiments of the present invention, can be after the target video object for determining to need to shoot, according to preset instructor in broadcasting
Strategy is filtered out from each video camera of instructor in broadcasting's camera system shoots the optimal target video camera of target video object effect,
And three-dimensional coordinate of the target video object under the corresponding coordinate system of target video camera is acquired, it is taken the photograph with controlling the target
Camera carries out camera parameters adjustment, and the view after output adjustment camera parameter according to the three-dimensional coordinate of the target video object
Frequency image enables instructor in broadcasting's camera system to be based on three-dimensional coordinate detection and preset instructor in broadcasting strategy, to improve the video object inspection
The precision surveyed and tracked, while the efficiency of camera parameters adjustment is improved, and effectively improve the shooting effect of video camera.
Fig. 6 is referred to, Fig. 6 is a kind of structural schematic diagram of instructor in broadcasting's camera system provided in an embodiment of the present invention.Specifically,
Instructor in broadcasting's camera system of the embodiment of the present invention may include the first video camera 1 and at least one second video camera 2, described first
Video camera 1 includes instructor in broadcasting's video camera 11 and binocular camera 12, between instructor in broadcasting's video camera 11 and the binocular camera 12,
It can be connected by wireline interface or wireless interface between first video camera 1 and second video camera 2；Wherein,
Instructor in broadcasting's video camera 11, the target video object shot for determining needs, and according to preset instructor in broadcasting's strategy
The target video camera for shooting the target video object is filtered out from the video camera of instructor in broadcasting's camera system；
The binocular camera 12, for obtaining the second three-dimensional coordinate of the target video object, and by described second
Three-dimensional coordinate is transferred to instructor in broadcasting's video camera 11；Wherein, second three-dimensional coordinate is the target video object described
Three-dimensional coordinate under corresponding second coordinate system of binocular camera 12；
Instructor in broadcasting's video camera 11, second three-dimensional coordinate transmitted for receiving the binocular camera 12；According to
The positional relationship of the binocular camera 12 and the target video camera demarcated in advance, second three-dimensional coordinate is converted to
First three-dimensional coordinate；The camera parameter of the target video camera is adjusted to camera shooting ginseng corresponding with first three-dimensional coordinate
Number, and the video image after output adjustment camera parameter；Wherein, first three-dimensional coordinate is the target video object in institute
State the three-dimensional coordinate under corresponding first coordinate system of target video camera.
Optionally, second video camera 2 may also comprise instructor in broadcasting's video camera and binocular camera, then the target video camera can
For any instructor in broadcasting's video camera in instructor in broadcasting's camera system；Alternatively, second video camera 2 is common Pan/Tilt/Zoom camera, then should
Target video camera can be instructor in broadcasting's video camera or common Pan/Tilt/Zoom camera.Further alternative, the binocular camera 12 is settable
It is connect on preset instructor in broadcasting's bracket, and by instructor in broadcasting's bracket with instructor in broadcasting's video camera 11.
Specifically, as shown in fig. 7, being a kind of structural schematic diagram of first video camera provided in an embodiment of the present invention.This
One video camera includes binocular camera and one or more instructor in broadcasting's video cameras.Assuming that the first video camera of this in the embodiment of the present invention is pacified
Equipped with 2 instructor in broadcasting's video cameras, for carrying out instructor in broadcasting's shooting and tracking, instructor in broadcasting's bracket (referred to as " bracket ") and binocular can be passed through
Video camera carries out wired or wireless connection.The binocular camera is mounted on the bracket, in addition, also mountable on the bracket have wheat
Gram wind, the microphone of installation can be array format, and the microphone of the array format can be used for realizing that auditory localization, sound source are known
The functions such as not, specifically may include the microphone of horizontal array and the microphone of orthogonal array.Further, instructor in broadcasting's video camera and
Bracket can be separation, is also possible to integrate, control interface can be used such as between instructor in broadcasting's video camera and bracket
Serial line interface is communicated.In some embodiments, above-mentioned instructor in broadcasting's video camera and instructor in broadcasting's bracket (including binocular camera, wheat
Gram wind etc.) it can be also integrated into instructor in broadcasting's equipment, for the type of attachment of each equipment in instructor in broadcasting's camera system, the embodiment of the present invention
Without limitation.
