CN108093213A - A kind of target trajectory Fuzzy Data Fusion method based on video monitoring - Google Patents
A kind of target trajectory Fuzzy Data Fusion method based on video monitoring Download PDFInfo
- Publication number
- CN108093213A CN108093213A CN201711329881.1A CN201711329881A CN108093213A CN 108093213 A CN108093213 A CN 108093213A CN 201711329881 A CN201711329881 A CN 201711329881A CN 108093213 A CN108093213 A CN 108093213A
- Authority
- CN
- China
- Prior art keywords
- fuzzy
- pending
- degree
- belief
- target
- 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
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/02—Services making use of location information
- H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/02—Services making use of location information
- H04W4/025—Services making use of location information using location based information parameters
- H04W4/027—Services making use of location information using location based information parameters using movement velocity, acceleration information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
- H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination
Abstract
The present invention relates to a kind of target trajectory Fuzzy Data Fusion methods based on video monitoring, and applied in actual police work, police can not obtain GPS location data by exception object active upload, are only capable of obtaining location information by monitoring sensor network.And all kinds of case suspicion objects have subjective concealment planning beforehand, implement, the behavior caused by different phases such as conceal more, it is mostly fuzzy data to cause the data collected by monitoring network, the designed target trajectory Fuzzy Data Fusion method based on video monitoring of the invention, it can help to recover blurring trajectorie, help brought for police work, target tracking.
Description
Technical field
The present invention relates to a kind of target trajectory Fuzzy Data Fusion methods based on video monitoring, belong to target trajectory tracking
Technical field.
Background technology
In Trace Formation field, common method is mainly by the fusion based on prediction model and based on filtering algorithm
Fusion, Doug Cox et al. are directed to the low frequency of sensor network, low fidelity data, it is proposed that a kind of new to be used for markov
The reorganization scheme of chain, and probability trajectories are generated based on initial data using new scheme.P Zhang et al. propose a kind of knot
Close settling position and the blending algorithm of Kalman filtering.AO B Wang et al. propose a kind of distributed multiple target tracking simultaneously
Algorithm, more Bernoulli Jacob (MB) wave filter based on general covariance intersection (G-CI).However in actual identification process, due to a variety of
Complicated field condition and the subjective hidden and interference for being tracked target so that image identification is difficult to draw accurate conclusion.Typical case
The problem of Fusion, is that sensor data values are not precisely high, and the Data Fusion of Sensor studied herein
The true or false for being directed to data obscures.Under video network in the data fusion of track, certain important attribute feature has been paid close attention to
Under matching technique, have ignored opinion under more attribute multi-angle fuzzy decisions and integrate, is i.e. the ambiguity of target identification conclusion, general
After Trace Formation problem reduction identifies for image, the sequential line of point is determined, for a small number of not true present in sequential line
Fixed point is filtered.And in real process, in order to expand search face, feature can be weakened, considers more suspect vehicles.Such as
What recovers real trace from the uncertain information of times truthful data becomes actual difficult point.For in the location information of low precision
Certain recovery can be made to being really distributed by redundancy, but any useful information will not be provided in wrong data, only meeting
Further increase error.In addition, the track to be tracked is often abnormal.Wave filter or prediction model are either based on, all
It is easy to lose the true exception information in fusion.
The content of the invention
The technical problems to be solved by the invention are to provide a kind of brand-new Logic Structure Design of use, can effectively improve rail
The target trajectory Fuzzy Data Fusion method based on video monitoring of mark tracking accuracy.
In order to solve the above-mentioned technical problem the present invention uses following technical scheme：The present invention devises a kind of based on video prison
The target trajectory Fuzzy Data Fusion method of control, for obtaining track data of the target object by start position, including walking as follows
Suddenly：
Step A. is for each monitor camera device is captured centered on start position, in pre-determined distance radius
Each fuzzy object, realizes the swarm intelligence decision-making with target object, and screening obtains each pending corresponding to target object
Fuzzy object, and enter step B；
Coordinate positions and corresponding time point of the step B. according to each pending fuzzy object, obtain by starting point
Position starts, and the velocity information of each pending fuzzy object coordinate position is reached successively according to sequential, as each pending mould
The corresponding arrival rate information of object difference is pasted, subsequently into step C；
Step C. is according to each pending fuzzy object corresponding arrival rate information respectively, with reference to start position, according to
Temporal order direction respectively for each pending fuzzy object, obtains respective numbers processing pair before pending fuzzy object
As respectively to its degree of belief, and based on the accounting of default degree of belief division, update each pending corresponding to target object
Fuzzy object, subsequently into step D；
Step D. is according to each pending fuzzy object corresponding arrival rate information respectively, with reference to start position, according to
Sequential backward direction respectively for each pending fuzzy object, obtains respective numbers processing pair after pending fuzzy object
As respectively to its degree of belief, and based on the accounting of default degree of belief division, update each pending corresponding to target object
Fuzzy object, subsequently into step E；
Step E. for each pending fuzzy object corresponding to target object, based on by start position successively chronologically
By the speed reasonability of each pending fuzzy object, using default score and degree of deviation method, realize target object by
The track data fusion of start position.
As a preferred technical solution of the present invention：It in the step A, is directed to respectively centered on start position, is pre-
If each monitor camera device in the range of distance radius, obtain monitor camera device and correspond in the range of preset time duration, institute
Each fuzzy object corresponding L angle information for presetting P attributive character respectively is captured, that is, obtains above-mentioned all monitoring camera dresses
Captured each fuzzy object corresponding L angle information for presetting P attributive character respectively is put, hesitation fuzzy set is built, and adopts
With default score and degree of deviation method, the swarm intelligence decision-making with target object is realized, screening is obtained corresponding to target object
Each pending fuzzy object.
As a preferred technical solution of the present invention, the default score and degree of deviation method are as follows：
Under any attributive character, if the attributive character score is determined as negating to have right of veto by one vote by force；Simultaneously similar
It spends on mode, higher similarity will have the increased effect certainly of stronger superlinearity.As both ends are decisive by force, intermediate
Weak suggestiveness；
Definition：Optimal selection obtains score based on Henan fuzzy set theory, highest scoring and must itemize there is no approximate, together
When meet under each attribute that score is more than confidence threshold or score is approximate with highest item, meet under each attribute score more than confidence threshold
Value, and the degree of deviation is larger；
Definition：Substantially credible, score meets score under each attribute and is more than confidence threshold close to highest item.
As a preferred technical solution of the present invention：In the step C, by start position and pending fuzzy pair each
As respectively as each process object, the arrival rate information corresponding according to each pending fuzzy object difference, according to sequential
Order direction, respectively for each pending fuzzy object corresponding to process object, using gaussian density method, handled
K deals with objects respectively with respect to the degree of belief of the process object before object, wherein, if dealt with objects before the process object
Number is less than K, then each process object is respectively with respect to the degree of belief of the process object before obtaining the process object, i.e. pin respectively
To each pending fuzzy object, respective numbers process object is obtained before pending fuzzy object respectively to its degree of belief,
Then obtain wherein less than the quantity accounting for the degree of belief for presetting degree of belief threshold value, the low letter as each pending fuzzy object
Appoint degree ratio, will be less than the pending fuzzy object corresponding to the low degree of belief ratio of pre-determined lower limit ratio and delete, more fresh target
Each pending fuzzy object corresponding to object.
As a preferred technical solution of the present invention：In the step D, by start position and pending fuzzy pair each
As respectively as each process object, the arrival rate information corresponding according to each pending fuzzy object difference, according to sequential
Inverted order direction, respectively for each pending fuzzy object corresponding to process object, using gaussian density method, handled
K deals with objects respectively with respect to the degree of belief of the process object after object, wherein, if dealt with objects after the process object
Number is less than K, then obtains after the process object each process object respectively with respect to the degree of belief of the process object, i.e. pin respectively
To each pending fuzzy object, respective numbers process object is obtained after pending fuzzy object respectively to its degree of belief,
Then obtain wherein less than the quantity accounting for the degree of belief for presetting degree of belief threshold value, the low letter as each pending fuzzy object
Appoint degree ratio, will be less than the pending fuzzy object corresponding to the low degree of belief ratio of pre-determined lower limit ratio and delete, more fresh target
Each pending fuzzy object corresponding to object.
A kind of application system of target trajectory Fuzzy Data Fusion method based on video monitoring of the present invention, use with
Upper technical solution compared with prior art, has following technique effect：The designed target trajectory based on video monitoring of the invention
Fuzzy Data Fusion method, applied in actual police work, police can not be obtained by exception object active upload
GPS location data is only capable of obtaining location information by monitoring sensor network.And all kinds of case suspicion objects premediation, implement,
The behavior caused by different phases such as conceal has subjective concealment more, and it is mostly mould to cause the data collected by monitoring network
Data are pasted, the designed target trajectory Fuzzy Data Fusion method based on video monitoring of the present invention can help to recover fuzzy rail
Mark brings help for police work, target tracking.
Description of the drawings
Fig. 1 a are real trace distribution schematic diagrams in the embodiment of the present invention；
Fig. 1 b are blurring trajectorie point distribution schematic diagrams in the embodiment of the present invention；
Fig. 2 is that single-sensor obscures letter in target trajectory Fuzzy Data Fusion method of the present invention design based on video monitoring
Cease data collection schematic diagram；
Fig. 3 is that track reasoning extension is shown in target trajectory Fuzzy Data Fusion method of the present invention design based on video monitoring
It is intended to.
Specific embodiment
The specific embodiment of the present invention is described in further detail with reference to Figure of description.
The present invention devises a kind of target trajectory Fuzzy Data Fusion method based on video monitoring, first against single prison
Sensor combination hesitation fuzzy set theory is controlled, realizes the more attribute various visual angles fuzzy decision modelings of target.Next indicates fuzzy determine
The invertibity that reliability is transferred in plan, with reference to sequential track serial correlation and fuzzy value, sequentially to judge to correct again with backward
Secondary structure hesitation fuzzy matrix, realizes the removal of ambiguity error data and path decision.
The present invention devises a kind of target trajectory Fuzzy Data Fusion method based on video monitoring, for obtaining target pair
As the track data by start position, among practical application, as shown in Figure 1a, red track is target object A0Track,
Remaining track is and A0Similar false target track when extracting all approximate trajectories points, can obtain scatterplot point as shown in Figure 1 b
Cloth, wherein suspicion data are at double in truthful data.Melt using target trajectory fuzzy data of the present invention design based on video monitoring
Conjunction method, actually specifically comprises the following steps：
Step A. realizes the more attributes of single-sensor fuzzy object based on hesitation fuzzy set theory perfect Xu Zeshui et al.
Multi-angle hesitation fuzzy set is built, and is specially：As shown in Fig. 2, presetting P attributive character, and according to preset time duration scope, draw
L angle is determined, in, respectively for each monitoring camera dress centered on start position, in pre-determined distance radius
It puts, obtains monitor camera device and correspond in the range of preset time duration, capture each fuzzy object corresponding default P category respectively
Property feature L angle information, that is, obtain above-mentioned all monitor camera devices and capture each fuzzy object corresponding default P respectively
L angle information of a attributive character builds hesitation fuzzy set.
In above-mentioned design, specific distributed pins are to each camera, for default P attributive character, in order to gather category comprehensively
Property information, each attribute divides L angle, i.e. L multi-angle acquisition image obtains the more attribute multi-angle information of present Fuzzy object
Collection, gathers and calculates and the more attribute multi-angle information collection MIS_A of fuzzy object every timei(i=0,1,2...n) each target phase in
Like situation.Known X={ x1,x2,...xPRepresenting P attribute, X is a fixed collection, then the collection that hesitates is that each element of X reflects
It is mapped to the function of subset.Hesitation fuzzy set may be constructed such that according to the similar situation of L angle under each attribute：
Wherein hA(x)={ p1,p2...pk, be some numerical value in [0,1] set, be respectively xi(i=1,2...p) belong to
Property obtained under k angle with the similar situation of target A, that is, hesitate obscuring element.
The more attribute informations of multi-angle collected under the camera and other targets A can similarly be obtainedi(i=1,2...n) it
Between the hesitation fuzzy set that forms, it is as shown in table 1 below that the fuzzy set matrix that must hesitate finally is built to current camera information.
Table 1
For above-mentioned obtained hesitation fuzzy set, using default score and degree of deviation method, the group with target object is realized
Intelligent decision, screening obtains each pending fuzzy object corresponding to target object, and enters step B.
Each pending fuzzy objects of the step B. corresponding to for above-mentioned target object builds substantially credible fuzzy time series
Data set according to the coordinate position of each pending fuzzy object and corresponding time point, is obtained and opened by start position
Begin, the velocity information of each pending fuzzy object coordinate position is reached successively according to sequential, as each pending fuzzy object
Corresponding arrival rate information respectively, subsequently into step C.
The corresponding arrival rate information of above-mentioned each pending fuzzy object difference, can build velocity information matrix such as
Shown in the following table 2：
| no1 | no2 | … | nol-k | … | nol-1 | |
| no2 | v1,2 | |||||
| no3 | … | v2,3 | ||||
| … | … | |||||
| nok | v1,k | … | … | vl-k,2+l-k | ||
| nok+1 | v2,k+1 | … | … | … | ||
| … | … | … | … | |||
| nol | … | vl-k,l | … | vl-1,l |
Table 2
Wherein, no0Represent start position, no2、…、nolRepresent each pending fuzzy object of chronologically order sequence,
L represents the length of substantially credible fuzzy time series data set.
After velocity information matrix and normal speed ranges are obtained, since normal speed border is inaccurate, border
Above containing the fuzzy value that just can be only distinguished with velocity information is difficult to, even if still there may be mistakes in normal speed nucleus
Data.In order to retain fuzzy message to the end during intermediate transfer, normal speed ranges can retain certain tolerance.
And tolerance can cause part falseness tracing point to be judged as very, and really putting can not realize that reliability connects by the False Intersection Points directly closed on
It is continuous, therefore perform following steps C.
Step C. waits to locate using start position and each pending fuzzy object as each process object according to each
The corresponding arrival rate information of fuzzy object difference is managed, according to temporal order direction, respectively for each pending fuzzy object
Corresponding process object, using gaussian density method, K process object is respectively with respect to the processing pair before being dealt with objects
The degree of belief of elephant, wherein, it is each before obtaining the process object if the number dealt with objects before the process object is less than K
Process object with respect to the degree of belief of the process object, i.e., obtains pending mould for each pending fuzzy object respectively respectively
Respective numbers process object is respectively to its degree of belief before pasting object, and further builds speed reliability decision matrix, as follows
Shown in table 3.
| no1 | no2… | … | nok | … | nol | |
| 0 | rv0,1 | rv0,2 | … | rv0,k | rvl-k,l | |
| 1 | rv1,2 | rv1,k | rvl-k+1,l | |||
| … | … | … | … | |||
| k-1 | rvk-1,k | rvl-1,l |
Table 3
Then dealt with objects respectively for each pending fuzzy object for respective numbers before pending fuzzy object
Respectively to its degree of belief, obtain wherein less than the quantity accounting for the degree of belief for presetting degree of belief threshold value, as each pending
The low degree of belief ratio of fuzzy object, will be less than the pending fuzzy object corresponding to the low degree of belief ratio of pre-determined lower limit ratio
It deletes, each pending fuzzy object corresponding to target object is updated, subsequently into step D.
In above-mentioned steps, the setting for K parameter when K is equal to credible fuzzy time series data set length substantially, that is, represents
Reasonability between fuzzy object to be handled can be examined；The effect of K i.e. check currently pending fuzzy object with it is K first
Between association reasonability, to calculate the reliability situation of current point.It, can there are larger redundancy, premature pending moulds when K is excessive
It pastes object and the soundness verification of currently pending fuzzy object is acted on little, while when K is too small, can cause to be more than K when existing
During a continuous mistake, the pending fuzzy object verification in track will be due to the missing of real trace point, and leads to not by true
Tracing point continues to transfer downwards and verification.Therefore the value of K need to only be slightly larger than the continuous wrong data length of maximum of estimation.
Same method is performed using with above-mentioned steps C, performs following steps D.
Step D. waits to locate using start position and each pending fuzzy object as each process object according to each
The corresponding arrival rate information of fuzzy object difference is managed, according to sequential inverted order direction, respectively for each pending fuzzy object
Corresponding process object, using gaussian density method, K process object is respectively with respect to the processing pair after being dealt with objects
The degree of belief of elephant, wherein, if the number dealt with objects after the process object is less than K, obtain each after the process object
Process object with respect to the degree of belief of the process object, i.e., obtains pending mould for each pending fuzzy object respectively respectively
Respective numbers process object is then obtained wherein respectively to its degree of belief less than the letter of default degree of belief threshold value after pasting object
Appoint the quantity accounting of degree, as the low degree of belief ratio of each pending fuzzy object, will be less than the low letter of pre-determined lower limit ratio
The pending fuzzy object corresponding to degree ratio is appointed to delete, updates each pending fuzzy object corresponding to target object, so
After enter step E.
Operation based on above-mentioned steps C, step D, as shown in Figure 3, it is assumed that 1,2,3 pair 4 of reliability judgement respectively 0.7,
0.5th, 0.9, in addition 2,3,4 pair 5 of judgement is 0.9,0.9,0.5.It when propagated forward, is only capable of obtaining 2 pair 4 of negative, when adding
Upper backward when correcting, i.e., 4 pair 5 of judgement, when 5 synthesis reliability is higher than 4,4 pair 5 of negative will react on 4, and 4 pair 5
Negative effect will significantly reduce.
Wherein RV (:, i) and represent reliability fuzzy value of all associations to i-th of track sets point, RV (i, j) represents i-th
The confidence level fuzzy value that a tracing point makes j-th of tracing point.F function is reliability aggregation function, to i-th of track sets point
The integrated final result of one all certainty values of row.
Finally, it is as shown in table 4 below by correcting to obtain matrix by backward：
Table 4
Step E. for each pending fuzzy object corresponding to target object, based on by start position successively chronologically
By the speed reasonability of each pending fuzzy object, using default score and degree of deviation method, realize target object by
The track data fusion of start position.
In above-mentioned steps implementation procedure, the application of the default score and degree of deviation method, under any attributive character, if should
Attributive character score is determined as thering is right of veto by one vote；Simultaneously on similarity mode, higher similarity will have
There is stronger superlinearity is increased to act on certainly.As both ends are decisive by force, intermediate weak suggestiveness；Definition：Optimal selection is based on
Henan fuzzy set theory obtains score, highest scoring and must itemize there is no approximate, while meets score under each attribute and be more than confidence
Threshold value or score are approximate with highest item, meet score under each attribute and are more than confidence threshold, and the degree of deviation is larger；Definition：Substantially may be used
Letter, score meet score under each attribute and are more than confidence threshold close to highest item.
Target trajectory Fuzzy Data Fusion method based on video monitoring designed by above-mentioned technical proposal, applied to actual police
It works in work, police can not obtain GPS location data by exception object active upload, be only capable of passing through monitoring sensor
Network obtains location information.And all kinds of case suspicion objects are being planned beforehand, implement, are concealing etc. behavior caused by different phases more to have
There is subjective concealment, it is mostly fuzzy data to cause the data collected by monitoring network, and the present invention is designed to be supervised based on video
The target trajectory Fuzzy Data Fusion method of control, can help to recover blurring trajectorie, and side is brought for police work, target tracking
It helps.
Embodiments of the present invention are explained in detail above in conjunction with attached drawing, but the present invention is not limited to above-mentioned implementations
Mode, within the knowledge of a person skilled in the art, can also be on the premise of present inventive concept not be departed from
Make various variations.
Claims (5)
1. a kind of target trajectory Fuzzy Data Fusion method based on video monitoring, for obtaining target object by start position
Track data, which is characterized in that include the following steps：
Step A. be directed to centered on start position, in pre-determined distance radius each monitor camera device captured it is each
Fuzzy object, realizes the swarm intelligence decision-making with target object, and screening obtains each pending fuzzy corresponding to target object
Object, and enter step B；
Coordinate positions and corresponding time point of the step B. according to each pending fuzzy object, obtain by start position
Start, reach the velocity information of each pending fuzzy object coordinate position successively according to sequential, as each pending fuzzy pair
As arrival rate information corresponding respectively, subsequently into step C；
Arrival rate information corresponding according to each pending fuzzy object difference step C., with reference to start position, according to sequential
Order direction respectively for each pending fuzzy object, obtains respective numbers process object point before pending fuzzy object
The other degree of belief to it, and based on the accounting of default degree of belief division, update each pending fuzzy corresponding to target object
Object, subsequently into step D；
Arrival rate information corresponding according to each pending fuzzy object difference step D., with reference to start position, according to sequential
Backward direction respectively for each pending fuzzy object, obtains respective numbers process object point after pending fuzzy object
The other degree of belief to it, and based on the accounting of default degree of belief division, update each pending fuzzy corresponding to target object
Object, subsequently into step E；
Step E. is for each pending fuzzy object corresponding to target object, based on chronologically being passed through successively by start position
The speed reasonability of each pending fuzzy object using default score and degree of deviation method, realizes target object by starting point
The track data fusion of position.
2. a kind of target trajectory Fuzzy Data Fusion method based on video monitoring according to claim 1, it is characterised in that：
In the step A, respectively for each monitor camera device centered on start position, in pre-determined distance radius, obtain
It obtains monitor camera device to correspond in the range of preset time duration, capturing each fuzzy object, corresponding default P attribute is special respectively
L angle information of sign obtains above-mentioned all monitor camera devices and captures each fuzzy object corresponding default P category respectively
Property feature L angle information, build hesitation fuzzy set, and using default score and degree of deviation method, realize and target object
Swarm intelligence decision-making, screening obtain target object corresponding to each pending fuzzy object.
3. a kind of target trajectory Fuzzy Data Fusion side based on video monitoring according to any one in claim 1 to 2
Method, which is characterized in that the default score and degree of deviation method are as follows：
Under any attributive character, if the attributive character score is determined as negating to have right of veto by one vote by force；It is many in similarity simultaneously
On number, higher similarity will have the increased effect certainly of stronger superlinearity.As both ends are decisive by force, and centre is weak to be built
View property；
Definition：Optimal selection obtains score based on Henan fuzzy set theory, highest scoring and must itemize there is no approximate, full simultaneously
Score is more than confidence threshold under each attribute of foot or score is approximate with highest item, meets score under each attribute and is more than confidence threshold, and
The degree of deviation is larger；
Definition：Substantially credible, score meets score under each attribute and is more than confidence threshold close to highest item.
4. a kind of target trajectory Fuzzy Data Fusion method based on video monitoring according to claim 1, it is characterised in that：
In the step C, using start position and each pending fuzzy object as each process object, according to each pending
The corresponding arrival rate information of fuzzy object difference, according to temporal order direction, respectively for each pending fuzzy object institute
Corresponding process object, using gaussian density method, K process object is respectively with respect to the process object before being dealt with objects
Degree of belief, wherein, if the process object before deal with objects number be less than K, each place before obtaining the process object
Object is managed respectively with respect to the degree of belief of the process object, i.e., is obtained pending fuzzy for each pending fuzzy object respectively
Respective numbers process object is then obtained wherein respectively to its degree of belief less than the trust of default degree of belief threshold value before object
The quantity accounting of degree as the low degree of belief ratio of each pending fuzzy object, will be less than the low trust of pre-determined lower limit ratio
Pending fuzzy object corresponding to degree ratio is deleted, and updates each pending fuzzy object corresponding to target object.
5. a kind of target trajectory Fuzzy Data Fusion method based on video monitoring according to claim 1, it is characterised in that：
In the step D, using start position and each pending fuzzy object as each process object, according to each pending
The corresponding arrival rate information of fuzzy object difference, according to sequential inverted order direction, respectively for each pending fuzzy object institute
Corresponding process object, using gaussian density method, K process object is respectively with respect to the process object after being dealt with objects
Degree of belief, wherein, if the process object after deal with objects number be less than K, obtain each place after the process object
Object is managed respectively with respect to the degree of belief of the process object, i.e., is obtained pending fuzzy for each pending fuzzy object respectively
Respective numbers process object is then obtained wherein respectively to its degree of belief less than the trust of default degree of belief threshold value after object
The quantity accounting of degree as the low degree of belief ratio of each pending fuzzy object, will be less than the low trust of pre-determined lower limit ratio
Pending fuzzy object corresponding to degree ratio is deleted, and updates each pending fuzzy object corresponding to target object.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711329881.1A CN108093213B (en) | 2017-12-13 | 2017-12-13 | Target track fuzzy data fusion method based on video monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711329881.1A CN108093213B (en) | 2017-12-13 | 2017-12-13 | Target track fuzzy data fusion method based on video monitoring |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108093213A true CN108093213A (en) | 2018-05-29 |
| CN108093213B CN108093213B (en) | 2020-05-08 |
Family
ID=62174223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711329881.1A Active CN108093213B (en) | 2017-12-13 | 2017-12-13 | Target track fuzzy data fusion method based on video monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108093213B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109298413A (en) * | 2018-09-01 | 2019-02-01 | 哈尔滨工程大学 | A kind of method that specific aim solves the problems, such as the multiple target tracking data correlation under complex electromagnetic environment |
| CN110958575A (en) * | 2019-12-02 | 2020-04-03 | 重庆邮电大学 | Positioning method and system based on WiFi fusion prediction |
| CN112033454A (en) * | 2020-08-28 | 2020-12-04 | 杭州海康威视数字技术股份有限公司 | Track data processing method and device |
| CN112102429A (en) * | 2020-08-03 | 2020-12-18 | 深圳拓邦股份有限公司 | Lawn mower, and map building method and storage medium thereof |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050185826A1 (en) * | 2004-02-20 | 2005-08-25 | Bogdan Georgescu | Method and system for multi-modal component-based tracking of an object using robust information fusion |
| US20090110236A1 (en) * | 2007-10-29 | 2009-04-30 | Ching-Chun Huang | Method And System For Object Detection And Tracking |
| CN103729859A (en) * | 2013-12-30 | 2014-04-16 | 西北工业大学 | Probability nearest neighbor domain multi-target tracking method based on fuzzy clustering |
| CN106101640A (en) * | 2016-07-18 | 2016-11-09 | 北京邮电大学 | Adaptive video sensor fusion method and device |
-
2017
- 2017-12-13 CN CN201711329881.1A patent/CN108093213B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050185826A1 (en) * | 2004-02-20 | 2005-08-25 | Bogdan Georgescu | Method and system for multi-modal component-based tracking of an object using robust information fusion |
| US20090110236A1 (en) * | 2007-10-29 | 2009-04-30 | Ching-Chun Huang | Method And System For Object Detection And Tracking |
| CN103729859A (en) * | 2013-12-30 | 2014-04-16 | 西北工业大学 | Probability nearest neighbor domain multi-target tracking method based on fuzzy clustering |
| CN106101640A (en) * | 2016-07-18 | 2016-11-09 | 北京邮电大学 | Adaptive video sensor fusion method and device |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109298413A (en) * | 2018-09-01 | 2019-02-01 | 哈尔滨工程大学 | A kind of method that specific aim solves the problems, such as the multiple target tracking data correlation under complex electromagnetic environment |
| CN110958575A (en) * | 2019-12-02 | 2020-04-03 | 重庆邮电大学 | Positioning method and system based on WiFi fusion prediction |
| CN110958575B (en) * | 2019-12-02 | 2020-12-18 | 重庆邮电大学 | Positioning method and system based on WiFi fusion prediction |
| CN112102429A (en) * | 2020-08-03 | 2020-12-18 | 深圳拓邦股份有限公司 | Lawn mower, and map building method and storage medium thereof |
| CN112033454A (en) * | 2020-08-28 | 2020-12-04 | 杭州海康威视数字技术股份有限公司 | Track data processing method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108093213B (en) | 2020-05-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108093213A (en) | A kind of target trajectory Fuzzy Data Fusion method based on video monitoring | |
| CN107330396B (en) | Pedestrian re-identification method based on multi-attribute and multi-strategy fusion learning | |
| CN106295594B (en) | A kind of across camera method for tracking target and device based on dynamic route tree | |
| CN103400391B (en) | A kind of multi-object tracking method based on the random forest improved and device | |
| CN106846374A (en) | The track calculating method of vehicle under multi-cam scene | |
| CN106885574A (en) | A kind of monocular vision robot synchronous superposition method based on weight tracking strategy | |
| CN109102531A (en) | A kind of target trajectory method for tracing and device | |
| KR102132722B1 (en) | Tracking method and system multi-object in video | |
| CN104010168B (en) | A kind of non-overlapping visual field multiple-camera monitors network topology adaptation learning method | |
| CN103605971B (en) | Method and device for capturing face images | |
| CN106204646A (en) | Multiple mobile object tracking based on BP neutral net | |
| CN106846378B (en) | A kind of across the video camera object matching and tracking of the estimation of combination topology of spacetime | |
| CN107145862A (en) | A kind of multiple features matching multi-object tracking method based on Hough forest | |
| CN108376406A (en) | A kind of Dynamic Recurrent modeling and fusion tracking method for channel blockage differentiation | |
| CN112581503B (en) | Multi-target detection and tracking method under multiple visual angles | |
| Vysotska et al. | Relocalization under substantial appearance changes using hashing | |
| CN104992453A (en) | Target tracking method under complicated background based on extreme learning machine | |
| CN108764269A (en) | A kind of cross datasets pedestrian recognition methods again based on space-time restriction incremental learning | |
| CN107230267A (en) | Intelligence In Baogang Kindergarten based on face recognition algorithms is registered method | |
| CN103150552B (en) | A kind of driving training management method based on number of people counting | |
| CN110428449A (en) | Target detection tracking method, device, equipment and storage medium | |
| CN111931582A (en) | Image processing-based highway traffic incident detection method | |
| CN106934817A (en) | Based on multiattribute multi-object tracking method and device | |
| CN104408741A (en) | Video global motion estimation method with sequential consistency constraint | |
| Li et al. | Simultaneous tracking and verification via sequential posterior estimation |
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 |