CN106534692A - Video image stabilization method and device - Google Patents
Video image stabilization method and device Download PDFInfo
- Publication number
- CN106534692A CN106534692A CN201611051216.6A CN201611051216A CN106534692A CN 106534692 A CN106534692 A CN 106534692A CN 201611051216 A CN201611051216 A CN 201611051216A CN 106534692 A CN106534692 A CN 106534692A
- Authority
- CN
- China
- Prior art keywords
- adjacent
- field pictures
- data
- field
- picture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/68—Control of cameras or camera modules for stable pick-up of the scene, e.g. compensating for camera body vibrations
- H04N23/681—Motion detection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/68—Control of cameras or camera modules for stable pick-up of the scene, e.g. compensating for camera body vibrations
- H04N23/682—Vibration or motion blur correction
Abstract
The invention discloses a video image stabilization method and device. The method comprises the steps of collecting two adjacent images and recording collection time points of the two adjacent images; calculating translation amount and a rotating angle between the two adjacent images according to motion data of a gyroscope in a time period between the collection time points, as first data; detecting whether corners exist in the two adjacent images or not through utilization of a corner detection algorithm, if the corners exist, calculating the translation amount and the rotating angle between the two adjacent images according to matching information of the corners in the two adjacent images, as second data; integrating the first data and the second data and filtering the integrated result data to obtain final motion data; and determining a next frame of image according to the final motion data and a last frame of image of the two adjacent images, thereby realizing image stabilization. According to the method and the device, through combination of the data of the gyroscope and an image processing method, image stabilization processing is carried out on a video, so the display effect of the video after the image stabilization is improved well.
Description
Technical field
The present invention relates to field of video processing, and in particular to a kind of video image stabilization method and device.
Background technology
Video Stabilization is that the dynamic image sequence acquired in a kind of video camera to randomized jitter or random motion carries out weight
It is new to change, arrangement so as to the technology for more smoothly showing over the display.It eliminates or reduces between image sequence not
The distortion situations such as the translation of rule, rotation, scaling, improve the quality of picture, so that picture is more suitable for intelligence regarding
Target detection, tracking and identification in frequency analysis etc. process operation.
There are various Video Stabilization technologies in prior art, for example machinery is steady as technology, based on image procossing
Electronic image stabilizing etc., but we are still devoted to constantly lifting the effect that Video Stabilization technology can reach.
The content of the invention
The invention provides a kind of video image stabilization method and device, for the video pictures that there is distortion situation, Neng Gouti
Rise its final display effect.
The invention provides a kind of video image stabilization method, methods described includes：
Adjacent two field pictures are gathered, and records the acquisition time of the adjacent two field pictures respectively；According in the phase
The exercise data of gyroscope in time period between the acquisition time of adjacent two field pictures, calculates between the adjacent two field pictures
Translational movement and the anglec of rotation, as the first data；
Detected using default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is, according to
The match information of angle point in the adjacent two field pictures, calculates the translational movement and the anglec of rotation between the adjacent two field pictures,
As the second data；
First data and second data are merged, fusion results data are obtained；To the fusion results
After data are filtered, the final exercise data of the adjacent two field pictures is obtained；
According to the final exercise data and the previous frame image of the adjacent two field pictures, the adjacent two frames figure is determined
The latter two field picture of picture, realizes the steady picture of the adjacent two field pictures.
Preferably, methods described also includes：
If being detected using default Corner Detection Algorithm and there is no angle point in the adjacent two field pictures, directly to described
First data are filtered, and obtain the final exercise data of the adjacent two field pictures；Continue executing with described according to described final
The previous frame image of exercise data and the adjacent two field pictures, obtains the latter two field picture of the adjacent two field pictures, realizes
The step of steady picture of the adjacent two field pictures.
Preferably, in time period of the basis between the acquisition time of the adjacent two field pictures gyroscope fortune
Dynamic data, calculate the translational movement and the anglec of rotation between the adjacent two field pictures, as the first data, including：
The angular velocity of rotation of the acquisition time gyroscope in the adjacent two field pictures is obtained respectively；
According to the angular velocity of rotation of the gyroscope, calculate between the acquisition time of the adjacent two field pictures when
Between in section gyroscope exercise data；
According to the exercise data and the camera focal length data for obtaining in advance of the gyroscope, the adjacent two frames figure is calculated
Translational movement and the anglec of rotation as between, as the first data.
Preferably, it is described that first data and second data are merged, fusion results data are obtained, is wrapped
Include：
Translational movement between the described adjacent two field pictures in first data and second data is calculated respectively
The mean value of the anglec of rotation between mean value, and the adjacent two field pictures, as fusion results data.
Preferably, it is described according to the final exercise data and the previous frame image of the adjacent two field pictures, determine institute
The latter two field picture of adjacent two field pictures is stated, the steady picture of the adjacent two field pictures is realized, including：
According to the final exercise data, line translation is entered to the previous frame image of the adjacent two field pictures, is converted
Image afterwards；
Image after the conversion is defined as into the latter two field picture of the adjacent two field pictures, the adjacent two frames figure is realized
The steady picture of picture.
Preferably, methods described also includes：
The fringe region of the latter two field picture of adjacent two field pictures described in cutting, to remove the black surround of image.
Present invention also offers a kind of Video Stabilization device, described device includes：
Acquisition module, for gathering adjacent two field pictures；
Logging modle, for recording the acquisition time of the adjacent two field pictures respectively；
First computing module, for gyro in time period of the basis between the acquisition time of the adjacent two field pictures
The exercise data of instrument, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the first data；
Detection module, for utilizing default Corner Detection Algorithm to detect in the adjacent two field pictures with the presence or absence of angle point；
Second computing module, for when the detection module detects angle point, according to angle in the adjacent two field pictures
The match information of point, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the second data；
Fusion Module, for first data and second data are merged, obtains fusion results data；
First filtration module, after being filtered to the fusion results data, obtains the adjacent two field pictures
Final exercise data；
Determining module, for the previous frame image according to the final exercise data and the adjacent two field pictures, it is determined that
The latter two field picture of the adjacent two field pictures, realizes the steady picture of the adjacent two field pictures.
Preferably, described device also includes：
Second filtration module, for when the detection module is not detected by angle point, being filtered to first data,
Obtain the final exercise data of the adjacent two field pictures.
Preferably, first computing module, including：
Acquisition submodule, for obtaining the anglec of rotation speed of the acquisition time gyroscope in the adjacent two field pictures respectively
Degree；
First calculating sub module, for the angular velocity of rotation according to the gyroscope, calculates in the adjacent two field pictures
Acquisition time between time period in gyroscope exercise data；
Second calculating sub module, the camera focal length number obtained for the exercise data according to the gyroscope and in advance
According to the translational movement and the anglec of rotation between the calculating adjacent two field pictures, as the first data.
Preferably, the Fusion Module, specifically for：
Translational movement between the described adjacent two field pictures in first data and second data is calculated respectively
The mean value of the anglec of rotation between mean value, and the adjacent two field pictures, as fusion results data.
Preferably, the determining module, including：
Transformation submodule, for according to the final exercise data, entering to the previous frame image of the adjacent two field pictures
Line translation, image after being converted；
Determination sub-module, it is for image after the conversion to be defined as the latter two field picture of the adjacent two field pictures, real
The steady picture of the existing adjacent two field pictures.
Preferably, described device also includes：
Cutting module, for the fringe region of the latter two field picture of adjacent two field pictures described in cutting, to remove image
Black surround.
In the video image stabilization method that the present invention is provided, first, adjacent two field pictures are gathered, and records described adjacent two respectively
The acquisition time of two field picture；According to the fortune of gyroscope in the time period between the acquisition time of the adjacent two field pictures
Dynamic data, calculate the translational movement and the anglec of rotation between the adjacent two field pictures, as the first data.Secondly, using default
Corner Detection Algorithm is detected, if it is, according to the adjacent two field pictures
The match information of upper angle point, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the second data.Again
It is secondary, first data and second data are merged, fusion results data are obtained；The fusion results data are entered
After row filtering, the final exercise data of the adjacent two field pictures is obtained.Finally, according to the final exercise data and the phase
The previous frame image of adjacent two field pictures, determines the latter two field picture of the adjacent two field pictures, realizes the adjacent two field pictures
Steady picture.The video image stabilization method that the present invention is provided, due to combining gyro data, and image processing method in video
Adjacent two field pictures carry out steady picture and process, so, the display effect for eventually passing through the steady video as after is preferably lifted.
Description of the drawings
For the technical scheme being illustrated more clearly that in the embodiment of the present application, below will be to making needed for embodiment description
Accompanying drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present application, for
For those of ordinary skill in the art, without having to pay creative labor, can be obtaining which according to these accompanying drawings
His accompanying drawing.
A kind of video image stabilization method flow chart that Fig. 1 is provided for the present invention；
A kind of exercise data of utilization gyroscope that Fig. 2 is provided for the present invention calculates the fortune between the adjacent two field pictures
The emotionally method flow diagram of condition；
Another kind of video image stabilization method flow chart that Fig. 3 is provided for the present invention；
A kind of Video Stabilization apparatus structure schematic diagram that Fig. 4 is provided for the present invention；
Another kind of Video Stabilization apparatus structure schematic diagram that Fig. 5 is provided for the present invention；
A kind of part-structure schematic diagram of computer that Fig. 6 is provided for the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present application, the technical scheme in the embodiment of the present application is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only some embodiments of the present application, rather than the embodiment of whole.It is based on
Embodiment in the application, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of the application protection.
A kind of video image stabilization method is embodiments provided, with reference to Fig. 1, is that one kind provided in an embodiment of the present invention is regarded
Frequency digital image stabilization method flow chart, methods described can specifically include：
S101：Gather adjacent two field pictures.
In the embodiment of the present invention, the adjacent two field pictures of Real-time Collection in the video that system is shot from camera.Specifically, originally
Inventive embodiments are by steady as processing between adjacent two field pictures, being finally reached the steady picture to the whole video for photographing
Process.
S102：The acquisition time of the adjacent two field pictures is recorded respectively.
In practical application, the every two field picture in the video that camera shoots all has shooting time point, the embodiment of the present invention
Described in acquisition time of adjacent two field pictures be camera and shoot the shooting time point of the adjacent two field pictures respectively.
S103：According to the motion number of gyroscope in the time period between the acquisition time of the adjacent two field pictures
According to the translational movement and the anglec of rotation between the calculating adjacent two field pictures, as the first data.
In the embodiment of the present invention, the equipment for shooting video has gyroscope, specifically, using the motion number of gyroscope
According to the motion conditions between the adjacent two field pictures can be calculated, such as the translational movement between the adjacent two field pictures and rotation
Gyration.
With reference to Fig. 2, it is that a kind of exercise data of utilization gyroscope provided in an embodiment of the present invention calculates adjacent two frame
The method flow diagram of the motion conditions between image.Specifically, methods described includes：
S201：The angular velocity of rotation of the acquisition time gyroscope in the adjacent two field pictures is obtained respectively.
In the embodiment of the present invention, when the adjacent two field pictures are gathered, acquisition time t is recorded respectively1And t2, and point
Do not obtain in t1The angular velocity of rotation of gyroscope during time pointAnd be in t2The rotation of gyroscope during time point
Angular speed
S202：According to the angular velocity of rotation of the gyroscope, calculate the acquisition time of the adjacent two field pictures it
Between time period in gyroscope exercise data.
In a kind of concrete implementation mode, to acquisition time t1And t2The angular velocity of rotation of gyroscope is averaged, and obtains
t1To t2Exercise data (the W of gyroscope in time periodx,Wy,Wz), wherein, computational methods are as follows：
S203：According to the exercise data and the camera focal length data for obtaining in advance of the gyroscope, calculate described adjacent
Translational movement and the anglec of rotation between two field pictures, as the first data.
The embodiment of the present invention obtains focal length data fx, the fy of the camera for shooting video in advance, burnt with reference to the camera
Away from data fx, fy, translational movement (Δ x, Δ y) and anglec of rotation θ, as the first number between the adjacent two field pictures is calculated
According to, wherein, computational methods are as follows：
θ x=Wx × (t2-t1)
θ y=Wy × (t2-t1)
θ z=Wz × (t2-t1)
θ=θ z
S104：Detected using default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is,
Perform S105.
The embodiment of the present invention is also based on image processing algorithm, calculate translational movement between the adjacent two field pictures and
The anglec of rotation.But, the basis of image processing algorithm is the acquisition of angle point on image, and only system is in the adjacent two field pictures
On be able to detect that angle point, system can carry out carrying out image procossing to the adjacent two field pictures, finally give the phase
Translational movement and angular velocity of rotation between adjacent two field pictures.
So, the embodiment of the present invention is after the adjacent two field pictures execution of collection that above-mentioned S101 includes terminates, you can profit
Detected with default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point.Utilizable angle point in practical application
Detection algorithm is more, including Moravec Corner Detection Algorithms, FAST Corner Detection Algorithms, Harris corner detection approach and shi_
Tomas corner detection approach etc..
System is detected using default Corner Detection Algorithm after there is angle point in the adjacent two field pictures, S105：According to
The match information of angle point in the adjacent two field pictures, calculates the translational movement and the anglec of rotation between the adjacent two field pictures,
As the second data.
It is a kind of preferred embodiment in, system is detected in the adjacent two field pictures using default Corner Detection Algorithm
After there is angle point, the angle point on the previous frame image of the adjacent two field pictures, for example, shi_tomas angle points are obtained first.
Secondly, in the latter two field picture of the adjacent two field pictures, search and record picture corresponding with the angle point on previous frame image
Vegetarian refreshments, specifically can be realized using existing pyramid LK optical flow methods.Then, according to angle point correspondence in the adjacent two field pictures
Pixel match information, calculate image change matrix, wherein, the match information includes the fortune of the adjacent two field pictures
Dynamic data, described image transformation matrices are used for the situation of change for representing the corresponding pixel of angle point in the adjacent two field pictures.
Finally, system can filter out global motion vector using RANSAC algorithms from described image transformation matrices, as described phase
Translational movement (Δ x1, Δ y1) and anglec of rotation θ 1 between adjacent two field pictures, as the second data.
In another embodiment, system is detected using default Corner Detection Algorithm and exist in the adjacent two field pictures
Secondly after angle point, the angle point in the adjacent two field pictures is obtained first, the angle point in the adjacent two field pictures is carried out
With obtaining match information, wherein, the match information includes the exercise data of the adjacent two field pictures.Finally, system can be with
Global motion vector is filtered out from described image transformation matrices using RANSAC algorithms, between as described adjacent two field pictures
Translational movement (Δ x1, Δ y1) and anglec of rotation θ 1, as the second data.
S106：First data and second data are merged, fusion results data are obtained.
It is a kind of preferred embodiment in, calculate described adjacent two in first data and second data respectively
The mean value of the anglec of rotation between the mean value of the translational movement between two field picture, and the adjacent two field pictures, as melting
Close result data.
Specifically, Δ x2=(Δ x+ Δ x1)/2；Δ y2=(Δ y+ Δ y1)/2；θ 2=(θ+θ 1)/2.The fusion knot
Fruit data include the translational movement (Δ x2, Δ y2) between the adjacent two field pictures, and between the adjacent two field pictures
Anglec of rotation θ 2.
S107：After being filtered to the fusion results data, the final exercise data of the adjacent two field pictures is obtained.
In the embodiment of the present invention after fusion results data are obtained, the fusion results data are filtered, wherein wrapping
Include the data to changing greatly in the fusion results data to delete or weaken.Specifically, can be to the fusion results data
Carry out Kalman filter.
In order to eliminate the impact of noise, the embodiment of the present invention can carry out denoising to the fusion results data after after filtering
Process.Finally, system can obtain the final exercise data of the adjacent two field pictures.Wherein, the final exercise data is used
In represent the adjacent two field pictures after filtering after exercise data.
S108：According to the final exercise data and the previous frame image of the adjacent two field pictures, determine described adjacent
The latter two field picture of two field pictures, realizes the steady picture of the adjacent two field pictures.
The embodiment of the present invention after the final exercise data for obtaining the adjacent two field pictures, according to the final motion number
According to entering line translation to the previous frame image of the adjacent two field pictures, image after being converted.And will be image after the conversion true
It is set to the latter two field picture of the adjacent two field pictures, realizes the steady picture of the adjacent two field pictures.That is, now obtaining
The adjacent two field pictures including the latter two field picture and the previous frame image be through the steady image as after.
In practical application, each the adjacent two field pictures in the video that camera is shot are carried through the embodiment of the present invention
For Video Stabilization process, the video for finally giving is as the video for processing through steady.