Further, Fig. 8 is referred to, is a kind of networking schematic diagram of instructor in broadcasting's camera system provided in an embodiment of the present invention.
As shown in figure 8, multiple seats in the plane can carry out networking, multimachine hyte net mode includes group between multiple seats in the plane for installing instructor in broadcasting's video cameras
Net installs seat in the plane+multiple common Pan/Tilt/Zoom camera networkings of instructor in broadcasting's video camera and instructor in broadcasting's bracket, installs instructor in broadcasting's video camera and instructor in broadcasting
Networking between the seat in the plane of bracket+seat in the plane (there was only instructor in broadcasting's bracket) without Pan/Tilt/Zoom camera, and the seat in the plane without Pan/Tilt/Zoom camera+multiple
Common Pan/Tilt/Zoom camera networking (i.e. without instructor in broadcasting's bracket).Between the video camera of each seat in the plane can by way of LAN or Wi-Fi into
With transmitting control message, which includes camera switching message, audio, video data such as the video object pattern number for row interconnection
According to etc..Further alternative, which can pass through Internet protocol (Internet Protocol, referred to as " IP ")
Transmission such as uses IP Camera protocol stack.Shooting overlapping region is required between the binocular camera in seat in the plane two-by-two.When
When a certain instructor in broadcasting's video camera needs to carry out multi-channel video output, the video square of group network system where may be connected to instructor in broadcasting's video camera
In battle array, output is switched over by video matrix.Optionally, the switchover policy of the video matrix can be by any specified in scene
Instructor in broadcasting's video camera be such as used as instructor in broadcasting's camera control of host bit, or controlled by third party device, the embodiment of the present invention
Without limitation.After the video image of video matrix output is encoded by coding/decoding apparatus, it can be transferred to distal end, to realize
Video conference.Specifically, video data, which can cascade, is handled (instructor in broadcasting's bracket branch if the number of cameras in networking is less
Hold video cascade)；If quantity is more, the video of multiple video cameras is all output to video matrix and is handled, by video matrix into
The switching or synthesis in row one or more camera video source.Further, bracket can externally provide video input/output and connect
Mouth, LAN/Wi-Fi network interface and serial line interface etc..Wherein, video input interface is used for the input video of external other video cameras；Depending on
Frequency output interface is for connecting the equipment such as terminal or video matrix with output video image；Serial line interface is then provided to bracket
Control and debugging interface；LAN/Wi-Fi network interface is used for the cascade of multiple video camera positions, can transmit audio, video data and control number
According to etc..
Further, under more instructor in broadcasting's video camera networking scenes, multiple instructor in broadcasting's video camera all has the video object detection
Ability and Pan/Tilt/Zoom camera function can be responsible for the selection of output seat in the plane and PTZ take the photograph using one of instructor in broadcasting's video camera as host bit
Camera control, other video cameras are used as from seat in the plane；The seat in the plane of instructor in broadcasting's video camera and instructor in broadcasting's bracket+multiple common Pan/Tilt/Zoom cameras field
Under scape, only one instructor in broadcasting's video camera has the video object detectability, is responsible for the selection of output seat in the plane and Pan/Tilt/Zoom camera control,
Common camera is only used as Pan/Tilt/Zoom camera use, and due to only having instructor in broadcasting's video camera that there is the video object detectability, it should
The data for obtaining slave digital video object model under scene not over network, carry out the matched of multimachine digital video object model
Process.
Specifically, the instructor in broadcasting's video camera and binocular camera in the embodiment of the present invention can refer to, above-mentioned Fig. 1-6 is corresponding to be implemented
The associated description of example, details are not described herein again.
Fig. 9 is referred to, Fig. 9 is a kind of structural schematic diagram of instructor in broadcasting's video camera provided in an embodiment of the present invention, for executing
Above-mentioned camera parameters method of adjustment.Specifically, as shown in figure 9, instructor in broadcasting's video camera of the embodiment of the present invention includes: logical
Believe interface 300, memory 200 and processor 100, the processor 100 respectively with the communication interface 300 and the memory
200 connections.The memory 200 can be high speed RAM memory, be also possible to non-labile memory (non-
Volatile memory), a for example, at least magnetic disk storage.The communication interface 300, memory 200 and processor
Data connection can be carried out by bus between 100, it can also data connection by other means.Connected in the present embodiment with bus
It connects and is illustrated.Device structure shown in Fig. 9 does not constitute the restriction to the embodiment of the present invention, can also include than illustrating more
More or less component perhaps combines certain components or different component layouts.Wherein:
Processor 100 is the control centre of equipment, using the various pieces of various interfaces and connection whole equipment, is led to
It crosses operation or executes the program and/or unit being stored in memory 200, and call the driving being stored in memory 200
Software, to execute the various functions and processing data of equipment.Processor 100 can be by integrated circuit (Integrated
Circuit, referred to as " IC ") composition, such as can be made of the IC of single encapsulation, it can also be by connecting more identical functions
Or different function encapsulation IC and form.For example, processor 100 can only include central processing unit (Central
Processing Unit, referred to as " CPU "), it is also possible to CPU, digital signal processor (Digital Signal
Processor, referred to as " DSP "), graphics processor (Graphic Processing Unit, referred to as " GPU ") and various
Control the combination of chip.In embodiments of the present invention, CPU can be single operation core, also may include multioperation core.
Communication interface 300 may include wireline interface, wireless interface etc..
Memory 200 can be used for storage driving software (or software program) and unit, processor 100, communication interface 300
By calling the drive software and unit that are stored in memory 200, thereby executing the various function applications and reality of equipment
Existing data processing.Memory 200 mainly includes program storage area and data storage area, wherein program storage area can store at least
Drive software needed for one function etc.；Data storage area can be stored according to the data in parameter tuning process, such as above-mentioned three
Tie up coordinate information.
Specifically, the processor 100 from the memory 200 reads the drive software and in the drive software
Effect is lower to be executed:
Determine the target video object for needing to shoot；
It is screened from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera according to preset instructor in broadcasting's strategy
Out for shooting the target video camera of the target video object；
The first three-dimensional coordinate of the target video object is obtained, first three-dimensional coordinate is the target video object
Three-dimensional coordinate under corresponding first coordinate system of the target video camera；
The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and defeated
The video image after camera parameter is adjusted out.
Optionally, the processor 100 from the memory 200 reads the drive software and in the drive software
Effect is lower to execute first three-dimensional coordinate for obtaining the target video object, specifically executes following steps:
The second three-dimensional of the binocular camera connecting with instructor in broadcasting's video camera transmission is obtained by the communication interface 300
Coordinate, second three-dimensional coordinate are three of the target video object under corresponding second coordinate system of the binocular camera
Tie up coordinate；
It is three-dimensional by described second according to the positional relationship of the binocular camera and the target video camera demarcated in advance
Coordinate is converted to the first three-dimensional coordinate.
Optionally, the processor 100 from the memory 200 reads the drive software and in the drive software
Effect is lower to execute the determining target video object for needing to shoot, and specifically executes following steps:
The shooting image of the binocular camera transmission is obtained, includes at least one the video object in the shooting image；
The video object model of at least one the video object described in including is established, and from least one described the video object
Determine target video object；
The processor 100 from the memory 200 reads the drive software and under the action of the drive software
It is screened from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera described in executing according to preset instructor in broadcasting's strategy
It is used to shoot the target video camera of the target video object out, specifically executes following steps:
The target video is determined from the shooting image that each video camera in instructor in broadcasting's camera system obtains respectively
Object, and the target video object is obtained in the shooting effect parameter of each video camera；
It is determined as the video camera that shooting effect parameter meets default instructor in broadcasting's strategy to be used to shoot the target video object
Target video camera.
Optionally, the processor 100 from the memory 200 reads the drive software and in the drive software
Lower execute of effect determines the target video object from the shooting image that video camera obtains, and specifically executes following steps:
According to the positional relationship of the binocular camera and current camera demarcated in advance, by second three-dimensional coordinate
Be converted to third three-dimensional coordinate, wherein the current camera be instructor in broadcasting's camera system in except the binocular camera with
Outer any video camera, the third three-dimensional coordinate are that the target video object is sat in the corresponding third of the current camera
Three-dimensional coordinate under mark system；
Judge that region and the current camera of the target video object under the third three-dimensional coordinate detect
The overlapping area in region of the video object under the three-dimensional coordinate of the video object whether be more than preset area threshold；
If being more than, which is determined as the target video object.
