CN103716594B - Panorama splicing linkage method and device based on moving target detecting - Google Patents
Panorama splicing linkage method and device based on moving target detecting Download PDFInfo
- Publication number
- CN103716594B CN103716594B CN201410008471.7A CN201410008471A CN103716594B CN 103716594 B CN103716594 B CN 103716594B CN 201410008471 A CN201410008471 A CN 201410008471A CN 103716594 B CN103716594 B CN 103716594B
- Authority
- CN
- China
- Prior art keywords
- ball machine
- panoramic mosaic
- unit
- original video
- splicing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention provides a panorama splicing linkage method and device based on moving target detecting. A large-view panorama image is obtained by panorama splicing, then view is adjusted, the position mapping relation of a ball machine and a panorama splicing video camera is established, then moving target detecting is carried out on an area of interest in the panorama image, after a moving target is detected, the corresponding ball machine is automatically triggered to track the target of the area of interest according to the mapping relation to achieve linkage, accordingly, the problem that the moving target (a human face, a license plate and the like) in the panorama image cannot be seen clearly under the situation of long distance is avoided, workload of monitoring personnel is reduced, a complete, large-view, and high-resolution continuous image is provided for an operator, the observing angle is similar to the view angle of a human eye, and switching back and forth between images during monitoring is avoided.
Description
Technical field
The present invention relates to technical field of video processing, refer in particular to a kind of linkage of the panoramic mosaic based on moving object detection
Method and device.
Background technology
In today of monitoring system fast development, system scale is increasing, and control point gets more and more, faced by staff
Increasing monitoring image is had no time to attend to, with the continuous expansion of monitoring range, needs to bed rearrangement bridge, airport, very long
One section of highway or skyscraper are monitored, and the visual field of most of video cameras does not reach so big monitoring range, therefore right
The requirement of big visual field video monitoring gradually increases, and has gradually formed a kind of high demand.Video panorama splicing solves as a kind of
Scheme, has also obtained increasing concern.Thing followed video monitoring system will play important role.
Up to the present, common monitoring device has 2 kinds, and the first is gunlock, is directed at certain scene using gunlock and carries out
Monitoring, weak point is when monitoring scene is more, needs to toggle between each scene, when video way is more, switching frequency
When numerous, monitoring personnel is easily tired, and the attention of people can only concentrate on part picture;Second is ball machine, and ball machine monitors
By rotating ball machine, peripheral region is monitored, but shortcoming is ball machine limited view, the same time can only monitor a scene, so
Still part important information can be missed.
In target following link, if singly be rely on ball machine, whether manually ball-handling or automatically track target mode its
Tracking effect is affected by many factors, and is susceptible to follow the tracks of unsuccessfully or target switches, and during tracking, monitored picture visual field collection
In follow the tracks of target proximity, easily miss other important informations.If dependence gunlock, target easily occurs in multiple pictures,
The switching impact seizure to video information for the monitoring personnel between picture, and gunlock installation direction is fixed, and monitoring scene will not be with
Target and change, be easier to lose follow the tracks of target.And in video monitoring, often due to target range farther out, leads to target(People
Face, car plate etc.)Detection recognition correct rate relatively low.
Content of the invention
It is an object of the invention to overcoming drawbacks described above, provide a kind of linkage of the panoramic mosaic based on moving object detection
Method and device.
The object of the present invention is achieved like this:A kind of panoramic mosaic interlock method based on moving object detection, it wraps
Include step,
A), in control point setting ball machine and panoramic mosaic video camera, in setting up procedure panoramic mosaic video camera using level/
The vertical connecting method of vertical/horizontal, realizes panorama splicing by adjusting each lens direction;
B), the visual field of adjustment panoramic mosaic video camera, be allowed to the ball machine visual field that the visual field in panoramic mosaic picture covers setting;
Because ball machine can be rotated down with horizontal 360-degree, level, the visual field that theoretical Apparatus for feeding balls as disintegrating members can cover is greater than panorama and spells
Connect video camera, but ball machine cannot the more than covering level visual field it is therefore desirable to this step adjustment panoramic mosaic video camera is
Wish the visual field of panorama all below horizontal, ball machine can cover after rotating.
C), demarcate ball machine and panoramic mosaic video camera, set up the position mapping relations of ball machine and panoramic mosaic video camera;
D), after in panoramic mosaic picture setting area-of-interest after, moving object detection is carried out to this area-of-interest;
E), when moving target is detected, corresponding moving target labelling tracking box, adjusted according to tracking box state modulator ball machine
Whole, focusing shoots tracking target image;
In above-mentioned steps, also comprise determining whether after described step C to demarcate successfully, be to continue, otherwise adjust ball machine
Picture be allowed to not splice with panoramic mosaic video camera before each raw frames center alignment after go back to the step of step C;
In above-mentioned steps, in described step F, tracking box parameter includes tracking box size, tracking box positional information;
In above-mentioned steps, also include after described step F ball machine image being carried out with feature detection and extracts feature then identifying
Step;
In above-mentioned steps, in described step, feature detection includes face, vehicle license plate characteristic detection;
In above-mentioned steps, it is identified by described step and the feature of extraction is identified with Data Comparison in data base.
The invention still further relates to a kind of panoramic mosaic linkage based on moving object detection, it includes:
Splicing setup module, in control point setting ball machine and panoramic mosaic video camera, panoramic mosaic in setting up procedure
Video camera adopts horizontal/vertical/horizontal vertical connecting method, realizes panorama splicing by adjusting each lens direction, then
Go to adjusting module;
Adjusting module, for adjusting the visual field of panoramic mosaic video camera, is allowed to the visual field in panoramic mosaic picture and covers setting
The ball machine visual field, then go to demarcating module;
Demarcating module, for demarcating ball machine and panoramic mosaic video camera, sets up the position of ball machine and panoramic mosaic video camera
Mapping relations, then go to setup module;
Setup module, is used for, after setting area-of-interest in panoramic mosaic picture, this area-of-interest being moved
Target detection, then goes to motion detection block;
Motion detection block, for when moving target is detected, corresponding moving target labelling tracking box, according to tracking box ginseng
The adjustment of numerical control marble forming machine, focusing shoot tracking target image;
In above-mentioned, described demarcating module specifically includes,
Broadcast unit, does not splice original video and ball machine video for play panoramic mosaic video camera, then goes to choosing
Select original video unit;
Select original video unit, select one of panoramic mosaic video camera not splice original video for unduplicated and
After go to characteristic point select unit;
Characteristic point select unit, for not splicing unduplicated selection one in original video by input equipment select
Individual characteristic point, the coordinate of record this feature point goes to grouped record unit;
Grouped record unit, for marking the center of picture in ball machine video, makes ball machine picture center fall by manipulation
Select in the selected characteristic point of original video unit, by ball machine level inclination value now, ball machine vertical dip angle value and ball machine
The coordinate grouped record of scaling multiplier value character pair point, goes to repetitives;
Repetitives, then go to demarcation list at least 6 times for repeat step characteristic point select unit, demarcation recording unit
Unit;
Demarcate unit, for the ball machine level inclination value in every group of record, ball machine vertical dip angle value and coupling point coordinates
Quadratic surface coefficient is gone out using least square fitting, using this quadratic surface coefficient as calibration coefficient together with select original video
In the identifier of splicing original video selecting in unit and grouped record unit, the identifier correspondence of the ball machine of manipulation has preserved
Become mapping relations to set up, then go to panorama repetitives;
Panorama repetitives, select original video unit, characteristic point select unit, grouped record list for repeat step
Unit, repetitives and demarcation unit, until all mapping relations not splicing original video completing panoramic mosaic video camera are built
Vertical;
In above-mentioned, described demarcating module specifically includes,
Manipulation unit, for manipulation make ball machine picture and panoramic mosaic video camera respectively not splice original video center consistent
Then go to feature extraction unit;
Feature extraction unit, for extracting the two field picture respectively not splicing original video, to this image zooming-out sift/
Surf/harris characteristic point, then characteristic matching, coupling then go to calculating perspective transformation matrix using method of least square
Margin estimation unit;
Margin estimation unit, for estimating its 4 edge angle points by perspective transformation matrix to respectively not splicing original video
Ball machine level inclination value and ball machine vertical dip angle value then go to ball machine bit cell;
Ball machine bit cell, does not splice original video for unduplicated select panoramic mosaic video camera one, by it
Changes in coordinates scope is divided into many parts, controls ball machine to arrive according to every part of ball machine level inclination value and ball machine vertical dip angle value respectively
Reach relevant position and then go to matching unit;
Matching unit, often reaches a position for ball machine, to now do not splice original video and ball machine picture is carried out
Feature extraction is simultaneously mated, if match point meets threshold value, preserves corresponding coordinate and ball machine ball machine level inclination value now and ball machine
Vertical dip angle value then goes to demarcation unit;
Demarcate unit, for repeating ball machine bit cell, matching unit, until complete panoramic mosaic video camera all not
The mapping relations of splicing original video are set up;
In above-mentioned, described demarcation unit goes to manual correction unit, and described manual correction unit is used for checking automatic Calibration
Coupling right, if there is error, controlling ball manually adjustment ball machine with reference to ball machine bit cell, matching unit, obtaining at least 6
Join a little, the ball machine level inclination value of the coordinate of match point and ball machine, ball machine vertical dip angle value and ball machine scaling multiplier value is corresponded to
Grouped record, adopts least square to the ball machine level inclination value in every group of record, ball machine vertical dip angle value and coupling point coordinates
Method simulates quadratic surface coefficient, using this quadratic surface coefficient as calibration coefficient together with select splicing original video mark
The corresponding preservation of the identifier of the ball machine of symbol and manipulation completes mapping relations and sets up;
In above-mentioned, the linkage of described motion detection block specifically has following units to realize,
Tracking box central point calculation unit, for according to tracking box information, the center point coordinate calculating tracking box then turns
To panoramic field positioning unit;
Panoramic field positioning unit, affiliated in panoramic mosaic picture for it is positioned according to tracking box center point coordinate value
Region then go to region decision unit;
Region decision unit, for judging the overlapping region whether above-mentioned zone is located in panoramic mosaic video camera, otherwise
The information obtaining region then executes coordinate calculating unit, is, coordinate calculating unit is compared in execution;
Coordinate calculating unit, is located for being calculated tracking box central point according to the splicing inverse transformation of the corresponding sequence number in region
Do not splice the coordinate position in original video, then go to ball machine linkage unit;
Relatively coordinate calculating unit, does not splice original video for comparing tracking box center point coordinate value with overlapping two
Central point distance, tracking box central point is calculated according to the splicing inverse transformation of the corresponding sequence number in closest region and is located
Do not splice the coordinate position in original video, then go to ball machine linkage unit;
Ball machine linkage unit, does not splice the mapping relations of original video for being located according to tracking box central point, to follow the tracks of
The coordinate position of frame central point obtains corresponding ball machine level inclination value and ball machine vertical dip angle value then goes to ball machine and executes list
Unit;
Ball machine performance element, drives ball machine adjustment according to ball machine level inclination value and ball machine vertical dip angle value.
The beneficial effects of the present invention is first passing through panoramic mosaic to obtain a wide-field panorama, then adjust
The visual field, set up ball machine and the position mapping relations of panoramic mosaic video camera after motion mesh is carried out to area-of-interest in panorama
Mark detection, triggers corresponding ball machine automatically according to mapping relations after moving target is detected and follows the tracks of area-of-interest realization of goal connection
Dynamic, thus the moving target not seeing in panorama in the case of avoiding at a distance(Face, car plate etc.), decrease monitoring personnel
Workload, area-of-interest in panorama can be manually selected and realize ball machine linkage, to manipulator provide one complete,
The big visual field, high-resolution continuous pictures, closer to the visual angle of eye-observation, it is to avoid toggle between each picture during monitoring.
Brief description
The concrete structure of the present invention is described in detail in detail below in conjunction with the accompanying drawings
Fig. 1 is that schematic diagram is spliced in the linkage of the present invention;
Fig. 2 is the horizontal spliced panoramic picture view of the present invention;
Fig. 3 is first embodiment of the present invention method flow diagram;
Fig. 4 is the vertical spliced panoramic picture view of the present invention;
Fig. 5 is second embodiment of the present invention method flow diagram;
Fig. 6 is the horizontal vertical spliced panoramic picture view of the present invention;
Fig. 7 is third embodiment of the present invention method flow diagram.
Specific embodiment
By the technology contents of the detailed description present invention, structural features, realized purpose and effect, below in conjunction with embodiment
And coordinate accompanying drawing to be explained in detail.
The invention provides a kind of panoramic mosaic interlock method based on moving object detection, it includes step,
A), in control point, ball machine and panoramic mosaic video camera are set, in setting up procedure, panoramic mosaic video camera adopts level
Or horizontal or vertical vertical connecting method, realize panorama splicing by adjusting each lens direction;
B), the visual field of adjustment panoramic mosaic video camera, be allowed to the ball machine visual field that the visual field in panoramic mosaic picture covers setting;
C), demarcate ball machine and panoramic mosaic video camera, set up the position mapping relations of ball machine and panoramic mosaic video camera;
D), after in panoramic mosaic picture setting area-of-interest after, moving object detection is carried out to this area-of-interest;
E), when moving target is detected, corresponding moving target labelling tracking box, according to tracking box parameter(Typically by
Tracking box parameter calculates corresponding pan, tilt and zoom)Control ball machine adjustment, focus on shooting tracking target image.
The video sequence that this method is had lap first each other carries out spatial match be aligned combination exposure fusion
Form the high-resolution consecutive image in big visual angle that a width comprises each Video sequence information afterwards, then set up ball machine and panoramic mosaic
Mapping relations between video camera, thus when operator directly hit or after frame favored area in panoramic mosaic midpoint picture, panorama is taken the photograph
Can achieve linkage between camera and ball machine, control the ball machine adjustment of relevant position, focus to this positioning area according to mapping relations
Domain.Further, operator also can arrange area-of-interest in panoramic mosaic picture, system automatically in this region whether
Moving object occurs detected, when splicing the appearance motion such as target such as vehicle, pedestrian in picture, system will link
Ball machine is controlled quickly to focus on this regional aim of positioning, so that operator can see the portion not seen in panoramic mosaic picture clearly
Point(Face, car plate etc.), be conducive to catching more details information.Thus whole monitoring method is by the linkage of panoramic mosaic picture
The mode of body ball machine, provide that one complete to manipulator, the big visual field, high-resolution continuous pictures, observe closer to people
The visual angle examined, it is to avoid toggle between each picture during monitoring.Coordinated using ball machine simultaneously, panorama analyzes motion mesh
Timestamp, calls ball machine quickly to focus on this target of positioning.Solve in original method frequent switching between multiple video pictures, lead to monitor
Personnel's distractibility, and check that multi-channel video is easily tired for a long time, pixel shared by target in panorama is less simultaneously, leads
Cause does not see target(Face, car plate etc.)Key message defect.
As an embodiment, above-mentioned steps C specifically include,
C1), play panoramic mosaic video camera do not splice original video and ball machine video;
C2), unduplicated select one of panoramic mosaic video camera and do not splice original video;
C3), by the input equipment such as mouse or touch select do not splice in original video unduplicated select one
Characteristic point i, records x/y axial coordinate pt [i] .x, pt [i] .y in picture for this feature point;
C4), mark the center of picture in ball machine video, make ball machine picture center fall step C2 by manipulation selected
In characteristic point, ball machine level inclination value now, ball machine vertical dip angle value and ball machine are scaled the seat of multiplier value character pair point
Mark grouped record;
Above-mentioned manipulation is generally manual manipulation, and ball machine level inclination value, ball machine vertical dip angle value and ball machine scaling multiplying power
Value then in ball machine respectively to should have ball machine level inclination function pan [i], vertical dip angle function tilt [i] of ball machine, scaling times
Rate function zoom [i], after being adjusted in place, reads the value of above-mentioned function.
C5), repeat step C3, C4 at least 6 times;
Do not splice unduplicated selection 6 characteristic points to each feature original video from panoramic mosaic video camera
The one-to-one ball machine level inclination value of center alignment characteristicses point record of picture of point manipulation ball machine video, ball machine vertical dip angle
Value and ball machine scaling multiplier value.
C6), the ball machine level inclination value in every group of record, ball machine vertical dip angle value(I.e. at least 6 groups of pan_a [i],
Tilt_a [i], i>=6)And coupling point coordinates(I.e. pt [i] .x, pt [i] .y)Quadratic surface system is gone out using least square fitting
Number, using this quadratic surface coefficient as calibration coefficient together with the splicing identifier of original video selecting in step C2 and step C4
The corresponding preservation of the identifier of the ball machine of middle manipulation completes to demarcate.
Above-mentioned quadratic surface Coefficient Fitting is gone out using least square fitting it is actually:
1st, set up the relation between ball machine level inclination value and original video coordinate x, y;
2nd, set up the relation between ball machine vertical dip angle value and original video coordinate x, y.
The above-mentioned relation principle set up ball machine and do not splice original video:
Assume that the PT value of ball machine and coupling meet following relation to coordinate:
pan=a1*X^2+a2*XY+a3*Y^2+a4*X+a5*Y+a6
tilt=b1*X^2+b2*XY+b3*Y^2+b4*X+b5*Y+b6
Mate to rear getting more than 6 groups, that is, pan, tilt and(X、Y)In the case of known, calculate a1, a2, a3,
A4, a5, a6 (quadratic surface coefficient)
b1、b2、b3、b4、b5、b6(Quadratic surface coefficient)
Herein, each of panoramic mosaic video camera is not spliced original video and comes from independent gunlock, and 1 gunlock generally has
One MAC Address, therefore identifier directly can adopt its MAC Address, and ball machine(Panorama hemisphere)Generally only one of which MAC ground
Location, therefore in order to distinguish, can using MAC+ passage as identifier, corresponding preserve after complete former to this not splicing of selecting
Beginning video and the staking-out work of ball machine(Set up mapping relations).
C7), repeat step C2-C6, until all mappings not splicing original video completing panoramic mosaic video camera are closed
System sets up.
Be can be seen that by the introduction of technique scheme, the present embodiment provides a kind of side being applied to manual demarcation
Case.The advantage of the program is that the precision of reconnaissance is higher, but shortcoming is then to use more inconvenience, and reconnaissance is cumbersome.
For the defect overcoming upper embodiment to demarcate manually, this patent provides a kind of embodiment of automatic Calibration, is equally
Corresponding step C specifically includes step:
C1), manipulation make ball machine picture and panoramic mosaic video camera respectively not splice original video center consistent;
The manipulation of this step typically by manual ball-handling make each ball machine picture and corresponding panoramic mosaic video camera respectively not
Splicing original video center is consistent, that is, do not splice picture center object of reference at ball machine picture center.
C2), extract the two field picture respectively not splicing original video, to this image zooming-out sift/surf/harris feature
Point, then characteristic matching, coupling calculate perspective transformation matrix to using method of least square;
C3), by perspective transformation matrix to respectively not splicing the ball machine level inclination that original video estimates its 4 edge angle points
Value and ball machine vertical dip angle value;
C4), be divided into many parts according to the excursion not splicing original video PT coordinate(General at least more than 16 parts), point
Ball machine is not controlled to reach relevant position according to every part of ball machine level inclination value and ball machine vertical dip angle value;
C5), ball machine often reach a position, to now do not splice original video and ball machine picture carries out feature extraction simultaneously
Coupling, if match point meets threshold value, preserves corresponding coordinate and ball machine ball machine level inclination value now and ball machine vertical dip angle
Value;
C6), repeat step C4-C5, until all mappings not splicing original video completing panoramic mosaic video camera are closed
System sets up.
Advantage of this embodiment is that easy and simple to handle, then precision is relatively low manually for shortcoming, and when using, ball machine picture will
Do not splice picture center consistent.Up to degree of precision after interpolation manual correction in automatic Calibration.
Not enough in order to solve the problems, such as above-mentioned precision, further, in one embodiment, walk in above-mentioned automatic calibration method
The step also including manual correction after rapid C5, including:
Checking that the coupling of automatic Calibration is right, if there is error, controlling ball manually adjustment ball machine with reference to step C4, C5, obtain to
Few 6 match points, the ball machine level inclination value of the coordinate of match point and ball machine, ball machine vertical dip angle value and ball machine are scaled multiplying power
The corresponding grouped record of value, to the ball machine level inclination value in every group of record, ball machine vertical dip angle value and coupling point coordinates using
Little square law simulates quadratic surface coefficient, using this quadratic surface coefficient as calibration coefficient together with select splicing original video
Identifier and the corresponding preservation of identifier of ball machine of manipulation complete mapping relations and set up.
In one embodiment, the linkage in described step E specifically includes step,
E1), according to tracking box information, calculate the center point coordinate of tracking box;
E2), its affiliated region in panoramic mosaic picture is positioned according to tracking box center point coordinate value;
E3), judge the overlapping region whether above-mentioned zone is located in panoramic mosaic video camera, otherwise obtain the information in region
Then execution step E4, is then execution step E5;
E4), be calculated tracking box central point and be located according to the splicing inverse transformation of the corresponding sequence number in region and do not splice original regarding
Coordinate position in frequency;
E5), compare the distance of the central point that tracking box center point coordinate value does not splice original video with overlapping two, root
It is calculated tracking box central point and is located according to the splicing inverse transformation of the corresponding sequence number in closest region and do not splice in original video
Coordinate position;
E6), be located according to tracking box central point and do not splice the mapping relations of original video, with the coordinate of tracking box central point
Position acquisition corresponding ball machine level inclination value and ball machine vertical dip angle value.
E7), drive ball machine adjustment according to ball machine level inclination value and ball machine vertical dip angle value.
, it comprises 4 secondary original images, and in figure A1, A2, A3, A4 are Non-overlapping Domain, and B1, B2, B3 are weights taking Fig. 1 as a example
Folded region, by certain fusion rule, eliminates splicing seam, when calculating splicing parameter, each region is in panorama sketch
In coordinate be known.
It is assumed that after above-mentioned steps E1, the center point coordinate calculating tracking box is (pt.x, pt.y).
When subsequently judging, if coordinate (pt.x, pt.y) belongs to A1 region, by the projection inverse transformation method of corresponding A 1
Then, according to the inverse process execution of splicing, if(Pt.x, pt.y) belong to A2, A3, A4 region, the inverse transformation of the corresponding sequence number of execution,
Calculate the coordinate position (mpt.x, mpt.y) in original image;
And if coordinate (pt.x, pt.y) belongs to B1 region, then this point need to be judged to the distance of original image 1 central point
D1, original image 2 central point apart from d2, near apart from which, just using the projection inverse transformation rule of corresponding sequence number, calculate former
Coordinate position (mpt.x, mpt.y) in beginning image;
Finally, the pan_a [6] and tilt_a [6] of corresponding identifier according to the result above demarcating preservation, are read,
Then according to formula:
PAN(Need the ball machine level inclination value of adjustment)=pan_a[1]*X^2+pan_a[2]*XY+pan_a[3]Y^2+
pan_a[4]*X+pan_a[5]*Y+pan_a[6](XY is the center following the tracks of target frame herein)
Calculate TILT using tilt_a [6] in the same manner(Need the ball machine vertical dip angle value of adjustment).
As an embodiment, also comprise determining whether after above-mentioned steps C to demarcate successfully, be to continue, otherwise adjust ball machine
Picture be allowed to not splice with panoramic mosaic video camera before each raw frames center alignment after go back to the step of step C.
Judging in the present embodiment demarcates whether success is critically important, by increasing this process so that follow-up operating entirely
Ensure that during scape picture that ball machine can complete the effect of rifle ball linkage.
And whether the demarcation in this step successfully judges, simplest mode is to rely on artificial judgment, by calibrated
In journey ball machine with splicing before single-image demarcate one by one, often demarcated a single-image, can by click on single-image scenario A, check
Whether ball machine can be responsive to A, judge to demarcate the successfully precision with demarcation with this.
As an embodiment, in above-mentioned steps E, tracking box parameter includes tracking box size, tracking box positional information.
Tracking box in existing video monitoring mostly is rectangle, and therefore, correspondence analysis typically obtains tracking box size, tracking box
Positional information, this two information can be described by the upper left of rectangle, lower-right most point feature.And the acquisition modes of parameter are divided into manually
With automatic, manual mode such as mouse draws rectangle in spliced panoramic picture, and automatic acquisition is then by spliced panoramic picture
Middle increase moving object detection, the common method of moving object detection has frame-to-frame differences method, Background difference, optical flow method etc., tracking box
Parameter depend on position and the size of moving target.
And the effect of tracking box size, tracking box positional information mainly supplies subsequent calculations ball machine to respond pan (horizontal field of view
Angle), tilt(Vertical field of view angle)、zoom(Scaling multiple)This 3 order parameters.
As an embodiment, after above-mentioned steps E, also include step F,
F), ball machine image carried out with feature detection and extracts feature and then identify.
The step for by increasing, obtain when moving target is detected and capturing moving target after positioning is focused on by ball machine
After image, system can be detected to image automatically according to feature to extract feature further, further carries out later stage correspondence
Identification.Because the extraction of feature is to be started according to the feature setting, therefore identification can be more targeted, thus effectively carrying
High discrimination.
Preferably, above-mentioned feature detection includes face, vehicle license plate characteristic detection.Wherein face is for personage's body in monitoring
The most effective information of part, and car plate is then to discriminate between the effective information of vehicle, therefore both features can obtain personage, vehicle respectively
Information so that follow-up identify.
Preferably, be identified by above-mentioned steps the feature of extraction is identified with Data Comparison in data base.Using
The mode comparing with data base can be the fastest guarantee identify the time comparing and accuracy.
As an embodiment, when detecting moving target in addition it is also necessary to judge whether target number is more than 1 in step E, that is, only
Have when moving target number is not 0, ball machine just can be responsive to target area and capture, and otherwise ball machine is motionless, at this, differentiate it is to need
Want, during target number more than one, can be sorted by target frame size, target frame centre distance ball machine position etc., grab one by one
Clap.
Embodiment one:
The present embodiment method as shown in figure 3,
A), in video camera installation process, panoramic mosaic video camera adopts horizontal connecting method, makes complete in installation process
Overlap between each picture of scape splicing video camera region, thus attempting splicing, if splicing is unsuccessful, adjusts panoramic mosaic
Each lens direction of video camera, increases lap until can achieve panorama splicing.The panorama obtaining such as Fig. 2 institute
Show, by multiple strip panoramas that individually picture splicing in the horizontal direction is constituted.
B), adjustment panoramic mosaic video camera the visual field, so that the visual field ball machine in panoramic mosaic picture can be covered.
C), camera calibration, set up the mapping relations between ball machine and the position of panoramic mosaic video camera;
D), judge to demarcate whether success, be then continuation step, otherwise adjust ball machine, make picture and the panoramic mosaic of ball machine
Video camera do not splice before each raw frames center alignment to basically identical;
Consistent adjustment herein can be made by selecting certain object of center not splicing single-image of panoramic mosaic video camera
For object of reference(One tree, a static car etc.), then adjust ball machine, make this object of reference also at ball machine picture center.
E), in panoramic mosaic picture, area-of-interest is set, and add moving object detection etc. in the region of interest
Video analysis function;
F), when area-of-interest analyzes moving target, corresponding ginseng is calculated according to tracking box position, size etc.
Number, ball machine response adjusts to moving target captures and follows the tracks of a period of time;
G), to ball machine capture preserve picture can by face, vehicle detection, extract face, car plate etc., Ran Houtong
Cross technology of identification and face, the contrast of car plate storehouse, identify suspicious object.
Embodiment two:
The present embodiment method as shown in figure 5,
A), in video camera installation process, panoramic mosaic video camera adopts vertical connecting method, makes complete in installation process
Overlap between each picture of scape splicing video camera region, thus attempting splicing, if splicing is unsuccessful, adjusts panoramic mosaic
Each lens direction of video camera, increases lap until can achieve panorama splicing.The panorama obtaining such as Fig. 4 institute
Show, by multiple strip panoramas that individually picture splicing vertically is constituted.
B), adjustment panoramic mosaic video camera the visual field, so that the most of visual field ball machine in panoramic mosaic picture can be covered.
C), camera calibration, set up the mapping relations between ball machine and the position of panoramic mosaic video camera;
D), judge to demarcate whether success, be then continuation step, otherwise adjust ball machine, make the picture of ball machine and the picture of gunlock
Face center is basically identical;
E), in panoramic mosaic picture, area-of-interest is set, and add moving object detection etc. in the region of interest
Video analysis function;
F), when area-of-interest analyzes moving target, corresponding ginseng is calculated according to tracking box position, size etc.
Number, ball machine response adjusts to moving target captures and follows the tracks of a period of time;
G), to ball machine capture preserve picture can by face, vehicle detection, extract face, car plate etc., Ran Houtong
Cross technology of identification and face, the contrast of car plate storehouse, identify suspicious object.
Embodiment three:
The present embodiment method as shown in fig. 7,
A), in video camera installation process, panoramic mosaic video camera adopts horizontal vertical connecting method, makes in installation process
Overlap between each picture obtaining panoramic mosaic video camera region, thus attempting splicing, if splicing is unsuccessful, adjusts panorama
Each lens direction of splicing video camera, increases lap until can achieve panorama splicing.The panorama obtaining such as Fig. 6
Shown, by multiple large scene panoramas individually constituting by the picture splicing along level, vertical direction of grid distribution.
B), adjustment panoramic mosaic video camera the visual field, so that the most of visual field ball machine in panoramic mosaic picture can be covered.
C), camera calibration, set up the mapping relations between ball machine and the position of panoramic mosaic video camera;
D), judge to demarcate whether success, be then continuation step, otherwise adjust ball machine, make the picture of ball machine and the picture of gunlock
Face center is basically identical;
E), in panoramic mosaic picture, area-of-interest is set, and add moving object detection etc. in the region of interest
Video analysis function;
F), when area-of-interest analyzes moving target, corresponding ginseng is calculated according to tracking box position, size etc.
Number, ball machine response adjusts to moving target captures and follows the tracks of a period of time;
G), to ball machine capture preserve picture can by face, vehicle detection, extract face, car plate etc., Ran Houtong
Cross technology of identification and face, the contrast of car plate storehouse, identify suspicious object.
The invention still further relates to a kind of panoramic mosaic linkage based on moving object detection, it includes:
Splicing setup module, in control point setting ball machine and panoramic mosaic video camera, panoramic mosaic in setting up procedure
Video camera adopts horizontal/vertical/horizontal vertical connecting method, realizes panorama splicing by adjusting each lens direction, then
Go to adjusting module;
Adjusting module, for adjusting the visual field of panoramic mosaic video camera, is allowed to the visual field in panoramic mosaic picture and covers setting
The ball machine visual field, then go to demarcating module;
Demarcating module, for demarcating ball machine and panoramic mosaic video camera, sets up the position of ball machine and panoramic mosaic video camera
Mapping relations, then go to setup module;
Setup module, is used for, after setting area-of-interest in panoramic mosaic picture, this area-of-interest being moved
Target detection, then goes to motion detection block;
Motion detection block, for when moving target is detected, corresponding moving target labelling tracking box, according to tracking box ginseng
The adjustment of numerical control marble forming machine, focusing shoot tracking target image.
As an embodiment, above-mentioned in, described demarcating module specifically includes,
Broadcast unit, does not splice original video and ball machine video for play panoramic mosaic video camera, then goes to choosing
Select original video unit;
Select original video unit, select one of panoramic mosaic video camera not splice original video for unduplicated and
After go to characteristic point select unit;
Characteristic point select unit, for not splicing unduplicated selection one in original video by input equipment select
Individual characteristic point, the coordinate of record this feature point goes to grouped record unit;
Grouped record unit, for marking the center of picture in ball machine video, makes ball machine picture center fall by manipulation
Select in the selected characteristic point of original video unit, by ball machine level inclination value now, ball machine vertical dip angle value and ball machine
The coordinate grouped record of scaling multiplier value character pair point, goes to repetitives;
Repetitives, then go to demarcation list at least 6 times for repeat step characteristic point select unit, demarcation recording unit
Unit;
Demarcate unit, for the ball machine level inclination value in every group of record, ball machine vertical dip angle value and coupling point coordinates
Quadratic surface coefficient is gone out using least square fitting, using this quadratic surface coefficient as calibration coefficient together with select original video
In the identifier of splicing original video selecting in unit and grouped record unit, the identifier correspondence of the ball machine of manipulation has preserved
Become mapping relations to set up, then go to panorama repetitives;
Panorama repetitives, select original video unit, characteristic point select unit, grouped record list for repeat step
Unit, repetitives and demarcation unit, until all mapping relations not splicing original video completing panoramic mosaic video camera are built
Vertical.
As an embodiment, above-mentioned in, described demarcating module specifically includes,
Manipulation unit, for manipulation make ball machine picture and panoramic mosaic video camera respectively not splice original video center consistent
Then go to feature extraction unit;
Feature extraction unit, for extracting the two field picture respectively not splicing original video, to this image zooming-out sift/
Surf/harris characteristic point, then characteristic matching, coupling then go to calculating perspective transformation matrix using method of least square
Margin estimation unit;
Margin estimation unit, for estimating its 4 edge angle points by perspective transformation matrix to respectively not splicing original video
Ball machine level inclination value and ball machine vertical dip angle value then go to ball machine bit cell;
Ball machine bit cell, does not splice original video for unduplicated select panoramic mosaic video camera one, by it
Changes in coordinates scope is divided into many parts, controls ball machine to arrive according to every part of ball machine level inclination value and ball machine vertical dip angle value respectively
Reach relevant position and then go to matching unit;
Matching unit, often reaches a position for ball machine, to now do not splice original video and ball machine picture is carried out
Feature extraction is simultaneously mated, if match point meets threshold value, preserves corresponding coordinate vertical with ball machine level inclination value now and ball machine
Inclination value then goes to demarcation unit;
Demarcate unit, for repeating ball machine bit cell, matching unit, until complete panoramic mosaic video camera all not
The mapping relations of splicing original video are set up.
As an embodiment, above-mentioned in, described demarcation unit goes to manual correction unit, and described manual correction unit is used for
Check that the coupling of automatic Calibration is right, if there is error, controlling ball manually adjustment ball machine with reference to ball machine bit cell, matching unit, obtaining
Take at least 6 match points, the ball machine level inclination value of the coordinate of match point and ball machine, ball machine vertical dip angle value and ball machine are scaled
Multiplier value corresponds to grouped record, adopts method of least square to the ball machine level inclination value in every group of record, ball machine vertical dip angle value
Simulate quadratic surface coefficient, using this quadratic surface coefficient as calibration coefficient together with select splicing original video identifier
And the corresponding preservation of identifier of the ball machine of manipulation completes mapping relations and sets up.
As an embodiment, above-mentioned in, the linkage of described motion detection block specifically has following units to realize,
Tracking box central point calculation unit, for according to tracking box information, the center point coordinate calculating tracking box then turns
To panoramic field positioning unit;
Panoramic field positioning unit, affiliated in panoramic mosaic picture for it is positioned according to tracking box center point coordinate value
Region then go to region decision unit;
Region decision unit, for judging the overlapping region whether above-mentioned zone is located in panoramic mosaic video camera, otherwise
The information obtaining region then executes coordinate calculating unit, is, coordinate calculating unit is compared in execution;
Coordinate calculating unit, is located for being calculated tracking box central point according to the splicing inverse transformation of the corresponding sequence number in region
Do not splice the coordinate position in original video, then go to ball machine linkage unit;
Relatively coordinate calculating unit, does not splice original video for comparing tracking box center point coordinate value with overlapping two
Central point distance, tracking box central point is calculated according to the splicing inverse transformation of the corresponding sequence number in closest region and is located
Do not splice the coordinate position in original video, then go to ball machine linkage unit;
Ball machine linkage unit, does not splice the mapping relations of original video for being located according to tracking box central point, to follow the tracks of
The coordinate position of frame central point obtains corresponding ball machine level inclination value and ball machine vertical dip angle value then goes to ball machine and executes list
Unit;
Ball machine performance element, drives ball machine adjustment according to ball machine level inclination value and ball machine vertical dip angle value.
As an embodiment, above-mentioned demarcating module goes to locating module by judge module, and described judge module is used for sentencing
Break and whether demarcate successfully, be to go to setup module, otherwise the picture of adjustment ball machine is allowed to the picture with panoramic mosaic video camera
Demarcating module is gone back to after center alignment.
As an embodiment, above-mentioned motion detection block goes to identification module, identification module after shooting tracking target image
For ball machine image being carried out with feature detection and extracting feature and then identify.
As an embodiment, in above-mentioned motion detection block, tracking box parameter includes tracking box size, tracking box position letter
Breath;The feature of extraction is identified by being identified by of described identification module with Data Comparison in data base;Special in identification module
Levy detection and include face, vehicle license plate characteristic detection.
It should be noted that as aforementioned 3 kinds of connecting methods such as Fig. 2, Fig. 4, Fig. 6, commonly horizontal connecting method at present,
It is mainly used in monitoring highway, square, warehouse etc., vertical connecting method can be used for the monitoring scenes such as building, bridge, works as monitoring
In local time that scene is that airport etc. is king-sized, individually horizontal and vertical connecting method all can not meet demand, now will
It is considered as horizontal vertical connecting method.
The foregoing is only embodiments of the invention, not thereby limit the present invention the scope of the claims, every using this
Equivalent structure or equivalent flow conversion that bright description and accompanying drawing content are made, or directly or indirectly it is used in the skills of other correlations
Art field, is included within the scope of the present invention.
Claims (6)
1. a kind of panoramic mosaic interlock method based on moving object detection it is characterised in that:It includes step,
A), in control point setting ball machine and panoramic mosaic video camera, in setting up procedure, panoramic mosaic video camera adopts level/vertical
Directly/horizontal vertical connecting method, realizes panorama splicing by adjusting each lens direction;
B), adjust the visual field of panoramic mosaic video camera, be allowed to the ball machine visual field that the visual field in panoramic mosaic picture covers setting;
C), demarcate ball machine and panoramic mosaic video camera, set up the position mapping relations of ball machine and panoramic mosaic video camera;
D), after setting area-of-interest in panoramic mosaic picture, moving object detection is carried out to this area-of-interest;
E), when moving target is detected, corresponding moving target labelling tracking box, according to tracking box state modulator linkage ball machine tune
Whole, focusing shoots tracking target image;
Described step C specifically includes,
C1), manipulation make ball machine picture and panoramic mosaic video camera respectively not splice original video center consistent;
C2), extract the two field picture respectively not splicing original video, to this image zooming-out s ift/surf/harris characteristic point, so
Characteristic matching, coupling calculate perspective transformation matrix to using method of least square afterwards;
C3), pass through perspective transformation matrix to respectively do not splice original video estimate its 4 edge angle points ball machine level inclination value and
Ball machine vertical dip angle value;
C4), unduplicated select panoramic mosaic video camera one does not splice original video, and its changes in coordinates scope is divided into
Many parts, ball machine is controlled to reach relevant position according to every part of ball machine level inclination value and ball machine vertical dip angle value respectively;
C5), ball machine often reaches a position, to now do not splice original video and ball machine picture carry out feature extraction and
Join, if match point meets threshold value, preserve corresponding coordinate and ball machine ball machine level inclination value now and ball machine vertical dip angle value;
C6), repeat step C4-C5, until all mapping relations not splicing original video completing panoramic mosaic video camera are built
Vertical.
2. the panoramic mosaic interlock method based on moving object detection as claimed in claim 1 it is characterised in that:Described step
The step also including manual correction after C5, including:
Check that the coupling of automatic Calibration is right, if there is error, controlling ball manually adjustment ball machine with reference to step C4, C5, obtaining at least 6
Individual match point, the ball machine level inclination value of the coordinate of match point and ball machine, ball machine vertical dip angle value and ball machine are scaled multiplier value
Corresponding grouped record, is gone out using least square fitting to the ball machine level inclination value in every group of record, ball machine vertical dip angle value
Quadratic surface coefficient, using this quadratic surface coefficient as calibration coefficient together with the splicing identifier of original video selecting and manipulation
The corresponding preservation of identifier of ball machine complete mapping relations and set up.
3. the panoramic mosaic interlock method based on moving object detection as claimed in claim 1 it is characterised in that:Described step
Linkage in E specifically includes step,
E1), according to tracking box information, calculate the center point coordinate of tracking box;
E2), its affiliated region in panoramic mosaic picture is positioned according to tracking box center point coordinate value;
E3), judge the overlapping region whether above-mentioned zone is located in panoramic mosaic video camera, otherwise obtain the information in region then
Execution step E4, is then execution step E5;
E4), it is calculated tracking box central point and is located according to the splicing inverse transformation of the corresponding sequence number in region and do not splice in original video
Coordinate position;
E5), compare the distance of the central point that tracking box center point coordinate value does not splice original video with overlapping two, according to away from
It is calculated the seat not spliced in original video that tracking box central point is located from the splicing inverse transformation of the corresponding sequence number in nearest region
Cursor position;
E6), it is located according to tracking box central point and does not splice the mapping relations of original video, with the coordinate position of tracking box central point
Obtain corresponding ball machine level inclination value and ball machine vertical dip angle value;
E7), ball machine adjustment is driven according to ball machine level inclination value and ball machine vertical dip angle value.
4. a kind of panoramic mosaic linkage based on moving object detection it is characterised in that:It includes,
Splicing setup module, in control point setting ball machine and panoramic mosaic video camera, panoramic mosaic shooting in setting up procedure
Machine adopts horizontal/vertical/horizontal vertical connecting method, realizes panorama splicing by adjusting each lens direction, then goes to
Adjusting module;
Adjusting module, for adjusting the visual field of panoramic mosaic video camera, is allowed to the ball that the visual field in panoramic mosaic picture covers setting
The machine visual field, then goes to demarcating module;
Demarcating module, for demarcating ball machine and panoramic mosaic video camera, sets up the position mapping of ball machine and panoramic mosaic video camera
Relation, then goes to setup module;
Setup module, is used for, after setting area-of-interest in panoramic mosaic picture, carrying out moving target to this area-of-interest
Detection, then goes to motion detection block;
Motion detection block, for when moving target is detected, corresponding moving target labelling tracking box, according to tracking box parameter control
Marble forming machine adjustment, focusing shoot tracking target image;
Described demarcating module specifically includes,
Manipulation unit, for manipulation make ball machine picture and panoramic mosaic video camera respectively not splice original video center consistent then
Go to feature extraction unit;
Feature extraction unit, for extracting the two field picture respectively not splicing original video, to this image zooming-out sift/surf/
Harris characteristic point, then characteristic matching, coupling then go to edge to calculating perspective transformation matrix using method of least square
Estimation unit;
Margin estimation unit, for by perspective transformation matrix to respectively not splicing the ball that original video estimates its 4 edge angle points
Machine level inclination value and ball machine vertical dip angle value then go to ball machine bit cell;
Ball machine bit cell, does not splice original video for unduplicated select panoramic mosaic video camera one, by its coordinate
Excursion is divided into many parts, controls ball machine to reach phase according to every part of ball machine level inclination value and ball machine vertical dip angle value respectively
Position is answered then to go to matching unit;
Matching unit, often reaches a position for ball machine, to now do not splice original video and ball machine picture carries out feature
Extract and mate, if match point meets threshold value, preserve corresponding coordinate and ball machine ball machine level inclination value now and ball machine is vertical
Inclination value then goes to demarcation unit;
Demarcate unit, for repeating ball machine bit cell, matching unit, until complete all of panoramic mosaic video camera not splicing
The mapping relations of original video are set up.
5. the panoramic mosaic linkage based on moving object detection as claimed in claim 4 it is characterised in that:Described demarcation
Unit goes to manual correction unit, and described manual correction unit is used for checking that the coupling of automatic Calibration is right, if there is error, joins
Examine ball machine bit cell, matching unit to control ball manually adjustment ball machine, obtain at least 6 match points, by the coordinate of match point and ball
The ball machine level inclination value of machine, ball machine vertical dip angle value and ball machine scaling multiplier value correspond to grouped record, in every group of record
Ball machine level inclination value, ball machine vertical dip angle value go out quadratic surface coefficient using least square fitting, by this quadratic surface system
Number completes together with the corresponding preservation of identifier of the splicing identifier of original video selecting and the ball machine of manipulation as calibration coefficient
Mapping relations are set up.
6. the panoramic mosaic linkage based on moving object detection as claimed in claim 4 it is characterised in that:Described motion
The linkage of detection module specifically has following units to realize,
Tracking box central point calculation unit, for according to tracking box information, the center point coordinate calculating tracking box then goes to entirely
Scene area positioning unit;
Panoramic field positioning unit, for positioning its affiliated area in panoramic mosaic picture according to tracking box center point coordinate value
Domain then goes to region decision unit;
Region decision unit, for judging the overlapping region whether above-mentioned zone is located in panoramic mosaic video camera, otherwise obtains
The information in region then executes coordinate calculating unit, is, coordinate calculating unit is compared in execution;
Coordinate calculating unit, does not spell for being calculated tracking box central point place according to the splicing inverse transformation of the corresponding sequence number in region
Connect the coordinate position in original video, then go to ball machine linkage unit;
Relatively coordinate calculating unit, is not spliced in original video with overlapping two for comparing tracking box center point coordinate value
The distance of heart point, is calculated tracking box central point place according to the splicing inverse transformation of the corresponding sequence number in closest region and does not spell
Connect the coordinate position in original video, then go to ball machine linkage unit;
Ball machine linkage unit, does not splice the mapping relations of original video for being located according to tracking box central point, with tracking box
The coordinate position of heart point obtains corresponding ball machine level inclination value and ball machine vertical dip angle value then goes to ball machine performance element;
Ball machine performance element, drives ball machine adjustment according to ball machine level inclination value and ball machine vertical dip angle value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410008471.7A CN103716594B (en) | 2014-01-08 | 2014-01-08 | Panorama splicing linkage method and device based on moving target detecting |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410008471.7A CN103716594B (en) | 2014-01-08 | 2014-01-08 | Panorama splicing linkage method and device based on moving target detecting |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103716594A CN103716594A (en) | 2014-04-09 |
CN103716594B true CN103716594B (en) | 2017-02-22 |
Family
ID=50409119
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410008471.7A Active CN103716594B (en) | 2014-01-08 | 2014-01-08 | Panorama splicing linkage method and device based on moving target detecting |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103716594B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108134923A (en) * | 2017-12-25 | 2018-06-08 | 天津天地伟业电子工业制造有限公司 | A kind of panorama intelligent integrated linkage |
Families Citing this family (50)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104167109B (en) * | 2014-07-17 | 2016-08-24 | 浙江宇视科技有限公司 | The detection method of vehicle location and detection device |
CN105354813B (en) * | 2014-08-18 | 2018-11-23 | 杭州海康威视数字技术股份有限公司 | Holder is driven to generate the method and device of stitching image |
CN104243849A (en) * | 2014-09-02 | 2014-12-24 | 苏州阔地网络科技有限公司 | Online classroom directed broadcasting control method and system |
CN104754231B (en) * | 2015-03-31 | 2019-02-19 | Oppo广东移动通信有限公司 | Shoot the method and device of personage's video |
CN104902236A (en) * | 2015-05-27 | 2015-09-09 | 深圳英飞拓科技股份有限公司 | Linkage control method and device for monitoring equipment |
CN105208323B (en) * | 2015-07-31 | 2018-11-27 | 深圳英飞拓科技股份有限公司 | A kind of panoramic mosaic picture monitoring method and device |
CN105208327A (en) * | 2015-08-31 | 2015-12-30 | 深圳市佳信捷技术股份有限公司 | Master/slave camera intelligent monitoring method and device |
CN106780613B (en) * | 2015-11-24 | 2021-03-16 | 杭州海康威视数字技术股份有限公司 | Visualization method and system for multi-camera calibration result |
CN105869330B (en) * | 2016-04-27 | 2018-03-23 | 深圳市万佳安物联科技股份有限公司 | A kind of safety monitoring device for office building anti-illegal-inbreak |
CN106210631A (en) * | 2016-07-16 | 2016-12-07 | 惠州学院 | The system for rapidly identifying of a kind of different angles video object and method |
CN108377407B (en) * | 2016-11-04 | 2021-03-12 | 株式会社理光 | Panoramic video processing method and device and electronic equipment |
CN108230348B (en) * | 2016-12-22 | 2022-01-21 | 杭州海康威视数字技术股份有限公司 | Target tracking method and device |
CN106791710B (en) * | 2017-02-10 | 2020-12-04 | 北京地平线信息技术有限公司 | Target detection method and device and electronic equipment |
CN106791715A (en) * | 2017-02-24 | 2017-05-31 | 深圳英飞拓科技股份有限公司 | Classification joint control intelligent control method and system |
CN109300084B (en) * | 2017-07-25 | 2023-07-04 | 杭州海康汽车技术有限公司 | Image stitching method and device, electronic equipment and storage medium |
CN107370994B (en) * | 2017-08-15 | 2018-11-02 | 深圳云天励飞技术有限公司 | Marine site overall view monitoring method, device, server and system |
CN107529021A (en) * | 2017-10-18 | 2017-12-29 | 北京伟开赛德科技发展有限公司 | The collection of tunnel type panoramic video, distribution, locating and tracking system and its method |
CN108038423B (en) * | 2017-11-22 | 2022-03-04 | 广东数相智能科技有限公司 | Automobile type identification method and device based on image identification |
CN109996055B (en) * | 2017-12-06 | 2022-08-12 | 杜比实验室特许公司 | Position zero time delay |
CN107995468A (en) * | 2017-12-19 | 2018-05-04 | 天津天地伟业信息系统集成有限公司 | The full-time monitoring device of panorama and monitoring method |
CN109982030B (en) * | 2017-12-27 | 2021-01-08 | 杭州海康威视数字技术股份有限公司 | Target object snapshot method and device, video monitoring equipment and storage medium |
CN108174152A (en) * | 2017-12-28 | 2018-06-15 | 深圳英飞拓科技股份有限公司 | A kind of target monitoring method and target monitor system |
CN108510497B (en) * | 2018-04-10 | 2022-04-26 | 四川和生视界医药技术开发有限公司 | Method and device for displaying focus information of retina image |
CN108437898B (en) * | 2018-05-12 | 2019-04-30 | 吉林大学 | A kind of environment sensing device based on eye-tracking |
CN111179305B (en) * | 2018-11-13 | 2023-11-14 | 晶睿通讯股份有限公司 | Object position estimation method and object position estimation device thereof |
CN109712193B (en) | 2018-12-04 | 2021-06-15 | 浙江大华技术股份有限公司 | Method and device for determining field angle of dome camera |
CN109613935A (en) * | 2018-12-05 | 2019-04-12 | 苏州博众机器人有限公司 | A kind of overall view monitoring method, system, equipment and storage medium |
CN109525816A (en) * | 2018-12-10 | 2019-03-26 | 北京智汇云舟科技有限公司 | A kind of more ball fusion linked systems of multiple gun based on three-dimensional geographic information and method |
CN111698455B (en) * | 2019-03-13 | 2022-03-11 | 华为技术有限公司 | Method, device and medium for controlling linkage of ball machine and gun machine |
CN111914592B (en) * | 2019-05-08 | 2023-09-05 | 杭州海康威视数字技术股份有限公司 | Multi-camera combined evidence obtaining method, device and system |
CN110351520B (en) * | 2019-07-11 | 2021-06-22 | 深圳市华方信息产业有限公司 | Moving body image tracking and acquiring method and system |
CN110445966B (en) * | 2019-08-09 | 2021-09-21 | 润博全景文旅科技有限公司 | Panoramic camera video shooting method and device, electronic equipment and storage medium |
CN112399033B (en) * | 2019-08-14 | 2022-04-15 | 杭州海康威视数字技术股份有限公司 | Camera assembly and monitoring camera |
CN110930437B (en) * | 2019-11-20 | 2023-06-23 | 深圳拙河科技有限公司 | Target tracking method and device |
CN111369684B (en) * | 2019-12-10 | 2023-09-01 | 杭州海康威视系统技术有限公司 | Target tracking method, device, equipment and storage medium |
CN111355884B (en) * | 2019-12-23 | 2021-11-02 | 深圳市鸿合创新信息技术有限责任公司 | Monitoring method, device, system, electronic equipment and storage medium |
CN111405196B (en) * | 2019-12-31 | 2022-08-02 | 智慧互通科技股份有限公司 | Vehicle management method and system based on video splicing |
CN111246172A (en) * | 2020-01-20 | 2020-06-05 | 深圳英飞拓科技股份有限公司 | Linkage image acquisition device and image acquisition method |
CN111581587A (en) * | 2020-04-29 | 2020-08-25 | 苏州新舟锐视信息技术科技有限公司 | Gun and ball linkage coordinate correlation method based on surface fitting |
CN111565299A (en) * | 2020-05-06 | 2020-08-21 | 苏州新舟锐视信息技术科技有限公司 | Method for capturing targets through linkage of multiple vehicle-mounted guns and one dome camera |
CN111491140B (en) * | 2020-05-15 | 2021-05-11 | 深圳金三立视频科技股份有限公司 | Power transmission cable video line patrol method and terminal |
CN111866437B (en) * | 2020-06-30 | 2022-04-08 | 厦门亿联网络技术股份有限公司 | Automatic switching method and device for double cameras of video conference, terminal equipment and storage medium |
CN112884803B (en) * | 2020-08-18 | 2023-11-24 | 眸芯科技(上海)有限公司 | Real-time intelligent monitoring target detection method and device based on DSP |
CN112967183A (en) * | 2021-01-30 | 2021-06-15 | 惠州华阳通用电子有限公司 | Panoramic picture splicing method and storage medium |
CN113038031B (en) * | 2021-03-02 | 2023-05-30 | 天津天地伟业智能安全防范科技有限公司 | Gun-ball linkage calibration method and device, electronic equipment and storage medium |
CN113301273B (en) * | 2021-05-24 | 2023-06-13 | 浙江大华技术股份有限公司 | Method and device for determining tracking mode, storage medium and electronic device |
CN113572960B (en) * | 2021-07-23 | 2023-11-14 | 武汉星环恒宇信息科技有限公司 | Video quick tag positioning method for water affair prevention and control |
CN114520879B (en) * | 2022-04-21 | 2022-09-06 | 浙江宇视科技有限公司 | Radar vision cooperative video acquisition method, device and system and storage medium |
CN116887057B (en) * | 2023-09-06 | 2023-11-14 | 北京立同新元科技有限公司 | Intelligent video monitoring system |
CN117255247B (en) * | 2023-11-14 | 2024-01-09 | 智广海联(天津)大数据技术有限公司 | Method and device for linkage of panoramic camera and detail dome camera |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7012637B1 (en) * | 2001-07-27 | 2006-03-14 | Be Here Corporation | Capture structure for alignment of multi-camera capture systems |
CN102902884A (en) * | 2012-09-24 | 2013-01-30 | 天津市亚安科技股份有限公司 | PTZ (pan/tilt/zoom) camera automatic positioning and angle calculating method |
CN103105858A (en) * | 2012-12-29 | 2013-05-15 | 上海安维尔信息科技有限公司 | Method capable of amplifying and tracking goal in master-slave mode between fixed camera and pan tilt zoom camera |
CN103198487A (en) * | 2013-04-15 | 2013-07-10 | 厦门博聪信息技术有限公司 | Automatic calibration method for video monitoring system |
CN103327238A (en) * | 2013-05-29 | 2013-09-25 | 山西绿色光电产业科学技术研究院(有限公司) | Space registering method based on panoramic all-in-one speed dome camera |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060028550A1 (en) * | 2004-08-06 | 2006-02-09 | Palmer Robert G Jr | Surveillance system and method |
CN101378501A (en) * | 2008-10-10 | 2009-03-04 | 天津市亚安科技电子有限公司 | Ganged monitoring apparatus for infrared panorama and control method thereof |
CN101707671A (en) * | 2009-11-30 | 2010-05-12 | 杭州普维光电技术有限公司 | Panoramic camera and PTZ camera combined control method and panoramic camera and PTZ camera combined control device |
CN101833231B (en) * | 2010-03-19 | 2012-05-23 | 中国人民解放军国防科学技术大学 | Device and method for adjusting view field of spliced panoramic camera |
CN102148965B (en) * | 2011-05-09 | 2014-01-15 | 厦门博聪信息技术有限公司 | Video monitoring system for multi-target tracking close-up shooting |
-
2014
- 2014-01-08 CN CN201410008471.7A patent/CN103716594B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7012637B1 (en) * | 2001-07-27 | 2006-03-14 | Be Here Corporation | Capture structure for alignment of multi-camera capture systems |
CN102902884A (en) * | 2012-09-24 | 2013-01-30 | 天津市亚安科技股份有限公司 | PTZ (pan/tilt/zoom) camera automatic positioning and angle calculating method |
CN103105858A (en) * | 2012-12-29 | 2013-05-15 | 上海安维尔信息科技有限公司 | Method capable of amplifying and tracking goal in master-slave mode between fixed camera and pan tilt zoom camera |
CN103198487A (en) * | 2013-04-15 | 2013-07-10 | 厦门博聪信息技术有限公司 | Automatic calibration method for video monitoring system |
CN103327238A (en) * | 2013-05-29 | 2013-09-25 | 山西绿色光电产业科学技术研究院(有限公司) | Space registering method based on panoramic all-in-one speed dome camera |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108134923A (en) * | 2017-12-25 | 2018-06-08 | 天津天地伟业电子工业制造有限公司 | A kind of panorama intelligent integrated linkage |
Also Published As
Publication number | Publication date |
---|---|
CN103716594A (en) | 2014-04-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103716594B (en) | Panorama splicing linkage method and device based on moving target detecting | |
CN103716595B (en) | Panoramic mosaic video camera and ball machine inter-linked controlling method and device | |
CN108419014B (en) | Method for capturing human face by linkage of panoramic camera and multiple capturing cameras | |
CN104754302B (en) | A kind of target detection tracking method based on rifle ball linked system | |
CN109151439B (en) | Automatic tracking shooting system and method based on vision | |
CN102291569B (en) | Double-camera automatic coordination multi-target eagle eye observation system and observation method thereof | |
CN105654471B (en) | Augmented reality AR system and method applied to internet video live streaming | |
CN110830756B (en) | Monitoring method and device | |
US8466913B2 (en) | User interface for accessing virtual viewpoint animations | |
US10515471B2 (en) | Apparatus and method for generating best-view image centered on object of interest in multiple camera images | |
CN103517041A (en) | Real-time full-view monitoring method and device based on multi-camera rotating scanning | |
CN105578044A (en) | Panoramic view adaptive teacher image analysis method | |
CN105898246A (en) | Smart home system | |
CN113838098B (en) | Intelligent tracking shooting system for long-distance high-speed moving target | |
CN110933316A (en) | Teacher tracking teaching system based on double-camera interactive mode | |
CN106022211B (en) | A method of utilizing gesture control multimedia equipment | |
CN111131713B (en) | Lens switching method, device, equipment and computer readable storage medium | |
CN112017210A (en) | Target object tracking method and device | |
CN107396059A (en) | A kind of panorama position adaptive student graphical analysis control method | |
CN109218587A (en) | A kind of image-pickup method and system based on binocular camera | |
CN111970434A (en) | Multi-camera multi-target athlete tracking shooting video generation system and method | |
KR101775153B1 (en) | Target Training System and Analysis Method | |
CN112766033B (en) | Method for estimating common attention targets of downlinks in scene based on multi-view camera | |
CN106998442A (en) | Intelligent domestic system | |
JP3393969B2 (en) | Method and apparatus for recognizing subject in automatic camera system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |