CN107507230A - Method for tracking target and panorama camera applied to panoramic picture - Google Patents
Method for tracking target and panorama camera applied to panoramic picture Download PDFInfo
- Publication number
- CN107507230A CN107507230A CN201710773963.9A CN201710773963A CN107507230A CN 107507230 A CN107507230 A CN 107507230A CN 201710773963 A CN201710773963 A CN 201710773963A CN 107507230 A CN107507230 A CN 107507230A
- Authority
- CN
- China
- Prior art keywords
- picture
- panoramic picture
- wide angle
- frame
- window
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/292—Multi-camera tracking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of method for tracking target and panorama camera applied to panoramic picture, and N number of wide angle picture is gathered using N number of wide-angle lens each moment, wherein, the corresponding frame panoramic picture of N number of wide angle picture of each moment collection；According to the corresponding relation between the pixel of wide angle picture and projection video in window, multiple projection video in windows are extracted from N number of wide angle picture；Using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, it is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to present frame panoramic picture；Wherein, an adjacent frame or multiframe panoramic picture are before present frame panoramic picture.Method for tracking target and panorama camera provided by the invention applied to panoramic picture, the real-time and accuracy of target following can be effectively improved.
Description
Technical field
The present invention relates to panoramic video technical field, more particularly to a kind of method for tracking target applied to panoramic picture and
Panorama camera.
Background technology
Panoramic video is made up of the panoramic image sequence associated with some having times, can be obtained from video true
Panorama scene, and positioning video frame can be carried out with passage time.Shoot the global fortune of the cube panoramic video of static scene
Move generally based on horizontal movement, cover 360 degree of scene domains, for the panoramic picture of continuous multiple frames, target object
Adjacent surface can be moved to from a face of cube panorama.Because it has dynamic background, be not suitable for the method using background subtraction
Carry out the detect and track of dynamic object, it is therefore desirable to design the target following in a kind of cube panoramic video rapidly and efficiently
Method.
The method that prior art tracks object in panoramic video, generally according to user viewing angle by panoramic video figure
As projecting into projection view, and as the input picture of target following.The model of target object is established during tracking, and
The probability for tracking target and appearing in different zones is established according to front and rear frame, the new position of target is used as using the position of greater probability；
It is not in missing inspection that premise using the above method, which is, and its direction of interest is going out necessarily comprising the object for needing to track
In the case of existing missing inspection, the accuracy that it is tracked is extremely low, so as to cause to occur existing method exist target in panoramic picture with
The problem of track accuracy is low.
And use the above method, it is necessary first to first obtain full-view video image and projected to obtain the throwing of preset field of view again
Shadow image, then carry out the probability that front and rear frame tracking target appears in different zones, determine to track target with the size of probability
New position, its calculation procedure is more and amount of calculation is larger, thus promotes the tracking of panoramic video less efficient, causes panorama occur
The problem of real-time of target following in image is poor.
The content of the invention
The invention provides a kind of method for tracking target and panorama camera applied to panoramic picture, mesh can be effectively improved
Mark the real-time and accuracy of tracking.
The application first aspect provides a kind of method for tracking target applied to panoramic picture, including：
N number of wide angle picture is gathered using N number of wide-angle lens each moment, wherein, N number of wide angle picture of each moment collection
A corresponding frame panoramic picture；
According to the corresponding relation between the pixel of wide angle picture and projection video in window, from N number of wide angle picture
Multiple projection video in windows are extracted, wherein, the projection video in window is in panoramic picture corresponding to N number of wide angle picture
The projection of parts of images, and overlapping region be present with least one projection video in window in each projection video in window；
Using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, in present frame panorama
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to image；Wherein, an adjacent frame or more
Frame panoramic picture is before present frame panoramic picture.
With reference in a first aspect, in the first possible embodiment of first aspect, according to wide angle picture and projection
Before corresponding relation between the pixel of video in window, methods described also includes：
Obtain the corresponding relation between wide angle picture and the pixel for projecting video in window.
In the first possible embodiment with reference to first aspect, in second of possible embodiment of first aspect
In, the corresponding relation between the pixel for obtaining wide angle picture and projecting video in window, specifically include：
According to the appearance between the mapping relations of latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle lens
State relation, calculate the first corresponding relation between the pixel of panoramic picture and N number of wide angle picture；
According to the division rule of projection video in window, the between panoramic picture and the pixel for projecting video in window is obtained
Two corresponding relations；
According to first corresponding relation and second corresponding relation, obtain wide angle picture and project the picture of video in window
Corresponding relation between vegetarian refreshments.
In second of possible embodiment with reference to first aspect, in the third possible embodiment of first aspect
In, the division rule includes the division rule based on default longitude, the division rule of default latitude and based on preset field of view
Division rule.
With reference to second of first aspect into the third possible embodiment, the 4th kind in first aspect is possible
In embodiment, the tracking target in using all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture is tracked
During, methods described also includes：
By in the currently displayed frame panoramic picture of the successful target following result of the tracking of previous frame.
In the 4th kind of possible embodiment with reference to first aspect, in the 5th kind of possible embodiment of first aspect
In, before by the currently displayed frame panoramic picture of the successful target following result of the tracking of previous frame, methods described is also wrapped
Include:
According to first corresponding relation, N number of wide angle picture corresponding to present frame is spliced into present frame panoramic picture.
With reference to second of first aspect into the third possible embodiment, the 6th kind in first aspect is possible
In embodiment, the tracking target using in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, working as
It is tracked, specifically includes in the multiple projection video in windows extracted in N number of wide angle picture corresponding to previous frame panoramic picture：
It is N number of wide according to corresponding in second corresponding relation and an adjacent frame or multiframe panoramic picture per frame panoramic picture
Angle image, obtain in adjacent a frame or multiframe panoramic picture per multiple projection video in windows corresponding to frame panoramic picture；
Based on preparatory condition, from multiple projection video in windows corresponding to N number of wide angle picture corresponding to every frame panoramic picture
Choose the tracking target；
Using the tracking target chosen from multiple projection video in windows corresponding to every frame panoramic picture, in present frame
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to panoramic picture.