In video image stabilization method provided in an embodiment of the present invention, first, adjacent two field pictures are gathered, and is recorded respectively described
The acquisition time of adjacent two field pictures；According to gyro in the time period between the acquisition time of the adjacent two field pictures
The exercise data of instrument, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the first data.Secondly, profit
Detected with default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is, according to described adjacent two
The match information of angle point on two field picture, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, used as the second number
According to.Again, first data and second data are merged, is obtained fusion results data；To the fusion results
After data are filtered, the final exercise data of the adjacent two field pictures is obtained.Finally, according to the final exercise data and
The previous frame image of the adjacent two field pictures, determines the latter two field picture of the adjacent two field pictures, realizes described adjacent two
The steady picture of two field picture.Video image stabilization method provided in an embodiment of the present invention, due to combining gyro data, and image procossing
Method carries out steady picture and processes to adjacent two field pictures in video, so, the display effect for eventually passing through the steady video as after is obtained
Preferably lifted.
When being able to detect that angle point in the adjacent two field pictures, video image stabilization method knot provided in an embodiment of the present invention
Gyro data and image processing algorithm are closed, and display effect are obtained preferably through the steady video as after.But, for detecting
There is no the situation of angle point in the adjacent two field pictures, video image stabilization method provided in an embodiment of the present invention is can not be using figure
As Processing Algorithm, so, the embodiment of the present invention is gone back for detecting the situation that there is no angle point in the adjacent two field pictures
There is provided a kind of video image stabilization method.
With reference to Fig. 3, it is another kind of video image stabilization method flow chart provided in an embodiment of the present invention.Methods described is specifically wrapped
Include：
S301：Gather adjacent two field pictures.
S302：The acquisition time of the adjacent two field pictures is recorded respectively.
S303：According to the motion number of gyroscope in the time period between the acquisition time of the adjacent two field pictures
According to the translational movement and the anglec of rotation between the calculating adjacent two field pictures, as the first data.
S304：Detected using default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is not, then
Perform S305.
In the embodiment of the present invention, there is no angle in the adjacent two field pictures if detected using default Corner Detection Algorithm
Point, then can only be carried out using calculated first data steady as processing.
If being detected using default Corner Detection Algorithm and there is angle point in the adjacent two field pictures, in above-mentioned S104
Implementation procedure it is identical.
S305：First data are filtered, the final exercise data of the adjacent two field pictures is obtained.
In the embodiment of the present invention, due to there is no angle point in adjacent two field pictures described in system detectio, so, system can only
First data are calculated by the exercise data of gyroscope, and first data are filtered, obtain described adjacent two
The final exercise data of two field picture.
S306：According to the final exercise data and the previous frame image of the adjacent two field pictures, obtain described adjacent
The latter two field picture of two field pictures, realizes the steady picture of the adjacent two field pictures.
In video image stabilization method provided in an embodiment of the present invention, if described adjacent using default Corner Detection Algorithm detection
There is no angle point in two field pictures, then system is directly filtered to first data, obtain the adjacent two field pictures
Final exercise data.Finally, according to the final exercise data and the previous frame image of the adjacent two field pictures, obtain described
The latter two field picture of adjacent two field pictures, realizes the steady picture of the adjacent two field pictures.The embodiment of the present invention is for can't detect
The image of angle point also can be carried out steady as processing by the exercise data of gyroscope.
Further, since the disappearance of image information, the embodiment of the present invention is according to the final exercise data and described adjacent
The previous frame image of two field pictures, after obtaining the latter two field picture of the adjacent two field pictures, after the adjacent two field pictures
The marginal portion of one two field picture typically has black surround appearance, so, the embodiment of the present invention can pass through adjacent two frame described in cutting
The fringe region of the latter two field picture of image, to remove the black surround of image.The mode of this cutting is fairly simple, can save and is
The expense of system.
In a kind of actual application scenarios, video image stabilization method provided in an embodiment of the present invention can be applied to unmanned plane neck
Domain.Specifically, embodiments provide a kind of UAV Video digital image stabilization method based on gyroscope and camera.Wherein,
In order to reduce the weight and cost of unmanned plane, the camera for shooting video can be the built-in camera of unmanned plane, and
Gyroscope can also be the built-in gyroscope of unmanned plane, without the need for extra hardware device.
The embodiment of the present invention additionally provides a kind of Video Stabilization device, with reference to Fig. 4, is one kind provided in an embodiment of the present invention
Video Stabilization apparatus structure schematic diagram, described device include：
Acquisition module 401, for gathering adjacent two field pictures；
Logging modle 402, for recording the acquisition time of the adjacent two field pictures respectively；
First computing module 403, in time period of the basis between the acquisition time of the adjacent two field pictures
The exercise data of gyroscope, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the first data；
Detection module 404, for utilizing default Corner Detection Algorithm to detect in the adjacent two field pictures with the presence or absence of angle
Point；
Second computing module 405, for when the detection module detects angle point, according in the adjacent two field pictures
The match information of angle point, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the second data；
Fusion Module 406, for first data and second data are merged, obtains fusion results number
According to；
First filtration module 407, after being filtered to the fusion results data, obtains the adjacent two field pictures
Final exercise data；
Determining module 408, for the previous frame image according to the final exercise data and the adjacent two field pictures, really
The latter two field picture of the fixed adjacent two field pictures, realizes the steady picture of the adjacent two field pictures.
In addition, the embodiment of the present invention is also provided for detecting the situation that there is no angle point in the adjacent two field pictures
A kind of Video Stabilization device, with reference to Fig. 5, is another kind of Video Stabilization apparatus structure schematic diagram provided in an embodiment of the present invention,
Described device includes：
Second filtration module 501, for when the detection module is not detected by angle point, filtering to first data
Ripple, obtains the final exercise data of the adjacent two field pictures.
In a kind of concrete implementation mode, first computing module, including：
Acquisition submodule, for obtaining the anglec of rotation speed of the acquisition time gyroscope in the adjacent two field pictures respectively
Degree；
First calculating sub module, for the angular velocity of rotation according to the gyroscope, calculates in the adjacent two field pictures
Acquisition time between time period in gyroscope exercise data；
Second calculating sub module, the camera focal length number obtained for the exercise data according to the gyroscope and in advance
According to the translational movement and the anglec of rotation between the calculating adjacent two field pictures, as the first data.
Specifically, the Fusion Module, specifically for：
Translational movement between the described adjacent two field pictures in first data and second data is calculated respectively
The mean value of the anglec of rotation between mean value, and the adjacent two field pictures, as fusion results data.
In practical application, the determining module, including：
Transformation submodule, for according to the final exercise data, entering to the previous frame image of the adjacent two field pictures
Line translation, image after being converted；
Determination sub-module, it is for image after the conversion to be defined as the latter two field picture of the adjacent two field pictures, real
The steady picture of the existing adjacent two field pictures.
In addition, described device also includes：
Cutting module, for the fringe region of the latter two field picture of adjacent two field pictures described in cutting, to remove image
Black surround.
Video Stabilization device provided in an embodiment of the present invention can realize following functions：Adjacent two field pictures are gathered, and is divided
The acquisition time of the adjacent two field pictures is not recorded；According between the acquisition time of the adjacent two field pictures when
Between in section gyroscope exercise data, calculate the translational movement and the anglec of rotation between the adjacent two field pictures, as the first number
According to.Detected using default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is, according to the phase
The match information of angle point in adjacent two field pictures, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, used as the
Two data.First data and second data are merged, fusion results data are obtained；To the fusion results number
After being filtered, the final exercise data of the adjacent two field pictures is obtained.According to the final exercise data and the phase
The previous frame image of adjacent two field pictures, determines the latter two field picture of the adjacent two field pictures, realizes the adjacent two field pictures
Steady picture.Video Stabilization device provided in an embodiment of the present invention, due to combining gyro data, and image processing method pair
In video, adjacent two field pictures carry out steady picture process, so, the display effect for eventually passing through the steady video as after is obtained preferably
Lifted.
Accordingly, the embodiment of the present invention also provides a kind of computer, shown in Figure 6, can include：
Processor 601, memory 602, input unit 603 and output device 604.Server and intelligent recommendation are collected in behavior
The quantity of the processor 601 in server can be with one or more, in Fig. 6 by taking a processor as an example.In some of the invention
In embodiment, processor 601, memory 602, input unit 603 and output device 604 can pass through bus or alternate manner connects
Connect, wherein, in Fig. 6 as a example by being connected by bus.
Memory 602 can be used to store software program and module, and processor 601 is stored in memory 602 by operation
Software program and module, so as to process performing collects the various function application and number of server and intelligent recommendation server
According to process.Memory 602 can mainly include storing program area and storage data field, wherein, storing program area can store operation system
Application program needed for system, at least one function etc..Additionally, memory 602 can include high-speed random access memory, may be used also
With including nonvolatile memory, for example, at least one disk memory, flush memory device or other volatile solid-states
Part.Input unit 603 can be used for the numeral or character information of receives input, and generation collects server with behavior and intelligence is pushed away
Recommend user's setting and the key signals input that function control is relevant of server.
Specifically in the present embodiment, processor 601 can be according to following instruction, by one or more application program
The corresponding executable file of process be loaded in memory 602, and by processor 601 come run storage in the memory 602
Application program, so as to realize various functions：
Adjacent two field pictures are gathered, and records the acquisition time of the adjacent two field pictures respectively；According in the phase
The exercise data of gyroscope in time period between the acquisition time of adjacent two field pictures, calculates between the adjacent two field pictures
Translational movement and the anglec of rotation, as the first data；
Detected using default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is, according to
The match information of angle point in the adjacent two field pictures, calculates the translational movement and the anglec of rotation between the adjacent two field pictures,
As the second data；
First data and second data are merged, fusion results data are obtained；To the fusion results
After data are filtered, the final exercise data of the adjacent two field pictures is obtained；
According to the final exercise data and the previous frame image of the adjacent two field pictures, the adjacent two frames figure is determined
The latter two field picture of picture, realizes the steady picture of the adjacent two field pictures.
For device embodiment, as which corresponds essentially to embodiment of the method, so related part is referring to method reality
Apply the part explanation of example.Device embodiment described above is only schematic, wherein described as separating component
The unit of explanation can be or may not be physically separate, as the part that unit shows can be or can also
It is not physical location, you can local to be located at one, or can also be distributed on multiple NEs.Can be according to reality
Need to select some or all of module therein to realize the purpose of this embodiment scheme.Those of ordinary skill in the art are not
In the case of paying creative work, you can to understand and implement.
It should be noted that herein, such as first and second or the like relational terms are used merely to a reality
Body or operation are made a distinction with another entity or operation, and are not necessarily required or implied these entities or deposit between operating
In any this actual relation or order.And, term " including ", "comprising" or its any other variant are intended to
Nonexcludability is included, so that a series of process, method, article or equipment including key elements not only will including those
Element, but also including other key elements being not expressly set out, or also include for this process, method, article or equipment
Intrinsic key element.In the absence of more restrictions, the key element for being limited by sentence "including a ...", it is not excluded that
Also there is other identical element in process, method, article or equipment including the key element.
A kind of video image stabilization method for being provided to the embodiment of the present invention above and device are described in detail, herein
Apply specific case to be set forth the principle and embodiment of the present invention, the explanation of above example is only intended to help
Understand the method for the present invention and its core concept；Simultaneously for one of ordinary skill in the art, according to the thought of the present invention,
Will change in specific embodiments and applications, in sum, this specification content is should not be construed as to this
The restriction of invention.
Claims (12)
1. a kind of video image stabilization method, it is characterised in that methods described includes：
Adjacent two field pictures are gathered, and records the acquisition time of the adjacent two field pictures respectively；According to described adjacent two
The exercise data of gyroscope in time period between the acquisition time of two field picture, calculates flat between the adjacent two field pictures
Shifting amount and the anglec of rotation, as the first data；
Detected using default Corner Detection Algorithm and whether there is in the adjacent two field pictures angle point, if it is, according to described
The match information of angle point in adjacent two field pictures, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as
Second data；
First data and second data are merged, fusion results data are obtained；To the fusion results data
After being filtered, the final exercise data of the adjacent two field pictures is obtained；
According to the final exercise data and the previous frame image of the adjacent two field pictures, the adjacent two field pictures are determined
Latter two field picture, realizes the steady picture of the adjacent two field pictures.
2. video image stabilization method according to claim 1, it is characterised in that methods described also includes：
If being detected using default Corner Detection Algorithm and there is no angle point in the adjacent two field pictures, directly to described first
Data are filtered, and obtain the final exercise data of the adjacent two field pictures；Continue executing with described according to the final motion
The previous frame image of data and the adjacent two field pictures, obtains the latter two field picture of the adjacent two field pictures, realizes described
The step of steady picture of adjacent two field pictures.
3. video image stabilization method according to claim 1, it is characterised in that the basis is in the adjacent two field pictures
The exercise data of gyroscope in time period between acquisition time, calculates the translational movement between the adjacent two field pictures and rotation
Gyration, as the first data, including：
The angular velocity of rotation of the acquisition time gyroscope in the adjacent two field pictures is obtained respectively；
According to the angular velocity of rotation of the gyroscope, the time period between the acquisition time of the adjacent two field pictures is calculated
The exercise data of interior gyroscope；
The camera focal length data obtained according to the exercise data of the gyroscope and in advance, calculate the adjacent two field pictures it
Between translational movement and the anglec of rotation, as the first data.
4. video image stabilization method according to claim 1, it is characterised in that described by first data and described second
Data are merged, and obtain fusion results data, including：
The average of translational movement between the described adjacent two field pictures in first data and second data is calculated respectively
Value, and the mean value of the anglec of rotation between the adjacent two field pictures, as fusion results data.
5. video image stabilization method according to claim 1, it is characterised in that described according to the final exercise data and institute
The previous frame image of adjacent two field pictures is stated, the latter two field picture of the adjacent two field pictures is determined, adjacent two frame is realized
The steady picture of image, including：
According to the final exercise data, line translation is entered to the previous frame image of the adjacent two field pictures, scheme after being converted
Picture；
Image after the conversion is defined as into the latter two field picture of the adjacent two field pictures, the adjacent two field pictures are realized
Steady picture.
6. video image stabilization method according to claim 1, it is characterised in that methods described also includes：
The fringe region of the latter two field picture of adjacent two field pictures described in cutting, to remove the black surround of image.
7. a kind of Video Stabilization device, it is characterised in that described device includes：
Acquisition module, for gathering adjacent two field pictures；
Logging modle, for recording the acquisition time of the adjacent two field pictures respectively；
First computing module, for gyroscope in time period of the basis between the acquisition time of the adjacent two field pictures
Exercise data, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the first data；
Detection module, for utilizing default Corner Detection Algorithm to detect in the adjacent two field pictures with the presence or absence of angle point；
Second computing module, for when the detection module detects angle point, according to angle point in the adjacent two field pictures
Match information, calculates the translational movement and the anglec of rotation between the adjacent two field pictures, as the second data；
Fusion Module, for first data and second data are merged, obtains fusion results data；
First filtration module, after being filtered to the fusion results data, obtains the final of the adjacent two field pictures
Exercise data；
Determining module, for the previous frame image according to the final exercise data and the adjacent two field pictures, it is determined that described
The latter two field picture of adjacent two field pictures, realizes the steady picture of the adjacent two field pictures.
8. Video Stabilization device according to claim 7, it is characterised in that described device also includes：
Second filtration module, for when the detection module is not detected by angle point, being filtered to first data, obtaining
The final exercise data of the adjacent two field pictures.
9. Video Stabilization device according to claim 7, it is characterised in that first computing module, including：
Acquisition submodule, for obtaining the angular velocity of rotation of the acquisition time gyroscope in the adjacent two field pictures respectively；
First calculating sub module, for the angular velocity of rotation according to the gyroscope, calculates adopting in the adjacent two field pictures
The exercise data of gyroscope in time period between collection time point；
Second calculating sub module, the camera focal length data obtained for the exercise data according to the gyroscope and in advance, meter
The translational movement and the anglec of rotation between the adjacent two field pictures is calculated, as the first data.
10. Video Stabilization device according to claim 7, it is characterised in that the Fusion Module, specifically for：
The average of translational movement between the described adjacent two field pictures in first data and second data is calculated respectively
Value, and the mean value of the anglec of rotation between the adjacent two field pictures, as fusion results data.
11. Video Stabilization devices according to claim 7, it is characterised in that the determining module, including：
Transformation submodule, for according to the final exercise data, becoming to the previous frame image of the adjacent two field pictures
Change, image after being converted；
Determination sub-module, for image after the conversion to be defined as the latter two field picture of the adjacent two field pictures, realizes institute
State the steady picture of adjacent two field pictures.
12. Video Stabilization devices according to claim 7, it is characterised in that described device also includes：
Cutting module, for the fringe region of the latter two field picture of adjacent two field pictures described in cutting, to remove the black surround of image.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611051216.6A CN106534692A (en) | 2016-11-24 | 2016-11-24 | Video image stabilization method and device |
| PCT/CN2017/111292 WO2018095262A1 (en) | 2016-11-24 | 2017-11-16 | Video stabilization method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611051216.6A CN106534692A (en) | 2016-11-24 | 2016-11-24 | Video image stabilization method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106534692A true CN106534692A (en) | 2017-03-22 |
Family
ID=58356968
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611051216.6A Pending CN106534692A (en) | 2016-11-24 | 2016-11-24 | Video image stabilization method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106534692A (en) |
| WO (1) | WO2018095262A1 (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106937053A (en) * | 2017-03-29 | 2017-07-07 | 维沃移动通信有限公司 | The digital image stabilization method and mobile terminal of a kind of video image |
| CN107220935A (en) * | 2017-05-25 | 2017-09-29 | 长光卫星技术有限公司 | A kind of in-orbit video image stabilization method of video satellite |
| CN107231526A (en) * | 2017-06-09 | 2017-10-03 | 联想(北京)有限公司 | Image processing method and electronic equipment |
| CN107770437A (en) * | 2017-09-08 | 2018-03-06 | 温州大学 | Unmanned plane photography and vedio recording system and its displacement compensating mechanism |
| CN107809590A (en) * | 2017-11-08 | 2018-03-16 | 青岛海信移动通信技术股份有限公司 | A kind of photographic method and device |
| WO2018095262A1 (en) * | 2016-11-24 | 2018-05-31 | 腾讯科技（深圳）有限公司 | Video stabilization method and device |
| CN108318506A (en) * | 2018-01-23 | 2018-07-24 | 深圳大学 | A kind of pipeline intelligent detection method and detecting system |
| WO2019205129A1 (en) | 2018-04-28 | 2019-10-31 | SZ DJI Technology Co., Ltd. | Motion estimation |
| CN111031200A (en) * | 2018-10-10 | 2020-04-17 | 三星电子株式会社 | Electronic device, camera and image stabilization method |
| CN112866556A (en) * | 2019-11-28 | 2021-05-28 | 南京理工大学 | Image stabilization method and system based on gyroscope and feature point matching |
| WO2021119982A1 (en) * | 2019-12-17 | 2021-06-24 | 深圳市大疆创新科技有限公司 | Image transmission system and method, control apparatus, and mobile platform |
| CN113243103A (en) * | 2018-12-26 | 2021-08-10 | 华为技术有限公司 | Imaging apparatus, image stabilization apparatus, imaging method, and image stabilization method |
| WO2022227040A1 (en) * | 2021-04-30 | 2022-11-03 | 深圳市大疆创新科技有限公司 | Video stability augmentation method, imaging apparatus, handheld gimbal, movable platform and storage medium |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112129317B (en) * | 2019-06-24 | 2022-09-02 | 南京地平线机器人技术有限公司 | Information acquisition time difference determining method and device, electronic equipment and storage medium |
| CN110956607B (en) * | 2019-07-11 | 2023-04-07 | 中国科学院国家天文台 | Method and device for testing ultimate signal-to-noise ratio and stability of ultimate signal-to-noise ratio |
| CN110991550B (en) * | 2019-12-13 | 2023-10-17 | 歌尔科技有限公司 | Video monitoring method and device, electronic equipment and storage medium |
| CN113923340B (en) * | 2020-07-09 | 2023-12-29 | 武汉Tcl集团工业研究院有限公司 | Video processing method, terminal and storage medium |
| CN112330618B (en) * | 2020-10-29 | 2023-09-01 | 浙江大华技术股份有限公司 | Image offset detection method, device and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104994304A (en) * | 2015-07-31 | 2015-10-21 | 小米科技有限责任公司 | Optical anti-vibration method and device, and terminal equipment |
| CN105096337A (en) * | 2014-05-23 | 2015-11-25 | 南京理工大学 | Image global motion compensation method based on hardware platform of gyroscope |
| CN105635588A (en) * | 2016-02-25 | 2016-06-01 | 杭州格像科技有限公司 | Image stabilization method and device |
| CN106027852A (en) * | 2016-06-24 | 2016-10-12 | 西北工业大学 | Video image stabilization method for micro/nano-satellite |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10322584A (en) * | 1997-05-20 | 1998-12-04 | Canon Inc | Shake correcting method, lens unit, image pickup device and computer readable recording medium |
| JP5859958B2 (en) * | 2010-11-11 | 2016-02-16 | パナソニック インテレクチュアル プロパティ コーポレーション オブアメリカＰａｎａｓｏｎｉｃ Ｉｎｔｅｌｌｅｃｔｕａｌ Ｐｒｏｐｅｒｔｙ Ｃｏｒｐｏｒａｔｉｏｎ ｏｆ Ａｍｅｒｉｃａ | Image processing apparatus, image processing method, and program |
| CN102780846B (en) * | 2012-07-11 | 2014-10-22 | 清华大学 | Electronic image stabilization method based on inertial navigation information |
| KR102282458B1 (en) * | 2015-03-23 | 2021-07-27 | 한화테크윈 주식회사 | Method and Device for dewobbling scene |
| CN106534692A (en) * | 2016-11-24 | 2017-03-22 | 腾讯科技（深圳）有限公司 | Video image stabilization method and device |
-
2016
- 2016-11-24 CN CN201611051216.6A patent/CN106534692A/en active Pending
-
2017
- 2017-11-16 WO PCT/CN2017/111292 patent/WO2018095262A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105096337A (en) * | 2014-05-23 | 2015-11-25 | 南京理工大学 | Image global motion compensation method based on hardware platform of gyroscope |
| CN104994304A (en) * | 2015-07-31 | 2015-10-21 | 小米科技有限责任公司 | Optical anti-vibration method and device, and terminal equipment |
| CN105635588A (en) * | 2016-02-25 | 2016-06-01 | 杭州格像科技有限公司 | Image stabilization method and device |
| CN106027852A (en) * | 2016-06-24 | 2016-10-12 | 西北工业大学 | Video image stabilization method for micro/nano-satellite |
Non-Patent Citations (2)
| Title |
|---|
| 张国栋: "基于电子稳像技术的视频稳像研究", 《中国优秀硕士学位论文全文数据库（电子期刊）》 * |
| 柴云峰: "基于光纤陀螺的稳像系统研究", 《中国优秀硕士学位论文全文数据库（电子期刊）》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018095262A1 (en) * | 2016-11-24 | 2018-05-31 | 腾讯科技（深圳）有限公司 | Video stabilization method and device |
| CN106937053A (en) * | 2017-03-29 | 2017-07-07 | 维沃移动通信有限公司 | The digital image stabilization method and mobile terminal of a kind of video image |
| CN107220935A (en) * | 2017-05-25 | 2017-09-29 | 长光卫星技术有限公司 | A kind of in-orbit video image stabilization method of video satellite |
| CN107220935B (en) * | 2017-05-25 | 2020-07-31 | 长光卫星技术有限公司 | Video satellite on-orbit video image stabilization method |
| CN107231526B (en) * | 2017-06-09 | 2020-02-21 | 联想(北京)有限公司 | Image processing method and electronic device |
| CN107231526A (en) * | 2017-06-09 | 2017-10-03 | 联想(北京)有限公司 | Image processing method and electronic equipment |
| CN107770437B (en) * | 2017-09-08 | 2020-03-17 | 温州大学 | Unmanned aerial vehicle photography and camera system and displacement compensation mechanism thereof |
| CN107770437A (en) * | 2017-09-08 | 2018-03-06 | 温州大学 | Unmanned plane photography and vedio recording system and its displacement compensating mechanism |
| CN107809590A (en) * | 2017-11-08 | 2018-03-16 | 青岛海信移动通信技术股份有限公司 | A kind of photographic method and device |
| CN107809590B (en) * | 2017-11-08 | 2020-04-28 | 青岛海信移动通信技术股份有限公司 | Photographing method and device |
| CN108318506A (en) * | 2018-01-23 | 2018-07-24 | 深圳大学 | A kind of pipeline intelligent detection method and detecting system |
| WO2019205129A1 (en) | 2018-04-28 | 2019-10-31 | SZ DJI Technology Co., Ltd. | Motion estimation |
| US11172218B2 (en) | 2018-04-28 | 2021-11-09 | SZ DJI Technology Co., Ltd. | Motion estimation |
| CN111279685A (en) * | 2018-04-28 | 2020-06-12 | 深圳市大疆创新科技有限公司 | Motion estimation |
| CN111031200A (en) * | 2018-10-10 | 2020-04-17 | 三星电子株式会社 | Electronic device, camera and image stabilization method |
| CN111031200B (en) * | 2018-10-10 | 2023-05-23 | 三星电子株式会社 | Electronic device, camera, and image stabilization method |
| CN113243103A (en) * | 2018-12-26 | 2021-08-10 | 华为技术有限公司 | Imaging apparatus, image stabilization apparatus, imaging method, and image stabilization method |
| CN113243103B (en) * | 2018-12-26 | 2022-11-22 | 华为技术有限公司 | Imaging apparatus, image stabilization apparatus, imaging method, and image stabilization method |
| CN112866556A (en) * | 2019-11-28 | 2021-05-28 | 南京理工大学 | Image stabilization method and system based on gyroscope and feature point matching |
| WO2021119982A1 (en) * | 2019-12-17 | 2021-06-24 | 深圳市大疆创新科技有限公司 | Image transmission system and method, control apparatus, and mobile platform |
| WO2022227040A1 (en) * | 2021-04-30 | 2022-11-03 | 深圳市大疆创新科技有限公司 | Video stability augmentation method, imaging apparatus, handheld gimbal, movable platform and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2018095262A1 (en) | 2018-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106534692A (en) | Video image stabilization method and device | |
| CN108898567B (en) | Image noise reduction method, device and system | |
| CN1979087B (en) | Image processing device and method | |
| CN101540046B (en) | Panoramagram montage method and device based on image characteristics | |
| CN101516001B (en) | Digital photographing apparatus, method of controlling the digital photographing apparatus, and recording medium | |
| CN104065854B (en) | A kind of image processing method and a kind of electronic equipment | |
| CN105635588B (en) | A kind of digital image stabilization method and device | |
| CN107146200B (en) | Unmanned aerial vehicle remote sensing image splicing method based on image splicing quality evaluation | |
| CN103985103A (en) | Method and device for generating panoramic picture | |
| KR102356448B1 (en) | Method for composing image and electronic device thereof | |
| US20120092559A1 (en) | Rolling Shutter Distortion Correction | |
| JP2009516257A (en) | Method and system for generating video summaries | |
| IL249739A (en) | System and method for secured capturing and authenticating of video clips | |
| CN103279952A (en) | Target tracking method and device | |
| CN111310759B (en) | Target detection inhibition optimization method and device for dual-mode cooperation | |
| CN109922267A (en) | Image stabilization processing method, computer installation and computer readable storage medium based on gyro data | |
| US20200160560A1 (en) | Method, system and apparatus for stabilising frames of a captured video sequence | |
| JP7043601B2 (en) | Methods and devices for generating environmental models and storage media | |
| CN107341787A (en) | Method, apparatus, server and the automobile that monocular panorama is parked | |
| CN108229281B (en) | Neural network generation method, face detection device and electronic equipment | |
| KR101868740B1 (en) | Apparatus and method for generating panorama image | |
| US9995905B2 (en) | Method for creating a camera capture effect from user space in a camera capture system | |
| CN104123716A (en) | Image stability detection method, device and terminal | |
| WO2022174603A1 (en) | Pose prediction method, pose prediction apparatus, and robot | |
| Rajakaruna et al. | Image deblurring for navigation systems of vision impaired people using sensor fusion data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |