CN204244385U - A kind of three layers of pluggable intelligent video information management system - Google Patents

Abstract

A kind of three layers of pluggable intelligent video information management system, the camera head of number of different types is comprised in this video information management system, local management server LUMS, dummy unit management server VUMS and global server GS, described camera head is evenly distributed on multiple geographical position, each camera head has a geographical position label G L and functional label FL, there is the public FL of multiple camera heads of identical function, the ground floor of this system is made up of the camera head of comprised number of different types and LUMS, the second layer is made up of described camera head and VUMS, third layer is by LUMS and VUMS, and GS composition.By three-decker, can ensure at any time and can effectively access Anywhere data and the information of any camera.

Description

A kind of three layers of pluggable intelligent video information management system

Technical field

The utility model relates to video information administrative skill field, particularly a kind of three layers of pluggable intelligent video information management system.

Background technology

A video system, such as, its fail safe is usually directed to camera dissimilar in a large number as shown in Figure 1.At one time, this system can have multiple use, such as, except fail safe also has automatic market survey function.Each camera can have different output formats (h264, MJPEG coding, MXPG etc.), and different signal type (analog signal, live wire 1394 signal, IP-based signal).The output of all these cameras can better go out detection anomalous event or hazard event continuously by automatic analysis.Once suspicious event be detected, record evidence and should related personnel be notified.The more important thing is, when event notifies, operator and the people offered help should collect and examine in a kind of efficient mode all evidences and information that event is occurring, so just can take suitable action within the shortest time.Therefore, efficient information and data re-examine should be very important aspects for a successful system for managing video.This depends on an effective architecture of camera management.

On the other hand, IP-based video management supplier, successfully sets up the open system for managing video that various video/audio can be supported to input.But, these systems and new video/audio parser incompatible.All video/audio analysis modules need to set up, and are merged in inside by corresponding supplier.If supplier can not provide required module, the requirement of user will be reached never.Recognize to there is this problem, the product supplier as HALCON and VisiQuest makes great efforts to provide a simple prototype environment vision algorithm for client.But if block merging is possible, existing code is integrated is very difficult all the time.MATLAB is a more complete and algorithm prototype environment for simple vision.But based on the object of prototype, user must buy expensive platform.And the module be shaped always is run very slow.In fact, in popular computer vision storehouse, there is many useful vision algorithms and data structure, as OpenCV and RAVL.All these prototype environments and processing module are mainly individual unit process.Recently, there are some to attempt, similar iceWing project, by with for the purpose of the prototype distributed vision system of Plugin Mechanism.This trend meets the development of cloud computing in Internet service.Bottom communication framework XCF success at European VAMPIRE project application in large-scale vision system.But above all effort concentrate on development vision algorithm, but more or less ignore their deployment.Need the platform that unified, can ideally the vision module of maturation be combined when development stops.

Therefore how to be combined by these vision modules, be deployed in a unified platform, be that people are in the problems of making great efforts to solve always.

Utility model content

In view of the foregoing defects the prior art has, the purpose of this utility model is to provide a kind of compatibility existing all types camera on the market, these vision modules combine, and are deployed in three layers of pluggable intelligent video information management system on a unified platform.

For reaching above-mentioned purpose, the utility model takes following scheme:

A kind of three layers of pluggable intelligent video information management system, the camera head of number of different types, local management server LUMS, dummy unit management server VUMS and global server GS is comprised in this video information management system, described camera head is evenly distributed on multiple geographical position, each camera head has a geographical position label G L and functional label FL, there is the public FL of multiple camera heads of identical function, it is characterized in that:

The ground floor of described three layers of pluggable intelligent video information management system is made up of the camera head of comprised number of different types and local management server LUMS, wherein said camera head divides into groups according to respective geographical position label G L, each group camera head is managed by a LUMS, the Data Post that LUMS manages the acquisition of each camera head is useful information, and controls the configuration operation of camera head;

The second layer of described three layers of pluggable intelligent video information management system is made up of the camera head of comprised number of different types and dummy unit management server VUMS, the described camera head wherein with identical function label F L is assigned to a dummy unit, each each dummy unit is managed by a dummy unit management server VUMS, and the data that VUMS is camera head export, store and information channel provides reference;

The third layer of described three layers of pluggable intelligent video information management system is by all local management server LUMS and dummy unit management server VUMS, and global server GS composition.

Preferably, this video information management system comprises multiple intelligent object IM, and this intelligent object IM comprises processing module and display module, and wherein processing module is used for Realization analysis task, and the video data from described camera head received is converted into information; Display module is for showing treated output information.

Preferably, described intelligent object IM using receive video data and configuration parameter as input information, video data is original rgb format;

Described intelligent object IM provides the parameter definition file of XML format, the parameter value of package types and use is defined in this parameter definition file, and each parameter has a unique ID associated with it and title, can automatically for user generates the parameter configuration page according to described parameter definition file video information management system, the data that processing module in described intelligent object IM exports encapsulate in xml format, and wherein the information of XML format can be transmitted by the interfacing between network of unified support http protocol;

Display module is the screen that directly can incorporate client.

Preferably, this video information management system comprises multiple intelligent object IM, and this intelligent object IM is made up of input module, output module, memory setting module and controling parameters module;

Input module, being arranged to obtain needs one or more image to be processed;

Output module, is arranged to and result is exported;

Controling parameters module, is configured to each module installation parameter;

Memory setting module, is arranged to the buffer memory allocated each module in advance and need to use, and issues the address of the buffer memory needing to use to each module.

Preferably, this video information management system comprises multiple intelligent object IM, and this intelligent object IM is by atom basic module ABM and form basic module CBM two generic module and form;

Atom basic module ABM, this generic module is not based upon in other modules;

Composition basic module CBM, this generic module is based upon on atom basic module ABM;

Described atom basic module ABM and composition basic module CBM derives from a basic module storehouse.

The utility model has the following advantages and good effect:

1) this management system can compatibility existing all types camera on the market.Although camera may have different types and different formatted outputs, the quantity of type and the form of output are confined.Therefore, this system all can provide support for them, even if new type or new form occur suddenly, can it be merged rapidly, all LUMS and VUMS are upgraded simultaneously, make end user and third party developer not need to be concerned about that this part works.

2) there is wide-open chip card.Third party developer can develop the plug-in unit of this system, as long as he/her adheres to that we are the consistency rule of Operation system setting.By the exploitation of plug-in unit, he/her can provide extra intelligent behaviour to system, and without the need to being concerned about that data are come wherefrom and where end product will be displayed on.

3) three-decker, can ensure at any time and can effectively access Anywhere data and the information of any camera.

Accompanying drawing explanation

Fig. 1 shows one and typically comprises multiple all kinds of camera video system;

Fig. 2 shows the schematic diagram being had effect spread after jumbo video information is converted into low capacity information by low broadband connection;

Fig. 3 shows all camera two-dimensional structure schematic diagrames with geographical position and functional mark;

Fig. 4 shows the total framework of three layer video management systems;

Fig. 5 shows the Video processing flow process of system for managing video;

Fig. 6 shows image and Video processing block diagram;

Fig. 7 shows " edge extracting ", and process reasonably realizes figure;

Fig. 8 is the schematic diagram in a basic module storehouse.

Embodiment

Below in conjunction with accompanying drawing, execution mode of the present utility model is described further.

As shown in Figure 1, although typical system for managing video may comprise multiple dissimilar video camera, they can gather together by certain mode by we.Its main thought is the space these video cameras being placed in various dimensions, and wherein each dimension has different tolerance, has had this to measure, and we can assemble all video cameras, and is divided into multiple group being easy to manage.Two most important indexs are:

1. geographical position: under normal circumstances, camera is evenly distributed on multiple geographical position.Camera can form an independently unit in a relatively intensive place.On the other hand, each camera can obtain a geographical position label G L, and it can be the unique ID of a correlation unit.Each unit is managed by local management server LUMS.LUMS manages all data that camera obtains, and is treated to useful information, controls position and the configuration of all cameras.Because broadband LAN LAN is available usually, the connection of each camera and LUMS is not only enough jumbo video data transmission.After analyzing data, the output information of LUMS is very little with regard to volume, therefore, it is possible to have effect spread by low broadband connection such as internet.

Local record voucher in corresponding LUMS, as dependent event information.The bridge communicated has been erected between each camera that LUMS also manages with it in the external world.Therefore, each LUMS should be controlled.This just means that each LUMS should be able to provide management and control to support the camera of all distribution.It is limited in one's ability that a usual LUMS runs on physical hardware, so each LUMS can manage at most NMAX camera.Each LUMS and its camera are combined closely by LAN.Different LUMS is by wide area network WAN loose couplings.Suppose that the cost of each LAN is CLAN, the cost of each LUMS is CLUMS, and the management cost of whole system is CM (nLUMS) (usually along with the quantity monotonic increase of LUMS, comprising variable cost, as WAN cost CWAN).So under the condition that each LUMS is controlled, the quantity nLUMS of required LUMS represents by with minor function:

$\underset{\mathrm{nLUMS}}{\arg \max }(n_{\mathrm{LUMS}}\·C_{\mathrm{LAM}}+C_M\left(n_{\mathrm{LUMS}}\right))$

Certainly in fact we also should consider the possibility that more restrictive condition connects as network.

2. function: each camera in video system exists wants certain reason.Have may for check ticketing service situation, have for detect doorway suspicious figure, also have can supervise tramway in case possible hazardous act (as electric car near time have people at runway).Therefore, each camera has a functional label FL.Multiple camera also public FL, a FL may be divided into multiple unit all cameras subsequently, and each unit is by identical FL.

According to above-mentioned two indices, we can form a two-dimensional structure as shown in Figure 3 to all cameras.Usually, when some causing danger property of function or the harmfulness event in a certain geographical position, just give the alarm.Therefore, we combine each event and composite attribute (GL, FL).According to this attribute, eject any event, we can be easy to review relevant camera by voucher and current state.

Based on this two-dimensional structure, we have proposed three layer video Governance frameworks, as shown in Figure 4:

Ground floor, divides into groups to all cameras according to GL attribute.LUMS manages the camera in a unit.

The second layer, the camera that FL is identical assigns to a dummy unit, is managed by dummy unit management server VUMS.All outputs of the camera data that each VUMS is it, storage and information channel provide reference.Its additional functionality is exactly that pure function relies on.

Third layer, all LUMS and VUMS are managed by a global server GS.By GS, various client (no matter be mobile phone, notebook or desktop computer) can locate any one LUMS or VUMS easily with a unified form.Behind location, two can realize communication by P2P technology or transmission server.

Certainly, in these three layers of frameworks, we can also use other based on the type of GL and FL index.Further, if necessary, whole framework can further expand to multilayer thus carry out more effective video management.

Plug-in unit is supported:

As mentioned above, a kind of effective and open system for managing video should support plug-in management.This not merely means the processing module that system can provide extra, is also easy to be enhanced simultaneously, also should give the more analysis module of third party's Development and Production under unified interface.In addition, system fast, seamlessly should be able to install new module.

Fortunately, although different cameras can have different signals to export, but all output signals can be converted to the signal of original RGB (or YUV) form, this signal again can as the input signal of each intelligent object IM (such as: recognition of face) as shown in Figure 5.In fact, each IM comprises two parts: a. process: this part carries out analysis task, and video data is converted into information; B. show: the output information of client process part is shown in this part.

IM ensures the hot plug of system by following four parts:

1) IM using receive video data and configuration parameter as input information.Video data is original rgb format (also can be YUV or other form).

2) provide parameter definition file by IM, this file is XML format, defines the probable value of the parameter of package types and use.At present, our limited field (integer in minimum and maximum scope) setting parameter type, and selective value (integer of limited field class).Each parameter has a unique ID associated with it and title.Had this parameter definition file, system automatically for user generates the parameter configuration page, and can revise the value of each parameter.

3) actual parameter is also encapsulated as XML format.This XML document can effectively by the transmission of unified interface (as XCF [11]) interconnection network.At present, we use HTTP service standard.

4) in addition, the information that processing section exports, encapsulates in xml format.This XML document can effectively by the transmission of unified interface (as XCF [11]) interconnection network.At present, we use HTTP service standard.

5) display section is unified window, directly can incorporate client screen.Internal logic program can process each independent IM completely., certainly, it may be displayed on the top of the display window that system provides, and avoids the basic function of repetition with this.

At present, each IM processing section is researched and developed as the DLL (dynamic link library) of windows platform, and display section shows as the container of Flash platform.Certainly these can by expansive approach to other platform.They are developed by third party and upload to GS to specify arrangement and deployment on LUMS or VUMS user.

IM modularization

In addition, each IM can by modularization.The process of clear and simple image and Video processing by modularization, as shown in Figure 6.Input and output module is all standard, because we only have the input channel of limited quantity and the possibility of output.In addition, so-called " process " module can be separated into some basic image processing techniquess.Such as, edge extracting process can realize the effect as Fig. 7.Wherein " convolution ", " algorithm calculations ", " thresholding " is the image processing techniques of all standards, and these can write as basic module in advance.These basic modules BM is the basis of a structure image procossing application program.

Each in them has the interface of 4 types to interact with outside:

Input: this is used to obtain needs image to be processed.Here 0 or multiple input picture may be had.Such as, the module of " generted noise image " does not have input picture, and " convolution " module has an input picture.As Fig. 7 " algorithm calculations " module has two input pictures.

Export: this is the passage that a transmission process result gives the next module in processing chain.Each module only has an output under normal circumstances.

Memory setting: in order to improve the validity of process, before whole process starts, each module needs the buffer memory used to be preassigned, and issues the available address of each module by this passage.

Controling parameters: each basic module is controlled by series of parameters, this series of parameters can the function of theatrical change module.And this series of parameters is by this channel setting.

In fact, its basic module can be divided into two classes:

Atom basic module (ABM): these basic modules are atoms.That is they are not based upon in other modules.They are just as leaf in the tree, and in the figure 7,3 kinds of all basic modules are these class basic modules

Composition basic module (CBM): this kind of basic module is based upon on other basic module.Only because they are very useful, therefore using them as basic module.They are just as the leaf node on one tree.Such as, in the figure 7, " edge extracting " process may be incorporated in together as a composition basic module.In our video frequency following system, " background estimating ", " foreground extraction ", " target following ", " target classification " can be built by as composition basic module.

All basic modules all derive from a basic module storehouse.Using all basic modules as node, its whole basic module storehouse is just as a unidirectional figure.

Fig. 8 illustrates the example in an image procossing basic module storehouse.As can be seen from this example we: 1.ABMs cannot inherit from other module; 2.CBM only can be based upon (as CBM01) on ABMs; 3.CBM can be based upon on ABMs and other CBM (as CBM03); 4. a CBM even can only be made up of (as CBM5) other CBMs; 5. in storehouse, there is not any circulation (such as: if A is based upon on B, so B is referred to as the child of A.Therefore, if D is the child of C offspring, so C should not be the child of D offspring).

Protection range of the present utility model is not limited to the above embodiments, and obviously, those skilled in the art can carry out various change and distortion to the utility model and not depart from scope and spirit of the present utility model.If these are changed and distortion belongs in the scope of the utility model claim and equivalent technologies thereof, then intention of the present utility model also comprises these changes and distortion.

Claims (5)

1. three layers of pluggable intelligent video information management system, the camera head of number of different types, local management server LUMS, dummy unit management server VUMS and global server GS is comprised in this video information management system, described camera head is evenly distributed on multiple geographical position, each camera head has a geographical position label G L and functional label FL, there is the public FL of multiple camera heads of identical function, it is characterized in that:

The ground floor of described three layers of pluggable intelligent video information management system is made up of the camera head of comprised number of different types and local management server LUMS, wherein said camera head divides into groups according to respective geographical position label G L, each group camera head is managed by a LUMS, the Data Post that LUMS manages the acquisition of each camera head is useful information, and controls the configuration operation of camera head;

The second layer of described three layers of pluggable intelligent video information management system is made up of the camera head of comprised number of different types and dummy unit management server VUMS, the described camera head wherein with identical function label F L is assigned to a dummy unit, each each dummy unit is managed by a dummy unit management server VUMS, the data that VUMS receives relevant camera head input and carry out the Intelligent treatment of respective settings, can store inputoutput data, inquiry and Intelligent Dynamic forward simultaneously;

The third layer of described three layers of pluggable intelligent video information management system is by all local management server LUMS and dummy unit management server VUMS, and global server GS composition.

2. according to layer pluggable intelligent video information management system of three described in claim 1, it is characterized in that: this video information management system comprises multiple intelligent object IM, this intelligent object IM comprises processing module and display module, wherein processing module is used for Realization analysis task, and the video data from described camera head received is converted into information; Display module is for showing treated output information.

3., according to layer pluggable intelligent video information management system of three described in claim 2, it is characterized in that:

Described intelligent object IM using receive video data and configuration parameter as input information, video data is original rgb format;

Described intelligent object IM provides the parameter definition file of XML format, the parameter value of package types and use is defined in this parameter definition file, and each parameter has a unique ID associated with it and title, can automatically for user generates the parameter configuration page according to described parameter definition file video information management system, the data that processing module in described intelligent object IM exports encapsulate in xml format, and wherein the information of XML format can be transmitted by the interfacing between network of unified support http protocol;

Display module is the screen that directly can incorporate client.

4. according to layer pluggable intelligent video information management system of three described in claim 1, it is characterized in that: this video information management system comprises multiple intelligent object IM, this intelligent object IM is made up of input module, output module, memory setting module and controling parameters module;

Input module, being arranged to obtain needs one or more image to be processed;

Output module, is arranged to and result is exported;

Controling parameters module, is configured to each module installation parameter;

Memory setting module, is arranged to the buffer memory allocated each module in advance and need to use, and issues the address of the buffer memory needing to use to each module.

5. according to layer pluggable intelligent video information management system of three described in claim 1, it is characterized in that: this video information management system comprises multiple intelligent object IM, this intelligent object IM is by atom basic module ABM and form basic module CBM two generic module and form;

Atom basic module ABM, this generic module is not based upon in other modules;

Composition basic module CBM, this generic module is based upon on atom basic module ABM;

Described atom basic module ABM and composition basic module CBM derives from a basic module storehouse.