Optionally, the shooting effect parameter may include the target video object in the corresponding coordinate system of current camera
Under eye to any one of scenario objects parameter of eye efficacy parameter, hiding relation parameter and shooting area or multinomial, institute
Stating current camera is any video camera in instructor in broadcasting's camera system in addition to the binocular camera.
Wherein, the eye may include the target video object relative to the corresponding seat of current camera to eye efficacy parameter
Mark system rotation angle, it is described rotation angle be according to the target video object second coordinate system rotation angle with
And the positional relationship of the binocular camera and the current camera demarcated in advance is determined.
Wherein, the hiding relation parameter and the scenario objects parameter can be takes the photograph according to the binocular demarcated in advance
Institute is thrown into the region for the scenario objects that the current camera detects by the positional relationship of camera and the current camera again
State what the imaging plane of current camera was determined.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, reference can be made to the related descriptions of other embodiments.
In several embodiments provided by the present invention, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or logical of device or unit
Letter connection can be electrical property, mechanical or other forms.
It is described this unit may or may not be physically separated as illustrated by the separation member, as unit
The component of display may or may not be physical unit, it can and it is in one place, or may be distributed over more
In a network unit.Some or all of unit therein can be selected to realize this embodiment scheme according to the actual needs
Purpose.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, referred to as " ROM "), random access memory (Random Access Memory, RAM), magnetic disk or light
The various media that can store program code such as disk.
Those skilled in the art can be understood that, for convenience and simplicity of description, only with above-mentioned each functional unit
Division progress for example, in practical application, can according to need and above-mentioned function distribution is complete by different functional units
At the internal structure of device being divided into different functional units, to complete all or part of the functions described above.On
The specific work process for stating the device of description, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to
So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into
Row equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution
The range of scheme.
Claims (13)
1. a kind of camera parameters method of adjustment, the method is applied in instructor in broadcasting's video camera characterized by comprising
Determine the target video object for needing to shoot；
Use is filtered out from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera according to preset instructor in broadcasting's strategy
In the target video camera for shooting the target video object；
The second three-dimensional coordinate of the binocular camera transmission connecting with instructor in broadcasting's video camera is obtained, second three-dimensional coordinate is
Three-dimensional coordinate of the target video object under corresponding second coordinate system of the binocular camera, second three-dimensional coordinate
It is that the binocular camera passes through left view and right view of the target video object obtained respectively in the binocular camera
What the inside and outside parameter evidence of the binocular camera of two-dimensional coordinate and acquisition in figure was calculated；
According to the positional relationship of the binocular camera and the target video camera demarcated in advance, by second three-dimensional coordinate
Be converted to the first three-dimensional coordinate, first three-dimensional coordinate is the target video object in the target video camera corresponding the
Three-dimensional coordinate under one coordinate system；
The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and exports tune
Video image after whole camera parameter.
2. the method according to claim 1, wherein the determining target video object for needing to shoot, comprising:
The shooting image of the binocular camera transmission is obtained, includes at least one the video object in the shooting image；
The video object model of at least one the video object described in including is established, and is determined from least one described the video object
Target video object out；
It is described to be screened from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera according to preset instructor in broadcasting's strategy
Out for shooting the target video camera of the target video object, comprising:
The target video object is determined from the shooting image that each video camera in instructor in broadcasting's camera system obtains respectively,
And the target video object is obtained in the shooting effect parameter of each video camera；
It is determined as the video camera that shooting effect parameter meets default instructor in broadcasting's strategy to be used to shoot the mesh of the target video object
Mark video camera.
3. according to the method described in claim 2, it is characterized in that, determining the mesh from the shooting image that video camera obtains
Mark the video object, comprising:
According to the positional relationship of the binocular camera and current camera demarcated in advance, second three-dimensional coordinate is converted
For third three-dimensional coordinate, wherein the current camera be instructor in broadcasting's camera system in addition to the binocular camera
Any video camera, the third three-dimensional coordinate are the target video object in the corresponding third coordinate system of the current camera
Under three-dimensional coordinate；
Judge the view that region and the current camera of the target video object under the third three-dimensional coordinate detect
Whether the overlapping area in region of the frequency object under the three-dimensional coordinate of the video object is more than preset area threshold；
If being more than, which is determined as the target video object.
4. according to the method described in claim 2, it is characterized in that, the shooting effect parameter includes the target video object
The scenario objects of eye under the corresponding coordinate system of current camera to eye efficacy parameter, hiding relation parameter and shooting area
Any one of parameter is multinomial, the current camera be instructor in broadcasting's camera system in addition to the binocular camera
Any video camera.
5. according to the method described in claim 4, it is characterized in that, the eye includes the target video pair to eye efficacy parameter
As the rotation angle relative to the corresponding coordinate system of current camera, the rotation angle is existed according to the target video object
It closes the position of the rotation angle of second coordinate system and the binocular camera demarcated in advance and the current camera
What system determined.
6. according to the method described in claim 4, it is characterized in that, the hiding relation parameter and the scenario objects parameter are
According to the positional relationship of the binocular camera and the current camera demarcated in advance, the current camera is detected
The regions of scenario objects throw into what the imaging plane of the current camera was determined again.
7. a kind of instructor in broadcasting's video camera, which is characterized in that including memory and processor, communication interface, the processor with it is described
Memory, communication interface connection；Wherein,
The memory is used for storage driving software；
The processor reads the drive software from the memory and executes under the action of the drive software:
Determine the target video object for needing to shoot；
Use is filtered out from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera according to preset instructor in broadcasting's strategy
In the target video camera for shooting the target video object；
The second three-dimensional coordinate of the binocular camera connecting with instructor in broadcasting's video camera transmission, institute are obtained by the communication interface
Stating the second three-dimensional coordinate is three-dimensional coordinate of the target video object under corresponding second coordinate system of the binocular camera,
Second three-dimensional coordinate is that the binocular camera passes through the target video object obtained respectively in the binocular camera
Left view and right view in two-dimensional coordinate and acquisition the binocular camera inside and outside parameter according to being calculated；
According to the positional relationship of the binocular camera and the target video camera demarcated in advance, by second three-dimensional coordinate
Be converted to the first three-dimensional coordinate, first three-dimensional coordinate is the target video object in the target video camera corresponding the
Three-dimensional coordinate under one coordinate system；
The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and exports tune
Video image after whole camera parameter.
8. instructor in broadcasting's video camera according to claim 7, which is characterized in that the processor is from described in memory reading
Drive software simultaneously executes the determining target video object for needing to shoot under the action of the drive software, it is specific execute with
Lower step:
The shooting image of the binocular camera transmission is obtained, includes at least one the video object in the shooting image；
The video object model of at least one the video object described in including is established, and is determined from least one described the video object
Target video object out；
The processor reads the drive software from the memory and presses described in execution under the action of the drive software
It is filtered out from each video camera of instructor in broadcasting's camera system where instructor in broadcasting's video camera for shooting according to preset instructor in broadcasting's strategy
The target video camera of the target video object, specifically executes following steps:
The target video object is determined from the shooting image that each video camera in instructor in broadcasting's camera system obtains respectively,
And the target video object is obtained in the shooting effect parameter of each video camera；
It is determined as the video camera that shooting effect parameter meets default instructor in broadcasting's strategy to be used to shoot the mesh of the target video object
Mark video camera.
9. instructor in broadcasting's video camera according to claim 8, which is characterized in that the processor is from described in memory reading
Drive software simultaneously executes under the action of the drive software and determines that the target is regarded from the shooting image that video camera obtains
Frequency object, specifically executes following steps:
According to the positional relationship of the binocular camera and current camera demarcated in advance, second three-dimensional coordinate is converted
For third three-dimensional coordinate, wherein the current camera be instructor in broadcasting's camera system in addition to the binocular camera
Any video camera, the third three-dimensional coordinate are the target video object in the corresponding third coordinate system of the current camera
Under three-dimensional coordinate；
Judge the view that region and the current camera of the target video object under the third three-dimensional coordinate detect
Whether the overlapping area in region of the frequency object under the three-dimensional coordinate of the video object is more than preset area threshold；
If being more than, which is determined as the target video object.
10. instructor in broadcasting's video camera according to claim 8, which is characterized in that the shooting effect parameter includes the target
Eye of the video object under the corresponding coordinate system of current camera is to eye efficacy parameter, hiding relation parameter and shooting area
Any one of scenario objects parameter is multinomial, and the current camera is that the binocular camera shooting is removed in instructor in broadcasting's camera system
Any video camera other than machine.
