CN113192125A - Multi-camera video concentration method and system in geographic scene with optimal virtual viewpoint - Google Patents

Multi-camera video concentration method and system in geographic scene with optimal virtual viewpoint Download PDF

Info

Publication number
CN113192125A
CN113192125A CN202110327605.1A CN202110327605A CN113192125A CN 113192125 A CN113192125 A CN 113192125A CN 202110327605 A CN202110327605 A CN 202110327605A CN 113192125 A CN113192125 A CN 113192125A
Authority
CN
China
Prior art keywords
camera
video
virtual viewpoint
geographic
scene
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.)
Granted
Application number
CN202110327605.1A
Other languages
Chinese (zh)
Other versions
CN113192125B (en
Inventor
解愉嘉
毛波
王崴
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Finance and Economics
Original Assignee
Nanjing University of Finance and Economics
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Finance and Economics filed Critical Nanjing University of Finance and Economics
Priority to CN202110327605.1A priority Critical patent/CN113192125B/en
Publication of CN113192125A publication Critical patent/CN113192125A/en
Application granted granted Critical
Publication of CN113192125B publication Critical patent/CN113192125B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/05—Geographic models
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Remote Sensing (AREA)
  • Computer Graphics (AREA)
  • Processing Or Creating Images (AREA)

Abstract

A multi-camera video concentration method and system in a geographic scene with a virtual viewpoint being optimized acquire coordinate data of a same-name point pair by acquiring video sequence image information and acquiring the same-name point pair from a video image and a three-dimensional geographic scene model; establishing a mapping relation between a video image and a geographic space according to coordinate data of the same-name point pair, and positioning a camera view; constructing a camera observation domain model by analyzing the observable distance and the sight line deflection angle, and generating an observable set of a camera group; optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group; and presetting display parameters of the moving target, and performing multi-camera video concentration according to the display parameters. The method has the remarkable effects that the mapping relation between the video target and the geographic scene is established, the effect of fusion expression of the monitoring video in the geographic scene is enhanced, and great convenience is provided for the quick retrieval and efficient understanding of the video geographic scene information integration.

Description

Multi-camera video concentration method and system in geographic scene with optimal virtual viewpoint
Technical Field
The invention relates to the technical field of real-time fusion of video streams and three-dimensional models, in particular to a method and a system for concentrating multi-camera videos in a geographic scene with optimal virtual viewpoints.
Background
With the requirement of the Virtual Geographic Environment (VGE) on the accuracy and the real-time performance of scene simulation, the multi-source heterogeneous data is introduced to enhance the visual expression and analysis functions of the VGE. The video data can not only realize the real scene display of the geographic environment, but also describe the space-time motion of moving objects (pedestrians, vehicles and the like) in the geographic scene. When a user views a video in a VGE, the virtual viewpoint is typically selected at a virtual location that is close to the original geographic location of the camera.
However, the conventional virtual viewpoint selecting method is often in actual deployment and use, and has the following difficulties and problems:
first, it is convenient to see a single camera and short duration videos in this way, but if a scene contains multiple videos, and the cameras have different sight directions and non-overlapping and discrete views, it is difficult for a user to view all videos through a single virtual viewpoint. If each path of video is individually arranged with one virtual viewpoint to be watched one by one, the watching duration can be greatly prolonged, and the user is not favorable for quickly watching the video content.
Secondly, video moving objects usually appear in cameras shooting different areas one after another, and each path of video image is viewed independently, so that the cross-camera global motion condition of the moving objects in a scene cannot be expressed.
Therefore, how to effectively select a small number of virtual viewpoints in a virtual scene to quickly view a multi-camera video target and show the cross-camera motion condition of the video target becomes a technical problem to be solved urgently.
Disclosure of Invention
Therefore, the invention provides a multi-camera video concentration method and system in a geographic scene with an optimal virtual viewpoint, which aims to solve the problem that in the prior art, when a user watches videos in a VGE, the virtual viewpoint is generally selected at a virtual position close to the original geographic position of a camera, so that the method is convenient for watching a single camera and short-duration videos, but if the scene contains multiple paths of videos, the sight directions of the cameras are different, and the sight fields are not overlapped and distributed discretely, the user is difficult to watch all the videos through a single virtual viewpoint; if each path of video is individually arranged with one virtual viewpoint to be watched one by one, the watching duration can be greatly prolonged, and the user is not facilitated to quickly view the video content; on the other hand, video moving objects usually appear in cameras shooting different areas in sequence, each path of video image is viewed independently, and the problem that the cross-camera global motion condition of the moving objects in a scene cannot be expressed is solved.
In order to achieve the above purpose, the invention provides the following technical scheme: in a first aspect, a method for concentrating multi-camera video in a geographic scene with a preferred virtual viewpoint is provided, which includes the following steps:
acquiring video sequence image information, acquiring homonymous point pairs in a video image and a three-dimensional geographic scene model, and acquiring coordinate data of the homonymous point pairs, wherein the coordinate data comprises image coordinates and geographic coordinates;
establishing a mapping relation between a video image and a geographic space according to coordinate data of the same-name point pair, and positioning a camera view;
constructing a camera observation domain model by analyzing the observable distance and the sight line deflection angle, and generating an observable set of a camera group;
optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group;
presetting display parameters of the moving target, and performing multi-camera video concentration according to the display parameters.
As a preferred scheme of the multi-camera video concentration method in the geographic scene with the preferred virtual viewpoint, the video image is a first frame image of the intercepted monitoring video.
As a preferred scheme of the multi-camera video concentration method in the geographic scene with the preferred virtual viewpoint, the three-dimensional geographic scene model is a three-dimensional scene model constructed according to the measurement information of the real geographic scene, the number of the same-name point pairs collected on the video image and the virtual geographic scene is not less than three pairs, and the three pairs of the same-name point pairs are not completely collinear.
As a preferred scheme of the multi-camera video condensing method in the geographic scene with the preferred virtual viewpoint, the establishing of the mapping relationship between the video image and the geographic space includes the following steps:
a1) presetting an object point geographic space coordinate Q corresponding to a given image point image space coordinate Q, and expressing Q and Q as a homogeneous coordinate form:
q＝[x y 1]T
Q＝[X Y Z 1]T；
when a homography matrix M is recorded, the relationship between Q and Q is:
q＝MQ；
the expression of the homography matrix M is:
Figure BDA0002995211980000031
a2) solving the geographic space coordinates of the object points corresponding to the image points in each image:
Figure BDA0002995211980000032
a3) assuming that there are L cameras in the current video camera network, the mapping matrix is labeled M for the k-th camera (k is 1,2 … L)k(ii) a Defining each camera position in geographic space as
Figure BDA0002995211980000033
The geospatial position of each camera field of view polygon is
Figure BDA0002995211980000034
Figure BDA0002995211980000035
Figure BDA0002995211980000036
Wherein the camera position
Figure BDA0002995211980000037
Considered as a point in the geographic space, the camera view polygon is bounded by o boundary points Pk,numSequentially concatenating the constituent polygon records.
As a preferred scheme of a multi-camera video concentration method in a geographic scene with a preferred virtual viewpoint, in the process of positioning the camera view, two factors of a virtual line-of-sight distance and a camera-virtual viewpoint included angle are selected as constraint conditions;
the virtual line-of-sight distance refers to the geographic space distance between a virtual viewpoint and a certain point in the view field; the camera-virtual viewpoint included angle is an included angle formed by projection of a given point in a visual field, the virtual viewpoint and a camera position point on a horizontal plane;
defining a distance threshold TdisAnd an angle threshold TangAs a constraint, and assuming a distance threshold TdisAnd an angle threshold TangIt has been given that finding areas in the scene model that meet the constraints serves as virtual viewpoint ranges.
As a preferred scheme of the multi-camera video concentration method in a geographical scene with a preferred virtual viewpoint, the method for constructing a camera observation domain model by analyzing an observable distance and a view deflection angle and generating an observable set of camera groups comprises the following steps:
b1) recording camera position
Figure BDA0002995211980000041
Polygon with camera view
Figure BDA0002995211980000042
And is
Figure BDA0002995211980000043
In the distance
Figure BDA0002995211980000044
Is Pk,n1Pk,n2The farthest line segment is Pk,n3Pk,n4；
b2) Respectively at point Pk,n3,Pk,n4As a circle center, a distance threshold value TdisIs a radius on the line segment Pk,n3Pk,n4Towards the camera position
Figure BDA0002995211980000045
Drawing a semicircle at one side, and taking two semicircles as a line segment Pk,n1Pk,n2Near camera position
Figure BDA0002995211980000046
One side intersection region as a virtual viewpoint distance reasonable region Ak,dis；
b3) Respectively at point Pk,n1,Pk,n2Being a corner point, an angle threshold TangFor deflection angle, clockwise and anticlockwise respectively deflect TangMaking four rays, and taking the four rays to be on the line segment Pk,n1Pk,n2Near camera position
Figure BDA0002995211980000047
One side intersection region as a virtual viewpoint angle reasonable region Ak,ang；
b4) Virtual viewpoint range A of camerakIs Ak,disAnd Ak,angThe intersection of (a);
b5) let Obj be the total set of all video moving objects in all video cameras; note that there is N in the kth camerakEach video moving object has a track marked as Ck,i，Ck,iThe expression of (a) is as follows:
Obj＝{Ck,i,(k＝1,2…L)}
Ck,i＝{Ik,i,j,Pk,i,j,(i＝1,2,…,Nk)(j＝1,2,…,n)}；
wherein L represents the number of cameras, Ik,i,jAnd Pk,i,jRepresenting the sub-image of the ith video moving object in the jth video frame in the kth camera and the geographic space position of the sub-image are analyzed by the cross-camera association of the video moving objects, and the tracks of the single-camera video moving object are merged to obtain the tracks of the multi-camera video moving object
Figure BDA00029952119800000410
Realizing the related organization of the multi-camera video moving target:
Figure BDA0002995211980000048
Cubeio＝{Ck1,i1,Ck2,i2…Cko,iL,(k1,k2,…ko)∈(1,2…L)}；
wherein L isoRepresenting the total number of video moving objects, Cube, after merging of cross-camera homonymous video moving objects in a surveillance video networkioRepresenting the global track of the video moving object with the sequence number io in the monitoring video network,
Figure BDA0002995211980000049
and (4) representing the sub-track of the video moving object with the sequence number io in the ko camera.
As a preferred scheme of a multi-camera video concentration method in a geographical scene with a preferred virtual viewpoint, the observable collection is preferred by constructing an evaluation model to obtain a virtual viewpoint group, and the method specifically comprises the following steps:
c1) the number of cameras is recorded as L, and a set formed by all combination modes of the cameras is recorded as M:
Figure BDA0002995211980000055
mi＝{ni,j}
Figure BDA0002995211980000051
wherein m isiThe ith camera combination mode comprises all camera sets in the combination mode; n isi,jDigital camera combination mode miA j-th group of cameras under the condition that all cameras of the group are included;
Figure BDA0002995211980000052
the first camera in the j camera set in the i camera set combination mode is referred to;
c2) by defining a distance threshold TdisAnd an angle threshold TangFor each cameraCombined mode miIn each camera group ni,jAll cameras in
Figure BDA0002995211980000053
Calculating observable domains and calculating intersection; if a certain camera combination mode miAll the camera sets ni,jIf the intersection of any number of the camera observation domains is not empty, the combination mode m of the camera is recordediIs an observable combination, otherwise, the camera combination mode m is recordediIs an unobservable combination;
c3) based on multi-camera video target trajectory data, the following video condensation optimization targets are specified to realize the optimization of the camera group:
firstly, expressing coherence of homonymous targets across cameras, namely expressing the video cameras with the appearance of a single target in a consistent manner by using as few virtual viewpoints as possible;
expressing that the total number of virtual viewpoints used by all video targets is as small as possible;
c4) comprehensively evaluating the multi-camera video target expression effect of the camera combination corresponding to the virtual viewpoint group by value:
Figure BDA0002995211980000054
wherein n iscRepresents the total number of cameras, nvRepresenting the number of virtual viewpoints, N representing the total number of video moving objects, miRepresenting the number of virtual viewpoints which are related and expressed by each video moving object, wherein mu is a weight parameter;
c5) at a distance threshold TdisAnd an angle threshold TangWhen the value is constant, value values of all current camera observable sets are calculated through defining a parameter alpha, the maximum value is taken as a camera combination selection result, and multi-camera video concentration in a virtual scene is carried out.
As a preferred scheme of the method for concentrating the multi-camera video in the geographic scene with the preferred virtual viewpoint, the method for concentrating the multi-camera video according to the display parameters of the preset moving object comprises the following steps:
d1) recording that W virtual viewpoints (W is less than or equal to L) are needed for viewing the video moving objects of all L cameras under the current camera combination; meanwhile, setting a frame rate fps displayed by the video moving target subgraph in a three-dimensional scene as the number of the subgraphs displayed by a single video moving target per second; setting an object display interval time t0As time intervals for additional display of new video moving objects;
d2) for a certain virtual viewpoint W (W ≦ W), the first occurring moving object O is first displayed0In geospatial track T0And identifying the sequence of appearance of the video object among different cameras;
screening the video object subgraph according to the frame rate fps, converting the plane coordinate corresponding to the screened video object subgraph into a geographic coordinate, and meanwhile, according to a proportionality coefficient Pw、PhScaling a video object sub-graph, Pw、PhThe calculation formula is as follows:
Figure BDA0002995211980000061
Figure BDA0002995211980000062
wherein
Figure BDA0002995211980000063
Randomly selecting an appropriate amount of average width and height of subgraphs from a video object subgraph library, mapping coordinates of three points, namely upper left, lower left and upper right, of the selected subgraphs in an original video frame to corresponding geographic positions in a virtual scene to obtain the length and height of the video object subgraphs in a three-dimensional space,
Figure BDA0002995211980000064
is the average length and height of the video object sub-graph displayed in the virtual scene;
d3) displaying O in the camera view field in the virtual scene according to the frame rate fps in the dynamic display process0The video object sub-graph of the current frame is in the corresponding geographic position, and the old video object sub-graph is not displayed any more;
at t0,2t0…,nt0At the moment, the video object O is added1,O2…OnDynamically expressed in a three-dimensional scene model to realize the concentration of multi-camera video objects.
As a preferred scheme of the multi-camera video concentration method in the geographic scene with the preferred virtual viewpoint, for the condition that the same section of object track generated by overlapping the shooting areas of the cameras is obtained by a plurality of cameras, determining the camera for obtaining the object sub-image by comparing the included angles between the virtual viewpoint and the object track point and the connecting line of three points at the two camera positions:
and a view overlapping part C exists between the camera a and the camera b, for the video object passing through the view C, the included angle among the three points of the camera position, the track point and the virtual viewpoint V, namely the sizes of alpha and beta, is compared, if alpha is less than or equal to beta, the video object sub-graph acquired by the camera a is used, and otherwise, the video object sub-graph acquired by the camera b is used.
In a second aspect, there is provided a multi-camera video concentration system in a geographical scene with a preferred virtual viewpoint, where the multi-camera video concentration method in a geographical scene with a preferred virtual viewpoint in the first aspect or any possible implementation manner thereof is adopted, and the concentration system includes:
the same-name point acquisition module: the system comprises a video sequence image acquisition module, a video acquisition module, a three-dimensional geographic scene model acquisition module, a coordinate acquisition module and a display module, wherein the video sequence image acquisition module is used for acquiring video sequence image information, acquiring homonymous point pairs in a video image and the three-dimensional geographic scene model and acquiring coordinate data of the homonymous point pairs, and the coordinate data comprises image coordinates and geographic coordinates;
the mapping relation construction module: the method comprises the steps of establishing a mapping relation between a video image and a geographic space according to coordinate data of a same-name point pair, and positioning a camera view;
a camera group observable collection generation module: the system comprises a camera observation domain model and a camera group observation set, wherein the camera observation domain model is constructed by analyzing an observable distance and a sight line deflection angle to generate a camera group observable set;
a virtual viewpoint group generation module: the system is used for optimizing the observable collection by constructing an evaluation model to obtain a virtual viewpoint group;
the video target space-time motion expression module: the method comprises the steps of presetting display parameters of a moving target, and carrying out multi-camera video concentration according to the display parameters.
The invention has the following advantages: acquiring coordinate data of a same-name point pair by acquiring video sequence image information and acquiring the same-name point pair in a video image and a three-dimensional geographic scene model, wherein the coordinate data comprises an image coordinate and a geographic coordinate; establishing a mapping relation between a video image and a geographic space according to coordinate data of the same-name point pair, and positioning a camera view; constructing a camera observation domain model by analyzing the observable distance and the sight line deflection angle, and generating an observable set of a camera group; optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group; and presetting display parameters of the moving target, and performing multi-camera video concentration according to the display parameters. The method has the remarkable effects that the mapping relation between the video target and the geographic scene is established, the effect of fusion expression of the monitoring video in the geographic scene is enhanced, and great convenience is provided for the quick retrieval and efficient understanding of the video geographic scene information integration.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It should be apparent that the drawings in the following description are merely exemplary, and that other embodiments can be derived from the drawings provided by those of ordinary skill in the art without inventive effort.
The structures, ratios, sizes, and the like shown in the present specification are only used for matching with the contents disclosed in the specification, so that those skilled in the art can understand and read the present invention, and do not limit the conditions for implementing the present invention, so that the present invention has no technical significance, and any structural modifications, changes in the ratio relationship, or adjustments of the sizes, without affecting the functions and purposes of the present invention, should still fall within the scope of the present invention.
Fig. 1 is a schematic diagram illustrating an integrated representation of multiple channels of video in a VGE according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of a method for concentrating multi-camera video in a geographic scene with a preferred virtual viewpoint according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a camera and a geospatial coordinate system and an image space coordinate system provided in an embodiment of the invention;
FIG. 4(a) is a schematic diagram of a virtual line-of-sight distance provided in an embodiment of the present invention;
fig. 4(b) is a schematic diagram of a camera-virtual viewpoint included angle provided in the embodiment of the present invention;
FIG. 5(a) is a schematic diagram of the closest/farthest line segment of a camera to a video field provided in an embodiment of the present invention;
fig. 5(b) is a schematic diagram of a virtual viewpoint distance reasonable region provided in the embodiment of the present invention;
fig. 5(c) is a schematic diagram of a virtual viewpoint angle reasonable region provided in the embodiment of the present invention;
fig. 5(d) is a schematic diagram of a virtual viewpoint angle and distance plausible region provided in the embodiment of the present invention;
FIG. 6 is a schematic view of a set of observable cameras provided in an embodiment of the present invention;
FIG. 7 is a schematic diagram illustrating the concentration of multiple camera video objects in a geographic scene provided in an embodiment of the present invention;
FIG. 8 is a schematic view of a camera view overlap process provided in an embodiment of the invention;
fig. 9 is a schematic diagram of a multi-camera video concentrating system in a geographic scene with a preferred virtual viewpoint according to an embodiment of the present invention.
Detailed Description
The present invention is described in terms of particular embodiments, other advantages and features of the invention will become apparent to those skilled in the art from the following disclosure, and it is to be understood that the described embodiments are merely exemplary of the invention and that it is not intended to limit the invention to the particular embodiments disclosed. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Referring to fig. 1, by introducing a video into the VGE and supporting video intelligent analysis with geospatial information, related functions such as video data organization management, spatial mapping, video-scene fusion expression and the like in the VGE can be realized.
Referring to fig. 2, a method for concentrating multi-camera video in a geographical scene with a preferred virtual viewpoint is provided, which comprises the following steps:
s1, acquiring video sequence image information, acquiring homonymy point pairs in a video image and a three-dimensional geographic scene model, and acquiring coordinate data of the homonymy point pairs, wherein the coordinate data comprises image coordinates and geographic coordinates;
s2, establishing a mapping relation between a video image and a geographic space according to coordinate data of the same-name point pair, and positioning a camera view;
s3, constructing a camera observation domain model by analyzing the observable distance and the sight line deflection angle, and generating an observable set of a camera group;
s4, optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group;
and S5, presetting display parameters of the moving target, and concentrating the multi-camera video according to the display parameters.
Specifically, in step S1, the video image is a first frame image of the captured monitoring video. In step S1, the three-dimensional geographic scene model is a three-dimensional scene model constructed according to the real geographic scene measurement information, the number of the same-name point pairs collected on the video image and the virtual geographic scene is not less than three pairs, and the three pairs of the same-name point pairs are not all collinear.
In particular, referring to fig. 3, the relationship between the camera and the image space coordinate system and the geospatial coordinate system is illustrated. Recording a camera station center as C and an image space coordinate system as OiXiYiThe geospatial coordinate system is marked as OgXgYgZg. In step S2, the establishing of the mapping relationship between the video image and the geographic space includes the following steps:
