CN109887098A - A kind of web AR data presentation mode based on distributed computing - Google Patents
A kind of web AR data presentation mode based on distributed computing Download PDFInfo
- Publication number
- CN109887098A CN109887098A CN201910116350.7A CN201910116350A CN109887098A CN 109887098 A CN109887098 A CN 109887098A CN 201910116350 A CN201910116350 A CN 201910116350A CN 109887098 A CN109887098 A CN 109887098A
- Authority
- CN
- China
- Prior art keywords
- model
- data
- client
- request
- server
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Information Transfer Between Computers (AREA)
Abstract
A kind of web AR data presentation mode based on distributed computing, the present invention relates to data presentation mode technical fields；The storage management that model and action data are distinguished in server end；According to the type of request, client sends request to neighbouring mobile Edge Server；Mobile Edge Server carries out demand parsing according to the request of user；Cloud computing end parses the interaction request of client, obtains matching movement and is sent to Edge Server；Data calculating is carried out at mobile edge calculations end；Mobile Edge Server is in the model for receiving the transmitting of cloud computing end；Calculation result data is returned into client, realizes the loading function of dynamic model.Greatly improve the operational efficiency of client；It services more flexible；The calculating response modes for more adapting to augmented reality mitigate the GPU operation caching pressure of mobile terminal, realize relative complex interaction models operation；The calculating pressure in cloud is alleviated, performance significantly improves.
Description
Technical field
The present invention relates to data presentation mode technical fields, and in particular to a kind of web AR data based on distributed computing
Presentation mode.
Background technique
With the propulsion of informatization and the development of intelligent mobile terminal equipment, the enhancing based on intelligent mobile terminal
Reality technology is that industry, business and tourist industry bring huge development opportunity.Meanwhile the mobile augmented reality based on web
The features such as technology is due to its portability, universality obtains more wide application in mobile augmented reality field.
But based on the mobile augmented reality of web compared with traditional mobile augmented reality technology, to users with
Come convenient simultaneously as the characteristic of its rendering based on browser, interacting operation is based on different from traditional mobile application
The rendering interacting operation of browser needs to transfer the operation of Intelligent mobile equipment bottom and real world devices by means of the kernel of browser
Interface；The browser of mobile terminal in the scheduling of memory and calculation resources there is certain limitation, this also indirect shadow
It has rung computing resource during mobile terminal is in augmented reality rendering, interacting operation to transfer, has influenced mobile web augmented reality
The building of interactive experience and service environment.
By above problem, at present in mobile augmented reality application, compression and transmission for model data are
Main research object, especially for the introducing of the network stream transmission in threedimensional model.It is being directed to network streaming at present
Transmission, domestic and international related researcher have been achieved for certain research achievement.Martin is proposed based on the adaptive of model
Star concept passes through different rate compact models respectively；Ren Huiling, Shen Yanchun are proposed under large scale scene for model data
Loading in terms of research and the improvement of streaming algorithm etc..
But current research is paid close attention to mostly based on the data compression of PC application environment drag and transmission loading mode,
The calculating pressure of client is mainly considered on being loaded into algorithm and transmission, loading mode, and mobile web augmented reality is answered
In, various factors such as user can consider flow cost during service application, be initially loaded the time, it is then desired to
Network streaming method is further optimized from initial loading efficiency, network transmission compression etc..
Summary of the invention
In view of the defects and deficiencies of the prior art, the present invention intends to provide a kind of structure is simple, design rationally, make
With the web AR data presentation mode easily based on distributed computing, the operational efficiency of client is greatly improved；Service is more
Add flexibly；The calculating response modes for more adapting to augmented reality mitigate the GPU operation caching pressure of mobile terminal, realize relative complex
Interaction models operation；The calculating pressure in cloud is alleviated, performance significantly improves.
To achieve the above object, the technical solution adopted by the present invention is that: its operating procedure is as follows:
1, the storage management for distinguishing model and action data in server end；First convert model data to pair
Image data needs to convert attribute pair for model data according to the definition for each key assignments in model in Transformation Application
The combination of elephant, is stored in server end；
2, according to the type of request, client sends request to neighbouring mobile Edge Server；Client is to service
When device end carries out page request, first by http agreement, association requests are sent to neighbouring mobile Edge Server；
3, mobile Edge Server carries out demand parsing according to the request of user；Whether decision model data carry out in advance
Routing, if model data is routed in advance, to cloud server request action data, carries out mould after wait-for-response
Type operation；If do not carried out content routing, then model and action data are requested to cloud, carries out model calculation, that is, cloud service
Device carries out the load of model, and the Multidimensional Data Model of received server-side to client is raw by model adaptation module and model
At module, rendering calculating is carried out；
4, the interaction request of cloud computing end parsing client obtains matching movement and is sent to Edge Server；In client
While sending interaction request, the interaction request of client is parsed, describes to carry out inquiry comparison in library in semantization,
Matched movement is obtained, mobile Edge Server is sent to；
5, data calculating is carried out at mobile edge calculations end；Mobile Edge Server is in the mould for receiving the transmitting of cloud computing end
Type, method of operating；By the data structure inside model, in a manner of program for the received model of institute and action data into
Row compound operation returns to model response data to server in data format；
6, calculation result data is returned into client, realizes the loading function of dynamic model；Edge calculations server with
New model generated is returned to client by the method quickly calculated by way of interface, realizes the business of response.
Further, the storage management for distinguishing model and action data in server end in the step 1
Particular content are as follows: the mass data storage model based on cloud computing, according to cloud computing core calculations mode MapReduce,
And the open source distributed parallel programming framework Hapdoop for realizing MapReduce calculating mode is relied on, by storage model and cloud computing
It is combined together, realizes the distributed storage of mass data, static models data required for user are deposited in this way
Data are mapped to different blocks by the program of Map function after data are divided, distribute to calculating by storage on cloud computing end
The processing of a machine group of planes achievees the effect that distributed arithmetic, to realize the effect of storage, and storage model in the database is dynamic
Make data, the action data of model is stored respectively with model.
Further, the request type particular content in the step 2 are as follows: in SmartClient request, client
Firstly the need of the bottom-up information of access equipment, the vision for obtaining equipment calculates relevant device parameter, as resolution ratio, screen size with
And the information such as CPU operation efficiency, construct the data model of request.
Further, the specific steps that request is sent to neighbouring mobile Edge Server in the step 2 are as follows: visitor
After family end sends request to server with http agreement, browser will monitor the solicited message of server-side return；Work as acquisition
After solicited message, the status code of return information is determined first, when status code is 200, it will grab the json format of request
Data, and store data in the memory of browser.
Further, the mobile Edge Server in the step 3 carries out the specific steps of demand parsing are as follows:
3.1, server is after acquisition request, the request data of analysis request client, carries out operation, determines client
Requesting state information, and determine the upper limit of model data presented under the conditions of optimal network；With under the conditions of the upper limit since M1,
It gradually to client transmission pattern back end set, and returns to receiving end request and receives to finish the spent time, calculate visitor
The network environment parameters at family end；This parameter passes to server end simultaneously, after server end receives this parameter, in determining M
Upper limit cutoff value is determined in sequence；
3.2, client parameter relevant to model load operation is constructed into a multi-Dimensional parameters model, carries characterization client
Operational capability, network bandwidth and the interactive display interface at end etc. parameter, are asked with physical parameter model to server-side
It asks, server-side will analyze the physical parameter of client, and the model that union goes out to be suitable for client returns to client
End carries out load and rendering operation, it will greatly improves the operational efficiency of client.
Further, the cloud server in the step 3 carries out the specific steps of the load of model are as follows: server-side connects
The multi-dimensional data model of client is received, model data parsing is carried out, obtains physics relevant to client current operating conditions
Parameter passes to model adaptation module, and model adaptation module parses the physical parameter of client, TRANSFER MODEL output
Weight gives model generation module；Model generation module will carry out model data generation processing according to weight, and by processing result
Client is returned in the form of interface, carries out rendering calculating.
Further, the specific steps of the interaction request of the cloud computing end parsing client in the step 4 are as follows: cloud meter
Calculation center will request to parse for the operation of client, obtain the physical environment model of client, and play cloud computing
High concurrent data edge, calculate the weight of client request model, the mobile side nearest with client be routed to by network
Edge server carries out corresponding model request calculating；Server end also needs the service request according to user simultaneously, is responded
Logical response, and the result of response is realized into the building of corresponding business scenario by returning to client.
Further, the acquisition matching movement in the step 4 is sent to the specific steps of Edge Server are as follows: in cloud
End, according to business demand, first static models required for storage service；Meanwhile the movement number of storage model in the database
According to；While for storage model action data, the description for carrying out semantization is acted for model, is sent and is interacted in client
While request, the interaction request of client is parsed, describes to carry out inquiry comparison in library in semantization, obtains matching
Movement, be sent to mobile Edge Server.
Further, the mobile edge calculations end in the step 5 carries out the specific steps of data calculating are as follows:
5.1, the mobile edge calculations under 5g application scenarios, relative to traditional with centralization, the large-scale cloud for turning to core
Calculating is compared, and is more adaptive to the calculating response modes of mobile augmented reality；Mobile edge calculations, by operation from traditional cloud
It calculates central server cluster and has been transformed into the edge closer from mobile terminal, in network circulation, avoid from cloud computing center
To the occupied Internet resources of network edge and consumed transmission time, meanwhile, by most operand from center service
Device has been transformed into mobile network edge, it will improves the operational capability of web augmented reality to a certain extent；Pass through cloud simultaneously
Calculating and mobile edge calculations carry out the mode of collaboration operation, mitigate the GPU operation caching pressure of mobile terminal, realize for again
The operational capability of miscellaneous interaction models；In interface operation framework mode based on mobile edge calculations, it will with traditional center
Function center of the server cluster as client traffic request response, to move the service of Edge Server as model calculation
Center；Client first sends the operation request for the physical parameter for carrying client to server-side under corresponding service environment,
Cloud computing center will request to parse for the operation of client, obtain the physical environment model of client, and play cloud
The high concurrent data edge of calculating calculates the weight of client request model, the shifting nearest with client is routed to by network
Dynamic Edge Server carries out corresponding model request calculating；Server end also needs the service request according to user simultaneously, carries out
The logical response of response, and the result of response is realized into corresponding business scenario building by returning to client；
5.2, be referenced to the operation mechanism of CDN network, by model data proactive routing to the closer edge of application site
Server end reduces the time of transmission；When client has request, sent after cloud computing end makes requests parsing
To mobile edge calculations end, mobile edge calculations server is immediately checked in local cache the content for whether having user to request, such as
Fruit has just directly service；If it is not, requesting model and action data to cloud, model calculation is carried out, and cache to this
Ground.Request time is reduced in this way, also solves the problems, such as network blockage；
5.3, model computation module is according to the weight received, the textures and three-dimensional imaging data for model data into
The further sampling of row and compression return to client in the form of new model generated is passed through interface by the method quickly calculated
End, realizes the business of response；About for model sampling and compressometer count in, in the realization of Edge Server, using base
In the calculation method of machine learning, certain pretreatment is carried out for model compression data first, model data is carried out first
Pretreatment, while according to the data probability of model request, the model of high request rates is subjected to the operation such as cloud computing and storage, in this way
Model can be made to carry out quick model data response in request process, match the physics and network condition of client.
Further, the specific steps of the loading function of the realization dynamic model in the step 6 are as follows: according to optimization side
The node data of case design, model will carry out response transmission, traditional streaming to client in such a way that dynamic controls
In transmission process, model data file needs in the mode transmitted after pre-processing upon request for file in time cost
On bring additional time loss, using the dynamic model transmission mode based on interface, by the data of different threshold M with independence
The form of data object is stored；
It is as follows that the model of client is specifically loaded into process: the entire loading procedure of a model is divided into two parts: one
It is the model data that client obtains JD format from server, that is, the process of network transmission；Second is to get Json
The usable data of three.js are converted by JDLoader plug-in unit after data, are then shown on the client, this mistake
Journey is unrelated with network, only the calculating of client；
The process that the model movement of client is loaded into is as follows: the case where being directed to model on-demand loading, the method that we realize
It is that model data and the animation data in JD file are separated in rear end, first stress model data and shows, user can be
Static model data information is first seen in client, then further according to the demand of user, is initiated demand to server, is got
Matched model acts, then on-demand loading animation data, through the method for a special load animation, i.e., from JDLoader
Extraction is converted into the usable animation data format of three.js and model is allowed to move up；Other than the load of single model, also
For the loading content of multi-model scene, traditional mode is that default will occupy network load institute simultaneously when page open
Have including model.One function of encapsulation has been selected now, model file can have been allowed to load by queue, and load one is shown simultaneously
One, carry out the display work of next model again after terminating a model.
After adopting the above method, the invention has the following beneficial effects: a kind of web based on distributed computing of the present invention
AR data presentation mode, greatly improves the operational efficiency of client；It services more flexible；The calculating for more adapting to augmented reality is rung
Mode is answered, mitigates the GPU operation caching pressure of mobile terminal, realizes relative complex interaction models operation；Alleviate the meter in cloud
The advantages that calculation pressure, performance significantly improve, and the present invention has structure simple, and setting is reasonable, low manufacture cost.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is that the model of client in the present invention is specifically loaded into flow chart.
Fig. 2 is that the model movement of client in the present invention is loaded into flow chart.
Fig. 3 be in specific embodiment data in browser flow chart stored in memory.
Specific embodiment
The present invention will be further described below with reference to the drawings.
Present embodiment the technical solution adopted is that: its operating procedure is as follows:
1, the storage management for distinguishing model and action data in server end；First convert model data to pair
Image data needs to convert attribute pair for model data according to the definition for each key assignments in model in Transformation Application
The combination of elephant, is stored in server end；The specific steps of model profile formula storage management are as follows:
By taking three dimensional urban model data stores as an example, the large scale map sheet division methods based on topological relation model, and
Uniform Name coding is carried out to three-dimensional modeling data after division；By the powerful mass data group of non-relational database MongoDB
It knits and efficient how concurrent access function, constructs MongoDB fragment cluster server；Three dimensional urban model data is carried out
Dividing elements, and modeled using rule modeling software City Engine, D Urban model is obtained, by non-relationship number
Data storage experiment is carried out according to library software MongoDB；
2, according to the type of request, client sends request to neighbouring mobile Edge Server；Client is to service
When device end carries out page request, first by http agreement, association requests are sent to neighbouring mobile Edge Server；With
For Android platform, Android client is wanted to establish communication mode with server, uses http communication mode, and
Http communication mode divides get and post two ways again, when client sends request, passes to server end
A data block, that is, solicited message have been sent, the status code of return information is determined according to solicited message, data has been carried out and is browsing
Storage (referring to Fig. 3) in device memory；
3, mobile Edge Server carries out demand parsing according to the request of user；Whether decision model data carry out in advance
Routing, if model data is routed in advance, to cloud server request action data, carries out mould after wait-for-response
Type operation；If do not carried out content routing, then model and action data are requested to cloud, carries out model calculation, specific method
As follows: due to the information source that is arranged close to of mobile edge calculations server, and mobile computing is serviced close to terminal device, works as shifting
Dynamic Edge Server has received the request of client, is locally carrying out simply data processing, it is not necessary to by all data or letter
Breath is all uploaded to cloud, and requested model data is preloaded with, corresponding model can be called directly, also not complete
At the pre-loaded of model, mobile Edge Server sends model data and the movement etc. that request cloud stores to cloud again
Information；If not carrying out content routing, cloud server carries out the load of model；Multidimensional data of the received server-side to client
Model carries out rendering calculating by model adaptation module and model generation module；
4, the interaction request of cloud computing end parsing client obtains matching movement and is sent to Edge Server；In client
While sending interaction request, the interaction request of client is parsed, describes to carry out inquiry comparison in library in semantization,
Matched movement is obtained, mobile Edge Server is sent to, in specific application example, cloud computing end parses client
Interaction request obtains the specific steps that matching movement is sent to Edge Server are as follows:
Client send HTTP request head, server according to request when GET/POST come according to the doGet () of response/
DoPost () method is handled；After processing request, PriterWriter output flow object out is obtained by response object,
By out.println () by data by the Accpt for the use submitted in client: in the form of middle it is a kind of such as according to
response.setcontentType("text/html；Charset=gb2312' format output stream)；Server is according to visitor
The request content at family end inquires the movement with Model Matching in the database, sends data information to Edge Server, then takes
Connection is closed at business device end, and client parsing postbacks head response, restores the page；
5, data calculating is carried out at mobile edge calculations end；Mobile Edge Server is in the mould for receiving the transmitting of cloud computing end
Type, method of operating；By the data structure inside model, in a manner of program for the received model of institute and action data into
Row compound operation returns to model response data to server in data format, with the applied field of " the big connection of low-power consumption " of " 5G "
For scape, mobile edge 93 is deployed in mobile edge, and both Wi-Fi and internet technology are effectively fused together,
And increase in wireless network side and calculate, storage, the functions such as processing, construct mobile edge cloud, provide information technology service environment and
Cloud computing ability.Reduce the time for forwarding and handling in data transmission, reduces time delay end to end, and then meet low time delay and want
It asks, reduces power consumption；
6, calculation result data is returned into client, realizes the loading function of dynamic model；Edge calculations server with
New model generated is returned to client by the method quickly calculated by way of interface, realizes the business of response；Specifically
Implementation method it is as follows:
According to Optimization Plan, the node data of model will be responded in such a way that dynamic controls to client
Transmission, in traditional streaming process, model data file needs transmit after pre-processing upon request for file
Mode on bring additional time loss on time cost, will be using dynamic based on interface in present embodiment
States model transmission mode stores the data of different threshold M in the form of independent data objects；
The model of client is specifically loaded into process (referring to Fig. 1), and the entire loading procedure of a model is divided into two parts:
One is model data of the client from server acquisition JD format, that is, the process of network transmission；Second is to get
The usable data of three.js are converted by JDLoader plug-in unit after Json data, are then shown on the client, this
A process is unrelated with network, only the calculating of client；
The process (referring to Fig. 2) that the model movement of client is loaded into: the case where being directed to model on-demand loading, what we realized
Method is that model data and the animation data in JD file are separated in rear end, first stress model data and is shown, Yong Huke
First to see static model data information on the client, then further according to the demand of user, demand is initiated to server, is obtained
Matched model movement is got, then on-demand loading animation data, by the method for a special load animation (from JDLoader
Middle extraction) it is converted into the usable animation data format of three.js and model is allowed to move up.
Other than the load of single model, there are also the loading contents for being directed to multi-model scene, and traditional mode is in the page
Default is by occupying network loads all models for including simultaneously when opening.One function of encapsulation, Neng Gourang have been selected now
Model file is loaded by queue, and the display one simultaneously of load one carries out next model after terminating a model again
Show work.
Present embodiment has the beneficial effect that:
1, under the application scenarios of augmented reality, operand and capacity etc. for model are strictly limited, and
And sent and requested to server-side with physical parameter model, the operational efficiency of client can be greatly improved；
2, the web augmented reality model service mode based on interface modes provides clothes relative to traditional with document form
The mode of business, will more flexibly in service；
3, the improvement of the excuse words operation based on mobile edge calculations, can be more adaptive to the calculating response mould of augmented reality
Formula mitigates the GPU operation caching pressure of mobile terminal, realizes relative complex interaction models operation；
4, in a manner of moving edge calculations, model response computation mode is dropped into the shifting closer with client from cloud
Dynamic marginal end, alleviates the calculating pressure in cloud；
5, with one action response modes, it can reduce the caching operation pressure of client in certain degree；Especially
It is the more movement interactive modes for being intended for complicated business scene, it will have in performance and be significantly improved.
The above is only used to illustrate the technical scheme of the present invention and not to limit it, and those of ordinary skill in the art are to this hair
The other modifications or equivalent replacement that bright technical solution is made, as long as it does not depart from the spirit and scope of the technical scheme of the present invention,
It is intended to be within the scope of the claims of the invention.
Claims (10)
1. a kind of web AR data presentation mode based on distributed computing, it is characterised in that: its operating procedure is as follows:
(1), the storage management for distinguishing model and action data in server end；Object is converted by model data first
Data need to convert attributes object for model data according to the definition for each key assignments in model in Transformation Application
Combination, be stored in server end；
(2), according to the type of request, client sends request to neighbouring mobile Edge Server；Client is to server
When end carries out page request, first by http agreement, association requests are sent to neighbouring mobile Edge Server；
(3), mobile Edge Server carries out demand parsing according to the request of user；Whether decision model data carry out road in advance
By to cloud server request action data, carrying out model after wait-for-response if model data is routed in advance
Operation；If do not carried out content routing, then model and action data are requested to cloud, carries out model calculation, that is, cloud server
The load of model is carried out, the Multidimensional Data Model of received server-side to client is generated by model adaptation module and model
Module carries out rendering calculating；
(4), the interaction request of cloud computing end parsing client obtains matching movement and is sent to Edge Server；It is sent out in client
While sending interaction request, the interaction request of client is parsed, describes to carry out inquiry comparison in library in semantization, obtain
Matched movement is taken, mobile Edge Server is sent to；
(5), data calculating is carried out at mobile edge calculations end；Mobile Edge Server is in the mould for receiving the transmitting of cloud computing end
Type, method of operating；By the data structure inside model, in a manner of program for the received model of institute and action data into
Row compound operation returns to model response data to server in data format；
(6), calculation result data is returned into client, realizes the loading function of dynamic model；Edge calculations server is with fast
New model generated is returned to client by the method that speed calculates by way of interface, realizes the business of response.
2. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
The particular content of the storage management for distinguishing model and action data in server end in the step of stating (1) are as follows: be based on
The mass data storage model of cloud computing, according to the core calculations mode MapReduce of cloud computing, and support realizes
MapReduce calculates the open source distributed parallel programming framework Hapdoop of mode, and storage model and cloud computing are combined together,
Static models data required for user, are stored in cloud computing end in this way by the distributed storage for realizing mass data
On, data are mapped to by the program of Map function by different blocks after data are divided, computer group processing is distributed to and reaches
To the effect of distributed arithmetic, so that the effect of storage is realized, and the action data of storage model in the database, model
Action data stored respectively with model.
3. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
Request type particular content in the step of stating (2) are as follows: in SmartClient request, client is firstly the need of access equipment
Bottom-up information, the vision for obtaining equipment calculate relevant device parameter, such as resolution ratio, screen size and CPU operation efficiency letter
Breath, constructs the data model of request.
4. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
The specific steps that request is sent to neighbouring mobile Edge Server in the step of stating (2) are as follows: client with http agreement to
After server sends request, browser will monitor the solicited message of server-side return；After acquisition request information, first
The status code for determining return information, when status code is 200, it will grab the json formatted data of request, and data are stored
In the memory of browser.
5. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
Mobile Edge Server in the step of stating (3) carries out the specific steps of demand parsing are as follows:
(3.1), server is after acquisition request, the request data of analysis request client, carries out operation, determines asking for client
Status information is sought, and determines the upper limit of model data presented under the conditions of optimal network；With under the conditions of the upper limit since M1, by
It walks to client transmission pattern back end set, and returns to receiving end request and receive to finish the spent time, calculate client
The network environment parameters at end；This parameter passes to server end simultaneously, after server end receives this parameter, in determining M sequence
Upper limit cutoff value is determined in column；
(3.2), client parameter relevant to model load operation is constructed into a multi-Dimensional parameters model, carries characterization client
Operational capability, network bandwidth and interactive display interface etc. parameter, made requests with physical parameter model to server-side,
Server-side will analyze the physical parameter of client, union go out be suitable for client model return client into
Row load and rendering operation.
6. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
Cloud server in the step of stating (3) carries out the specific steps of the load of model are as follows: the various dimensions of server-side reception client
Data model carries out model data parsing, obtains physical parameter relevant to client current operating conditions, and it is suitable to pass to model
With module, model adaptation module parses the physical parameter of client, and TRANSFER MODEL exports weight and generates mould to model
Block；Model generation module will carry out model data generation processing according to weight, and processing result is returned in the form of interface
To client, rendering calculating is carried out.
7. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
The specific steps of the interaction request of cloud computing end parsing client in the step of stating (4) are as follows: cloud computing center will be for visitor
The operation request at family end is parsed, and obtains the physical environment model of client, and play the high concurrent data edge of cloud computing,
The weight for calculating client request model is routed to corresponding with the nearest mobile Edge Server progress of client by network
Model request calculates；Server end also needs the service request according to user, the logical response responded simultaneously, and will respond
Result by returning to client, realize the building of corresponding business scenario.
8. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
Acquisition matching movement in the step of stating (4) is sent to the specific steps of Edge Server are as follows: beyond the clouds, according to business demand,
Static models required for storage service first；Meanwhile the action data of storage model in the database；For storage model
While action data, the description for carrying out semantization is acted for model, while client sends interaction request, for visitor
The interaction request at family end is parsed, and is described to carry out inquiry comparison in library in semantization, is obtained matched movement, be sent to movement
Edge Server.
9. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that: institute
Mobile edge calculations end in the step of stating (5) carries out the specific steps of data calculating are as follows:
(5.1), the mobile edge calculations under 5g application scenarios are turned in terms of the cloud of core relative to traditional by centralization, large size
It compares, is more adaptive to the calculating response modes of mobile augmented reality；Mobile edge calculations, by operation from traditional cloud meter
Calculate central server cluster be transformed into the edge closer from mobile terminal, network circulation on, avoid from cloud computing center to
The occupied Internet resources of network edge and consumed transmission time, meanwhile, by most operand from central server
Mobile network edge has been transformed into it, it will improve the operational capability of web augmented reality to a certain extent；Pass through cloud meter simultaneously
Calculation and mobile edge calculations carry out the mode of collaboration operation, mitigate the GPU operation caching pressure of mobile terminal, realize for complexity
The operational capability of interaction models；In interface operation framework mode based on mobile edge calculations, it will with genuinely convinced in traditional
It is engaged in the function center that device cluster responds as client traffic request, to move Edge Server as in the service of model calculation
The heart；Client first sends the operation request for the physical parameter for carrying client, cloud to server-side under corresponding service environment
Calculating center will request to parse for the operation of client, obtain the physical environment model of client, and play cloud meter
The high concurrent data edge of calculation calculates the weight of client request model, the movement nearest with client is routed to by network
Edge Server carries out corresponding model request calculating；Server end also needs the service request according to user simultaneously, is rung
The logical response answered, and the result of response is realized into corresponding business scenario building by returning to client；
(5.2), it is referenced to the operation mechanism of CDN network, model data proactive routing is taken to the closer edge of application site
It is engaged in device end, reducing the time of transmission；When client has request, it is sent to after cloud computing end makes requests parsing
Mobile edge calculations end, mobile edge calculations server are immediately checked in local cache the content for whether having user to request, if
There is just directly service；If it is not, requesting model and action data to cloud, model calculation is carried out, and cache to local；
(5.3), according to the weight received, textures and three-dimensional imaging data for model data carry out model computation module
Further sampling and compression return to client in the form of new model generated is passed through interface by the method quickly calculated
End, realizes the business of response；About for model sampling and compressometer count in, in the realization of Edge Server, using base
In the calculation method of machine learning, certain pretreatment is carried out for model compression data first, model data is carried out first
Pretreatment, while according to the data probability of model request, the model of high request rates is subjected to the operation such as cloud computing and storage, in this way
Model can be made to carry out quick model data response in request process, match the physics and network condition of client.
10. a kind of web AR data presentation mode based on distributed computing according to claim 1, it is characterised in that:
The specific steps of the loading function of realization dynamic model in the step (6) are as follows: according to Optimization Plan, the section of model
Point data will carry out response transmission, in traditional streaming process, model to client in such a way that dynamic controls
Data file needs to bring on time cost in the mode transmitted after upon request pre-processing file additional
Time loss, using the dynamic model transmission mode based on interface, by the data of different threshold M in the form of independent data objects
It is stored；
It is as follows that the model of client is specifically loaded into process: the entire loading procedure of a model is divided into two parts: one is visitor
Family end obtains the model data of JD format, that is, the process of network transmission from server；Second is to get Json data
The usable data of three.js are converted by JDLoader plug-in unit afterwards, are then shown on the client, this process with
Network is unrelated, only the calculating of client；
Client model movement be loaded into process it is as follows: be directed to model on-demand loading the case where, we realize method be
Rear end separates model data and the animation data in JD file, first stress model data and shows, user can be in client
Static model data information is first seen on end, then further according to the demand of user, is initiated demand to server, is got matching
Model movement, then on-demand loading animation data extracted by the method for a special load animation from JDLoader
It is converted into the usable animation data format of three.js and model is allowed to move up.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910116350.7A CN109887098B (en) | 2019-02-13 | 2019-02-13 | Web AR data presentation mode based on distributed computing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910116350.7A CN109887098B (en) | 2019-02-13 | 2019-02-13 | Web AR data presentation mode based on distributed computing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109887098A true CN109887098A (en) | 2019-06-14 |
| CN109887098B CN109887098B (en) | 2023-01-17 |
Family
ID=66928133
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910116350.7A Active CN109887098B (en) | 2019-02-13 | 2019-02-13 | Web AR data presentation mode based on distributed computing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109887098B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110677683A (en) * | 2019-09-30 | 2020-01-10 | 北京奇艺世纪科技有限公司 | Video storage and video access method and distributed storage and video access system |
| CN111212128A (en) * | 2019-12-30 | 2020-05-29 | 北京邮电大学 | Edge-assisted-based collaborative Web AR multi-user service system and method |
| CN111212107A (en) * | 2019-12-10 | 2020-05-29 | 中移（杭州）信息技术有限公司 | Service processing method for CDN platform and CDN system |
| CN111260540A (en) * | 2020-01-13 | 2020-06-09 | 成都卓影科技股份有限公司 | 2D-3D 2.5D conversion engine under 5G network |
| CN111339192A (en) * | 2020-02-21 | 2020-06-26 | 深圳供电局有限公司 | Distributed edge computing data storage system |
| CN111400381A (en) * | 2020-02-28 | 2020-07-10 | 北京致医健康信息技术有限公司 | Data storage method and device, computer equipment and storage medium |
| CN114042311A (en) * | 2021-11-15 | 2022-02-15 | 中国联合网络通信集团有限公司 | Information processing method, edge server, electronic device, and computer medium |
| WO2022111444A1 (en) * | 2020-11-25 | 2022-06-02 | 摩赛恩科技(苏州)有限公司 | Mass spectrum data processing system and method |
| CN114745260A (en) * | 2022-03-09 | 2022-07-12 | 优刻得科技股份有限公司 | Computing power enhancement method, device, equipment and storage medium for content distribution network |
| WO2023109924A1 (en) * | 2021-12-15 | 2023-06-22 | Lei Xu | Serialization system for random data access |
| CN117453377A (en) * | 2023-12-21 | 2024-01-26 | 腾讯科技（深圳）有限公司 | Model scheduling method, terminal equipment and server |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103999434A (en) * | 2011-12-21 | 2014-08-20 | 瑞典爱立信有限公司 | Devices and methods using network load data in mobile cloud accelerator context to optimize network usage by selectively deferring content delivery |
| CN107222468A (en) * | 2017-05-22 | 2017-09-29 | 北京邮电大学 | Augmented reality processing method, terminal, cloud server and edge server |
| US9912718B1 (en) * | 2011-04-11 | 2018-03-06 | Viasat, Inc. | Progressive prefetching |
| CN108847981A (en) * | 2018-06-26 | 2018-11-20 | 咸宁职业技术学院 | Distributed computer cloud computing processing method |
-
2019
- 2019-02-13 CN CN201910116350.7A patent/CN109887098B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9912718B1 (en) * | 2011-04-11 | 2018-03-06 | Viasat, Inc. | Progressive prefetching |
| CN103999434A (en) * | 2011-12-21 | 2014-08-20 | 瑞典爱立信有限公司 | Devices and methods using network load data in mobile cloud accelerator context to optimize network usage by selectively deferring content delivery |
| CN107222468A (en) * | 2017-05-22 | 2017-09-29 | 北京邮电大学 | Augmented reality processing method, terminal, cloud server and edge server |
| CN108847981A (en) * | 2018-06-26 | 2018-11-20 | 咸宁职业技术学院 | Distributed computer cloud computing processing method |
Non-Patent Citations (1)
| Title |
|---|
| 吕强等: "移动云计算――移动增强现实技术和服务", 《中兴通讯技术》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110677683A (en) * | 2019-09-30 | 2020-01-10 | 北京奇艺世纪科技有限公司 | Video storage and video access method and distributed storage and video access system |
| CN111212107A (en) * | 2019-12-10 | 2020-05-29 | 中移（杭州）信息技术有限公司 | Service processing method for CDN platform and CDN system |
| CN111212128A (en) * | 2019-12-30 | 2020-05-29 | 北京邮电大学 | Edge-assisted-based collaborative Web AR multi-user service system and method |
| CN111260540B (en) * | 2020-01-13 | 2023-06-13 | 成都卓影科技股份有限公司 | 2.5D conversion engine of 2D-3D under 5G network |
| CN111260540A (en) * | 2020-01-13 | 2020-06-09 | 成都卓影科技股份有限公司 | 2D-3D 2.5D conversion engine under 5G network |
| CN111339192A (en) * | 2020-02-21 | 2020-06-26 | 深圳供电局有限公司 | Distributed edge computing data storage system |
| CN111400381A (en) * | 2020-02-28 | 2020-07-10 | 北京致医健康信息技术有限公司 | Data storage method and device, computer equipment and storage medium |
| CN111400381B (en) * | 2020-02-28 | 2024-05-10 | 北京致医健康信息技术有限公司 | Data storage method, device, computer equipment and storage medium |
| WO2022111444A1 (en) * | 2020-11-25 | 2022-06-02 | 摩赛恩科技(苏州)有限公司 | Mass spectrum data processing system and method |
| CN114042311A (en) * | 2021-11-15 | 2022-02-15 | 中国联合网络通信集团有限公司 | Information processing method, edge server, electronic device, and computer medium |
| WO2023109924A1 (en) * | 2021-12-15 | 2023-06-22 | Lei Xu | Serialization system for random data access |
| CN114745260A (en) * | 2022-03-09 | 2022-07-12 | 优刻得科技股份有限公司 | Computing power enhancement method, device, equipment and storage medium for content distribution network |
| CN114745260B (en) * | 2022-03-09 | 2024-04-02 | 优刻得科技股份有限公司 | Method, device, equipment and storage medium for enhancing computing power of content distribution network |
| CN117453377A (en) * | 2023-12-21 | 2024-01-26 | 腾讯科技（深圳）有限公司 | Model scheduling method, terminal equipment and server |
| CN117453377B (en) * | 2023-12-21 | 2024-04-26 | 腾讯科技（深圳）有限公司 | Model scheduling method, terminal equipment and server |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109887098B (en) | 2023-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109887098A (en) | A kind of web AR data presentation mode based on distributed computing | |
| Chen et al. | Intelligent resource allocation management for vehicles network: An A3C learning approach | |
| Qiao et al. | A new era for web AR with mobile edge computing | |
| He et al. | Software-defined networks with mobile edge computing and caching for smart cities: A big data deep reinforcement learning approach | |
| Wang et al. | A survey on deploying mobile deep learning applications: A systemic and technical perspective | |
| Zhang et al. | A novel edge computing architecture based on adaptive stratified sampling | |
| Zhang et al. | Toward transcoding as a service: energy-efficient offloading policy for green mobile cloud | |
| US10521485B1 (en) | Measuring page value | |
| Zheng et al. | MEC-enabled wireless VR video service: A learning-based mixed strategy for energy-latency tradeoff | |
| Hu et al. | If-rans: intelligent traffic prediction and cognitive caching toward fog-computing-based radio access networks | |
| Ren et al. | Multi-objective optimization for task offloading based on network calculus in fog environments | |
| Huang et al. | A geospatial hybrid cloud platform based on multi-sourced computing and model resources for geosciences | |
| Chen et al. | A hybrid task scheduling scheme for heterogeneous vehicular edge systems | |
| Li et al. | An intelligent collaborative inference approach of service partitioning and task offloading for deep learning based service in mobile edge computing networks | |
| Young et al. | Dynamic collaboration of centralized & edge processing for coordinated data management in an IoT paradigm | |
| Huang et al. | Computation offloading for multimedia workflows with deadline constraints in cloudlet-based mobile cloud | |
| Zhang et al. | Dependent task offloading mechanism for cloud–edge-device collaboration | |
| Shi et al. | A framework for mobile cloud computing selective service system | |
| CN109617960A (en) | A kind of web AR data presentation method based on attributed separation | |
| Song et al. | Dram: A drl-based resource allocation scheme for mar in mec | |
| Xia et al. | The interplay between artificial intelligence and fog radio access networks | |
| Liu et al. | Service chain caching and workload scheduling in mobile edge computing | |
| Huang et al. | Service reservation and pricing for green metaverses: A stackelberg game approach | |
| Chen et al. | VCMaker: Content-aware configuration adaptation for video streaming and analysis in live augmented reality | |
| Huang et al. | Joint optimization of task scheduling and computing resource allocation for VR video services in 5G‐advanced networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |