CN110191424A - A kind of orbit generation method and device of specific suspect - Google Patents

Abstract

The present invention relates to public safety technical fields, and in particular to a kind of orbit generation method and device of specific suspect, method include: the fence and camera in conjunction with crime ground, are occurred with determining crime but the target person of non-carrying mobile phone；Each target person is tracked using multiple cameras around crime ground, obtains the movement routine of each target person after the incident；The camera in each movement routine is grouped respectively according to the shooting feature of camera, obtains preferred group corresponding to each target person and time choosing group camera；, to the priority orders of secondary choosing group, figure information is inquired according to from preferred group in each movement routine respectively, until determining the identity or movement track of suspect.The suspect of the present invention with can filtering out crime non-carrying mobile phone to carry out special investigation analysis, and multiple cameras in movement routine is grouped in advance, improves efficiency and accuracy that merit is studied and judged.

Description

A kind of orbit generation method and device of specific suspect

[technical field]

The present invention relates to public safety technical fields, and in particular to a kind of orbit generation method and dress of specific suspect It sets.

[background technique]

As the rapid development of technology of Internet of things has had face camera shooting in the actual internet of things equipment environment of public security Head, a variety of awareness apparatus such as video camera, fence, WIFI sniff, therefore personnel activity track, vehicle driving trace letter The perception datas such as breath, mobile phone location information can be recorded, these data accumulations get up to form the data of flood tide, so that In public security practical business, when investigation suspected target is with clue is found, workload becomes abnormal bulky complex, if merely Manually go clue is excavated from the data of these magnanimity according to merit, be tantamount to search for a needle in a haystack, also need to spend much more very Time and police strength, often the effect is unsatisfactory, also delays the detection best opportunity of case.How to be found out in these data for Case studies and judges valuable clue, quickly navigates to suspected target, improves the working efficiency of investigation, becomes urgently to be resolved and asks Topic.

Due to present most people go out all can carrying mobile phone, according to this feature, the police can during case is studied and judged The cellphone information of suspect is tracked by fence equipment, to determine the movement track after suspect's crime, is found and is checked The identity of suspect is locked after suspected target.But some suspects may have certain counterreconnaissance consciousness, to hide the police Tracking, intentional in flagrant delict not carrying mobile phone, therefore this part suspect can not be tracked by cellphone information, also It is difficult to quickly and effectively be checked, this undoubtedly will have a direct impact on the efficiency and accuracy that police's case is studied and judged.

In consideration of it, overcoming defect present in the above-mentioned prior art is the art urgent problem to be solved.

[summary of the invention]

The technical problem to be solved in the invention is:

When suspect tracks by cellphone information in the police, there is some suspects certain counterreconnaissance to realize, and commit a crime When deliberately carrying mobile phone does not cause to be difficult to quickly be checked for this part suspicion object, influences police's case to hide tracking The efficiency and accuracy that part is studied and judged.

The present invention reaches above-mentioned purpose by following technical solution:

In a first aspect, the present invention provides the orbit generation methods of specific suspect a kind of, comprising:

In conjunction with the fence equipment and camera in crime place, determine that the crime time occurred in crime place but do not took One or more target persons with mobile phone；

In conjunction with the characteristics of image of each target person, each target person is carried out using multiple cameras around crime place Target following, and then movement routine of each target person after the incident in preset time period is obtained respectively；

According to the shooting feature of camera, multiple cameras in each movement routine are grouped respectively, to obtain Preferred group of camera and time choosing group camera corresponding to each target person；

, to the priority orders of secondary choosing group, corresponding camera acquisition is inquired according to from preferred group in each movement routine respectively Figure information, until determine suspect identity or from multiple movement routines determine suspect movement track.

Preferably, the fence equipment and camera in combination crime place, determines the crime time in crime place Occurred but one or more target persons of non-carrying mobile phone, specifically included:

By the fence equipment in crime place, multiple mobile phone IMSI that the crime time occurred in crime place are determined Number corresponding mobile data；

By the camera device in crime place, the multiple personages for determining that the crime time occurred in crime place are corresponding Mobile data；

The mobile data of the multiple mobile phone IMSI number is matched with the mobile data of the multiple personage, is filtered out The crime time occurred but one or more target persons of non-carrying mobile phone in crime place；

Wherein, the mobile data includes the moving direction and/or movement speed in crime place.

Preferably, the shooting feature of the camera include: the shooting direction of camera, coverage, shooting resolution ratio, It is one or more in shooting height and shooting angle.

Preferably, the shooting feature according to camera, respectively divides multiple cameras in each movement routine Group specifically includes to obtain the preferred group of camera and time choosing group camera corresponding to each target person:

It obtains the shooting direction of each camera in each movement routine, and is carried out with the movement routine where each camera Match；

For each camera in each movement routine, shooting direction is met to the positive camera of personage and is classified as preferably Shooting direction is met and is classified as time selecting group to the non-frontal camera of personage by group.

Preferably, when being grouped to multiple cameras in each movement routine, the grouping further includes alternative group and takes the photograph As head；Wherein, compared with the preferred group of camera and time choosing group camera, the installation site of the alternative group of camera is special Very, coverage is hidden, and the priority orders being queried in three groups of cameras are minimum.

Preferably, described, to the priority orders of secondary choosing group, to inquire phase according to from preferred group in each movement routine respectively The figure information for answering camera to acquire, until determining the identity of suspect or determining the row of suspect from multiple movement routines Dynamic rail mark, specifically includes:

For each movement routine, the figure information of the corresponding preferably group camera acquisition of preferential inquiry, to obtain corresponding shifting The face characteristic of target person on dynamic path；

If face characteristic can not be got by preferably group camera, continue corresponding choosing group camera acquisition of inquiry Figure information；

If can not still get face characteristic by time choosing group camera, corresponding target person is tentatively classified as suspicion People is doubted, corresponding movement routine primarily determines as the movement track after suspect's crime.

Preferably, for any movement routine, if can not still get face spy by time choosing group camera described Sign, then after corresponding target person being tentatively classified as suspect, the method also includes:

The information of vehicles of corresponding suspect is obtained according to the camera in corresponding movement routine, and then true according to information of vehicles Surely correspond to the identity of suspect, and with there is the people of criminal record to match in police's information bank；It is corresponding to dislike if successful match The identity for doubting people is verified.

Preferably, for any movement routine, if getting face by preferably group camera or secondary choosing group camera Feature then will have the people of criminal record to match in face characteristic and police's information bank；If successful match, by mutually should the criminal of having The people of guilty previous conviction is classified as suspect, and finally determines the identity of suspect.

Preferably, for any movement routine, the method also includes:

The historical data acquired by camera obtains the historical movement track of target person in corresponding movement routine；

Determine the relevance between corresponding target person current movement routine and historical movement track；If relevance is unsatisfactory for Corresponding target person is then classified as suspect, and determines the identity of suspect by face characteristic by preset requirement.

Second aspect, the present invention provides the track generating means of specific suspect a kind of, including at least one processor And memory, it is connected between at least one described processor and memory by data/address bus, the memory is stored with can quilt The instruction that at least one described processor executes, described instruction are used for above-mentioned first aspect institute after being executed by the processor The orbit generation method of the specific suspect stated.

The beneficial effects of the present invention are:

In the orbit generation method of specific suspect provided by the invention, in conjunction with the fence equipment and camera shooting on crime ground Head apparatus, can filter out those may have counterreconnaissance to realize and the suspect of non-carrying mobile phone, and then be directed to this part suspicion People carries out special path analysis and identity investigation；Multiple cameras in movement routine are grouped in advance simultaneously, and are arranged each The priority orders that group camera is queried can carry out looking into for figure information according to priority orders when carrying out case and studying and judging It askes, improves efficiency and accuracy that merit is studied and judged.

[Detailed description of the invention]

In order to illustrate the technical solution of the embodiments of the present invention more clearly, will make below to required in the embodiment of the present invention Attached drawing is briefly described.It should be evident that drawings described below is only some embodiments of the present invention, for For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other Attached drawing.

Fig. 1 is provided in an embodiment of the present invention a kind of for determining the perception data analysis method flow chart of suspect；

Fig. 2 is that the movement track of suspect and perception on the way are set in a kind of police map provided in an embodiment of the present invention Standby schematic diagram；

Fig. 3 is a kind of schematic diagram for constructing space-time on police map and studying and judging region provided in an embodiment of the present invention；

Fig. 4 is a kind of stream that the movement track after suspect's crime is determined on police map provided in an embodiment of the present invention Cheng Tu；

Fig. 5 is the movement track after another suspect's crime determining on police map provided in an embodiment of the present invention Flow chart；

Fig. 6 is a kind of stream that determining suspect's identity is compared by perception data intersection-type collision provided in an embodiment of the present invention Cheng Tu；

Fig. 7 be it is provided in an embodiment of the present invention it is a kind of compared by intersection-type collision after obtain perception data list signal Figure；

Fig. 8 is a kind of process that " people-vehicle-IMSI " corresponding relationship is established by perception data provided in an embodiment of the present invention Figure；

Fig. 9 is a kind of method flow diagram that can determine suspect's movement track provided in an embodiment of the present invention；

Figure 10 is a kind of crime provided in an embodiment of the present invention ground mobile road of multiple target persons for occurring after the incident Diameter schematic diagram；

Figure 11 is the group technology flow chart of camera in a kind of each movement routine provided in an embodiment of the present invention；

Figure 12 is a kind of orbit generation method flow chart of specific suspect provided in an embodiment of the present invention；

Figure 13 is that a kind of determination provided in an embodiment of the present invention in crime occurred but the target person of non-carrying mobile phone Method flow diagram；

Figure 14 is a kind of architecture diagram of the track generating means of specific suspect provided in an embodiment of the present invention.

[specific embodiment]

In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and It is not used in the restriction present invention.

In the description of the present invention, term "inner", "outside", " longitudinal direction ", " transverse direction ", "upper", "lower", "top", "bottom", The orientation or positional relationship of the instructions such as "left", "right", "front", "rear" is to be based on the orientation or positional relationship shown in the drawings, and is only For ease of description the present invention rather than require the present invention that must be constructed and operated in a specific orientation, therefore be not construed as Limitation of the present invention.

In various embodiments of the present invention, symbol "/" indicate simultaneously tool there are two types of function meaning, and for symbol " A with/ Or B " then shows that the combination between the front and back object connected by the symbol includes " A ", " B ", " A and B " three kinds of situations.

In addition, as long as technical characteristic involved in the various embodiments of the present invention described below is each other not Constituting conflict can be combined with each other.Just with reference to drawings and examples, in conjunction with coming, the present invention will be described in detail below.

Embodiment 1:

The embodiment of the present invention provides firstly a kind of for determining the perception data analysis method of suspect, can help the police Personnel quickly determine suspect's identity after the incident, and the tracking for being conducive to suspect is arrested.As shown in Figure 1, the embodiment of the present invention The perception data analysis method of offer specifically includes:

Step 201, comprehensive analysis is carried out by each perception data acquired to awareness apparatus each around crime place, in police With the movement track after suspect's crime determining on map.

The present invention is based on police maps to be tracked analysis, and the point information of the various awareness apparatus in involved range is equal It can be displayed on the police map.Wherein, the awareness apparatus includes camera and/or fence, and the camera is again Including one of portrait camera, video camera and vehicle camera or a variety of, correspondingly, the perception data includes people One of picture data, vehicle data and fence data are a variety of.

On each street around crime place, it is usually fitted with corresponding awareness apparatus, each awareness apparatus is real-time The perception data in corresponding sensing range is acquired, and is stored in perception data system, therefore has magnanimity in perception data system View structure data (being acquired by various cameras) and fence data.After crime, by perception data system Mass data after crime time point in preset time period carries out comprehensive analysis, the course of action of suspect can be obtained, in turn Using the track marking Function on police map, the movement track of suspect is marked and portrays, such as the arrow line institute in Fig. 2 Show.After the movement track for determining suspect, to subsequent tracking and arrests suspect and have great importance.Wherein, described default Period can be configured according to the demand of studying and judging of the police, for example, if the police need to obtain after crime suspect in 2 hours Activity condition, then the preset time period is set as 2 hours.

Step 202, it according to the movement track of suspect, is matched on police map and the multiple perception marked out on the way is set It is standby, and calculate case and study and judge required space-time unique condition.

It usually will be installed multiple awareness apparatus, after determining movement track, related platform on the way in suspect's movement track (such as cloud server where each awareness apparatus) can obtain and match automatically all kinds of perception of the movement track on the way and set It is standby, so that getting case studies and judges required multiple awareness apparatus and the point information of each awareness apparatus, sensing range, sense Know direction etc..Alternatively, police personnel can also be relevant according to keyword search in police map along the movement track Front end awareness apparatus, such as fence filter out the awareness apparatus list for meeting keyword, then by selective listing Equipment is quickly positioned on police map.

By taking Fig. 2 as an example, five awareness apparatus of A, B, C, D, E are successively matched in the corresponding movement track of arrow (the solid small triangle i.e. in figure).Accordingly, it can determine that case is ground according to the status information of each awareness apparatus, sensing range etc. The spatial dimension sentenced.Again based on the action modes of suspect, the action institute of suspect within the scope of additional space can be calculated It takes time, accordingly, the time range that case is studied and judged can be specified.According to the spatial dimension and time range studied and judged, then may be used Required space-time unique condition is studied and judged to calculate case.

Step 203, it is based on space-time unique condition and the multiple awareness apparatus, is confined on police map along movement track Multiple space-times study and judge region out, and study and judge region to each space-time and carry out node setting；Wherein, each space-time studies and judges region Nei Bao Containing at least one awareness apparatus.

It is typically provided on police map and studies and judges tool box accordingly, it, can be directly using grinding when confining space-time and studying and judging region Sentence that " circle choosing " in tool box, " quadrilateral frame choosing ", " choosing of polygon frame " etc. are a variety of to study and judge regioselective way, in police map On confine out multiple space-times along the movement track and study and judge region and corresponding awareness apparatus, as shown in figure 3, being with quadrangle Frame is selected as example；It is of course also possible to directly carry out manual frame choosing according to the sensing range of awareness apparatus by police personnel, do not do herein It limits.In Fig. 3 for including an awareness apparatus in each region, but not to limit the present invention；In practical operation, It also may include two or more awareness apparatus in one region, for example, for being set at a distance of closer at least two perception At least two awareness apparatus on standby or same street, can confine into same space-time and study and judge region.

After confining out each space-time and studying and judging region and corresponding awareness apparatus, also need to study and judge each space-time region into The more node settings of row, so as to the subsequent statistics for carrying out perception data to awareness apparatus in each region, such as each region is studied and judged Time range studies and judges the perception data (such as cell-phone number ownership place) that need to be selected and associated docket etc.；Meanwhile in order to Facilitate Division identification on police map, the Node color difference that each space-time studies and judges region can also be set.

Step 204, intersection-type collision comparison is carried out by studying and judging the perception data acquired in region to each space-time, determines suspicion The corresponding perception data information of people, and then determine the identity of suspect.

It is same by the way that multiple space-times to be studied and judged to the carry out such as the portrait data in region, fence data and/or vehicle data Corresponding intersection-type collision compares between item perception data, such as carries out intersection contrast conting and go out the sense occurred jointly in these regions Primary data, corresponding personage can be determined as suspicion object.This is because usually in the specific time (when committing a crime and escaping Between), in specific route and specific space, while the object that perceived equipment captures can't be particularly more.Therefore, lead to Crossing the suspicion object that data collision obtains is to compare accurately, is analyzed by the perception data obtained to collision, can be true Determine the identity of suspect；It may also be combined with the existing information bank of the police to be matched, more further navigate to suspicion after screening Doubt people.

Wherein, the identity for determining suspect by the perception data that collision obtains here, specifically can be and passes through camera The portrait data of acquisition, analysis matching obtains the ID card information of suspect, and then determines the identity of suspect；It is also possible to lead to The vehicle data of camera acquisition is crossed, analysis obtains corresponding information of vehicles, such as license plate number, vehicle color, vehicle model, into And the identity for determining suspect is matched according to information of vehicles；The fence data that can also be acquired by fence, analysis Corresponding cellphone information, such as mobile phone IMSI number, mobile phone ownership place are obtained, and then is matched according to cellphone information and determines suspect's Identity.It is, of course, also possible to be to combine two or three of perception data therein, the identity of suspect is determined after comprehensive analysis.

In the above-mentioned perception data analysis method for determining suspect provided in an embodiment of the present invention, suspect can be based on Multiple awareness apparatus in movement track neatly confine out multiple space-times on police map and study and judge region, to construct Node is studied and judged in multiple space-time comparisons, by the perception data progress intersection-type collision comparison to region is respectively studied and judged, determines suspect, The efficiency and accuracy that merit is studied and judged are improved, can helping police personnel, quickly, accurately investigation arrives suspected target, determines suspicion Personal part.

From the foregoing, it will be observed that the awareness apparatus includes camera and/or fence, then the movement track after suspect's crime It can be obtained according to the perception data (such as portrait data, vehicle data) that camera each around crime place acquires, it can also be according to case The perception data (fence data) of each fence acquisition obtains around hair place.When movement track is obtained according to camera When, each perception data by acquiring to awareness apparatus each around crime place carries out comprehensive analysis, on police map After determining suspect's crime movement track (i.e. step 201) can refer to Fig. 4, specifically include:

Step 2011, the camera in conjunction with crime time, crime place acquires information, and/or, it victim and/or witnesses The information that person provides, determines the characteristics of image of suspect.

Based on the universalness of current monitoring camera setting, usually can all be installed near crime place or crime place There is camera, the camera can be portrait camera, video camera.It is supervised by the camera for transferring crime time point Control, and/or, suspect's information that victim and/or witness provide can quickly determine the characteristics of image of suspect.Wherein, institute It states the figure and features feature that characteristics of image refers mainly to suspect, such as height, figure, hair style and dresses；In addition, if suspect Action modes be by bus or drive, then described image feature may also include the vehicle characteristics of suspect, such as license plate number, vehicle type Number and vehicle color etc..

Step 2012, according to the characteristics of image of suspect, to multiple view structure data in perception data system into Row inquiry, and then detect the movement track after suspect's crime.

Function is looked into fastly using perception data in the perception data system in conjunction with the characteristics of image of suspect, from crime View structure data (the i.e. all kinds of collected numbers of camera of collected magnanimity are corresponded to after time point in preset time period According to) in, with the response time of Millisecond, detect movement track of the suspect after the crime time.It can specifically operate as follows: first Necessary querying condition is first inputted in perception data system by police personnel, characteristics of image, crime time, case such as suspect Send out place, inquiry time segment limit (i.e. preset time period, such as be set as within 3 hours after the crime time), then by Related platform is inquired automatically according to input condition, carries out secondary people according to query result by police personnel again after the completion of inquiry To study and judge, the final movement track for determining suspect.In addition to this, related platform can also be based on the view in perception data system Structural data directly carries out target following to the suspect of respective image feature, and then determines the action after suspect's crime Track.

Step 2013, the movement track after the suspect's crime detected is labeled on police map.It is specific available The track marking Function in tool box is studied and judged, the movement track of suspect is marked and depict on police map, such as Fig. 2 and figure Shown in arrow line in 3.

Further, invading in wealth case in mobile phone plunder one kind, suspect grab short time in the interior also future after mobile phone It obtains and to theft handset processes, the mobile phone of victim must be also carried with it, and corresponding cellphone information and mobile phone are mobile Situation can be obtained by the fence around crime place, therefore the movement track after suspect's crime can also pass through fence It determines.At this point, each perception data by acquiring to awareness apparatus each around crime place carries out comprehensive analysis, police On map determine suspect's crime after movement track (i.e. step 201) can refer to Fig. 5, specifically include:

Step 2011 ', the cellphone information or the fence of crime time, crime place provided according to victim is adopted Collect information, obtains the corresponding IMSI number of victim's mobile phone.