a1) presetting an object point geographic space coordinate Q corresponding to a given image point image space coordinate Q, and expressing Q and Q as a homogeneous coordinate form:
q＝[x y 1]T
Q＝[X Y Z 1]T；
when a homography matrix M is recorded, the relationship between Q and Q is:
q＝MQ；
the expression of the homography matrix M is:
Figure BDA0002995211980000101
a2) because M has 6 unknowns, at least 3 groups of known image point image space coordinates and object point geographic space coordinates are needed to realize the solution of M. After M is determined, solving the geographic space coordinates of the object points corresponding to the image points in each image:
Figure BDA0002995211980000102
a3) assuming that there are L cameras in the current video camera network, the mapping matrix is labeled M for the k-th camera (k is 1,2 … L)k(ii) a On this basis, each camera position in the geographic space is defined as
Figure BDA0002995211980000103
The geospatial position of each camera field of view polygon is
Figure BDA0002995211980000104
Figure BDA0002995211980000105
Figure BDA0002995211980000106
Wherein the camera position
Figure BDA0002995211980000107
Considered as a point in the geographic space, the camera view polygon is bounded by o boundary points Pk,numSequentially concatenating the constituent polygon records.
In this embodiment, in step S2, in the process of positioning the camera view, two factors, namely, the virtual line-of-sight distance and the camera-virtual viewpoint included angle, are selected as constraint conditions;
referring specifically to fig. 4, the virtual line-of-sight distance refers to a geospatial distance between a virtual viewpoint and a certain point in the view field; the camera-virtual viewpoint included angle is an included angle formed by projection of a given point in a visual field, the virtual viewpoint and a camera position point on a horizontal plane;
defining a distance threshold TdisAnd an angle threshold TangAs a constraint, and assuming a distance threshold TdisAnd an angle threshold TangIt has been given that on this basis, areas meeting the constraints are found in the scene model as virtual viewpoint ranges.
Specifically, in step S3, the step of constructing a camera observation domain model by analyzing the observable distance and the gaze deflection angle, and generating the observable set of the camera group includes the following steps:
b1) referring specifically to FIG. 5(a), camera position is recorded
Figure BDA0002995211980000108
Polygon with camera view
Figure BDA0002995211980000109
And is
Figure BDA00029952119800001010
In the distance
Figure BDA00029952119800001011
Is Pk,n1Pk,n2The farthest line segment is Pk,n3Pk,n4；
b2) With specific reference to FIG. 5(b), at points Pk,n3,Pk,n4As a circle center, a distance threshold value TdisIs a radius on the line segment Pk, n3Pk,n4Towards the camera position
Figure BDA0002995211980000111
Drawing a semicircle at one side, and taking two semicircles as a line segment Pk,n1Pk,n2Near camera position
Figure BDA0002995211980000112
One side intersection region as a virtual viewpoint distance reasonable region Ak,dis；
b3) Referring specifically to FIG. 5(c), at points Pk,n1,Pk,n2Being a corner point, an angle threshold TangFor deflection angle, clockwise and anticlockwise respectively deflect TangMaking four rays, and taking the four rays to be on the line segment Pk,n1Pk,n2Near camera position
Figure BDA0002995211980000113
One side intersection region as a virtual viewpoint angle reasonable region Ak,ang；
b4) Referring specifically to fig. 5(d), the virtual viewpoint range a of the camerakIs Ak,disAnd Ak,angThe intersection of (a);
b5) let Obj be the total set of all video moving objects in all video cameras; note N in the kth camerakEach video moving object has a track marked as Ck,i，Ck,iThe expression of (a) is as follows:
Obj＝{Ck,i,(k＝1,2…L)}
Ck,i＝{Ik,i,j,Pk,i,j,(i＝1,2,…,Nk)(j＝1,2,…,n)}；
wherein L represents the number of cameras, Ik,i,jAnd Pk,i,jAnd (3) representing the sub-image of the ith video moving object in the jth video frame in the kth camera and the geographic spatial position of the sub-image are subjected to cross-camera correlation analysis by the video moving object, and the single camera view is mergedObtaining multi-camera video motion target track by frequency motion target track
Figure BDA0002995211980000114
Realizing the related organization of the multi-camera video moving target:
Figure BDA0002995211980000115
Cubeio＝{Ck1,i1,Ck2,i2…Cko,iL,(k1,k2,…ko)∈(1,2…L)}；
wherein L isoRepresenting the total number of video moving objects, Cube, after merging of cross-camera homonymous video moving objects in a surveillance video networkioRepresenting the global track of the video moving object with the sequence number io in the monitoring video network,
Figure BDA0002995211980000116
and (4) representing the sub-track of the video moving object with the sequence number io in the ko camera.
Specifically, in step S4, the optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group specifically includes the following steps:
c1) the number of cameras is recorded as L, and a set formed by all combination modes of the cameras is recorded as M:
Figure BDA0002995211980000117
mi＝{ni,j}
Figure BDA0002995211980000118
see in particular fig. 6, where miThe ith camera combination mode comprises all camera sets in the combination mode; n isi,jDigital camera combination mode miA j-th group of cameras under the condition that all cameras of the group are included;
Figure BDA0002995211980000121
the first camera in the j camera set in the i camera set combination mode is referred to;
c2) by defining a distance threshold TdisAnd an angle threshold TangAssembling the patterns m for each cameraiIn each camera group ni,jAll cameras in
Figure BDA0002995211980000122
Calculating observable domains and calculating intersection; if a certain camera combination mode miAll the camera sets ni,jIf the intersection of any number of the camera observation domains is not empty, the combination mode m of the camera is recordediIs an observable combination, otherwise, the camera combination mode m is recordediIs an unobservable combination;
c3) based on multi-camera video target trajectory data, the following video condensation optimization targets are specified to realize the optimization of the camera group:
firstly, expressing coherence of homonymous targets across cameras, namely expressing the video cameras with the appearance of a single target in a consistent manner by using as few virtual viewpoints as possible;
expressing that the total number of virtual viewpoints used by all video targets is as small as possible;
c4) comprehensively evaluating the multi-camera video target expression effect of the camera combination corresponding to the virtual viewpoint group by value:
Figure BDA0002995211980000123
wherein n iscRepresents the total number of cameras, nvRepresenting the number of virtual viewpoints, N representing the total number of video moving objects, miRepresenting the number of virtual viewpoints which are related and expressed by each video moving object, wherein mu is a weight parameter;
c5) at a distance threshold TdisAnd an angle threshold TangWhen the value is taken to be one, the value values of all current camera observable sets are calculated by defining a parameter alpha, and the maximum value is taken as the camera combination for selectionAnd taking the result, and concentrating the multi-camera video in the virtual scene.
Referring to fig. 7 in particular, step S4 sets the moving object display parameters to perform multi-camera video concentration under the condition that the camera display combination is selected, with the central point of the observable domain of each camera group as the virtual viewpoint, based on the observable set preference result.
In step S5, the preset display parameters of the moving object, according to the display parameters, perform multi-camera video concentration, including the following steps:
d1) recording that W virtual viewpoints (W is less than or equal to L) are needed for viewing the video moving objects of all L cameras under the current camera combination; meanwhile, setting a frame rate fps displayed by the video moving target subgraph in a three-dimensional scene as the number of the subgraphs displayed by a single video moving target per second; setting an object display interval time t0As time intervals for additional display of new video moving objects;
d2) for a certain virtual viewpoint W (W ≦ W), the first occurring moving object O is first displayed0In geospatial track T0And identifying the sequence of appearance of the video object among different cameras;
screening the video object subgraph according to the frame rate fps, converting the plane coordinate corresponding to the screened video object subgraph into a geographic coordinate, and meanwhile, according to a proportionality coefficient Pw、PhScaling a video object sub-graph, Pw、PhThe calculation formula is as follows:
Figure BDA0002995211980000131
Figure BDA0002995211980000132
wherein
Figure BDA0002995211980000133
Randomly selecting the average width and height of a proper number of subgraphs from a video object subgraph library, and selectingThe extracted subgraph is mapped to the corresponding geographic position in the virtual scene corresponding to the three-point coordinates of the upper left, the lower left and the upper right in the original video frame to obtain the length and the height of the video object subgraph in the three-dimensional space,
Figure BDA0002995211980000134
is the average length and height of the video object sub-graph displayed in the virtual scene;
d3) displaying O in the camera view field in the virtual scene according to the frame rate fps in the dynamic display process0The video object sub-graph of the current frame is in the corresponding geographic position, and the old video object sub-graph is not displayed any more;
on the other hand, at t0,2t0…,nt0At the moment, the video object O is added1,O2…OnDynamically expressed in a three-dimensional scene model to realize the concentration of multi-camera video objects.
Specifically, in step S5, for the case where the same object trajectory generated by overlapping the camera shooting areas is obtained by multiple cameras, the camera that obtains the object sub-graph is determined by comparing the included angles between the virtual viewpoint and the connecting line between the object trajectory point and the three points at the two camera positions:
specifically referring to fig. 8, a view overlapping part C exists between the camera a and the camera b, for a video object passing through the view C, the included angles between three points, i.e., the sizes of α and β, of the camera position, the track point and the virtual viewpoint V are compared, if α is not greater than β, the video object sub-graph acquired by the camera a is used, otherwise, the video object sub-graph acquired by the camera b is used.
Example 2
Referring to fig. 9, the present invention further provides a multi-camera video concentration system in a geographical scene with a preferred virtual viewpoint, which employs the multi-camera video concentration method in a geographical scene with a preferred virtual viewpoint in embodiment 1 or any possible implementation manner thereof, and the concentration system includes:
the homologous point acquisition module 1: the system comprises a video sequence image acquisition module, a video acquisition module, a three-dimensional geographic scene model acquisition module, a coordinate acquisition module and a display module, wherein the video sequence image acquisition module is used for acquiring video sequence image information, acquiring homonymous point pairs in a video image and the three-dimensional geographic scene model and acquiring coordinate data of the homonymous point pairs, and the coordinate data comprises image coordinates and geographic coordinates;
the mapping relation building module 2: the method comprises the steps of establishing a mapping relation between a video image and a geographic space according to coordinate data of a same-name point pair, and positioning a camera view;
the camera group observable collection generation module 3: the system comprises a camera observation domain model and a camera group observation set, wherein the camera observation domain model is constructed by analyzing an observable distance and a sight line deflection angle to generate a camera group observable set;
virtual viewpoint group generation module 4: the system is used for optimizing the observable collection by constructing an evaluation model to obtain a virtual viewpoint group;
specifically, a video image visual domain model is constructed to describe the range in which each camera in a virtual scene can be effectively observed, then a virtual viewpoint is generated, and the global motion condition of a video motion target in a multi-camera geographic scene is checked; on the basis of a camera observation domain model, exhaustively exhausting a camera observable set, and preferably selecting a combination with the best video target information expression effect as a virtual viewpoint generation region;
the video target space-time motion expression module 5: the method comprises the steps of presetting display parameters of a moving target, and carrying out multi-camera video concentration according to the display parameters.
Specifically, based on the optimal result of the observable set, under the condition that the camera display combination is selected, the central point of the observable domain of each camera group is used as a virtual viewpoint, the moving object display parameters are set, and multi-camera video concentration is performed.
It should be noted that, for the information interaction, execution process and other contents between the modules/units of the multi-camera video concentrating system in the geographic scene with the preferred virtual viewpoint, the same concept is based on as the method embodiment in the embodiment of the present application, and the technical effect brought by the same concept is the same as the method embodiment of the present application, and specific contents can be referred to the description in the foregoing method embodiment of the present application.
Acquiring video sequence image information, acquiring homonymous point pairs in a video image and a three-dimensional geographic scene model, and acquiring coordinate data of the homonymous point pairs, wherein the coordinate data comprises image coordinates and geographic coordinates; establishing a mapping relation between a video image and a geographic space according to coordinate data of the same-name point pair, and positioning a camera view; constructing a camera observation domain model by analyzing the observable distance and the sight line deflection angle, and generating an observable set of a camera group; optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group; and presetting display parameters of the moving target, and performing multi-camera video concentration according to the display parameters. The method has the remarkable effects that the mapping relation between the video target and the geographic scene is established, the effect of fusion expression of the monitoring video in the geographic scene is enhanced, and great convenience is provided for the quick retrieval and efficient understanding of the video geographic scene information integration.
As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method or program product. Thus, various aspects of the invention may be embodied in the form of: an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.) or an embodiment combining hardware and software aspects that may all generally be referred to herein as a "module" or "platform.
Although the invention has been described in detail above with reference to a general description and specific examples, it will be apparent to one skilled in the art that modifications or improvements may be made thereto based on the invention. Accordingly, such modifications and improvements are intended to be within the scope of the invention as claimed.

Claims (10)

1. A method for concentrating multi-camera video in a geographic scene with a preferred virtual viewpoint is characterized by comprising the following steps:
acquiring video sequence image information, acquiring homonymous point pairs in a video image and a three-dimensional geographic scene model, and acquiring coordinate data of the homonymous point pairs, wherein the coordinate data comprises image coordinates and geographic coordinates;
establishing a mapping relation between a video image and a geographic space according to coordinate data of the same-name point pair, and positioning a camera view;
constructing a camera observation domain model by analyzing the observable distance and the sight line deflection angle, and generating an observable set of a camera group;
optimizing the observable set by constructing an evaluation model to obtain a virtual viewpoint group;
presetting display parameters of the moving target, and performing multi-camera video concentration according to the display parameters.
2. The method for concentrating multi-camera video in geographic scene with preferred virtual viewpoint as claimed in claim 1, wherein the video image is the first frame image of the intercepted surveillance video.
3. The method as claimed in claim 2, wherein the three-dimensional geo-scene model is a three-dimensional scene model constructed according to real geo-scene measurement information, the number of the same-name point pairs collected on the video image and the virtual geo-scene is not less than three, and the three same-name point pairs are not all collinear.
4. The method of claim 3, wherein the step of establishing a mapping relationship between the video images and the geographic space and positioning the camera view comprises the steps of:
a1) presetting an object point geographic space coordinate Q corresponding to a given image point image space coordinate Q, and expressing Q and Q as a homogeneous coordinate form:
q＝[x y 1]T
Q＝[X Y Z 1]T；
when a homography matrix M is recorded, the relationship between Q and Q is:
q＝MQ；
the expression of the homography matrix M is:
Figure FDA0002995211970000021
a2) solving the geographic space coordinates of the object points corresponding to the image points in each image:
Figure FDA0002995211970000022
a3) assuming that there are L cameras in the current video camera network, the mapping matrix of the kth camera (k 1, 2.. L) is labeled Mk(ii) a Defining each camera position in geographic space as
Figure FDA0002995211970000023
The geospatial position of each camera field of view polygon is
Figure FDA0002995211970000024
Figure FDA0002995211970000025
Figure FDA0002995211970000026
Wherein the camera position
Figure FDA0002995211970000027
Considered as a point in the geographic space, the camera view polygon is bounded by o boundary points Pk，numSequentially concatenating the constituent polygon records.
5. The method for concentrating multi-camera video in geographic scene with a preferred virtual viewpoint as claimed in claim 4, wherein the process of positioning the camera view selects two factors of virtual line-of-sight distance and camera-virtual viewpoint angle as constraints;
the virtual line-of-sight distance refers to the geographic space distance between a virtual viewpoint and a certain point in the view field; the camera-virtual viewpoint included angle is an included angle formed by projection of a given point in a visual field, the virtual viewpoint and a camera position point on a horizontal plane;
defining a distance threshold TdisAnd an angle threshold TangAs a constraint, and assuming a distance threshold TdisAnd an angle threshold TangIt has been given that finding areas in the scene model that meet the constraints serves as virtual viewpoint ranges.
6. The method for concentrating multiple cameras in a geographic scene with a preferred virtual viewpoint as claimed in claim 5, wherein the step of constructing a camera observation domain model by analyzing an observable distance and a view deflection angle to generate a camera group observable set comprises the following steps:
b1) recording camera position
Figure FDA0002995211970000028
Polygon with camera view
Figure FDA0002995211970000029
And is
Figure FDA00029952119700000210
In the distance
Figure FDA00029952119700000211
Is Pk，n1Pk，n2The farthest line segment is Pk，n3Pk，n4；
b2) Respectively at point Pk，n3，Pk，n4As a circle center, a distance threshold value TdisIs a radius on the line segment Pk，n3Pk，n4Towards the camera position
Figure FDA0002995211970000031
Drawing a semicircle at one side, and taking two semicircles as a line segment Pk，n1Pk，n2Near camera position
Figure FDA0002995211970000032
One side intersection region as a virtual viewpoint distance reasonable region Ak，dis；
b3) Respectively at point Pk，n1，Pk，n2Being a corner point, an angle threshold TangFor deflection angle, clockwise and anticlockwise respectively deflect TangMaking four rays, and taking the four rays to be on the line segment Pk，n1Pk，n2Near camera position
Figure FDA0002995211970000033
One side intersection region as a virtual viewpoint angle reasonable region Ak，ang；
b4) Virtual viewpoint range A of camerakIs Ak，disAnd Ak，angThe intersection of (a);
b5) let Obj be the total set of all video moving objects in all video cameras; note that there is N in the kth camerakEach video moving object has a track marked as Ck，i，Ck，iThe expression of (a) is as follows:
Obj＝{Ck，i，(k＝1，2...L)}
Ck，i＝{Ik，i，j，Pk，i，j，(i＝1，2，...，Nk)(j＝1，2，...，n)}；
wherein L represents the number of cameras, Ik，i，jAnd Pk，i，jRepresenting the sub-image of the ith video moving object in the jth video frame in the kth camera and the geographic space position of the sub-image are analyzed by the cross-camera association of the video moving objects, and the tracks of the single-camera video moving object are merged to obtain the tracks of the multi-camera video moving object
Figure FDA0002995211970000034
Realizing the related organization of the multi-camera video moving target:
Figure FDA0002995211970000035
Cubeio＝{Ck1，i1，Ck2，i2...Cko，iL，(k1，k2，...ko)∈(1，2...L)}；
wherein L isoRepresenting the total number of video moving objects, Cube, after merging of cross-camera homonymous video moving objects in a surveillance video networkioRepresenting the global track of the video moving object with the sequence number io in the monitoring video network,
Figure FDA0002995211970000036
and (4) representing the sub-track of the video moving object with the sequence number io in the ko camera.
7. The method for concentrating the multi-camera video in the geographic scene with the preferred virtual viewpoint as claimed in claim 6, wherein the observable set is preferred by constructing an evaluation model to obtain a virtual viewpoint group, specifically comprising the following steps:
c1) the number of cameras is recorded as L, and a set formed by all combination modes of the cameras is recorded as M:
Figure FDA0002995211970000037
mi＝{ni，j}
Figure FDA0002995211970000038
wherein m isiThe ith camera combination mode comprises all camera sets in the combination mode; n isi，jDigital camera combination mode miA j-th group of cameras under the condition that all cameras of the group are included;
Figure FDA0002995211970000041
the first camera in the j camera set in the i camera set combination mode is referred to;
c2) by defining a distance threshold TdisAnd an angle threshold TangAssembling the patterns m for each cameraiIn each camera group ni，jAll cameras in
Figure FDA0002995211970000042
Calculating observable domains and calculating intersection; if a certain camera combination mode miAll the camera sets ni，jIf the intersection of any number of the camera observation domains is not empty, the combination mode m of the camera is recordediIs an observable combination, otherwise, the camera combination mode m is recordediIs an unobservable combination;
c3) based on multi-camera video target trajectory data, the following video condensation optimization targets are specified to realize the optimization of the camera group:
firstly, expressing coherence of homonymous targets across cameras, namely expressing the video cameras with the appearance of a single target in a consistent manner by using as few virtual viewpoints as possible;
expressing that the total number of virtual viewpoints used by all video targets is as small as possible;
c4) comprehensively evaluating the multi-camera video target expression effect of the camera combination corresponding to the virtual viewpoint group by value:
Figure FDA0002995211970000043
wherein n iscRepresents the total number of cameras, nvRepresenting the number of virtual viewpoints, N representing the total number of video moving objects, miRepresenting the number of virtual viewpoints which are related and expressed by each video moving object, wherein mu is a weight parameter;
c5) at a distance threshold TdisAnd an angle threshold TangWhen the value is constant, value values of all current camera observable sets are calculated through defining a parameter alpha, the maximum value is taken as a camera combination selection result, and multi-camera video concentration in a virtual scene is carried out.
8. The method of claim 7, wherein the method for concentrating the multi-camera video in the geographic scene with the virtual viewpoint being preferred comprises the following steps:
d1) recording that W virtual viewpoints (W is less than or equal to L) are needed for viewing the video moving objects of all L cameras under the current camera combination; meanwhile, setting a frame rate fps displayed by the video moving target subgraph in a three-dimensional scene as the number of the subgraphs displayed by a single video moving target per second; setting an object display interval time t0As time intervals for additional display of new video moving objects;
d2) for a certain virtual viewpoint W (W ≦ W), the first occurring moving object O is first displayed0In geospatial track T0And identifying the sequence of appearance of the video object among different cameras;
screening the video object subgraph according to the frame rate fps, converting the plane coordinate corresponding to the screened video object subgraph into a geographic coordinate, and meanwhile, according to a proportionality coefficient Pw、PhScaling a video object sub-graph, Pw、PhThe calculation formula is as follows:
Figure FDA0002995211970000051
Figure FDA0002995211970000052
wherein
Figure FDA0002995211970000053
Randomly selecting an appropriate amount of average width and height of subgraphs from a video object subgraph library, mapping coordinates of three points, namely upper left, lower left and upper right, of the selected subgraphs in an original video frame to corresponding geographic positions in a virtual scene to obtain the length and height of the video object subgraphs in a three-dimensional space,
Figure FDA0002995211970000054
is the average length and height of the video object sub-graph displayed in the virtual scene;
d3) displaying O in the camera view field in the virtual scene according to the frame rate fps in the dynamic display process0The video object sub-graph of the current frame is in the corresponding geographic position, and the old video object sub-graph is not displayed any more;
at t0，2t0...，nt0At the moment, the video object O is added1，O2...OnDynamically expressed in a three-dimensional scene model to realize the concentration of multi-camera video objects.
9. The method according to claim 8, wherein for the case that the same object track generated by overlapping camera shooting areas is obtained by a plurality of cameras, the cameras for obtaining the object subgraph are determined by comparing the included angles between the virtual viewpoint and the object track point and the connecting line of three points at the two camera positions:
and a view overlapping part C exists between the camera a and the camera b, for the video object passing through the view C, the included angle among the three points of the camera position, the track point and the virtual viewpoint V, namely the sizes of alpha and beta, is compared, if alpha is less than or equal to beta, the video object sub-graph acquired by the camera a is used, and otherwise, the video object sub-graph acquired by the camera b is used.
10. A multi-camera video concentration system in a geographical scene with a preferred virtual viewpoint, which adopts the multi-camera video concentration method in a geographical scene with a preferred virtual viewpoint as claimed in any one of claims 1 to 9, wherein the concentration system comprises:
the same-name point acquisition module: the system comprises a video sequence image acquisition module, a video acquisition module, a three-dimensional geographic scene model acquisition module, a coordinate acquisition module and a display module, wherein the video sequence image acquisition module is used for acquiring video sequence image information, acquiring homonymous point pairs in a video image and the three-dimensional geographic scene model and acquiring coordinate data of the homonymous point pairs, and the coordinate data comprises image coordinates and geographic coordinates;
the mapping relation construction module: the method comprises the steps of establishing a mapping relation between a video image and a geographic space according to coordinate data of a same-name point pair, and positioning a camera view;
a camera group observable collection generation module: the system comprises a camera observation domain model and a camera group observation set, wherein the camera observation domain model is constructed by analyzing an observable distance and a sight line deflection angle to generate a camera group observable set;
a virtual viewpoint group generation module: the system is used for optimizing the observable collection by constructing an evaluation model to obtain a virtual viewpoint group;
the video target space-time motion expression module: the method comprises the steps of presetting display parameters of a moving target, and carrying out multi-camera video concentration according to the display parameters.
CN202110327605.1A 2021-03-26 2021-03-26 Multi-camera video concentration method and system in virtual viewpoint-optimized geographic scene Active CN113192125B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110327605.1A CN113192125B (en) 2021-03-26 2021-03-26 Multi-camera video concentration method and system in virtual viewpoint-optimized geographic scene

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110327605.1A CN113192125B (en) 2021-03-26 2021-03-26 Multi-camera video concentration method and system in virtual viewpoint-optimized geographic scene

Publications (2)

Publication Number Publication Date
CN113192125A true CN113192125A (en) 2021-07-30
CN113192125B CN113192125B (en) 2024-02-20

Family

ID=76974146

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110327605.1A Active CN113192125B (en) 2021-03-26 2021-03-26 Multi-camera video concentration method and system in virtual viewpoint-optimized geographic scene

Country Status (1)

Country Link
CN (1) CN113192125B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113840159A (en) * 2021-09-26 2021-12-24 北京沃东天骏信息技术有限公司 Video processing method, device, computer system and readable storage medium
CN114067071A (en) * 2021-11-26 2022-02-18 湖南汽车工程职业学院 High-precision map making system based on big data

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2013033442A1 (en) * 2011-08-30 2013-03-07 Digimarc Corporation Methods and arrangements for identifying objects
US20130088592A1 (en) * 2011-09-30 2013-04-11 OOO "ITV Group" Method for searching for objects in video data received from a fixed camera
CN110009561A (en) * 2019-04-10 2019-07-12 南京财经大学 A kind of monitor video target is mapped to the method and system of three-dimensional geographical model of place
CN110148223A (en) * 2019-06-03 2019-08-20 南京财经大学 Monitor video target concentration expression and system in three-dimensional geography model of place
CN110516014A (en) * 2019-01-18 2019-11-29 南京泛在地理信息产业研究院有限公司 A method of two-dimensional map is mapped to towards urban road monitor video
CN111161130A (en) * 2019-11-25 2020-05-15 北京智汇云舟科技有限公司 Video correction method based on three-dimensional geographic information
CN111582022A (en) * 2020-03-26 2020-08-25 深圳大学 Fusion method and system of mobile video and geographic scene and electronic equipment
CN112381935A (en) * 2020-09-29 2021-02-19 西安应用光学研究所 Synthetic vision generation and multi-element fusion device

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2013033442A1 (en) * 2011-08-30 2013-03-07 Digimarc Corporation Methods and arrangements for identifying objects
US20130088592A1 (en) * 2011-09-30 2013-04-11 OOO "ITV Group" Method for searching for objects in video data received from a fixed camera
CN110516014A (en) * 2019-01-18 2019-11-29 南京泛在地理信息产业研究院有限公司 A method of two-dimensional map is mapped to towards urban road monitor video
CN110009561A (en) * 2019-04-10 2019-07-12 南京财经大学 A kind of monitor video target is mapped to the method and system of three-dimensional geographical model of place
CN110148223A (en) * 2019-06-03 2019-08-20 南京财经大学 Monitor video target concentration expression and system in three-dimensional geography model of place
CN111161130A (en) * 2019-11-25 2020-05-15 北京智汇云舟科技有限公司 Video correction method based on three-dimensional geographic information
CN111582022A (en) * 2020-03-26 2020-08-25 深圳大学 Fusion method and system of mobile video and geographic scene and electronic equipment
CN112381935A (en) * 2020-09-29 2021-02-19 西安应用光学研究所 Synthetic vision generation and multi-element fusion device

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JIANQING ZHU; SHENGCAI LIAO; STAN Z. LI: "Multicamera Joint Video Synopsis", IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY *
解愉嘉: "地理场景中监控视频浓缩方法研究", 中国博士学位论文全文数据库 信息科技辑 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113840159A (en) * 2021-09-26 2021-12-24 北京沃东天骏信息技术有限公司 Video processing method, device, computer system and readable storage medium
CN114067071A (en) * 2021-11-26 2022-02-18 湖南汽车工程职业学院 High-precision map making system based on big data
CN114067071B (en) * 2021-11-26 2022-08-30 湖南汽车工程职业学院 High-precision map making system based on big data

Also Published As

Publication number Publication date
CN113192125B (en) 2024-02-20

Similar Documents

Publication Publication Date Title
Morse et al. UAV video coverage quality maps and prioritized indexing for wilderness search and rescue
US7522186B2 (en) Method and apparatus for providing immersive surveillance
CN103795976B (en) A kind of full-time empty 3 d visualization method
CN104156972B (en) Perspective imaging method based on laser scanning distance measuring instrument and multiple cameras
RU2612378C1 (en) Method of replacing objects in video stream
Sturm et al. Evaluating egomotion and structure-from-motion approaches using the TUM RGB-D benchmark
US8180107B2 (en) Active coordinated tracking for multi-camera systems
US20120147152A1 (en) 3d image generation
US20100103173A1 (en) Real time object tagging for interactive image display applications
US20180075590A1 (en) Image processing system, image processing method, and program
US20110172556A1 (en) System And Method For Quantifying And Mapping Visual Salience
US20220141425A1 (en) Target Tracking in a Multi-Camera Surveillance System
TW200818916A (en) Wide-area site-based video surveillance system
CN113192125B (en) Multi-camera video concentration method and system in virtual viewpoint-optimized geographic scene
JP2006197373A (en) Viewer information measuring instrument
TW201025193A (en) Method for automatic detection and tracking of multiple targets with multiple cameras and system therefor
WO2008132741A2 (en) Apparatus and method for tracking human objects and determining attention metrics
CN114387679B (en) System and method for realizing sight estimation and attention analysis based on recurrent convolutional neural network
Cui et al. Fusing surveillance videos and three‐dimensional scene: A mixed reality system
CN111753112B (en) Information generation method, device and storage medium
den Hollander et al. Automatic inference of geometric camera parameters and inter-camera topology in uncalibrated disjoint surveillance cameras
US20050071105A1 (en) Method and system for calibrating relative fields of view of multiple cameras
KR20160099933A (en) Method for analyzing a visible area of a closed circuit television considering the three dimensional features
KR101051355B1 (en) 3D coordinate acquisition method of camera image using 3D spatial data and camera linkage control method using same
Bagdanov et al. Acquisition of high-resolution images through on-line saccade sequence planning

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
GR01 Patent grant
GR01 Patent grant