11. instructor in broadcasting's video camera according to claim 10, which is characterized in that the eye includes the mesh to eye efficacy parameter
Rotation angle of the video object relative to the corresponding coordinate system of current camera is marked, the rotation angle is regarded according to the target
Rotation angle and the in advance binocular camera and the current camera demarcated of the frequency object in second coordinate system
Positional relationship determine.
12. instructor in broadcasting's video camera according to claim 10, which is characterized in that the hiding relation parameter and the scene pair
As parameter is according to the positional relationship for the binocular camera and the current camera demarcated in advance, by the current camera shooting
What the imaging plane that the current camera is thrown into the region for the scenario objects that machine examination measures again was determined.
13. a kind of instructor in broadcasting's camera system, which is characterized in that including the first video camera and at least one second video camera, described
One video camera includes instructor in broadcasting's video camera and binocular camera, between instructor in broadcasting's video camera and the binocular camera, described
It is connected between one video camera and second video camera by wireline interface or wireless interface；Wherein,
Instructor in broadcasting's video camera, for determining the target video object that shoots of needs, and according to preset instructor in broadcasting's strategy from described
The target video camera for shooting the target video object is filtered out in the video camera of instructor in broadcasting's camera system；
The binocular camera is sat for obtaining the second three-dimensional coordinate of the target video object, and by second three-dimensional
Mark is transferred to instructor in broadcasting's video camera；Wherein, second three-dimensional coordinate is the target video object in the binocular camera shooting
Three-dimensional coordinate under corresponding second coordinate system of machine, second three-dimensional coordinate are the binocular cameras by obtaining respectively
Described pair of two-dimensional coordinate and acquisition of the target video object in the left view and right view of the binocular camera
What the inside and outside parameter evidence of lens camera was calculated；
Instructor in broadcasting's video camera, for receiving second three-dimensional coordinate of the binocular camera transmission；According to preparatory calibration
The binocular camera and the target video camera positional relationship, second three-dimensional coordinate is converted to first and three-dimensional is sat
Mark；The camera parameter of the target video camera is adjusted to camera parameter corresponding with first three-dimensional coordinate, and exports tune
Video image after whole camera parameter；Wherein, first three-dimensional coordinate is that the target video object is imaged in the target
Three-dimensional coordinate under corresponding first coordinate system of machine.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610562671.6A CN106251334B (en) | 2016-07-18 | 2016-07-18 | A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system |
| PCT/CN2017/091863 WO2018014730A1 (en) | 2016-07-18 | 2017-07-05 | Method for adjusting parameters of camera, broadcast-directing camera, and broadcast-directing filming system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610562671.6A CN106251334B (en) | 2016-07-18 | 2016-07-18 | A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106251334A CN106251334A (en) | 2016-12-21 |
| CN106251334B true CN106251334B (en) | 2019-03-01 |
Family
ID=57613157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610562671.6A Active CN106251334B (en) | 2016-07-18 | 2016-07-18 | A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106251334B (en) |
| WO (1) | WO2018014730A1 (en) |
Families Citing this family (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106251334B (en) * | 2016-07-18 | 2019-03-01 | 华为技术有限公司 | A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system |
| US10091412B1 (en) * | 2017-06-30 | 2018-10-02 | Polycom, Inc. | Optimal view selection method in a video conference |
| CN109413359B (en) * | 2017-08-16 | 2020-07-28 | 华为技术有限公司 | Camera tracking method, device and equipment |
| CN109922251B (en) * | 2017-12-12 | 2021-10-22 | 华为技术有限公司 | Method, device and system for quick snapshot |
| WO2019206247A1 (en) * | 2018-04-27 | 2019-10-31 | Shanghai Truthvision Information Technology Co., Ltd | System and method for camera calibration |
| CN109031201A (en) * | 2018-06-01 | 2018-12-18 | 深圳市鹰硕技术有限公司 | The voice localization method and device of Behavior-based control identification |
| CN108900860A (en) * | 2018-08-23 | 2018-11-27 | 佛山龙眼传媒科技有限公司 | A kind of instructor in broadcasting's control method and device |
| CN110969662B (en) * | 2018-09-28 | 2023-09-26 | 杭州海康威视数字技术股份有限公司 | Method and device for calibrating internal parameters of fish-eye camera, calibration device controller and system |
| CN111243029B (en) * | 2018-11-28 | 2023-06-23 | 驭势（上海）汽车科技有限公司 | Calibration method and device of vision sensor |
| CN109360250A (en) * | 2018-12-27 | 2019-02-19 | 爱笔（北京）智能科技有限公司 | Scaling method, equipment and the system of a kind of pair of photographic device |
| CN109712188A (en) * | 2018-12-28 | 2019-05-03 | 科大讯飞股份有限公司 | A kind of method for tracking target and device |
| CN111325790B (en) * | 2019-07-09 | 2024-02-20 | 杭州海康威视系统技术有限公司 | Target tracking method, device and system |
| CN111787243B (en) * | 2019-07-31 | 2021-09-03 | 北京沃东天骏信息技术有限公司 | Broadcasting guide method, device and computer readable storage medium |
| CN110456829B (en) * | 2019-08-07 | 2022-12-13 | 深圳市维海德技术股份有限公司 | Positioning tracking method, device and computer readable storage medium |
| CN112468680A (en) * | 2019-09-09 | 2021-03-09 | 上海御正文化传播有限公司 | Processing method of advertisement shooting site synthesis processing system |
| CN110737798B (en) * | 2019-09-26 | 2022-10-14 | 万翼科技有限公司 | Indoor inspection method and related product |
| CN111080698B (en) * | 2019-11-27 | 2023-06-06 | 上海新时达机器人有限公司 | Method, system and storage device for calibrating position of long plate |
| CN111131697B (en) * | 2019-12-23 | 2022-01-04 | 北京中广上洋科技股份有限公司 | Multi-camera intelligent tracking shooting method, system, equipment and storage medium |
| CN111080679B (en) * | 2020-01-02 | 2023-04-18 | 东南大学 | Method for dynamically tracking and positioning indoor personnel in large-scale place |
| CN113516717A (en) * | 2020-04-10 | 2021-10-19 | 富华科精密工业（深圳）有限公司 | Camera device external parameter calibration method, electronic equipment and storage medium |
| CN111698467B (en) * | 2020-05-08 | 2022-05-06 | 北京中广上洋科技股份有限公司 | Intelligent tracking method and system based on multiple cameras |
| CN113808199B (en) * | 2020-06-17 | 2023-09-08 | 华为云计算技术有限公司 | Positioning method, electronic equipment and positioning system |
| CN111800590B (en) * | 2020-07-06 | 2022-11-25 | 深圳博为教育科技有限公司 | Broadcasting-directing control method, device and system and control host |
| CN112887653B (en) * | 2021-01-25 | 2022-10-21 | 联想(北京)有限公司 | Information processing method and information processing device |
| CN112819770B (en) * | 2021-01-26 | 2022-11-22 | 中国人民解放军陆军军医大学第一附属医院 | Iodine contrast agent allergy monitoring method and system |
| CN113453021B (en) * | 2021-03-24 | 2022-04-29 | 北京国际云转播科技有限公司 | Artificial intelligence broadcasting guide method, system, server and computer readable storage medium |
| CN113129376A (en) * | 2021-04-22 | 2021-07-16 | 青岛联合创智科技有限公司 | Checkerboard-based camera real-time positioning method |
| CN113271482A (en) * | 2021-05-17 | 2021-08-17 | 广东彼雍德云教育科技有限公司 | Portable full-width image scratching blackboard |
| CN113587895B (en) * | 2021-07-30 | 2023-06-30 | 杭州三坛医疗科技有限公司 | Binocular distance measuring method and device |
| CN113610932A (en) * | 2021-08-20 | 2021-11-05 | 苏州智加科技有限公司 | Binocular camera external parameter calibration method and device |
| CN113838146A (en) * | 2021-09-26 | 2021-12-24 | 昆山丘钛光电科技有限公司 | Method and device for verifying calibration precision of camera module and method and device for testing camera module |
| CN114025107B (en) * | 2021-12-01 | 2023-12-01 | 北京七维视觉科技有限公司 | Image ghost shooting method, device, storage medium and fusion processor |
| CN116389660B (en) * | 2021-12-22 | 2024-04-12 | 广州开得联智能科技有限公司 | Recorded broadcast guiding method, recorded broadcast guiding device, recorded broadcast guiding equipment and storage medium |
| CN117523431A (en) * | 2023-11-17 | 2024-02-06 | 中国科学技术大学 | Firework detection method and device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101630406A (en) * | 2008-07-14 | 2010-01-20 | 深圳华为通信技术有限公司 | Camera calibration method and camera calibration device |
| CN102843540B (en) * | 2011-06-20 | 2015-07-29 | 宝利通公司 | Automatic camera for video conference is selected |
| CN104869365A (en) * | 2015-06-02 | 2015-08-26 | 阔地教育科技有限公司 | Direct recording and broadcasting system based mouse tracking method and device |
| CN105049764A (en) * | 2015-06-17 | 2015-11-11 | 武汉智亿方科技有限公司 | Image tracking method and system for teaching based on multiple positioning cameras |
| 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 |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5733775B2 (en) * | 2008-06-06 | 2015-06-10 | 日本電気株式会社 | Object image display system |
| US8537195B2 (en) * | 2011-02-09 | 2013-09-17 | Polycom, Inc. | Automatic video layouts for multi-stream multi-site telepresence conferencing system |
| CN106251334B (en) * | 2016-07-18 | 2019-03-01 | 华为技术有限公司 | A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system |
-
2016
- 2016-07-18 CN CN201610562671.6A patent/CN106251334B/en active Active
-
2017
- 2017-07-05 WO PCT/CN2017/091863 patent/WO2018014730A1/en active Application Filing
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101630406A (en) * | 2008-07-14 | 2010-01-20 | 深圳华为通信技术有限公司 | Camera calibration method and camera calibration device |
| CN102843540B (en) * | 2011-06-20 | 2015-07-29 | 宝利通公司 | Automatic camera for video conference is selected |
| CN104869365A (en) * | 2015-06-02 | 2015-08-26 | 阔地教育科技有限公司 | Direct recording and broadcasting system based mouse tracking method and device |
| CN105049764A (en) * | 2015-06-17 | 2015-11-11 | 武汉智亿方科技有限公司 | Image tracking method and system for teaching based on multiple positioning cameras |
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2018014730A1 (en) | 2018-01-25 |
| CN106251334A (en) | 2016-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106251334B (en) | A kind of camera parameters method of adjustment, instructor in broadcasting's video camera and system | |
| WO2017215295A1 (en) | Camera parameter adjusting method, robotic camera, and system | |
| US10665261B2 (en) | Camera array including camera modules | |
| US20200275079A1 (en) | Generating three-dimensional video content from a set of images captured by a camera array | |
| US20160191815A1 (en) | Camera array removing lens distortion | |
| JP7179515B2 (en) | Apparatus, control method and program | |
| JP6625745B2 (en) | Image processing device | |
| KR20150050172A (en) | Apparatus and Method for Selecting Multi-Camera Dynamically to Track Interested Object | |
| JP2019083402A (en) | Image processing apparatus, image processing system, image processing method, and program | |
| US11108971B2 (en) | Camera array removing lens distortion | |
| US11048099B2 (en) | Communication system generating a floating image of a remote venue | |
| JP2015506030A (en) | System for shooting video movies | |
| KR20180125878A (en) | A system comprising a video camera and a client device and a method performed by the same | |
| CN108307183A (en) | Virtual scene method for visualizing and system | |
| KR101853269B1 (en) | Apparatus of stitching depth maps for stereo images | |
| JP6403862B1 (en) | Three-dimensional model generation apparatus, generation method, and program | |
| CN103500471A (en) | Method for realizing high-resolution augmented reality system | |
| CN107172413A (en) | Method and system for displaying video of real scene | |
| WO2022127747A1 (en) | Method and system for real social using virtual scene | |
| JP2020178237A (en) | Video processing device, video processing method, and video processing program | |
| CN103546680B (en) | A kind of deformation-free omni-directional fisheye photographic device and a method for implementing the same | |
| JP2021119459A (en) | Generation device and generation method | |
| CN108563410A (en) | A kind of display control method and electronic equipment | |
| KR102298047B1 (en) | Method of recording digital contents and generating 3D images and apparatus using the same | |
| US20200389722A1 (en) | Processing of data of a video sequence in order to zoom to a speaker detected in the sequence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |