CN108093213B - Target track fuzzy data fusion method based on video monitoring - Google Patents
Target track fuzzy data fusion method based on video monitoring Download PDFInfo
- Publication number
- CN108093213B CN108093213B CN201711329881.1A CN201711329881A CN108093213B CN 108093213 B CN108093213 B CN 108093213B CN 201711329881 A CN201711329881 A CN 201711329881A CN 108093213 B CN108093213 B CN 108093213B
- Authority
- CN
- China
- Prior art keywords
- fuzzy
- processed
- processing
- preset
- score
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 35
- 238000007500 overflow downdraw method Methods 0.000 title claims abstract description 18
- 238000012545 processing Methods 0.000 claims description 76
- 238000000034 method Methods 0.000 claims description 21
- 230000004927 fusion Effects 0.000 claims description 12
- 238000012216 screening Methods 0.000 claims description 4
- 230000008092 positive effect Effects 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 abstract description 3
- 239000011159 matrix material Substances 0.000 description 6
- 238000013461 design Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000012937 correction Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 230000010354 integration Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
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
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a video monitoring-based target track fuzzy data fusion method, which is applied to actual police work, police officers cannot actively upload abnormal objects to obtain GPS position data and can only obtain positioning information through a monitoring sensor network. And the behavior of various case suspicion objects generated in different stages such as prefusion, implementation, hiding and the like is mostly subjective hiding, so that data collected by a monitoring network are mostly fuzzy data.
Description
Technical Field
The invention relates to a video monitoring-based target track fuzzy data fusion method, and belongs to the technical field of target track tracking.
Background
In the field of trajectory fusion, a common method mainly comprises fusion based on a prediction model and fusion based on a filtering algorithm, and the Doug Cox et al provides a new recombination scheme for a markov chain aiming at low-frequency and low-fidelity data of a sensor network, and generates a probability trajectory based on original data by using the new scheme. P Zhang et al propose a fusion algorithm combining stable position and kalman filtering. Meanwhile, AO B Wang et al propose a distributed multi-target tracking algorithm, which is a generalized covariance intersection (G-CI) -based multi-Bernoulli (MB) filter. However, in the actual recognition process, due to various complex field conditions and subjective concealment and interference of the tracked target, accurate conclusions are difficult to draw in image recognition. The problem with typical multi-sensor data fusion is that the sensor data values are not accurate, whereas the sensor data fusion studied herein is directed to the ambiguity of the authenticity of the data. In track data fusion in a video network, a matching technology under certain important attribute characteristics is focused, opinion integration under multi-attribute multi-angle fuzzy decision is ignored, namely the ambiguity of a target identification conclusion, the track fusion problem is simplified into the problems of determining time sequence connecting lines of points after image identification, and filtering is carried out on a few uncertain points in the time sequence connecting lines. In practice, in order to enlarge the search area, the features are weakened, and more suspicious vehicles are considered. It is a practical difficulty how to recover a real track from uncertain information that is multiple of the real data. Some recovery can be made for the true distribution by redundancy in the position information with low precision, but any useful information is not provided in the error data, and only the error is further increased. Furthermore, the trajectory to be tracked is often anomalous. Either filter-based or predictive model-based, true anomaly information in the fusion is easily lost.
Disclosure of Invention
The invention aims to solve the technical problem of providing a target track fuzzy data fusion method based on video monitoring, which adopts a brand-new logic structure design and can effectively improve the track tracking precision.