The cellphone informations such as the user identity that the specific cell-phone number provided in combination with victim, mobile phone model, cell-phone number are bound, It is converted to corresponding mobile phone IMSI number.Either, when crime place is provided with fence just, can remember in the crime time Data in mobile phone under record in corresponding sensing range, therefore the corresponding mobile phone IMSI of victim can be directly acquired by fence Number.

Step 2012 ', according to the IMSI number of mobile phone, multiple fence data in perception data system are looked into It askes, and then detects the movement track after suspect's crime.

Function is looked into fastly using perception data in the perception data system in conjunction with the mobile phone IMSI number of victim, from case In the fence data for corresponding to collected magnanimity after hair time point in preset time period, with the response time of Millisecond, visit Trend of the corresponding mobile phone IMSI number after the crime time is found, and then quickly determines action rail of the suspect after the crime time Mark.It can specifically operate as follows: necessary querying condition be inputted in perception data system by police personnel first, such as victim Mobile phone IMSI number, the crime time, crime place, inquiry time segment limit (i.e. preset time period, after being such as set as the crime time 3 hours within) etc., then inquired automatically by related platform according to input condition, inquiry after the completion of again by police personnel It is artificially studied and judged according to query result progress is secondary, the final movement track for determining suspect.In addition to this, related platform can also base Fence data in perception data system directly carry out target following to the suspect of corresponding mobile phone IMSI number, in turn Movement track after determining suspect's crime.

Step 2013 ', the movement track after the suspect's crime detected is labeled on police map.It is specific available The track marking Function in tool box is studied and judged, the movement track of suspect is marked and depict on police map, such as Fig. 2 and figure Shown in arrow line in 3.

In the step 202, case is studied and judged required space-time unique and can specifically be determined in accordance with the following methods: when matching The multiple awareness apparatus of suspect's movement track on the way, and after the point information of determining each awareness apparatus and sensing range, from case Hair place start or the movement track on first awareness apparatus (i.e. awareness apparatus A) start, onto the movement track Region between the last one awareness apparatus (i.e. awareness apparatus E), can be used as the spatial dimension that case is studied and judged.Determining suspicion It, can be according to the traveling speed of corresponding action modes (such as walking, ride a bicycle, ride electric vehicle, drive) after the action modes of people Degree calculates within the scope of additional space t the time required to the action of suspect, i.e., from crime place or awareness apparatus A to perception Equipment E, t the time required to the action of suspect, then since crime time point T after time period t in (i.e. T~T+t) The time range studied and judged as case.Furthermore, it is contemplated that determining that time range may have certain mistake according to action modes Difference can also determine by the following method time range: in step 201, determine suspicion by carrying out comprehensive analysis to perception data After the movement track of people, inquiry suspect passes through the time point of the last one awareness apparatus (i.e. awareness apparatus E) in movement track T ' then passes through the period (i.e. T~T ') in awareness apparatus E time point to suspect since crime time point T, can be used as The time range that case is studied and judged.

Further, in the step 203, region, police map upper ledge choosing are studied and judged for each individual space-time Area size out can be used as the spatial dimension of studying and judging of corresponding region, and the area size that frame is selected can be according in the region Total sensing range of each awareness apparatus determines.Region is studied and judged for each space-time, studies and judges the determination method of time range accordingly Specifically:

Action modes based on suspect and the time for leaving crime place, it is each on the way that estimation suspect passes through movement track Time of awareness apparatus, and then study and judge the division in region in conjunction with space-time determines that each space-time studied and judged region studies and judges time range.Ginseng Fig. 3 is examined, it, can root after determining the action modes of suspect by taking each space-time is studied and judged in region comprising awareness apparatus as an example Distance and suspect according to the travel speed of corresponding action modes, crime place away from awareness apparatus A leave crime place when Between point, time point T1 when calculating suspect by awareness apparatus A, and giving on the basis of corresponding time point certain upper Lower fluctuation range Δ t, then T1 ± Δ t can be used as the time range of studying and judging that corresponding space-time studies and judges region, and Δ t can be according to studying and judging need Flexible choice is sought, herein without limitation.Similarly, calculate in the same way suspect pass through respectively awareness apparatus B, C, D, Time point when E, and then obtain that corresponding space-time studies and judges region studies and judges time range.

In another optional scheme, for reduce estimated with action modes caused by error, area is studied and judged for each space-time Domain, the determination method for studying and judging time range accordingly can also be as follows: obtaining suspect by carrying out comprehensive analysis to perception data Movement track after, determine that suspect studies and judges region by the time of movement track each awareness apparatus on the way, and then in conjunction with space-time Division, determine that each space-time studied and judged region studies and judges time range.With reference to Fig. 3, still studied and judged in region with each space-time comprising one For a awareness apparatus, in step 201, after carrying out the movement track that comprehensive analysis determines suspect to perception data, Time point when inquiry suspect is by awareness apparatus A, B, C, D, E respectively, and given centainly on the basis of corresponding time point Fluctuation range up and down (i.e. corresponding time point ± Δ t) studies and judges time range in this, as what corresponding space-time studied and judged region.

It is described to be carried out by studying and judging the perception data acquired in region to each space-time during actually studying and judging in conjunction with Fig. 6 Intersection-type collision compares, and determines the corresponding perception data information of suspect, and then determines identity (i.e. step 204), the tool of suspect Body again the following steps are included:

Step 2041, thinking is studied and judged according to the police, in each awareness apparatus and each space-time of the movement track on the way Interregional match ratio of establishing is studied and judged to relationship, model is studied and judged in formation.

It, can be by way of dragging, in each awareness apparatus and each space-time of the movement track on the way in police map Interregional match ratio of establishing is studied and judged to relationship, so as to which to be converted into a personal-machine thinking consistent by the thinking of studying and judging of police personnel Flow chart is studied and judged, that is, studies and judges model.Wherein, establish match ratio here to relationship primarily to carry out perception data intersection Collision compares, and the match ratio can be intersection, union and exclusion etc. to relationship.For example, for containing at least two perception The space-time of equipment studies and judges region, establishes union relationship between at least two awareness apparatus, grinds so as to correspond to space-time The perception data for sentencing each awareness apparatus acquisition in region does union processing, and the perception data in region is studied and judged as corresponding space-time；It is right Converging relation is established between region in entire movement track, each space-time on the way is studied and judged, so as to which each space-time is studied and judged area The perception data of corresponding acquisition does intersection processing in domain；And there are the awareness apparatus of failure, the perception datas of acquisition for certain Accuracy can decrease, then can be excluded.

Further, after building and studying and judging model, related platform can also judge automatically the legitimacy for studying and judging model, For example judge that each awareness apparatus and each space-time study and judge whether interregional match ratio constructs correct, reasonable, each space-time to relationship Whether whether the node setting for studying and judging region is reasonable, can smoothly execute etc. according to the model is arranged at present.It is legal only to judge, The model of studying and judging can just continue to execute；If it is determined that it is illegal, then it needs to be handled according to the problem of finding out, Jin Erchong Established model.

Step 2042, model is studied and judged described in execution, so that each space-time is studied and judged the perception data in region and carries out intersection-type collision ratio It is right, obtain the suspicion list comprising one or more suspects.

According to the above match ratio to relationship, the same item perception data that each space-time studies and judges region is subjected to intersection-type collision comparison Afterwards, the perception data occurred jointly in each region can be obtained, corresponding personage can be determined as suspicion object.This part is common The perception data occurred can be shown in the form of a list, and each perception data corresponds to a suspicion object.With electricity For sub- fence perception data, the fence data occurred jointly in each region, including mobile phone can be obtained after comparing by intersecting IMSI number, cell-phone number ownership place, that corresponding cell-phone number studies and judges capture number, corresponding cell-phone number in region in corresponding space-time is nearest Motion frequency (such as active days in nearly 30 days), corresponding cell-phone number corresponding space-time study and judge activity time in region and One or more in moving position, the list of fence data shows that result can refer to Fig. 7.

Wherein, to show convenient for list, capture number can use minimum number, nearly 30 days active days after doing intersection processing It can use maximum activity number of days after doing intersection processing.When being checked, it is contemplated that actual conditions are targetedly checked, For example, it is contemplated that may be to flee about to commit crimes to some suspects, a possibility that stranger is larger therefore non-for cell-phone number ownership place Local suspicion object can emphasis investigation；For the nearly 30 days lower objects of motion frequency in monitoring range, it may be possible to interim It commits a crime with reaching crime, it is usually substantially movable not around crime ground, therefore can also emphasis investigation；Etc..

By taking Fig. 7 as an example, 4 common perception datas are obtained in intersection-type collision after comparing, and correspond to 4 suspicion objects, according to The cellphone information that fence data provide, can determine respectively the identity of 4 suspicion objects, and then obtains one and include 4 suspicion The suspicion list of people.Wherein, when using portrait data and/or vehicle data, sense that intersection-type collision occurs jointly after comparing Primary data can also carry out list displaying in the way of Fig. 7, further according to figure information and/or information of vehicles, determine suspicion object Identity.

Step 2043, in conjunction with the previous conviction background of each suspect in case feature and police's information bank, in suspicion list Each suspect carry out screening investigation, the final identity for determining suspect.

For there is the suspect of criminal record background, can all there is corresponding record in police's information bank, and with no crime The people of previous conviction compares, and has people's crime possibility of criminal record usually larger, therefore after obtaining suspicion list, can tie It closes case feature and police's information bank further screens investigation to each suspect, to be more accurately located final suspect. For example, current case is to steal case, by finding after matching inquiry, before having a people once to have theft crime in suspicion list Section, therefore can tentatively conclude that the suspicion object is the suspect of the case.By this investigation method, the police will be greatly reduced The investigation range of personnel also can significantly reduce the investigation time, be greatly improved and study and judge efficiency, promote the accuracy for studying and judging result.

Further, it is contemplated that during actual monitored, each awareness apparatus by point position, pull-in time interval, catch It catches the movement speed of object, capture the influence of the factors such as surrounding enviroment, capture rate is not 100%, may have capture and lose The case where leakage.For example, 5 space-times for studying and judging model in embodiments of the present invention are studied and judged in region, it is assumed that only 4 space-times are studied and judged The trace information of suspicion object X is captured in region, at this time if the perception data for still taking this 5 space-times to study and judge region carries out Intersection-type collision compares, and can be omitted due to awareness apparatus capture causes the track of suspicion object X not ground in some space-time in the middle Sentence in region, so that the comparison impact conditions for studying and judging model specification cannot be fully met.Therefore, suspicion object X would not occur In the suspicion list for comparing collision, suspected target, which is missed, to be caused to study and judge not precisely.

To solve the problems, such as that above-mentioned this capture is omitted, it can use " mode is studied and judged in leak source analysis " and carry out secondary study and judge. Specifically: before studying and judging model described in the execution, each space-time is analyzed comprehensively and studies and judges perception data in region, and then is studied and judged each Situation is omitted in the capture of suspected target；The space-time omitted if there is capture studies and judges the region (awareness apparatus i.e. in corresponding region There is capture to omit), then it is studied and judged in region in the space-time that appearance capture is omitted, the perception data of corresponding suspected target has been supplemented It is whole, it is assumed that these suspected targets also exist in the awareness apparatus that capture is omitted at these and occurred, so as to make up crash data Integrality, collision result is rectified a deviation, the knot for studying and judging fault is led to more accurate corrections awareness apparatus capture omission Fruit.

Wherein, it not carrying out during studying and judging of leak source analysis, after the comparison of perception data intersection-type collision, actually obtains It is to study and judge the object occurred jointly in region in all space-times, that is, is classified as suspect.Process is studied and judged in progress leak source analysis In, after perception data intersection-type collision compares, what is actually obtained is studied and judged in region and occurred jointly in the space-time of preset ratio Object generally exactly study and judge the object all occurred in region in most space-time, that is, be classified as suspect；Wherein, The preset ratio is typically larger than 50%, such as optional 80%.Thus, it is assumed that some object is not studied and judged in all space-times All occur in region, but study and judge in region in 80% or 80% or more space-time and all occurred, then the object may be in a small portion Timesharing sky, which is studied and judged, has the case where capture is omitted in region, but still the object can be classified as suspect.

In conclusion can be first passed through in the perception data analysis method provided in an embodiment of the present invention that can determine suspect The movement track that suspect is determined to the comprehensive analysis of perception data, the multiple perception being then based in suspect's movement track are set It is standby, multiple space-times are made in police map upper ledge and study and judge region, so that constructing multiple space-time comparisons studies and judges node, by each The perception data for studying and judging region carries out intersection-type collision comparison, can screen and determine suspect, improve efficiency that merit is studied and judged and Accuracy, can helping police personnel, quickly, accurately investigation arrives suspected target.In addition, by combining case feature and the police to believe Investigation is further screened in breath library to suspect, can be more accurately located from the people for have criminal record to final suspect, substantially Degree reduces the investigation range of police personnel, also can significantly reduce the investigation time.Meanwhile being omitted to solve awareness apparatus capture " mode is studied and judged in leak source analysis " also can be used in problem, to make up the integrality of crash data, collision result is rectified a deviation, is kept away Exempt to study and judge fault.

Embodiment 2:

It mentions in the above-described embodiments, the identity of suspect can be determined by perception data information, that is, utilizes portrait data Any one of (camera acquisition), vehicle data (camera acquisition) and fence data (fence acquisition) are appointed Two or all three, to analyze the identity of determining suspect.Generally, suspicion is determined by all three perception datas The identity of people is easier to accomplish, and obtains any one perception data by wherein wantonly two to determine that the identity of suspect can There can be certain difficulty, when especially only obtaining a perception data, it is desirable to determine that the identity of suspect, difficulty are to compare Big.

In the actual business of public security, since many suspicion objects have certain anti-reconnaissance capability, in flagrant delict It may be hidden using various gimmicks and be found or identified by public security organ.For example, some suspects are motor-driven in utilization When vehicle is committed a crime, the conscious baffle using above automobile driver seat of meeting deliberately shelters from face, then portrait data can not Effectively obtain；Some suspects, can be deliberately without mobile phone when crime, then fence data can not be obtained effectively；Deng Deng.In this way, above-mentioned all three perception data may finally can not be got, and it is only capable of getting one such or two Kind perception data, therefore need to accomplish when actually studying and judging, according to any one of them perception data, can fast and effeciently lock The identity of suspect, without three kinds of perception datas must be relied on.

To realize above-mentioned requirements, before carrying out practical case and studying and judging, the method also includes: it is perceived using before crime The magnanimity perception data that historical accumulation gets up in data system carries out the mixing together of multidimensional space-time, establishes " people-vehicle-IMSI " Between corresponding relationship；Wherein, the perception data includes portrait data, vehicle data and fence data.Specifically may be used With reference to Fig. 8, comprising:

Step 101, the point data of the perception data, each awareness apparatus that are acquired before the case happened using various awareness apparatus and The static data having by oneself in police's information bank, constructs multiple data sets.

First is that building face characteristic and portrait track data collection: when usually without crime situation, acquiring multiple people in advance As the data of camera, and import perception data system, clustered using human Facial Image Recognition Algorithm, thus generate face characteristic and The structured data sets of portrait track.

Second is that building vehicle characteristics and track of vehicle data set: when usually without crime situation, acquiring multiple vehicles in advance Data of camera, and import perception data system, are clustered using vehicle identification algorithm, thus generate vehicle characteristics and The structured data sets of track of vehicle.

Third is that building IMSI and IMSI track data collection: when usually without crime situation, acquiring multiple electronics in advance and enclose The data on column, and perception data system is imported, generate the structured data sets of fence, the i.e. structure of the track IMSI and IMSI Change data set.

Fourth is that the point and point corresponding relation data collection of building awareness apparatus: when usually without crime situation, in advance The point and point feature for determining multiple awareness apparatus, based on the location information of each awareness apparatus, using GIS algorithm, to perception Corresponding relationship between equipment point and each awareness apparatus point is calculated, to form the point and point pair of awareness apparatus Answer relational dataset.

Fifth is that building static relation data set: being closed based on the static data being had by oneself in police's information bank, such as household register of personnel Coefficient is according to, the registration data of vehicle and processing data violating the regulations of vehicle etc., using the algorithm logic of chart database, construct people, The static relation data set of the objects such as vehicle.

Step 102, in conjunction with the information in the multiple data set, " people-vehicle " relationship, " vehicle-IMSI " relationship are constructed respectively " people-IMSI " relationship.

First is that building " people-vehicle " relationship: in the register information of motor vehicle or processing information violating the regulations etc., generally can all have The relevant information of people, such as identification card number, name, telephone number real name information；There are also the information of vehicle, such as license plate numbers simultaneously Code, number plate color, vehicle brand model etc..It therefore, can be fast by the register information of motor vehicle or processing information violating the regulations etc. The direct relation of speed building people and Che.But in real life, not necessarily car owner is driving vehicle, it is also possible to be The friends and family of car owner can then be carried out one by the awareness apparatus of front end at this time and grab triple time (shooting vehicle, shooting driver, bat Take the photograph copilot), to obtain the figure information of driver, it is aided with Identification of Images and comparison technology, practical driving vehicle can be found Personal information, such as identification card number, so that the people and the vehicle are associated.By above method and technology, can by people and The relation excavation of vehicle comes out.

Second is that building " vehicle-IMSI " relationship: specifically using (be provided with vehicle camera) at vehicle bayonet and near Fence equipment, by vehicle bayonet capture vehicle data and fence capture mobile phone IMSI number carry out intersection ratio It is right, to extrapolate the potential relationship between vehicle and mobile phone IMSI number.In actual operation, vehicle bayonet would generally nearby be built Some fence equipment are found, while vehicle bayonet captures vehicle pass-through information, fence equipment can also be worked at the same time, The mobile phone IMSI in range (can be specifically arranged) in capture certain distance according to the demand of producer and business.Due to perception data system In save vehicle bayonet, the mass data that fence accumulates for a long time, then, those there are the vehicle of certain relationship and IMSI can calculate index by colleague's number, colleague's point digit, number of days of colleague etc., infer between vehicle and IMSI the two With the presence or absence of incidence relation.Wherein, the colleague specifically refers to, in (such as in 30 seconds) certain time difference, all by apart from phase Close vehicle bayonet and fence equipment captures.For example, certain vehicle and certain mobile phone IMSI number have super between multiple points 5 days are spent, total 30 colleagues (being almost captured to corresponding information simultaneously), so that it may infer this vehicle and this mobile phone There are strong relationships between IMSI.Accordingly, the relationship between Che He IMSI can be established.

Third is that building " people-IMSI " relationship: specifically using (be provided with portrait camera) at portrait bayonet and near Fence equipment, by portrait bayonet capture portrait data and fence capture mobile phone IMSI number carry out intersection ratio It is right, to extrapolate the potential relationship between people and mobile phone IMSI.In actual operation, the position of portrait bayonet is disposed generally also Some fence equipment can be disposed accordingly, when the capture of portrait bayonet passes by the looks information of pedestrian, fence Also it works at the same time, the letter of the mobile phone IMSI in range (can be specifically arranged) in capture certain distance according to the demand of producer and business Breath.Due to saving mass data for a long time in perception data system, then, those there are the personnel of certain relationship and IMSI can calculate index by colleague's number, colleague's point digit, number of days of colleague etc., and inferring between people and IMSI the two is It is no that there are incidence relations.Wherein, the colleague specifically refers to, all closely located in (such as in 30 seconds) certain time difference Portrait bayonet and fence equipment capture.For example, someone has more than 5 between multiple points with certain mobile phone IMSI It, total 30 colleagues (being almost captured to corresponding information simultaneously), so that it may infer this people and this mobile phone IMSI Between there are strong relationships.Accordingly, the relationship between people and IMSI can be established.

Step 103, it according to " people-vehicle " relationship, " vehicle-IMSI " relationship and " people-IMSI " relationship, establishes " people-vehicle-IMSI " Between corresponding relationship.

Based on the building of three fundamental relations in the step 102, basic data collection has been constructed for whole relationship, The corresponding relationship between " people-vehicle-IMSI " three can further be got through.Further, " people-vehicle-IMSI " that can will also obtain Corresponding relationship is integrated with the static relation in the static relation data set, tamps the relationship between object.For example, passing through The data of awareness apparatus can calculate the potential relationship between vehicle owner and the practical driver of vehicle, but if vehicle institute Simultaneously again there are if household register relationship between someone and the practical driver of vehicle, these relationships can be superimposed construct it is more Series of fortified passes system, by the further abundantization of corresponding relationship and generalization between " people-vehicle-IMSI ".

In conclusion in embodiments of the present invention, making full use of the magnanimity portrait number to accumulate over a long period all the year round before crime It is established " people-vehicle-IMSI " according to, vehicle data and fence data etc. based on the rule or information found when not committing a crime usually Between corresponding relationship, study and judge foundation for use in the crime period.In this way, when actually being studied and judged, as long as obtaining To the wherein a certain item perception data of suspect, other senses can be excavated according to the corresponding relationship between " people-vehicle-IMSI " Primary data, so comprehensive analysis study and judge after determine suspect identity, meet case and study and judge demand.

Embodiment 3:

It in embodiment 1, is to carry out synthesis by each perception data for acquiring awareness apparatus each around crime place to grind Sentence, movement track (the i.e. step 201) after determining suspect's crime.When the perception data that is acquired according to camera on the way determines When, need first to determine the characteristics of image of suspect (with specific reference to step 2011), and if the graphic feature of suspect is difficult to directly Connect determination, for example, crime be not provided with camera or camera shooting is unintelligible, and provide without witness corresponding Information, or since the crowd is crowded, crime personnel can not be also directly locked in camera, it is also just difficult to directly pass through at this time Camera obtains the movement track of suspect.

In view of mobile phone is all carried in most people's trip, and the motion track of mobile phone is can to pass through fence Equipment is tracked detection, therefore, can in the embodiment of the present invention in the case where suspect's characteristics of image is difficult to directly acquire First fence equipment is combined to find out possible multiple movement routines, in conjunction with the camera of each movement routine on the way, comprehensive point The movement track of suspect is determined in analysis, or directly determines the identity of suspect in the process.Simultaneously it is considered that camera Quantity setting is more, the workload of the police can be made to increase if transferring inquiry one by one, therefore the embodiment of the present invention is according to camera Shooting feature camera is grouped in advance and the priority orders of browsing are set, to improve search efficiency.

As shown in figure 9, being specifically included the embodiment of the invention provides a kind of method that can determine suspect's movement track:

Step 301, by fence equipment, multiple mobile phone IMSI that the crime time occurred in crime place are determined Number, and movement routine of each mobile phone IMSI number after the incident in preset time period is obtained respectively.

At crime time point, crime place usually will appear multiple people, and since most people can carry mobile phone, As long as being therefore provided with fence equipment around crime place and crime place, everyone corresponding mobile phone can be detected Movement routine after IMSI number and tracing detection to each mobile phone IMSI number crime, that is, crime occur everyone Movement routine after crime.Concrete operations are as follows: firstly, the electronics acquired by inquiry by the fence equipment in crime place Fence data determine the corresponding multiple mobile phone IMSI numbers of multiple target persons that the crime time occurred in crime place.Then, Using multiple fence equipment around crime place, target following is carried out respectively to the multiple mobile phone IMSI number, in turn Movement routine of each mobile phone IMSI number after the incident in preset time period is obtained respectively.

For example, with reference to Figure 10, crime place in the crime time passes through after fence equipment detects and finds 3 hands Machine IMSI number respectively corresponds three target person first, second, third；Further pass through multiple fence equipment around crime ground After carrying out target following to this 3 mobile phone IMSI numbers, it is as shown in the figure to obtain corresponding three movement routines a, b, c difference.Wherein, The preset time period can be configured according to the demand of studying and judging of the police, for example, if the police needed to obtain after crime in 2 hours The activity condition of suspect, then the preset time period is set as 2 hours.

Step 302, according to the shooting feature of camera, multiple cameras in each movement routine are grouped respectively, To obtain the preferred group of camera and time choosing group camera corresponding to each mobile phone IMSI number.

In the corresponding movement routine of each target person, multiple cameras usually can be all provided with, determine movement routine Afterwards, related platform can obtain and match automatically each camera of corresponding movement routine on the way；By being carried out to these cameras Specific figure information inquiry, can further check out the information of suspect.And specific portrait is inquired transferring each camera Before information, to study and judge the time convenient for saving police personnel, the camera in each movement routine is divided in advance Group.Wherein, group basis is the shooting feature of camera, and the shooting feature includes: the shooting direction of camera, shooting model Enclose, shoot it is one or more in resolution ratio, shooting height and shooting angle, it can only according to an one of them feature be grouped, Also it is grouped in combination with multinomial feature therein.In embodiments of the present invention, mainly for being divided into two groups, i.e., preferred group and time choosing Group after being finally completed grouping, is theoretically preferably organized corresponding overall image quality and should be better than and time select group, can correspondingly obtain effectively The probability of information selects group also greater than secondary, therefore ought to preferentially be inquired.

Step 303, it, to the priority orders of secondary choosing group, inquires and accordingly takes the photograph according to from preferred group in each movement routine respectively As the figure information that head acquires, until determining the identity of suspect or determining the action rail of suspect from multiple movement routines Mark.

In above step, the movement routine of multiple target persons is obtained based on multiple mobile phone IMSI numbers, but not Be each target person it is suspect, also needs further to screen investigation by the camera in each movement routine；Pass through screening Investigation, suspect can be primarily determined out from multiple target persons, has also determined that corresponding movement track, to it is subsequent tracking and Suspect is arrested to have great importance.The probability that effective information can be obtained due to preferably organizing camera selects group greater than secondary, Preferred group can be first inquired, is not necessarily to continue to inquire time choosing group camera if it can get effective information；Only obtain less than having Just continue to inquire time choosing group camera when imitating information.In this way, without being adjusted to each camera on each path It takes, can greatly save the police studies and judges the time.

In the above method provided in an embodiment of the present invention, multiple cameras in suspect's movement routine are divided in advance Group, such as be divided into preferred group and select group with secondary, and the priority orders that each group camera is queried are set, then it is studied and judged in progress case When, the inquiry of figure information can be carried out according to priority orders, so that efficiency and accuracy that merit is studied and judged greatly are improved, it can Police personnel are helped quickly, accurately to determine suspect's identity or the movement track of suspect.

In the step 302, it is assumed that be only that foundation is grouped with the shooting direction of camera.This is allowed for, from The shooting direction of camera is analyzed, if the front of target person can be taken, obtains the probability of effective face characteristic more It is larger, therefore shooting direction is had into important reference significance as group basis.In this case, the basis is taken the photograph As the shooting feature of head, multiple cameras in each movement routine are grouped respectively, to obtain corresponding to each mobile phone Preferred group of camera of IMSI number and time choosing group camera (i.e. step 302) specifically refers to Figure 11, comprising:

Step 3021, the shooting direction of each camera in each movement routine is obtained, and the mobile road where with each camera Diameter is matched.

With reference to Figure 10, by taking the movement routine c where target person third as an example, the multiple cameras being arranged on the path are as schemed Shown in middle dot (including black circle and soft dot).It determines on the path after the shooting direction of each camera, with movement Path c matching, can determine, during target person third normally moves forward, each camera is for shooting after matching Front, side or the back side of target person third.For example, target person third is to travel eastwards on movement routine c, then if The shooting direction of camera is towards west, or has the subtle deviation of directivity, can theoretically take the front of target person third just, And the camera of other directions then can not theoretically take the front of target person third, may take side or the back side.

Step 3022, for each camera in each movement routine, shooting direction is met to the positive camera of personage It is classified as preferred group, shooting direction is met and is classified as time selecting group to the non-frontal camera of personage.

Continue to refer to figure 10, still by taking the movement routine c where target person third as an example, in figure solid dot expression meet to The positive camera of personage is classified as preferably organizing camera；The camera non-frontal to personage is met in hollow dot expression, is classified as secondary The priority that choosing group camera, preferably group are queried selects group higher than secondary.It is to be divided into camera using shooting direction as foundation herein For two groups, it certainly, in other alternative-embodiments, may also be combined with more foundations, camera be divided into more multiple groups.Example Such as, it is assumed that using shooting direction and resolution ratio as foundation, the camera on each path is divided into four groups, then can be arranged as follows: the One group of camera is high-resolution, and shooting direction is positive towards target person；Second group of camera is low resolution, but is shot Direction is positive towards target person；Third group camera is high-resolution, but shooting direction is non-frontal towards target person；4th Group camera is low resolution, and shooting direction is non-frontal towards target person；It theoretically analyzes, the image matter of four groups of cameras Amount successively reduces, and the probability that can obtain effective information also successively reduces, therefore the priority orders being queried successively reduce.Its In, the differentiation of high-resolution and low resolution specifically can be using a certain preset value as line of demarcation, be higher than the preset value High-resolution is low resolution lower than the preset value, it is not limited here.

In another optional scheme, it is contemplated that can coverage also have a major impact to get effective information, Therefore when being grouped to multiple cameras in each movement routine, in addition to preferred group camera and time choosing group camera, institute Stating grouping may also include alternative group camera.Wherein, described alternative compared with the preferred group of camera and time choosing group camera The installation site of group camera is special, coverage is hidden, and the priority orders being queried in three groups of cameras are minimum.Example Such as, the preferred group of camera and time choosing group camera, which are normally located on street, waits visibility points, it is easier to take The target person passed by, and the alternative group of camera may be on some enclosure wall being arranged in by street, or by grove etc. Barrier blocks, therefore coverage is more covert, when target person is on the way difficult to be photographed when driving, obtains effective information A possibility that it is very low, therefore the priority orders being queried are minimum.But in some special cases, it is possible to taking Target person, for example, when target person climb enclosure wall just or be hidden in grove it is medium may just can by alternatively group camera clap It takes the photograph, may be got at this time than preferably organizing camera and time more valuable information of choosing group camera, therefore still have certain Reference significance.

Further, described respectively in each movement routine according to from preferred group to the priority orders of secondary choosing group, inquiry The figure information of corresponding camera acquisition, until determining the identity of suspect or determining suspect's from multiple movement routines Movement track (i.e. step 303), specifically: for each movement routine, the people of the corresponding preferably group camera acquisition of preferential inquiry As information, to obtain the face characteristic of target person in corresponding movement routine.If can not be got by preferably group camera Face characteristic then continues the figure information of corresponding choosing group camera acquisition of inquiry；If still can not by time choosing group camera Face characteristic is got, then corresponding target person is tentatively classified as suspect, corresponding movement routine primarily determines as suspicion Movement track after people's crime.

With reference to Figure 10, by taking the movement routine c where target person third as an example, if target person third is normal along movement routine Traveling, theoretically positive effective face characteristic can be collected by preferably organizing camera (i.e. black circle), at this time just without Time choosing group camera need to be transferred again.But if target person third exist under steam face block, bow, the feelings such as side rotary head Farther out etc., preferably group camera can not may just collect face characteristic for condition or shooting distance, just transfer time group is selected to take the photograph at this time As head further confirms that.Theoretical last time choosing group camera (i.e. soft dot) can not collect positive face characteristic, but if There is situations such as side rotary head, turn one's head in target person third, may take front by choosing group camera just, therefore still under steam There is certain reference value.If can not still get effective face characteristic, very possible mesh by time choosing group camera Mark personage third has deliberately blocked face or has avoided camera, it is believed that there are certain crime suspicion, therefore can tentatively be classified as suspicion People is doubted, corresponding movement routine primarily determines as the movement track after suspect's crime.

Wherein, for any movement routine, if can not still get face characteristic by time choosing group camera described, After corresponding target person is tentatively then classified as suspect, the method also includes: according to the mobile phone in corresponding movement routine IMSI number, determines the identity of corresponding suspect, and with there is the people of criminal record to match in police's information bank；If matching at Function, the then identity for corresponding to suspect are verified.In conjunction with above-mentioned analysis it is found that if can not get people by each group camera Face feature, it is believed that there are certain crime suspicion for the target person in corresponding movement routine, although face characteristic can not at this time Know, but mobile phone IMSI number has determined that, therefore can determine the identity of corresponding suspect according to mobile phone IMSI number.And it So this is allowed for having the people of criminal record to match in police's information bank, compared with the personnel of not criminal record, There is a possibility that people of criminal record commits a crime bigger, therefore this groups of people is preferentially checked.Wherein, according to mobile phone IMSI number The method for determining identity can refer to embodiment 2, be based on the corresponding relationship of " people-vehicle-IMSI ", only it is to be understood that mobile phone IMSI number, i.e., It can determine other perception data information of suspect, and then determine identity.

Further, for any movement routine, if got by preferably group camera or secondary choosing group camera The face characteristic of effect, then can be directly by the face characteristic progress for the people for having criminal record in face characteristic and police's information bank Match；If successful match, the people of Xiang Yingyou criminal record is classified as suspect, and finally determines the identity of suspect.Or first The identity of corresponding target person is determined according to the face characteristic got, then is carried out with the people for having criminal record in police's information bank Matching；If successful match, corresponding target person is classified as suspect, and finally determine the identity of suspect.This is equally examined Consider, compared with the personnel of not criminal record, have criminal record people commit a crime a possibility that it is bigger, therefore to this groups of people into Row major investigation.

In the above scheme, for any movement routine, determining whether corresponding target person is possible to as suspect When, the method can also further comprise: the historical data acquired by fence equipment or camera obtains corresponding movement The historical movement track of target person on path；It determines between corresponding target person current movement routine and historical movement track Relevance；If relevance is unsatisfactory for preset requirement, corresponding target person is classified as suspect, and pass through face characteristic or mobile phone IMSI number determines the identity of suspect.

Wherein, the magnanimity perception data accumulated all the year round in each movement routine is stored in perception data system, In also have recorded the historical movement track of the pedestrian occurred in each movement routine, historical movement track can be based on face The target following of feature is realized, is also possible to the target following based on mobile phone IMSI number and is realized.For any movement routine, if The face characteristic of target person can obtain, then can go in perception data system to match based on face characteristic corresponding with target person Multiple historical movement tracks；If the face characteristic of target person can not obtain, but since mobile phone IMSI number can obtain, then may be used It goes in perception data system to match multiple historical movement tracks corresponding with target person based on mobile phone IMSI number.Obtain history After motion profile, can relevance between the current movement routine and historical movement track of goal in research personage, that is, judge Whether target person is in history often along current movement routine activity, if being more than certain time along the movable number of current movement routine Number, the then it is believed that relevance of the two meets preset requirement；At this time in view of target person in history often along current movement routine A possibility that activity, this is the regular course of target person for deducibility, crime escape is smaller, therefore can be tentatively by corresponding target person The possibility that object excludes as suspect.On the contrary, if target person is in history never or almost without living along current movement routine Moved, then the relevance between current movement routine and historical movement track is unsatisfactory for preset requirement, deducibility this be not target The regular course of personage, it is likely to the route selected after interim crime escape, therefore can tentatively be classified as corresponding target person Suspect, and the face characteristic by obtaining or mobile phone IMSI number determine the identity of suspect.

In conclusion in perception data analysis method provided in an embodiment of the present invention, crime place awareness apparatus without In the case that method effectively locks suspect, fence equipment around crime ground to crime all target person cases are first passed through Movement routine after hair carries out trace analysis, then cooperate the camera device in each movement routine further determine that suspect with And the movement track of suspect.Wherein, before being analyzed using camera, previously according to shooting feature to taking the photograph on each path As head is grouped, so that the police can carry out information inquiry according to priority orders when studying and judging, i.e., picture quality is best is taken the photograph It is checked as head can be transferred preferentially, more targetedly, to greatly improve efficiency and accuracy that merit is studied and judged, can help to warn Square personnel quickly, accurately determine suspect's identity or the movement track of suspect.

Embodiment 4:

In above-described embodiment 3, it is contemplated that mobile phone is all carried in most people's trip, and the police can be set by fence It is standby, target following is carried out to each cell phone apparatus that crime time crime place occurred, and then determine the movement of each target person Path.But, it is contemplated that some suspects may have certain counterreconnaissance consciousness, for the tracking for hiding the police, deliberately make Not carrying mobile phone when case, therefore target tracking can not be carried out to this part suspect by cellphone information.For above situation, originally The orbit generation method that inventive embodiments provide a kind of specific suspect specifically includes as shown in figure 12:

Step 401, in conjunction with the fence equipment and camera in crime place, determine that the crime time occurs in crime place It crosses but one or more target persons of non-carrying mobile phone.

At crime time point, crime place usually will appear multiple people, and since most people can carry mobile phone, As long as being therefore provided with fence equipment to crime, with can be detected crime the corresponding cellphone information of the people of carrying mobile phone. The camera on crime ground everyone image information crime then can be detected, if image information only can be detected in someone, But without corresponding cellphone information, it may be determined that the people can be determined as one that this case is studied and judged by the non-carrying mobile phone of the people Target person.

Step 402, in conjunction with the characteristics of image of each target person, using multiple cameras around crime place to each target Personage carries out target following, and then obtains movement routine of each target person after the incident in preset time period respectively.

Herein, the characteristics of image of target person refers mainly to the figure and features feature reflected in image, as height, figure, It hair style and dresses, described image feature is mainly determined by the camera device in crime place.Characteristics of image is true After fixed, multiple cameras around crime place can carry out target following to each target person according to corresponding characteristics of image, To obtain the movement routine after each target person crime in preset time period.

For example, with reference to Figure 10, crime place in the crime time passes through fence equipment and camera device combines There are three the non-carrying mobile phones of people for discovery after analysis, are classified as target person first, second, third respectively；Further by around crime ground After multiple cameras carry out target following to these three target persons, corresponding three movement routines a, b, c are obtained respectively as schemed It is shown.Wherein, the preset time period can be configured according to the demand of studying and judging of the police, for example, if the police need to obtain case After hair in 2 hours suspect activity condition, then the preset time period is set as 2 hours.

Step 403, according to the shooting feature of camera, multiple cameras in each movement routine are grouped respectively, To obtain the preferred group of camera and time choosing group camera corresponding to each target person.

In the corresponding movement routine of each target person, multiple cameras usually can be all provided with, determine movement routine Afterwards, related platform can obtain and match automatically each camera of corresponding movement routine on the way；By being carried out to these cameras Specific figure information inquiry, can further check out the information of suspect.And specific portrait is inquired transferring each camera Before information, to study and judge the time convenient for saving police personnel, the camera in each movement routine is divided in advance Group.Wherein, group basis is the shooting feature of camera, and the shooting feature includes: the shooting direction of camera, shooting model Enclose, shoot it is one or more in resolution ratio, shooting height and shooting angle, it can only according to an one of them feature be grouped, Also it is grouped in combination with multinomial feature therein.In embodiments of the present invention, mainly for being divided into two groups, i.e., preferred group and time choosing Group after being finally completed grouping, is theoretically preferably organized corresponding overall image quality and should be better than and time select group, can correspondingly obtain effectively The probability of information selects group also greater than secondary, therefore ought to preferentially be inquired.

Step 404, it, to the priority orders of secondary choosing group, inquires and accordingly takes the photograph according to from preferred group in each movement routine respectively As the figure information that head acquires, until determining the identity of suspect or determining the action rail of suspect from multiple movement routines Mark.

In above step, multiple target persons based on non-carrying mobile phone obtain a plurality of movement routine, but not The target person for being each non-carrying mobile phone is suspect, is also needed through the further row of screening of the camera in each movement routine It looks into；It is checked by screening, suspect can be determined from multiple target persons, corresponding movement track has also been determined that, to rear Continue tracking and arrest suspect and has great importance.The probability that effective information can be obtained due to preferably organizing camera is greater than time choosing Group, therefore preferred group can be first inquired, it is not necessarily to continue to inquire time choosing group camera if getting effective information；It only obtains not Just continue to inquire time choosing group camera when to effective information.In this way, without to each camera on each path into Row is transferred, and can greatly save the police studies and judges the time.

In the orbit generation method of specific suspect provided in an embodiment of the present invention, in conjunction with the fence equipment on crime ground And camera device, can filter out those may have counterreconnaissance to realize and the suspect of non-carrying mobile phone, and then be directed to this portion Suspect is divided to carry out special path analysis and identity investigation；Multiple cameras in movement routine are grouped in advance simultaneously, and The priority orders that each group camera is queried are set, figure information can be carried out according to priority orders when carrying out case and studying and judging Inquiry, improve the merit efficiency and accuracy studied and judged.

It, specifically can be by studying and matching the mobile data of mobile phone IMSI number and the mobile number of people in the step 401 According to come the target person of with finding out crime non-carrying mobile phone.The then fence equipment and camera in combination crime place, Determine the crime time crime place occurred but one or more target persons of non-carrying mobile phone (i.e. step 401) can join Figure 13 is examined, is specifically included:

Step 4011, by the fence equipment in crime place, it is more to determine that the crime time occurred in crime place The corresponding mobile data of a mobile phone IMSI number.

By fence equipment, the mobile phone IMSI number within sensing range can be detected, and track out each mobile phone IMSI Number corresponding mobile data；Wherein, the mobile data includes the moving direction and/or movement speed in crime place.

Step 4012, by the camera device in crime place, it is multiple to determine that the crime time occurred in crime place The corresponding mobile data of personage.

By camera device, the character image within sensing range can be detected, and it is corresponding to track out each character image Mobile data；Wherein, the mobile data includes the moving direction and/or movement speed in crime place.

Step 4013, by the mobile data of the multiple mobile phone IMSI number and the progress of the mobile data of the multiple personage Match, filter out the crime time occurred but one or more target persons of non-carrying mobile phone in crime place.

After the mobile data that two kinds of awareness apparatus are got is matched, for the people of carrying mobile phone, character image Mobile data can be with the mobile data successful match of corresponding mobile phone IMSI number；And for the people of non-carrying mobile phone, then only pair The mobile data of character image is answered, the mobile data of corresponding mobile phone IMSI number can not be matched to.According to the above rule, can sieve Select the crime time occurred but the people of non-carrying mobile phone in crime place, and the target person as follow-up study.

In the step 402, it is assumed that be only that foundation is grouped with the shooting direction of camera.This is allowed for, from The shooting direction of camera is analyzed, if the front of target person can be taken, obtains the probability of effective face characteristic more It is larger, therefore shooting direction is had into important reference significance as group basis.In this case, the basis is taken the photograph As the shooting feature of head, multiple cameras in each movement routine are grouped respectively, to obtain corresponding to each target person (i.e. step 402) specifically refers to associated description and figure in embodiment 3 for preferred group of camera of object and time choosing group camera 11, comprising the following steps:

Step 3021, the shooting direction of each camera in each movement routine is obtained, and the mobile road where with each camera Diameter is matched.

With reference to Figure 10, by taking the movement routine c where target person third as an example, the multiple cameras being arranged on the path are as schemed Shown in middle dot (including black circle and soft dot).It determines on the path after the shooting direction of each camera, with movement Path c matching, can determine, during target person third normally moves forward, each camera is for shooting after matching Front, side or the back side of target person third.For example, target person third is to travel eastwards on movement routine c, then if The shooting direction of camera is towards west, or has the subtle deviation of directivity, can theoretically take the front of target person third just, And the camera of other directions then can not theoretically take the front of target person third, may take side or the back side.

Step 3022, for each camera in each movement routine, shooting direction is met to the positive camera of personage It is classified as preferred group, shooting direction is met and is classified as time selecting group to the non-frontal camera of personage.

Continue to refer to figure 10, still by taking the movement routine c where target person third as an example, in figure solid dot expression meet to The positive camera of personage is classified as preferably organizing camera；The camera non-frontal to personage is met in hollow dot expression, is classified as secondary The priority that choosing group camera, preferably group are queried selects group higher than secondary.It is to be divided into camera using shooting direction as foundation herein For two groups, it certainly, in other alternative-embodiments, may also be combined with more foundations, camera be divided into more multiple groups.Example Such as, it is assumed that using shooting direction and resolution ratio as foundation, the camera on each path is divided into four groups, then can be arranged as follows: the One group of camera is high-resolution, and shooting direction is positive towards target person；Second group of camera is low resolution, but is shot Direction is positive towards target person；Third group camera is high-resolution, but shooting direction is non-frontal towards target person；4th Group camera is low resolution, and shooting direction is non-frontal towards target person；It theoretically analyzes, the image matter of four groups of cameras Amount successively reduces, and the probability that can obtain effective information also successively reduces, therefore the priority orders being queried successively reduce.Its In, the differentiation of high-resolution and low resolution specifically can be using a certain preset value as line of demarcation, be higher than the preset value High-resolution is low resolution lower than the preset value, it is not limited here.

In another optional scheme, it is contemplated that can coverage also have a major impact to get effective information, Therefore when being grouped to multiple cameras in each movement routine, in addition to preferred group camera and time choosing group camera, institute Stating grouping may also include alternative group camera.Wherein, described alternative compared with the preferred group of camera and time choosing group camera The installation site of group camera is special, coverage is hidden, and the priority orders being queried in three groups of cameras are minimum.Example Such as, the preferred group of camera and time choosing group camera, which are normally located on street, waits visibility points, it is easier to take The target person passed by, and the alternative group of camera may be on some enclosure wall being arranged in by street, or by grove etc. Barrier blocks, therefore coverage is more covert, when target person is on the way difficult to be photographed when driving, obtains effective information A possibility that it is very low, therefore the priority orders being queried are minimum.But in some special cases, it is possible to taking Target person, for example, when target person climb enclosure wall just or be hidden in grove it is medium may just can by alternatively group camera clap It takes the photograph, may be got at this time than preferably organizing camera and time more valuable information of choosing group camera, therefore still have certain Reference significance.

Further, described respectively in each movement routine according to from preferred group to the priority orders of secondary choosing group, inquiry The figure information of corresponding camera acquisition, until determining the identity of suspect or determining suspect's from multiple movement routines Movement track (i.e. step 404), specifically: for each movement routine, the people of the corresponding preferably group camera acquisition of preferential inquiry As information, to obtain the face characteristic of target person in corresponding movement routine.If can not be got by preferably group camera Face characteristic then continues the figure information of corresponding choosing group camera acquisition of inquiry；If still can not by time choosing group camera Face characteristic is got, then corresponding target person is tentatively classified as suspect, corresponding movement routine primarily determines as suspicion Movement track after people's crime.

With reference to Figure 10, by taking the movement routine c where target person third as an example, if target person third is normal along movement routine Traveling, theoretically positive effective face characteristic can be collected by preferably organizing camera (i.e. black circle), at this time just without Time choosing group camera need to be transferred again.But if target person third exist under steam face block, bow, the feelings such as side rotary head Farther out etc., preferably group camera can not may just collect face characteristic for condition or shooting distance, just transfer time group is selected to take the photograph at this time As head further confirms that.Theoretical last time choosing group camera (i.e. soft dot) can not collect positive face characteristic, but if There is situations such as side rotary head, turn one's head in target person third, may take front by choosing group camera just, therefore still under steam There is certain reference value.If can not still get effective face characteristic, very possible target by choosing group camera Personage third has deliberately blocked face or has avoided camera, in addition the non-carrying mobile phone of target person third, it is believed that there are certain works Case suspicion, therefore it can be tentatively classified as suspect, corresponding movement routine primarily determines as the movement track after suspect's crime.

Wherein, for any movement routine, if can not still get face characteristic by time choosing group camera described, After corresponding target person is tentatively then classified as suspect, the method also includes: according to the camera shooting in corresponding movement routine Head obtains the information of vehicles of corresponding suspect, and then the identity of corresponding suspect is determined according to information of vehicles, and with police's information There is the people of criminal record to match in library；If successful match, the identity of corresponding suspect is verified.It can in conjunction with the above analysis Know, for any movement routine, if can not get face characteristic by each group camera, it is believed that corresponding movement routine On target person there are certain crime suspicion, and face characteristic can not be known at this time, and cellphone information, which can not also be known, (not to be taken Band mobile phone), therefore, the identity of corresponding suspect can be further determined by information of vehicles.Certainly, this method is suitable for disliking The case where doubtful people drives to go on a journey (specifically can be the vehicle at vehicle bayonet by the cam device in corresponding movement routine Camera), information of vehicles of corresponding suspect, such as license plate number, vehicle color, vehicle model etc. can be obtained, and then determine mesh Mark identity of personage.Wherein, embodiment 2 can refer to according to the method that information of vehicles determines identity, based on " people-vehicle-IMSI " Corresponding relationship, only it is to be understood that information of vehicles, that is, can determine the other information of suspect, and then determine identity.And why with police There is the people of criminal record to match in square information bank, this is allowed for, compared with the personnel of not criminal record, before having crime A possibility that people's crime of section, is bigger, therefore is preferentially checked to this groups of people.

Further, for any movement routine, if got by preferably group camera or secondary choosing group camera The face characteristic of effect, then can be directly by the face characteristic progress for the people for having criminal record in face characteristic and police's information bank Match；If successful match, the people of Xiang Yingyou criminal record is classified as suspect, and finally determines the identity of suspect.Or first The identity of corresponding target person is determined according to the face characteristic got, then is carried out with the people for having criminal record in police's information bank Matching；If successful match, corresponding target person is classified as suspect, and finally determine the identity of suspect.This is equally examined Consider, compared with the personnel of not criminal record, have criminal record people commit a crime a possibility that it is bigger, therefore to this groups of people into Row major investigation.

In the above scheme, for any movement routine, determining whether corresponding target person is possible to as suspect When, the method can also further comprise: the historical data acquired by camera, obtain target person in corresponding movement routine Historical movement track；Determine the relevance between corresponding target person current movement routine and historical movement track；If association Property be unsatisfactory for preset requirement, then corresponding target person is classified as suspect, and the identity of suspect is determined by face characteristic.

Wherein, the magnanimity perception data accumulated all the year round in each movement routine is stored in perception data system, In also have recorded the historical movement track of the people occurred in each movement routine, historical movement track can be based on face spy The target following of sign is realized, is also possible to the target following based on mobile phone IMSI number and is realized.For any movement routine, if mesh The face characteristic of mark personage can obtain, then can go in perception data system to match based on face characteristic corresponding with target person Multiple historical movement tracks.After obtaining historical movement track, can goal in research personage current movement routine and historical movement Relevance between track, that is, target person is judged in history whether often along current movement routine activity, if along when Forward The number of dynamic path activity is more than certain number, then it is believed that the relevance of the two meets preset requirement；At this time in view of target person A possibility that object is in history often along current movement routine activity, this is the regular course of target person for deducibility, crime escape It is smaller, therefore corresponding target person tentatively can be excluded into the possibility for suspect.On the contrary, if target person in history never Or almost without along current movement routine activity, then the relevance between current movement routine and historical movement track is unsatisfactory for pre- If it is required that deducibility this be not the regular course of target person, it is likely to the route selected after interim crime escape, therefore Corresponding target person tentatively can be classified as suspect, and the face characteristic by obtaining determines the identity of suspect.

Wherein, perception data analysis method provided in an embodiment of the present invention is the equal of the supplement to method in embodiment 3, Embodiment 3 carries out analysis investigation mainly for the people of carrying mobile phone, and embodiment 4 is then carried out mainly for the people of non-carrying mobile phone Investigation, mutual cooperation can check all possible suspects.

In conclusion awareness apparatus of the embodiment of the present invention in crime place can not effectively lock suspect, and suspect For hide the police trace and deliberately without mobile phone in the case where, first in conjunction with the fence equipment and camera device on crime ground, Filtering out those may have counterreconnaissance to realize and the suspect of non-carrying mobile phone, and then obtain the mobile road of this part suspect Then diameter cooperates the camera device in each movement routine to determine the movement track of suspect and suspect.Wherein, it is utilizing Before camera is analyzed, the camera on each path is grouped previously according to shooting feature, so that the police are when studying and judging Information inquiry can be carried out according to priority orders, i.e. the best camera of picture quality can be transferred preferentially and be checked, is more directed to Property, to greatly improve efficiency and accuracy that merit is studied and judged, police personnel can be helped quickly, accurately to determine the suspicion person The movement track of part or suspect.

Embodiment 5:

On the basis of each method that above-described embodiment 1- embodiment 4 provides, the present invention also provides one kind can be used for reality The device of the existing above method is as shown in figure 14 the device architecture schematic diagram of the embodiment of the present invention, one or more processors 21 And memory 22.Wherein, in Figure 14 by taking a processor 21 as an example.

The processor 21 can be connected with the memory 22 by bus or other modes, by total in Figure 14 For line connection.

The memory 22 is used as a kind of non-volatile computer readable storage medium storing program for executing, can be used for storing non-volatile software Program, non-volatile computer executable program and module, such as the orbit generation method of the specific suspect in embodiment 4. Non-volatile software program, instruction and the module that the processor 21 is stored in the memory 22 by operation, thus Application and data processing are performed various functions, that is, realizes each method of embodiment 1- embodiment 4.

The memory 22 may include high-speed random access memory, can also include nonvolatile memory, such as At least one disk memory, flush memory device or other non-volatile solid state memory parts.In some embodiments, described Optional memory 22 includes the memory remotely located relative to the processor 21, these remote memories can pass through network It is connected to the processor 21.The example of above-mentioned network includes but is not limited to internet, intranet, local area network, moves and lead to Letter net and combinations thereof.

Described program instruction/module is stored in the memory 22, is held when by one or more of processors 21 When row, each method in above-described embodiment 1- embodiment 4 is executed, for example, executing each step shown in the drawings described above.

Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of embodiment is can to lead to Program is crossed to instruct relevant hardware and complete, which can be stored in a computer readable storage medium, storage medium It may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, RandomAccess Memory), disk or CD etc..

The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.

Claims (10)

1. a kind of orbit generation method of specific suspect characterized by comprising

In conjunction with the fence equipment and camera in crime place, determine that the crime time occurred in crime place but do not carried hand One or more target persons of machine；

In conjunction with the characteristics of image of each target person, target is carried out to each target person using multiple cameras around crime place Tracking, and then movement routine of each target person after the incident in preset time period is obtained respectively；

According to the shooting feature of camera, multiple cameras in each movement routine are grouped respectively, to be corresponded to In the preferred group of camera and time choosing group camera of each target person；

, to the priority orders of secondary choosing group, the people of corresponding camera acquisition is inquired according to from preferred group in each movement routine respectively As information, until determining the identity of suspect or determining the movement track of suspect from multiple movement routines.

2. the orbit generation method of specific suspect according to claim 1, which is characterized in that combination crime place Fence equipment and camera, determine that the crime time occurred but one or more mesh of non-carrying mobile phone in crime place Personage is marked, is specifically included:

By the fence equipment in crime place, multiple mobile phone IMSI numbers pair that the crime time occurred in crime place are determined The mobile data answered；

By the camera device in crime place, the corresponding movement of multiple personages that the crime time occurred in crime place is determined Data；

The mobile data of the multiple mobile phone IMSI number is matched with the mobile data of the multiple personage, filters out crime Time occurred but one or more target persons of non-carrying mobile phone in crime place；

Wherein, the mobile data includes the moving direction and/or movement speed in crime place.

3. the orbit generation method of specific suspect according to claim 1, which is characterized in that the shooting of the camera Feature includes: one or more in shooting direction, coverage, shooting resolution ratio, shooting height and the shooting angle of camera ?.

4. the orbit generation method of specific suspect according to claim 1, which is characterized in that described according to camera Feature is shot, multiple cameras in each movement routine are grouped respectively, to obtain corresponding to the excellent of each target person Choosing group camera and time choosing group camera, specifically include:

The shooting direction of each camera in each movement routine is obtained, and is matched with the movement routine where each camera；

For each camera in each movement routine, shooting direction is met to the positive camera of personage and is classified as preferred group, it will Shooting direction, which is met, to be classified as time selecting group to the non-frontal camera of personage.

5. the orbit generation method of specific suspect according to claim 1, which is characterized in that on to each movement routine Multiple cameras when being grouped, the grouping further includes alternative group camera；Wherein, with the preferred group of camera and time Choosing group camera is compared, and the coverage of the alternative group of camera is hidden, and the priority being queried in three groups of cameras It is sequentially minimum.

6. the orbit generation method of specific suspect according to claim 1, which is characterized in that described respectively in each movement , to the priority orders of secondary choosing group, the figure information of corresponding camera acquisition is inquired, until determining according to from preferred group on path The identity of suspect or the movement track that suspect is determined from multiple movement routines, specifically include:

For each movement routine, the figure information of the corresponding preferably group camera acquisition of preferential inquiry, to obtain corresponding movement road The face characteristic of target person on diameter；

If face characteristic can not be got by preferably group camera, continue the people of corresponding choosing group camera acquisition of inquiry As information；

If can not still get face characteristic by time choosing group camera, corresponding target person is tentatively classified as suspicion People, corresponding movement routine primarily determine as the movement track after suspect's crime.

7. the orbit generation method of specific suspect according to claim 6, which is characterized in that for any mobile road Corresponding target person is tentatively classified as by diameter if can not still get face characteristic by time choosing group camera described After suspect, the method also includes:

The information of vehicles of corresponding suspect is obtained according to the camera in corresponding movement routine, and then according to determining pair of information of vehicles Answer the identity of suspect, and with there is the people of criminal record to match in police's information bank；If successful match, suspect is corresponded to Identity be verified.

8. the orbit generation method of specific suspect according to claim 6, which is characterized in that for any mobile road Diameter, if face characteristic is got by preferably group camera or secondary choosing group camera, by face characteristic and police's information bank In there is the people of criminal record to match；If successful match, the people of Xiang Yingyou criminal record is classified as suspect, and final true Determine the identity of suspect.

9. the orbit generation method of specific suspect according to claim 6, which is characterized in that for any mobile road Diameter, the method also includes:

The historical data acquired by camera obtains the historical movement track of target person in corresponding movement routine；

Determine the relevance between corresponding target person current movement routine and historical movement track；If relevance is unsatisfactory for presetting It is required that corresponding target person is then classified as suspect, and the identity of suspect is determined by face characteristic.

10. a kind of track generating means of specific suspect, which is characterized in that including at least one processor and memory, institute State between at least one processor and memory by data/address bus connect, the memory be stored with can by it is described at least one The instruction that processor executes, described instruction by the processor after being executed, for completing any spy of claim 1-9 Determine the orbit generation method of suspect.