According to the second aspect of the invention, there is provided a kind of panorama camera, including N number of wide-angle lens and readable storage medium storing program for executing,
Had program stored therein on the readable storage medium storing program for executing, following steps are realized when described program is executed by processor：
N number of wide angle picture is gathered using N number of wide-angle lens each moment, wherein, N number of wide-angle of each moment collection
The corresponding frame panoramic picture of image；
According to the corresponding relation between the pixel of wide angle picture and projection video in window, from N number of wide angle picture
Multiple projection video in windows are extracted, wherein, the projection video in window is in panoramic picture corresponding to N number of wide angle picture
The projection of parts of images；
Using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, in present frame panorama
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to image；Wherein, an adjacent frame or more
Frame panoramic picture is before present frame panoramic picture.
With reference to second aspect, in the first possible embodiment of second aspect, described program is executed by processor
When according to wide angle picture and project video in window pixel between corresponding relation before, also realize following steps：Obtain
Corresponding relation between wide angle picture and the pixel for projecting video in window.
In the first possible embodiment with reference to second aspect, in second of possible embodiment of second aspect
In, the step of corresponding relation when described program is executed by processor between acquisition wide angle picture and the pixel for projecting video in window
When rapid, following steps are specifically included：
According to the appearance between the mapping relations of latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle lens
State relation, calculate the first corresponding relation between the pixel of panoramic picture and N number of wide angle picture；
According to the division rule of projection video in window, the between panoramic picture and the pixel for projecting video in window is obtained
Two corresponding relations；
According to first corresponding relation and second corresponding relation, obtain wide angle picture and project the picture of video in window
Corresponding relation between vegetarian refreshments.
In second of possible embodiment with reference to second aspect, in the third possible embodiment of second aspect
In, when described program is executed by processor in using all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture with
During track target is tracked, following steps are also realized：The successful target following result of the tracking of previous frame is included working as
In previous frame panoramic picture.
In second of possible embodiment with reference to second aspect, in the 4th kind of possible embodiment of second aspect
In, the tracking in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture is utilized when described program is executed by processor
Target, the step being tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to present frame panoramic picture
When rapid, following steps are specifically included：
It is N number of wide according to corresponding in second corresponding relation and an adjacent frame or multiframe panoramic picture per frame panoramic picture
Angle image, obtain in adjacent a frame or multiframe panoramic picture per multiple projection video in windows corresponding to frame panoramic picture；
Based on preparatory condition, from multiple projection video in windows corresponding to N number of wide angle picture corresponding to every frame panoramic picture
Choose the tracking target；
Using the tracking target chosen from multiple projection video in windows corresponding to every frame panoramic picture, in present frame
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to panoramic picture.
Beneficial effects of the present invention are as follows：
Due to the embodiment of the present application be according to wide angle picture and project video in window pixel between corresponding relation, from
Multiple projection video in windows are extracted in N number of wide angle picture of collection；Then target following is carried out in multiple projection video in windows；
The embodiment of the present invention by N number of wide angle picture without being converted into panoramic picture；And the amount of calculation of the track algorithm used is also obvious
Less than the amount of calculation for the new position that the track algorithm that prior art uses determines tracking target.Therefore, compared with prior art, its
Calculation procedure is reduced and amount of calculation reduces, thus promotes the tracking efficiency of panoramic video to improve, and can effectively improve panoramic video
In target following real-time.
Further, because projection video in window is the parts of images in panoramic picture corresponding to N number of wide angle picture
Projection, promote the distortion for projecting video in window minimum or be not present；And in minimum or in the absence of distortion the projection window that distorts
When target following is carried out in image, its accuracy tracked can greatly improve, so as to effectively improve the accuracy of tracking.
Brief description of the drawings
Fig. 1 is the flow chart for the method for tracking target for being applied to panoramic picture in the embodiment of the present invention；
Fig. 2 is the corresponding diagram of panoramic picture and projection video in window in the embodiment of the present invention.
Embodiment
The invention provides a kind of method for tracking target and panorama camera applied to panoramic picture, mesh can be effectively improved
Mark the real-time and accuracy of tracking.
As shown in figure 1, a kind of method for tracking target applied to panoramic picture is provided in first aspect present invention, including
Following steps：
S101, using N number of wide-angle lens each moment gather N number of wide angle picture, wherein, each moment collection it is N number of wide
The corresponding frame panoramic picture of angle image；
S102, according to wide angle picture and project video in window pixel between corresponding relation, from N number of wide-angle figure
Multiple projection video in windows are extracted as in, wherein, the projection video in window is panoramic picture corresponding to N number of wide angle picture
In parts of images projection；
S103, using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, current
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to frame panoramic picture；Wherein, described adjacent one
Frame or multiframe panoramic picture are before present frame panoramic picture.
In the embodiment of the present invention, N number of wide-angle lens is arranged in a panorama camera or described N number of wide
Angle mirror head connects same main frame, and the application is not specifically limited.Lower mask body is arranged on a panorama with N number of wide-angle lens
Illustrated exemplified by camera.
In the embodiment of the present invention, N number of wide-angle lens is by common wide-angle lens, bugeye lens and fish eye lens
One or more camera lenses composition, the application is not restricted.
Wherein, in step S101, during panoramic video is recorded by N number of wide-angle lens, N number of wide-angle lens is utilized
Each moment gathers N number of wide angle picture, then can be spliced with the real-time or non real-time N number of wide angle picture for gathering each moment
Into a frame panoramic picture；Finally all panoramic pictures being spliced into are ranked up sequentially in time, obtain panoramic video.
In the embodiment of the present invention, it can be spliced into by N number of wide angle picture that merging algorithm for images gathers each moment
One frame panoramic picture.Described image stitching algorithm for example can be related to feature based based on the related stitching algorithm in region
Joining method etc..
Preferably, can be by the R1 that is calculated in step S102, N number of wide angle picture that each moment is gathered splices
Into a frame panoramic picture.
Next step S102 is performed, in N number of wide angle picture corresponding to one frame panoramic picture of every collection, according to wide-angle figure
Corresponding relation between picture and the pixel for projecting video in window, multiple projection window figures are extracted from N number of wide angle picture
Picture.
In specific implementation process, the corresponding relation between according to the pixel of wide angle picture and projection video in window,
Before multiple projection video in windows are extracted from N number of wide angle picture, also need to obtain wide angle picture and project video in window
Corresponding relation between pixel.
Specifically, mapping relations that can be according to latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle
Posture relation between camera lens, calculate the first corresponding relation between the pixel of panoramic picture and N number of wide angle picture；
Further according to the division rule of projection video in window, the second correspondence between panoramic picture and the pixel for projecting video in window is obtained
Relation；Finally according to first corresponding relation and second corresponding relation, obtain wide angle picture and project video in window
Corresponding relation between pixel.
Specifically, the picture of first corresponding relation, second corresponding relation and wide angle picture and projection video in window
Corresponding relation between vegetarian refreshments can be before the method in performing the embodiment of the present invention, just have calculated that and store panorama
In the memory of camera.When needing to use, directly read from the relevant position of memory need to use one or
Then multiple corresponding relations are called.Certainly, above three corresponding relation can also perform the embodiment of the present invention
By being calculated accordingly while method, the application is not specifically limited.
In the embodiment of the present invention, when obtaining first corresponding relation and second corresponding relation, can synchronously it obtain
Take, asynchronous can also obtain, the application is not specifically limited.Synchronous obtain refers to perform acquisition first correspondence simultaneously
The step of relation and second corresponding relation；The asynchronous acquisition refers to first carry out the step of obtaining first corresponding relation
Perform and obtain after the step of performing the step of obtaining second corresponding relation afterwards, or, first carrying out acquisition second corresponding relation
The step of taking first corresponding relation.
Specifically, when obtaining first corresponding relation, because each camera lens is fixed in N number of wide-angle lens
Constant, so as to understand the posture relation between each wide-angle lens.N number of wide angle picture meeting of N number of wide-angle lens collection
Panoramic picture is spliced into, because the posture relation between each wide-angle lens is constant, promotes each pixel in panoramic picture
The first corresponding relation with the pixel in N number of wide angle picture is constant.
If the panoramic picture PanoImg tables that the wide angle picture of each wide-angle lens collection is represented and is spliced into camImg
Show, then can determine the first couple between the pixel of the panoramic picture that is spliced into and the wide angle picture of each wide-angle lens collection
It should be related to that R1 can use { panoImg->CamImg } represent.
Specifically, can be with the first corresponding relation between determining the pixel of panoramic picture and N number of wide angle picture
According to the posture relation between the mapping relations of latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle lens, meter
Calculate first corresponding between panoramic picture and the pixel for the wide angle picture (i.e. N number of wide angle picture) that each wide-angle lens gathers
Relation.
In actual applications, a pixel in panoramic picture may correspond to some in one or more wide angle pictures
Pixel so that some pixel in panoramic picture may have corresponding pixel points in multiple wide angle pictures, it is also possible to
Only corresponding pixel points be present in a wide angle picture.In this way, the pixel in order to ensure each pixel in the panoramic picture
The accuracy of value, the pixel value of each pixel in the panoramic picture can be determined using following manner, be specially：If institute
It is corresponding with the c1 pixels in the c pixels and b1 wide angle pictures in b wide angle pictures respectively to state a pixels in panoramic picture, then
According to c pixels and the pixel value and weight of c1 pixels, the pixel value of a pixels in the panoramic picture is determined；It is if described
A1 pixels are only corresponding with the c2 pixels in b wide angle pictures in panoramic picture, it is determined that the a1 pixels in the panoramic picture
The pixel value of point is the pixel value of c2 pixels.Wherein, { panoImg->CamImg } include corresponding relation (a->(b, c),
(b1, c1)) and (a1->(b, c2)).
Specifically, when obtaining the second relation between panoramic picture and the pixel for projecting video in window, Ke Yigen
According to the division rule of projection video in window, obtain panoramic picture and project the second corresponding relation between window image pixel point.
The division rule can be the division rule based on default longitude or default latitude or based on default longitude and latitude, can also be pre-
If division rule of visual field etc..
Specifically, panoramic picture is mapped to spherical diagram picture first, one or more throw then is chosen in spherical diagram picture
Shadow point, then choose multiple projection video in windows of projection from spherical diagram picture based on division rule again.In this way, according to each throwing
The corresponding relation of shadow video in window, spherical diagram picture and panoramic picture, each pixel in each projection video in window can be obtained
The corresponding relation of point and the pixel in panoramic picture.If multiple projection video in windows are represented with ProjImg, can determine complete
Corresponding relation R2 between the pixel of scape image and multiple projection video in windows can use { ProjImg->PanoImg } represent.
For example, the e pixels that a pixels in panoramic picture are projected with d in video in window are corresponding, the a1 in panoramic picture
The e1 pixels that pixel is projected with d1 in video in window are corresponding, then can determine { ProjImg->PanoImg } include correspondingly
Relation (a->(d, e)) and (a1->(d1, e1)).
In actual applications, after R1 and R2 is obtained, N number of wide angle picture and can be calculated according to R1 and R2 more
The corresponding relation R3 of pixel in individual projection video in window can use { ProjImg->CamImg } represent.
For example, { panoImg->CamImg } include corresponding relation (a->(b, c), (b1, c1)) and (a1->(b, c2)),
And { ProjImg->PanoImg } include corresponding relation (a->(d, e)) and (a1->(d1, e1)), then it can derive
{ProjImg->CamImg } include by corresponding relation ((b, c), (b1, c1)->(d, e)) and ((b, c2)->(d1, e1)).
Preferably due to the embodiment of the present invention needs to carry out target following, it is desirable, therefore, to assure that each projecting video in window
Overlapping region be present with least one projection video in window, tracking target occur with reduction enters separately from a projection video in window
There is the probability of situation about disappearing during one projection video in window.If each projection video in window and any one projection video in window
In the absence of overlapping region, when in tracking target from a projection video in window into another projection video in window, due to not
Overlapping region be present, so as to which the problem of tracking target disappears from projection video in window occur；And in each projection window figure
As overlapping region be present with least one projection video in window so that there is very maximum probability to be thrown by overlapping region from one in target
Shadow video in window enters another projection video in window.So, it is possible effectively reduces target enters from a projection video in window
There is the probability of situation about disappearing during another projection video in window, improve the accuracy of target following.
During the present invention is implemented, after R3 is calculated by R1 and R2, R1, R2 and R3 can be directly stored so that needing
When calling R1, R2 and R3, R1, R2 and R3 directly are read from the position of storage, R1, R2 and R3 are calculated again to avoid the occurrence of
Cause to consume the more time.So, it is possible to shorten panoramic video it is live during carry out target following data processing it is time-consuming
Duration.
In actual applications, as shown in Fig. 2 panoramic picture is mapped to spherical diagram as 20, then selected in comprehensive image 20
Take subpoint 200, wherein, based on division rule from spherical diagram as 20 in choose projection video in window 30, project video in window 30
With spherical diagram as on 20 this four pixels of pixel 201, pixel 202, pixel 203 and pixel 204 it is corresponding；Based on
The same mode of projection video in window 30 is chosen, remaining all projection video in windows are chosen from panoramic picture 20.
Wherein, video in window 30 is projected compared with spherical diagram spherical area as corresponding to aforementioned four pixel in 20, its
Distortion is smaller or is not present, and making it, data are more accurate when projecting video in window 30 and being tracked, and promote the accurate of target following
Property improve.
Next step S103 is performed, before the step is performed, it is also necessary to obtain adjacent a frame or multiframe panoramic picture
Tracking target in corresponding all wide angle pictures.Obtaining all wide angle pictures corresponding to adjacent a frame or multiframe panoramic picture
In tracking target when, preparatory condition can be based on, from all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture
Choose tracking target.
Specifically, can be according to every frame panorama sketch in second corresponding relation and an adjacent frame or multiframe panoramic picture
N number of wide angle picture as corresponding to, obtain in adjacent a frame or multiframe panoramic picture per multiple projection windows corresponding to frame panoramic picture
Mouth image；Preparatory condition is based on again, from multiple projection video in windows corresponding to N number of wide angle picture corresponding to every frame panoramic picture
Choose the tracking target；Utilize the tracking mesh chosen from multiple projection video in windows corresponding to every frame panoramic picture
Mark, be tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to present frame panoramic picture.
Specifically, the preparatory condition can be that user voluntarily chooses or be detected by object detection method
Arrive, can also be what is traced into by method for tracking target, at least two methods in above-mentioned three kinds of methods can also be used
Obtain.The object detection method can be method for detecting human face, the deep learning target inspection based on Region Proposal
Method of determining and calculating and biometric discrimination method etc., the application is not specifically limited.
Further, the method for tracking target can the tracking based on active profile, the track side of feature based
Method, the tracking based on region and tracking based on model etc..
In the embodiment of the present invention, the quantity of the tracking target can be with 1 or more than 1, and the application is not specifically limited.
In the embodiment of the present invention, the tracking target in using all wide angle pictures corresponding to adjacent frame panoramic picture
State, to the appearance of the tracking target in multiple projection video in windows for being extracted in N number of wide angle picture corresponding to present frame panoramic picture
, can first choice N number of wide angle picture, meter according to corresponding to the R3 and previous frame panoramic picture that step S102 is obtained when state is tracked
Calculation obtains multiple projection video in windows corresponding to previous frame panoramic picture；Then it is complete from previous frame further according to the preparatory condition
Tracking target is chosen in multiple projection video in windows corresponding to scape image；, equally can root after the tracking target is chosen
According to N number of wide angle picture corresponding to R3 and present frame panoramic picture, multiple projection windows corresponding to present frame panoramic picture are calculated
Mouth image；The tracking target chosen is recycled to be tracked in multiple projection video in windows corresponding to previous frame panoramic picture.
For example, 4 fish eye lenses are respectively F, G, H and K fish eye lens, the fish eye images that each fish eye lens is shot respectively
It is followed successively by (F1, F2 ... Fm), (G1, G2 ... Gm), (H1, H2 ... Hm) and (K1, K2 ... Km).It follows that per frame
4 fish eye images corresponding to panoramic picture be followed successively by (F1, G1, H1, K1), (F2, G2, H2, K2) ... (Fm, Gm, Hm,
Km) }, wherein m is the integer not less than 3；If 4 fish eye images are (F3, G3, H3, K3) corresponding to present frame panoramic picture,
4 fish eye images are (F2, G2, H2, K2) corresponding to adjacent frame panoramic picture before present frame panoramic picture can be determined.
Wherein, tracking is chosen from multiple window projected images corresponding to (F2, G2, H2, K2) according to preparatory condition first
Target, such as tracking can be chosen in multiple window projected images according to corresponding to method for detecting human face from (F2, G2, H2, K2)
Target；After the tracking target is chosen, according to the step S102 R3 being calculated and (F3, G3, H3, K3), it is calculated
Multiple projection video in windows corresponding to (F3, G3, H3, K3)；Then by the tracking target of selection by optical flow method (F3, G3,
H3, K3) corresponding to it is multiple projection video in windows in be tracked.
During the present invention is implemented, the tracking target in using all wide angle pictures corresponding to adjacent multiframe panoramic picture,
, can be preferred when being tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to present frame panoramic picture
According to all wide angle pictures corresponding to the step S102 R3 obtained and adjacent multiframe panoramic picture, adjacent multiframe panorama is calculated
Per multiple projection video in windows corresponding to frame panoramic picture in image；Then further according to the preparatory condition, from every frame panorama sketch
Tracking target is chosen in multiple projection video in windows as corresponding to；, equally can be according to R3 after the tracking target is chosen
With present frame panoramic picture corresponding to N number of wide angle picture, be calculated corresponding to present frame panoramic picture it is multiple projection window figures
Picture；The tracking target chosen is recycled to be tracked in multiple projection video in windows corresponding to previous frame panoramic picture.Under
Mask body illustrates so that the tracking target in all wide angle pictures corresponding to adjacent two frames panoramic picture is tracked as an example.
For example, 4 fish eye lenses are respectively F, G, H and K fish eye lens, the fish eye images that each fish eye lens is shot respectively
It is followed successively by (F1, F2 ... Fm), (G1, G2 ... Gm), (H1, H2 ... Hm) and (K1, K2 ... Km).It follows that per frame
4 fish eye images corresponding to panoramic picture be followed successively by (F1, G1, H1, K1), (F2, G2, H2, K2) ... (Fm, Gm, Hm,
Km) }, wherein m is the integer not less than 3；If 4 fish eye images are (F3, G3, H3, K3) corresponding to present frame panoramic picture,
4 fish eye images are (F1, G1, H1, K1) corresponding to adjacent two frames panoramic picture before present frame panoramic picture can be determined
(F2, G2, H2, K2).
Wherein, first according to preparatory condition, respectively from multiple windows corresponding to (F1, G1, H1, K1) and (F2, G2, H2, K2)
Choose tracking target in mouthful projected image, for example, can according to corresponding to method for detecting human face from (F1, G1, H1, K1) multiple windows
The first tracking target is chosen in mouth projected image, and from multiple corresponding to (F2, G2, H2, K2) by way of subscriber frame choosing
The second tracking target is chosen in window projected image；After the first tracking target and the second tracking target is chosen,
According to the step S102 R3 being calculated and (F3, G3, H3, K3), multiple projection windows corresponding to (F3, G3, H3, K3) are calculated
Mouth image；Then target and the second tracking target are tracked by optical flow method corresponding to (F3, G3, H3, K3) by the first of selection
It is tracked in multiple projection video in windows.
In the embodiment of the present invention, due to being to carry out target following in multiple projection video in windows, and with panoramic picture
Distortion is compared, and the distortion of multiple projection video in windows is minimum or is not present, in minimum or in the absence of distortion the multiple projections that distort
The accuracy of target following is carried out in video in window to be increased substantially.
In the embodiment of the present invention, above-mentioned data can be handled in the internal processor of the panorama camera so that on
State that data are shorter from the path for collecting data transfer in processing procedure, reduce the time loss of data transfer, further improve
The real-time of target following during panoramic video is live.It is of course also possible to by the above-mentioned data transfer of collection to it is described complete
Handled in the connected main frame of scape camera, now, the time loss according to transmission can be greatly improved, its real-time can drop therewith
It is low.
The present invention implement in, due to multiple projection video in windows be present, it is multiple projection video in windows in carry out target with
Can be with parallel processing, to reduce the time loss of object tracking process, mesh during further raising panoramic video is live during track
Mark the real-time of tracking.Certainly, it is multiple projection video in windows in carry out target following when can also serial process, compared to simultaneously
Row processing, the time loss of its object tracking process may increase.
In the embodiment of the present invention, in order to further improve the real-time and tracking result that the above method carries out target following
Accuracy, it is while step S103 is performed, the successful currently displayed frame of target following result of the tracking of previous frame is complete
In scape image.
Specifically, because every frame panoramic picture is required for being spliced according to corresponding N number of wide angle picture so that inciting somebody to action
Corresponding N number of wide angle picture is required to consume the regular hour when being spliced into panoramic picture；And entered by step S101~S103
It is also required to consume the regular hour during row target following.And carry out target following compared with being spliced into panoramic picture, its need into
The data volume calculated of going is bigger and also more the step of processing so that is spliced into the time far smaller than progress that panoramic picture is consumed
The time that target following is consumed, and be spliced into panoramic picture and carry out target following and handled simultaneously so that splicing
Differed greatly into panoramic picture and progress target following each self-corresponding moment.
In actual applications, N number of wide angle picture splicing corresponding with next frame at the time of corresponding during target following success
Gap very little at the time of into corresponding to panoramic picture, in this way, the successful target following result of the tracking of previous frame can be shown
In present frame panoramic picture, to improve the real-time and accuracy that target following is shown.
Specifically, by N number of wide angle picture corresponding to each frame, N number of wide angle picture can be spliced according to stitching algorithm
Into panoramic picture；Then successful target following result, which will be tracked, to be included in the present frame panoramic picture being spliced into.
Preferably, according to R1, N number of wide angle picture corresponding to present frame is spliced into present frame panoramic picture；Then will be upper
One frame tracks successful target following result and is shown in the present frame panoramic picture being calculated.
Specifically, when previous frame being tracked in the currently displayed frame panoramic picture of successful target following result, it is necessary to
According to R2, previous frame is tracked into successful target following result and carries out Coordinate Conversion, then by the target following after Coordinate Conversion
As a result in currently displayed frame panoramic picture.
For example, the first tracking target of selection and the second tracking target is right at (F3, G3, H3, K3) by optical flow method
When tracking successfully in the multiple projection video in windows answered, N number of wide angle picture corresponding to present frame is (F4, G4, H4, K4).In this way,
According to the R1 of storage, present frame panoramic picture corresponding to (F4, G4, H4, K4) is calculated；Then will track successful first,
Second tracking target is shown in the present frame panoramic picture being calculated after carrying out Coordinate Conversion.
Because the embodiment of the present invention is that previous frame is tracked into successful target following result to show in present frame panoramic picture.
Need to be spaced a frame it follows that target following successfully carries out display, that is, to carry out target during panoramic video is live
Tracking only exists the delay of a frame.So, it is possible to effectively improve panoramic video it is live during target following real-time.
Further, due to being tracked in multiple projection video in windows, and project and do not distorted in video in window,
Its accuracy tracked in panoramic picture than having the accuracy of the tracking of distortion higher.
Based on the method identical technical concept provided with first aspect present invention, second aspect of the present invention additionally provides one
Panorama camera, including N number of wide-angle lens and readable storage medium storing program for executing are planted, is had program stored therein on the readable storage medium storing program for executing, the journey
Following steps are realized when sequence is executed by processor：
N number of wide angle picture is gathered using N number of wide-angle lens each moment, wherein, N number of wide-angle of each moment collection
The corresponding frame panoramic picture of image；
According to the corresponding relation between the pixel of wide angle picture and projection video in window, from N number of wide angle picture
Multiple projection video in windows are extracted, wherein, the projection video in window is in panoramic picture corresponding to N number of wide angle picture
The projection of parts of images；
Using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, in present frame panorama
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to image；Wherein, an adjacent frame or more
Frame panoramic picture is before present frame panoramic picture.
Preferably, when described program is executed by processor between according to the pixel of wide angle picture and projection video in window
Corresponding relation before, also realize following steps：Obtain the corresponding pass between wide angle picture and the pixel for projecting video in window
System.
Preferably, each there is overlapping region in projection video in window with least one projection video in window.
Preferably, obtained when described program is executed by processor between wide angle picture and the pixel for projecting video in window
During the step of corresponding relation, following steps are specifically included：
According to the appearance between the mapping relations of latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle lens
State relation, calculate the first corresponding relation between the pixel of panoramic picture and N number of wide angle picture；
According to the division rule of projection video in window, the between panoramic picture and the pixel for projecting video in window is obtained
Two corresponding relations；
According to first corresponding relation and second corresponding relation, obtain wide angle picture and project the picture of video in window
Corresponding relation between vegetarian refreshments.
Preferably, the division rule includes the division rule based on default longitude, the division rule and base of default latitude
In the division rule of preset field of view.
Preferably, all wide corresponding to the adjacent frame of utilization or multiframe panoramic picture when described program is executed by processor
During tracking target in the image of angle is tracked, following steps are also realized：By the successful target following of the tracking of previous frame
As a result in currently displayed frame panoramic picture.
Preferably, all wide-angles corresponding to an adjacent frame or multiframe panoramic picture are utilized when described program is executed by processor
Tracking target in image, in the multiple projection video in windows extracted in N number of wide angle picture corresponding to present frame panoramic picture
During the step being tracked, following steps are specifically included：
It is N number of wide according to corresponding in second corresponding relation and an adjacent frame or multiframe panoramic picture per frame panoramic picture
Angle image, obtain in adjacent a frame or multiframe panoramic picture per multiple projection video in windows corresponding to frame panoramic picture；
Based on preparatory condition, from multiple projection video in windows corresponding to N number of wide angle picture corresponding to every frame panoramic picture
Choose the tracking target；
Using the tracking target chosen from multiple projection video in windows corresponding to every frame panoramic picture, in present frame
It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to panoramic picture.
Preferably, the preparatory condition can be that user voluntarily chooses or be detected by object detection method
Arrive, can also be what is traced into by method for tracking target, at least two methods in above-mentioned three kinds of methods can also be used
Obtain.The object detection method can be method for detecting human face, the deep learning target inspection based on Region Proposal
Method of determining and calculating and biometric discrimination method etc., the application is not specifically limited.
Preferably, the method for tracking target can the tracking based on active profile, feature based tracking,
Tracking based on region and the tracking based on model etc..
Beneficial effects of the present invention are as follows：
Due to the embodiment of the present application be according to wide angle picture and project video in window pixel between corresponding relation, from
Multiple projection video in windows are extracted in N number of wide angle picture of collection；Then target following is carried out in multiple projection video in windows；
The embodiment of the present invention by N number of wide angle picture without being converted into panoramic picture；And the amount of calculation of the track algorithm used is also obvious
Less than the amount of calculation for the new position that the track algorithm that prior art uses determines tracking target.Therefore, compared with prior art, its
Calculation procedure is reduced and amount of calculation reduces, thus promotes the tracking efficiency of panoramic video to improve, and can effectively improve panoramic video
In target following real-time.
Further, because projection video in window is the parts of images in panoramic picture corresponding to N number of wide angle picture
Projection, promote the distortion for projecting video in window minimum or be not present；And in minimum or in the absence of distortion the projection window that distorts
When target following is carried out in image, its accuracy tracked can greatly improve, so as to effectively improve the accuracy of tracking.
All features disclosed in this specification, or disclosed all methods or during the step of, except mutually exclusive
Feature and/or step beyond, can combine in any way.
Any feature disclosed in this specification (including any accessory claim, summary and accompanying drawing), except non-specifically chatting
State, can alternative features equivalent by other or with similar purpose replaced.I.e., unless specifically stated otherwise, each feature
It is an example in a series of equivalent or similar characteristics.
The invention is not limited in foregoing embodiment.The present invention, which expands to, any in this manual to be disclosed
New feature or any new combination, and disclose any new method or process the step of or any new combination.
Claims (12)
- A kind of 1. method for tracking target applied to panoramic picture, it is characterised in that including：N number of wide angle picture is gathered using N number of wide-angle lens each moment, wherein, N number of wide angle picture of each moment collection is corresponding One frame panoramic picture；According to the corresponding relation between the pixel of wide angle picture and projection video in window, extracted from N number of wide angle picture Multiple projection video in windows, wherein, the projection video in window is the part in panoramic picture corresponding to N number of wide angle picture The projection of image；Using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, in present frame panoramic picture It is tracked in the multiple projection video in windows extracted in corresponding N number of wide angle picture；Wherein, an adjacent frame or multiframe are complete Scape image is before present frame panoramic picture.
- 2. the method as described in claim 1, it is characterised in that according to wide angle picture and projection video in window pixel it Between corresponding relation before, methods described also includes：Obtain the corresponding relation between wide angle picture and the pixel for projecting video in window.
- 3. method as claimed in claim 2, it is characterised in that the pixel for obtaining wide angle picture and projecting video in window Between corresponding relation, specifically include：Closed according to the posture between the mapping relations of latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle lens System, calculates the first corresponding relation between the pixel of panoramic picture and N number of wide angle picture；According to the division rule of projection video in window, the second couple between panoramic picture and the pixel for projecting video in window is obtained It should be related to；According to first corresponding relation and second corresponding relation, obtain wide angle picture and project the pixel of video in window Between corresponding relation.
- 4. method as claimed in claim 3, it is characterised in that the division rule includes the division rule based on default longitude Then, the division rule of latitude and the division rule based on preset field of view are preset.
- 5. such as the method described in claim any one of 2-4, it is characterised in that utilizing an adjacent frame or multiframe panoramic picture During tracking target in corresponding all wide angle pictures is tracked, methods described also includes：By in the currently displayed frame panoramic picture of the successful target following result of the tracking of previous frame.
- 6. method as claimed in claim 5, it is characterised in that shown by the successful target following result of the tracking of previous frame Before in present frame panoramic picture, methods described also includes:According to first corresponding relation, N number of wide angle picture corresponding to present frame is spliced into present frame panoramic picture.
- 7. such as the method described in claim any one of 2-4, it is characterised in that described to utilize an adjacent frame or multiframe panorama sketch Tracking target in all wide angle pictures as corresponding to, that is extracted in N number of wide angle picture corresponding to present frame panoramic picture is more It is tracked, specifically includes in individual projection video in window：N number of wide-angle figure according to corresponding in second corresponding relation and an adjacent frame or multiframe panoramic picture per frame panoramic picture Picture, obtain in adjacent a frame or multiframe panoramic picture per multiple projection video in windows corresponding to frame panoramic picture；Based on preparatory condition, chosen from multiple projection video in windows corresponding to N number of wide angle picture corresponding to every frame panoramic picture The tracking target；Using the tracking target chosen from multiple projection video in windows corresponding to every frame panoramic picture, in present frame panorama It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to image.
- A kind of 8. panorama camera, it is characterised in that including N number of wide-angle lens and readable storage medium storing program for executing, the readable storage medium storing program for executing On have program stored therein, following steps are realized when described program is executed by processor：N number of wide angle picture is gathered using N number of wide-angle lens each moment, wherein, N number of wide angle picture of each moment collection A corresponding frame panoramic picture；According to the corresponding relation between the pixel of wide angle picture and projection video in window, extracted from N number of wide angle picture Multiple projection video in windows, wherein, the projection video in window is the part in panoramic picture corresponding to N number of wide angle picture The projection of image, and overlapping region be present with least one projection video in window in each projection video in window；Using the tracking target in all wide angle pictures corresponding to an adjacent frame or multiframe panoramic picture, in present frame panoramic picture It is tracked in the multiple projection video in windows extracted in corresponding N number of wide angle picture；Wherein, an adjacent frame or multiframe are complete Scape image is before present frame panoramic picture.
- 9. panorama camera as claimed in claim 8, it is characterised in that according to wide-angle figure when described program is executed by processor Before corresponding relation between picture and the pixel for projecting video in window, following steps are also realized：Obtain wide angle picture and projection Corresponding relation between the pixel of video in window.
- 10. panorama camera as claimed in claim 9, it is characterised in that wide-angle figure is obtained when described program is executed by processor During the step of the corresponding relation between picture and the pixel for projecting video in window, following steps are specifically included：Closed according to the posture between the mapping relations of latitude and longitude coordinates, spherical coordinate and plane coordinates and each wide-angle lens System, calculates the first corresponding relation between the pixel of panoramic picture and N number of wide angle picture；According to the division rule of projection video in window, the second couple between panoramic picture and the pixel for projecting video in window is obtained It should be related to；According to first corresponding relation and second corresponding relation, obtain wide angle picture and project the pixel of video in window Between corresponding relation.
- 11. panorama camera as claimed in claim 10, it is characterised in that adjacent in utilization when described program is executed by processor During tracking target in all wide angle pictures corresponding to one frame or multiframe panoramic picture is tracked, following walk also is realized Suddenly：By in the currently displayed frame panoramic picture of the successful target following result of the tracking of previous frame.
- 12. panorama camera as claimed in claim 10, it is characterised in that adjacent one is utilized when described program is executed by processor Tracking target in all wide angle pictures corresponding to frame or multiframe panoramic picture, in N number of wide-angle corresponding to present frame panoramic picture During the step being tracked in the multiple projection video in windows extracted in image, following steps are specifically included：N number of wide-angle figure according to corresponding in second corresponding relation and an adjacent frame or multiframe panoramic picture per frame panoramic picture Picture, obtain in adjacent a frame or multiframe panoramic picture per multiple projection video in windows corresponding to frame panoramic picture；Based on preparatory condition, chosen from multiple projection video in windows corresponding to N number of wide angle picture corresponding to every frame panoramic picture The tracking target；Using the tracking target chosen from multiple projection video in windows corresponding to every frame panoramic picture, in present frame panorama It is tracked in the multiple projection video in windows extracted in N number of wide angle picture corresponding to image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710773963.9A CN107507230A (en) | 2017-08-31 | 2017-08-31 | Method for tracking target and panorama camera applied to panoramic picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710773963.9A CN107507230A (en) | 2017-08-31 | 2017-08-31 | Method for tracking target and panorama camera applied to panoramic picture |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107507230A true CN107507230A (en) | 2017-12-22 |
Family
ID=60693558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710773963.9A Pending CN107507230A (en) | 2017-08-31 | 2017-08-31 | Method for tracking target and panorama camera applied to panoramic picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107507230A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108195378A (en) * | 2017-12-25 | 2018-06-22 | 北京航天晨信科技有限责任公司 | It is a kind of based on the intelligent vision navigation system for looking around camera |
| CN108848304A (en) * | 2018-05-30 | 2018-11-20 | 深圳岚锋创视网络科技有限公司 | A kind of method for tracking target of panoramic video, device and panorama camera |
| CN109035327A (en) * | 2018-06-25 | 2018-12-18 | 北京大学 | Panorama camera Attitude estimation method based on deep learning |
| CN109670462A (en) * | 2018-12-24 | 2019-04-23 | 北京天睿空间科技股份有限公司 | Continue tracking across panorama based on the aircraft of location information |
| CN111200750A (en) * | 2018-11-16 | 2020-05-26 | 北京字节跳动网络技术有限公司 | Multi-window playing method and device of panoramic video, electronic equipment and storage medium |
| CN111860051A (en) * | 2019-04-27 | 2020-10-30 | 北京初速度科技有限公司 | Vehicle-based loop detection method and device and vehicle-mounted terminal |
| CN117197195A (en) * | 2023-09-15 | 2023-12-08 | 南京芯传汇电子科技有限公司 | Video processing method and system based on target recognition and tracking |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101119482A (en) * | 2007-09-28 | 2008-02-06 | 北京智安邦科技有限公司 | Overall view monitoring method and apparatus |
| CN101702939A (en) * | 2007-03-16 | 2010-05-05 | 科尔摩根公司 | System for panoramic image processing |
| CN102999918A (en) * | 2012-04-19 | 2013-03-27 | 浙江工业大学 | Multi-target object tracking system of panorama video sequence image |
| CN103763479A (en) * | 2013-12-31 | 2014-04-30 | 深圳英飞拓科技股份有限公司 | Splicing device for real-time high speed high definition panoramic video and method thereof |
| CN104767911A (en) * | 2015-04-28 | 2015-07-08 | 腾讯科技（深圳）有限公司 | Method and device for processing image |
| CN106296589A (en) * | 2016-08-26 | 2017-01-04 | 北京疯景科技有限公司 | The processing method and processing device of panoramic picture |
| CN107038686A (en) * | 2017-04-21 | 2017-08-11 | 成都观界创宇科技有限公司 | A kind of method and apparatus of image mosaic processing |
-
2017
- 2017-08-31 CN CN201710773963.9A patent/CN107507230A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101702939A (en) * | 2007-03-16 | 2010-05-05 | 科尔摩根公司 | System for panoramic image processing |
| CN101119482A (en) * | 2007-09-28 | 2008-02-06 | 北京智安邦科技有限公司 | Overall view monitoring method and apparatus |
| CN102999918A (en) * | 2012-04-19 | 2013-03-27 | 浙江工业大学 | Multi-target object tracking system of panorama video sequence image |
| CN103763479A (en) * | 2013-12-31 | 2014-04-30 | 深圳英飞拓科技股份有限公司 | Splicing device for real-time high speed high definition panoramic video and method thereof |
| CN104767911A (en) * | 2015-04-28 | 2015-07-08 | 腾讯科技（深圳）有限公司 | Method and device for processing image |
| CN106296589A (en) * | 2016-08-26 | 2017-01-04 | 北京疯景科技有限公司 | The processing method and processing device of panoramic picture |
| CN107038686A (en) * | 2017-04-21 | 2017-08-11 | 成都观界创宇科技有限公司 | A kind of method and apparatus of image mosaic processing |
Non-Patent Citations (1)
| Title |
|---|
| OM K. GUPTA 等: "Robust pose estimation and tracking system for a mobile robot using a panoramic camera", 《2010 IEEE CONFERENCE ON ROBOTICS》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108195378A (en) * | 2017-12-25 | 2018-06-22 | 北京航天晨信科技有限责任公司 | It is a kind of based on the intelligent vision navigation system for looking around camera |
| JP2021527865A (en) * | 2018-05-30 | 2021-10-14 | 影石創新科技股▲ふん▼有限公司Ａｒａｓｈｉ Ｖｉｓｉｏｎ Ｉｎｃ． | Panorama video target tracking method and panoramic camera |
| CN108848304A (en) * | 2018-05-30 | 2018-11-20 | 深圳岚锋创视网络科技有限公司 | A kind of method for tracking target of panoramic video, device and panorama camera |
| US11509824B2 (en) | 2018-05-30 | 2022-11-22 | Arashi Vision Inc. | Method for tracking target in panoramic video, and panoramic camera |
| JP7048764B6 (en) | 2018-05-30 | 2022-05-16 | 影石創新科技股▲ふん▼有限公司 | Panorama video target tracking method and panoramic camera |
| JP7048764B2 (en) | 2018-05-30 | 2022-04-05 | 影石創新科技股▲ふん▼有限公司 | Panorama video target tracking method and panoramic camera |
| CN108848304B (en) * | 2018-05-30 | 2020-08-11 | 影石创新科技股份有限公司 | Target tracking method and device of panoramic video and panoramic camera |
| CN109035327A (en) * | 2018-06-25 | 2018-12-18 | 北京大学 | Panorama camera Attitude estimation method based on deep learning |
| CN109035327B (en) * | 2018-06-25 | 2021-10-29 | 北京大学 | Panoramic camera attitude estimation method based on deep learning |
| CN111200750B (en) * | 2018-11-16 | 2021-09-14 | 北京字节跳动网络技术有限公司 | Multi-window playing method and device of panoramic video, electronic equipment and storage medium |
| CN111200750A (en) * | 2018-11-16 | 2020-05-26 | 北京字节跳动网络技术有限公司 | Multi-window playing method and device of panoramic video, electronic equipment and storage medium |
| CN109670462B (en) * | 2018-12-24 | 2019-11-01 | 北京天睿空间科技股份有限公司 | Continue tracking across panorama based on the aircraft of location information |
| CN109670462A (en) * | 2018-12-24 | 2019-04-23 | 北京天睿空间科技股份有限公司 | Continue tracking across panorama based on the aircraft of location information |
| CN111860051A (en) * | 2019-04-27 | 2020-10-30 | 北京初速度科技有限公司 | Vehicle-based loop detection method and device and vehicle-mounted terminal |
| CN117197195A (en) * | 2023-09-15 | 2023-12-08 | 南京芯传汇电子科技有限公司 | Video processing method and system based on target recognition and tracking |
| CN117197195B (en) * | 2023-09-15 | 2024-04-26 | 南京芯传汇电子科技有限公司 | Video processing method and system based on target recognition and tracking |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107507230A (en) | Method for tracking target and panorama camera applied to panoramic picture | |
| CN108416754A (en) | A kind of more exposure image fusion methods automatically removing ghost | |
| CN109151439A (en) | A kind of the automatic tracing camera system and method for view-based access control model | |
| CN110855972B (en) | Image processing method, electronic device, and storage medium | |
| CN107516319A (en) | A kind of high accuracy simple interactive stingy drawing method, storage device and terminal | |
| CN101916455B (en) | Method and device for reconstructing three-dimensional model of high dynamic range texture | |
| CN105957008A (en) | Panoramic image real-time stitching method and panoramic image real-time stitching system based on mobile terminal | |
| CN103856727A (en) | Multichannel real-time video splicing processing system | |
| CN102668537A (en) | Multocular image pickup apparatus and multocular image pickup method | |
| CN109453517B (en) | Virtual character control method and device, storage medium and mobile terminal | |
| CN101180653A (en) | Method and device for three-dimensional rendering | |
| CN109949347A (en) | Human body tracing method, device, system, electronic equipment and storage medium | |
| CN103729860B (en) | A kind of method and apparatus of tracking image target | |
| CN105763829A (en) | Image processing method and electronic device | |
| CN104063871B (en) | The image sequence Scene Segmentation of wearable device | |
| CN106997579A (en) | The method and apparatus of image mosaic | |
| CN110349163A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN107564039A (en) | Multi-object tracking method and panorama camera applied to panoramic video | |
| CN107205626A (en) | The focusing control method of endoscope apparatus and endoscope apparatus | |
| CN106023073A (en) | Image splicing system | |
| US20230289982A1 (en) | Methods and systems to track a moving objects trajectory using a single camera | |
| CN103544696B (en) | A kind of suture line real-time searching method realized for FPGA | |
| CN110910449B (en) | Method and system for identifying three-dimensional position of object | |
| CN108010058A (en) | A kind of method and system that vision tracking is carried out to destination object in video flowing | |
| CN113192125B (en) | Multi-camera video concentration method and system in virtual viewpoint-optimized geographic scene |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20201013 |
|
| AD01 | Patent right deemed abandoned |