The invention adopts the following technical scheme for solving the technical problems: the invention designs a video monitoring-based target track fuzzy data fusion method for obtaining track data of a target object from a starting point position, which comprises the following steps:
step A, aiming at each fuzzy object captured by each monitoring camera device within a preset distance radius range by taking a starting point position as a center, realizing group intelligent decision with a target object, screening to obtain each fuzzy object to be processed corresponding to the target object, and entering step B;
b, acquiring speed information which starts from the starting position and sequentially reaches the coordinate positions of the fuzzy objects to be processed according to the coordinate positions of the fuzzy objects to be processed and the corresponding time points, wherein the speed information is used as the arrival speed information corresponding to each fuzzy object to be processed, and then entering the step C;
step C, according to the arrival speed information corresponding to each fuzzy object to be processed, combining the starting point position, aiming at each fuzzy object to be processed respectively according to the time sequence direction, obtaining the respective trust degrees of the corresponding number of processing objects before the fuzzy object to be processed on the fuzzy object to be processed, updating each fuzzy object to be processed corresponding to the target object based on the preset trust degree division ratio, and then entering the step D;
d, according to the arrival speed information corresponding to each fuzzy object to be processed, combining the starting point position, and aiming at each fuzzy object to be processed respectively in the time sequence reverse direction, obtaining the respective credibility of the corresponding number of processing objects behind the fuzzy object to be processed, updating each fuzzy object to be processed corresponding to the target object based on the ratio of the preset credibility division, and entering the step E;
and E, aiming at each fuzzy object to be processed corresponding to the target object, based on the speed reasonableness of sequentially passing through each fuzzy object to be processed according to the time sequence from the starting point position, and realizing the track data fusion of the target object from the starting point position by adopting a preset score and deviation degree method.
As a preferred technical scheme of the invention: in the step a, L angle information of each captured fuzzy object corresponding to P preset attribute features in a preset time duration range corresponding to the monitoring camera device is obtained for each monitoring camera device within a preset distance radius range with the starting point position as a center, that is, L angle information of each captured fuzzy object corresponding to P preset attribute features is obtained for each fuzzy object captured by all the monitoring camera devices, a hesitation fuzzy set is constructed, a preset score and deviation method is adopted to realize group intelligent decision with a target object, and each fuzzy object to be processed corresponding to the target object is obtained through screening.
As a preferred embodiment of the present invention, the method of the preset score and the degree of deviation is as follows:
under any attribute characteristic, if the score of the attribute characteristic is judged to be strong negative, the method has a vote rejection power; meanwhile, on the similarity mode, the higher similarity has a stronger positive effect of super-linear increase, namely strong determinism at two ends and weak advisability in the middle;
obtaining a score based on a hesitation fuzzy set theory, and defining the best choice as follows: the score is highest, no approximate score item exists, and the score is larger than a confidence threshold value under the condition that all attributes are met; or the score is similar to the highest item, the score is greater than the confidence threshold value under the condition of meeting the requirements of all attributes, and the deviation degree is greater than the preset deviation threshold value;
obtaining a score based on the hesitation fuzzy set theory, and defining the basic credibility as follows: the score is close to the highest item, and the score is larger than the confidence threshold value under the condition that each attribute is met.
As a preferred technical scheme of the invention: in the step C, the starting point position and each fuzzy object to be processed are respectively used as each processing object, and according to the arrival speed information corresponding to each fuzzy object to be processed, the confidence level of each processing object before the processing object with respect to the processing object corresponding to each fuzzy object to be processed is obtained by adopting a gaussian density method according to the time sequence order direction, wherein if the number of the processing objects before the processing object is less than K, the confidence level of each processing object before the processing object with respect to the processing object is obtained, that is, the confidence level of each processing object before the processing object with respect to the processing object is obtained for each fuzzy object to be processed, and then the number ratio of the confidence levels lower than the preset confidence level threshold is obtained as the low confidence level ratio of each fuzzy object to be processed, and deleting the fuzzy objects to be processed corresponding to the low confidence ratio lower than the preset lower limit ratio, and updating each fuzzy object to be processed corresponding to the target object.
As a preferred technical scheme of the invention: in the step D, the starting point position and each fuzzy object to be processed are respectively used as each processing object, according to the arrival speed information corresponding to each fuzzy object to be processed, respectively, in the time sequence reverse direction, the gaussian density method is adopted for the processing object corresponding to each fuzzy object to be processed, to obtain the trust of K processing objects after the processing object with respect to the processing object, respectively, wherein, if the number of the processing objects after the processing object is less than K, the trust of each processing object after the processing object with respect to the processing object is obtained, that is, the trust of the corresponding number of the processing objects after the fuzzy object to be processed with respect to the processing object is obtained for each fuzzy object to be processed, respectively, then the number ratio of the trust which is lower than the preset trust threshold is obtained as the low trust ratio of each fuzzy object to be processed, and deleting the fuzzy objects to be processed corresponding to the low confidence ratio lower than the preset lower limit ratio, and updating each fuzzy object to be processed corresponding to the target object.
Compared with the prior art, the application system of the video monitoring-based target track fuzzy data fusion method has the following technical effects by adopting the technical scheme: the target track fuzzy data fusion method based on video monitoring is applied to actual police work, police officers cannot actively upload the target track fuzzy data through abnormal objects to obtain GPS position data, and positioning information can be obtained only through a monitoring sensor network. And the behavior of various case suspicion objects generated in different stages such as prefusion, implementation, hiding and the like is mostly subjective hiding, so that data collected by a monitoring network are mostly fuzzy data.
Drawings
FIG. 1a is a diagram illustrating a distribution of real tracks in an embodiment of the present invention;
FIG. 1b is a schematic diagram of the distribution of fuzzy track points in the embodiment of the present invention;
FIG. 2 is a schematic diagram of single-sensor fuzzy information data collection in a video monitoring-based target track fuzzy data fusion method according to the present invention;
FIG. 3 is a schematic diagram of track reasoning extension in the video monitoring-based target track fuzzy data fusion method.
Detailed Description
The following description will explain embodiments of the present invention in further detail with reference to the accompanying drawings.
The invention designs a target track fuzzy data fusion method based on video monitoring, which firstly aims at a single monitoring sensor and combines the hesitation fuzzy set theory to realize target multi-attribute multi-view fuzzy decision modeling. And secondly, indicating the reversibility of reliability transmission in fuzzy decision, and combining the correlation and fuzzy value of the time sequence track sequence to reconstruct a hesitation fuzzy matrix by sequential judgment and reverse order correction so as to realize fuzzy error data removal and path decision.
The invention designs a video monitoring-based target track fuzzy data fusion method for obtaining track data of a target object from a starting point, wherein in practical application, as shown in figure 1a, a red track is a target object A0The rest of the tracks are AND0Similar false target trajectories, when all approximate trajectory points are extracted, a scatter distribution can be obtained as shown in fig. 1b, where the suspect data is many times the true data. The invention relates to a video monitoring-based target track fuzzy data fusion method, which practically comprises the following steps:
step A, based on the optimized hesitation fuzzy set theory of Xuezui and the like, realizing the multi-attribute multi-angle hesitation fuzzy set construction of the fuzzy target of the single sensor, which specifically comprises the following steps: as shown in fig. 2, P attribute features are preset, L angles are defined according to a preset time duration range, in application, L angle information corresponding to the preset P attribute features of each captured fuzzy object is obtained in the preset time duration range corresponding to the monitoring camera device for each monitoring camera device within a preset distance radius range with a starting point position as a center, that is, L angle information corresponding to the preset P attribute features of each fuzzy object captured by all the monitoring camera devices is obtained, and a hesitation fuzzy set is constructed.
In the design, specific distribution aims at each camera, P attribute characteristics are preset, in order to comprehensively collect attribute information, each attribute is divided into L angles, namely L times of multi-angle collected images, a current fuzzy object multi-attribute multi-angle information set is obtained, and the multi-attribute multi-angle information set MIS _ A of the fuzzy object is collected and calculated every timeiEach object in (i ═ 0,1,2.. n) is similar. Known as X ═ X1,x2,...xPDenotes P attributes, X is a fixed set, and the hesitation set is a function of each element of X mapped to a subset. The similarity according to the L angles under each property can be constructed as a hesitation fuzzy set:
wherein h isA(x)＝{p1,p2...pkIs [0,1 ]]A set of some of the values, each xi(i ═ 1,2.. p) the attribute achieves a similar situation to that of target a at k angles, i.e. hesitant fuzzy elements.
In the same way, the multi-angle multi-attribute information and other targets A collected under the camera can be obtainediAnd (i ═ 1,2.. n), and finally constructing a hesitation fuzzy set matrix for the current camera information as shown in the following table 1.
TABLE 1
And aiming at the obtained hesitation fuzzy set, adopting a preset score and deviation degree method to realize group intelligent decision with the target object, screening to obtain each fuzzy object to be processed corresponding to the target object, and entering the step B.
And step B, aiming at each fuzzy object to be processed corresponding to the target object, constructing a basic credible fuzzy time sequence data set, acquiring speed information which starts from a starting point position and sequentially reaches the coordinate position of each fuzzy object to be processed according to the coordinate position of each fuzzy object to be processed and the corresponding time point, taking the speed information as the arrival speed information corresponding to each fuzzy object to be processed, and then entering the step C.
The arrival speed information corresponding to each fuzzy object to be processed may be configured as a speed information matrix shown in table 2 below:
| 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
Therein, no0Denotes the starting position, no2、…、nolRepresenting the respective fuzzy objects to be processed sorted in time sequence order, and l representing the length of the basic credible fuzzy time sequence data set.
After obtaining the speed information matrix and the normal speed range, because the normal speed boundary is not accurate, the boundary contains fuzzy values which are difficult to distinguish only by the speed information, and error data can still exist in the normal speed core area. In order to retain the fuzzy information to the end in the intermediate transfer process, the normal speed range is retained with a certain tolerance. And the tolerance can cause part of the false track points to be judged as true, and the true points can not realize reliability continuity by the directly adjacent false points, so the following step C is executed.
And step C, respectively taking the starting point position and each fuzzy object to be processed as each processing object, and respectively aiming at the processing object corresponding to each fuzzy object to be processed according to the arrival speed information corresponding to each fuzzy object to be processed respectively and in the time sequence direction, adopting a Gaussian density method to obtain the trust of K processing objects before the processing object relative to the processing object respectively, wherein if the number of the processing objects before the processing object is less than K, the trust of each processing object before the processing object relative to the processing object is obtained, namely, the trust of the corresponding number of the processing objects before the fuzzy object to be processed to the processing objects is obtained respectively aiming at each fuzzy object to be processed respectively, and further constructing a speed reliability decision matrix as shown in the following 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
And then, respectively aiming at each fuzzy object to be processed, respectively aiming at the trust degrees of the corresponding number of processing objects before the fuzzy object to be processed on the fuzzy object to be processed, obtaining the number ratio of the trust degrees which are lower than a preset trust degree threshold value, taking the number ratio as the low trust degree ratio of each fuzzy object to be processed, deleting the fuzzy object to be processed corresponding to the low trust degree ratio which is lower than a preset lower limit ratio, updating each fuzzy object to be processed corresponding to the target object, and then entering the step D.
In the above step, for the setting of the parameter K, when K is equal to the length of the basic trusted fuzzy time series data set, it means that the rationality among all fuzzy objects to be processed is checked; and the function of K is to check the relevance rationality between the current fuzzy object to be processed and the previous K items so as to calculate the reliability condition of the current point. When K is too large, great redundancy exists, the rationality verification effect of the premature fuzzy object to be processed on the current fuzzy object to be processed is not great, and when K is too small, the situation that when continuous errors larger than K exist can be caused, the track fuzzy object to be processed is verified, and the fact that the real track points cannot be continuously transmitted and verified downwards can be caused due to the fact that the real track points are lacked. Therefore, the value of K only needs to be slightly larger than the estimated maximum consecutive error data length.
The following step D is performed in the same manner as the step C described above.
Step D, respectively taking the starting point position and each fuzzy object to be processed as each processing object, respectively obtaining the trust of K processing objects behind the processing object relative to the processing object by adopting a Gaussian density method according to the arrival speed information corresponding to each fuzzy object to be processed respectively in the time sequence reverse direction and aiming at the processing object corresponding to each fuzzy object to be processed respectively, wherein if the number of the processing objects behind the processing object is less than K, the trust of each processing object behind the processing object relative to the processing object is obtained, namely, the trust of the corresponding number of the processing objects behind the fuzzy object to be processed respectively is obtained aiming at each fuzzy object to be processed respectively, then the number ratio of the trust which is lower than the preset trust threshold is obtained and is used as the low trust ratio of each fuzzy object to be processed, and E, deleting the fuzzy objects to be processed corresponding to the low confidence ratio lower than the preset lower limit ratio, updating each fuzzy object to be processed corresponding to the target object, and then entering the step E.
Based on the operations in step C and step D, as shown in fig. 3, it is assumed that the reliability determinations of 1,2, and 3 to 4 are 0.7, 0.5, and 0.9, respectively, and the determinations of 2, 3, and 4 to 5 are 0.9, and 0.5, respectively. When propagating forward, only 2 to 4 negatives can be obtained, and when adding the backward correction, i.e. 4 to 5 judgments, 4 to 5 negatives will react on 4 and 4 to 5 negatives will decrease significantly when the overall confidence of 5 is higher than 4.
Wherein RV (i) represents the reliability fuzzy value of the related item to the ith track sequence point, and RV (i, j) represents the reliability fuzzy value of the ith track point to the jth track point. And f, the function is a reliability integration function, and the final result of integration of all the reliability values of the ith track sequence point in a row is obtained.
Finally, the matrix obtained by reverse order correction is shown in the following table 4:
TABLE 4
And E, aiming at each fuzzy object to be processed corresponding to the target object, based on the speed reasonableness of sequentially passing through each fuzzy object to be processed according to the time sequence from the starting point position, and realizing the track data fusion of the target object from the starting point position by adopting a preset score and deviation degree method.
In the execution process of the steps, if the score of the attribute characteristic is judged to be strong negative under any attribute characteristic, the preset score and the deviation degree method are applied, and a vote rejection power is provided; meanwhile, on the similarity mode, the higher similarity has a stronger positive effect of super-linear increase, namely strong determinism at two ends and weak advisability in the middle; obtaining a score based on a hesitation fuzzy set theory, and defining the best choice as follows: the score is highest, no approximate score item exists, and the score is larger than a confidence threshold value under the condition that all attributes are met; or the score is similar to the highest item, the score is greater than the confidence threshold value under the condition of meeting the requirements of all attributes, and the deviation degree is greater than the preset deviation threshold value; obtaining a score based on the hesitation fuzzy set theory, and defining the basic credibility as follows: the score is close to the highest item, and the score is larger than the confidence threshold value under the condition that each attribute is met.
The target track fuzzy data fusion method based on video monitoring is applied to actual police work, police officers cannot actively upload the target track fuzzy data to obtain GPS position data through abnormal objects, and positioning information can only be obtained through a monitoring sensor network. And the behavior of various case suspicion objects generated in different stages such as prefusion, implementation, hiding and the like is mostly subjective hiding, so that data collected by a monitoring network are mostly fuzzy data.
The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention.
Claims (4)
1. A target track fuzzy data fusion method based on video monitoring is used for obtaining track data of a target object from a starting point position, and is characterized by comprising the following steps:
step A, respectively aiming at each monitoring camera device which takes a starting point position as a center and is within a preset distance radius range, acquiring L angle information of each captured fuzzy object corresponding to preset P attribute characteristics in a corresponding preset time duration range of the monitoring camera device, namely acquiring the L angle information of each captured fuzzy object corresponding to the preset P attribute characteristics of each fuzzy object captured by all the monitoring camera devices, constructing a hesitation fuzzy set, realizing group intelligent decision with a target object by adopting a preset score and deviation method, screening to acquire each fuzzy object to be processed corresponding to the target object, and entering the step B;
b, acquiring speed information which starts from the starting position and sequentially reaches the coordinate positions of the fuzzy objects to be processed according to the coordinate positions of the fuzzy objects to be processed and the corresponding time points, wherein the speed information is used as the arrival speed information corresponding to each fuzzy object to be processed, and then entering the step C;
step C, according to the arrival speed information corresponding to each fuzzy object to be processed, combining the starting point position, aiming at each fuzzy object to be processed respectively according to the time sequence direction, obtaining the respective trust degrees of the corresponding number of processing objects before the fuzzy object to be processed on the fuzzy object to be processed, updating each fuzzy object to be processed corresponding to the target object based on the preset trust degree division ratio, and then entering the step D;
d, according to the arrival speed information corresponding to each fuzzy object to be processed, combining the starting point position, and aiming at each fuzzy object to be processed respectively in the time sequence reverse direction, obtaining the respective credibility of the corresponding number of processing objects behind the fuzzy object to be processed, updating each fuzzy object to be processed corresponding to the target object based on the ratio of the preset credibility division, and entering the step E;
and E, aiming at each fuzzy object to be processed corresponding to the target object, based on the speed reasonableness of sequentially passing through each fuzzy object to be processed according to the time sequence from the starting point position, and realizing the track data fusion of the target object from the starting point position by adopting a preset score and deviation degree method.
2. The video monitoring-based target track fuzzy data fusion method according to claim 1, wherein the preset score and deviation degree method comprises the following steps:
under any attribute characteristic, if the score of the attribute characteristic is judged to be strong negative, the method has a vote rejection power; meanwhile, on the similarity mode, the higher similarity has a stronger positive effect of super-linear increase, namely strong determinism at two ends and weak advisability in the middle; obtaining a score based on a hesitation fuzzy set theory, and defining the best choice as follows: the score is highest, no approximate score item exists, and the score is larger than a confidence threshold value under the condition that all attributes are met; or the score is similar to the highest item, the score is greater than the confidence threshold value under the condition of meeting the requirements of all attributes, and the deviation degree is greater than the preset deviation threshold value;
obtaining a score based on the hesitation fuzzy set theory, and defining the basic credibility as follows: the score is close to the highest item, and the score is larger than the confidence threshold value under the condition that each attribute is met.
3. The video monitoring-based target track fuzzy data fusion method according to claim 1, characterized in that: in the step C, the starting point position and each fuzzy object to be processed are respectively used as each processing object, and according to the arrival speed information corresponding to each fuzzy object to be processed, the confidence level of each processing object before the processing object with respect to the processing object corresponding to each fuzzy object to be processed is obtained by adopting a gaussian density method according to the time sequence order direction, wherein if the number of the processing objects before the processing object is less than K, the confidence level of each processing object before the processing object with respect to the processing object is obtained, that is, the confidence level of each processing object before the processing object with respect to the processing object is obtained for each fuzzy object to be processed, and then the number ratio of the confidence levels lower than the preset confidence level threshold is obtained as the low confidence level ratio of each fuzzy object to be processed, and deleting the fuzzy objects to be processed corresponding to the low confidence ratio lower than the preset lower limit ratio, and updating each fuzzy object to be processed corresponding to the target object.
4. The video monitoring-based target track fuzzy data fusion method according to claim 1, characterized in that: in the step D, the starting point position and each fuzzy object to be processed are respectively used as each processing object, according to the arrival speed information corresponding to each fuzzy object to be processed, respectively, in the time sequence reverse direction, the gaussian density method is adopted for the processing object corresponding to each fuzzy object to be processed, to obtain the trust of K processing objects after the processing object with respect to the processing object, respectively, wherein, if the number of the processing objects after the processing object is less than K, the trust of each processing object after the processing object with respect to the processing object is obtained, that is, the trust of the corresponding number of the processing objects after the fuzzy object to be processed with respect to the processing object is obtained for each fuzzy object to be processed, respectively, then the number ratio of the trust which is lower than the preset trust threshold is obtained as the low trust ratio of each fuzzy object to be processed, and deleting the fuzzy objects to be processed corresponding to the low confidence ratio lower than the preset lower limit ratio, and updating each fuzzy object to be processed corresponding to the 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 CN108093213A (en) | 2018-05-29 |
| CN108093213B true 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) |
Families Citing this family (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 |
| CN110958575B (en) * | 2019-12-02 | 2020-12-18 | 重庆邮电大学 | Positioning method and system based on WiFi fusion prediction |
| CN112102429B (en) * | 2020-08-03 | 2024-07-26 | 深圳拓邦股份有限公司 | Graph building method of mower, storage medium and mower |
| CN112033454B (en) * | 2020-08-28 | 2022-07-12 | 杭州海康威视数字技术股份有限公司 | Track data processing method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7072494B2 (en) * | 2004-02-20 | 2006-07-04 | Siemens Corporate Research, Inc. | Method and system for multi-modal component-based tracking of an object using robust information fusion |
| TWI348659B (en) * | 2007-10-29 | 2011-09-11 | Ind Tech Res Inst | Method and system for object detection and tracking |
-
2017
- 2017-12-13 CN CN201711329881.1A patent/CN108093213B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108093213A (en) | 2018-05-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108093213B (en) | Target track fuzzy data fusion method based on video monitoring | |
| Tsintotas et al. | Assigning visual words to places for loop closure detection | |
| CN109459045B (en) | Improved interactive voting matching method for low-frequency GPS track | |
| Tartakovsky et al. | Change-point detection in multichannel and distributed systems with applications | |
| CN111161315B (en) | Multi-target tracking method and system based on graph neural network | |
| CN101567087A (en) | Method for detecting and tracking small and weak target of infrared sequence image under complex sky background | |
| WO2012009947A1 (en) | Device and method of signature-driven multi-target tracking | |
| WO2023197232A1 (en) | Target tracking method and apparatus, electronic device, and computer readable medium | |
| CN104699755A (en) | Intelligent multi-target comprehensive identification method based on data mining | |
| CN115830075A (en) | Hierarchical association matching method for pedestrian multi-target tracking | |
| CN111160212A (en) | Improved tracking learning detection system and method based on YOLOv3-Tiny | |
| Makris et al. | A comparison of trajectory compression algorithms over AIS data | |
| Han et al. | A method based on multi-convolution layers joint and generative adversarial networks for vehicle detection | |
| CN116071667A (en) | Method and system for detecting abnormal aircraft targets in specified area based on historical data | |
| CN111126504A (en) | Multi-source incomplete information fusion image target classification method | |
| CN114090718A (en) | Bi-LSTM prediction and fuzzy analysis based interrupted track correlation method | |
| CN108241150B (en) | Method for detecting and tracking moving object in three-dimensional sonar point cloud environment | |
| Gong et al. | Multi-target trajectory tracking in multi-frame video images of basketball sports based on deep learning | |
| Casagrande et al. | Abnormal motion analysis for tracking-based approaches using region-based method with mobile grid | |
| Wang et al. | Vision‐Based Lane Departure Detection Using a Stacked Sparse Autoencoder | |
| CN117115752A (en) | Expressway video monitoring method and system | |
| CN113012193A (en) | Multi-pedestrian tracking method based on deep learning | |
| CN112257750B (en) | Distributed fusion system for composite detection node | |
| Volk et al. | Environment-aware optimization of track-to-track fusion for collective perception | |
| EP4064120B1 (en) | Method for determining images plausible to have a false negative object detection |
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 |