CN102857565B - Intelligent clothes trying-on system based on cloud computing - Google Patents
Intelligent clothes trying-on system based on cloud computing Download PDFInfo
- Publication number
- CN102857565B CN102857565B CN201210320051.3A CN201210320051A CN102857565B CN 102857565 B CN102857565 B CN 102857565B CN 201210320051 A CN201210320051 A CN 201210320051A CN 102857565 B CN102857565 B CN 102857565B
- Authority
- CN
- China
- Prior art keywords
- cloud server
- file
- client
- data
- datanode
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention requests the protection for an intelligent clothes trying-on system based on cloud computing. The system adopts intelligent terminals to collect data and transmits the data to a cloud server terminal through a wireless network, and the cloud server terminal collects file information collected by a plurality of intelligent terminals and stores the file information to all nodes. When a plurality of users send requests, the cloud server terminal divides a task into blocks, allocates the blocks of task to different idle resources for processing, obtains a clothes trying-on effect picture, a matching suggestion, color and price and the like through the comprehensive analysis on the data and returns the processed result to the intelligent terminals to be displayed, so as to realize the mass data transmission and the mass data processing between the intelligent terminals and a cloud server.
Description
Technical field
The present invention relates to cloud computing and communication technical field, particularly relate to the data storage of cloud terminal equipment and cloud server end, data analysis and data transmission technology.
Background technology
Along with the development of cloud computing and technology of Internet of things, people are more quick to the acquisition of demand information, these data are very easy to the life of people simultaneously, as Smart Home progressively realizes, realize Smart Home a kind of good method beyond doubt in conjunction with cloud computing and technology of Internet of things.In life, network is ubiquitous, and connected with Cloud Server by intelligent terminal by network, the real time communication of both realizations has been a kind of demand.In market, fitting mirror is a kind of indispensable instrument, and it is for providing pattern reference when people buy clothes, and the fitting effects that user obtains clothes by fitting mirror is selected.Fitting mirror common at present can only provide the simplest visual information for user, can not meet the demand of client's multi-angle, and existing Intelligent fitting mirror is mainly made up of display device, picture pick-up device, computer equipment and associated software thereof.Although this Intelligent fitting mirror can contrast clothes effect by the picture on display device, but majority is unit operation system, and depend on simple network service, very large deficiency is there is in storage, and under traditional IT architecture, also the flexible dispatching realized computing power and resource is difficult to, when fitting mirror use amount increases, require also to improve constantly to mass data processing, the storage resources of server and computational resource are encountered by powerful challenge, have impact on response speed and computing capability, so along with the increase of demand, such fitting mirror also differs greatly from the requirement of Smart Home.
Because client is numerous, image information and the data of magnanimity will be produced in client's fitting process, for existing network storing and processing mode, due to limited storage space, operational capability is not strong, the defect of each side such as scheduling of resource is dumb, will pose a big pressure to server end and client to the transmission of large nuber of images information and process.Easily cause the problem such as traffic congestion, picture delay.
Summary of the invention
The present invention is directed to current Intelligent fitting mirror in limited storage space, operational capability is not strong, and the defect of each side such as scheduling of resource is dumb, in conjunction with cloud computing and communication technical field, devise a kind of Intelligent fitting mirror based on cloud computing, this fitting mirror utilizes cloud platform, and view data is divided into block, the process of mapping abbreviation is carried out to block, carries out mass memory and magnanimity calculating.
The technical scheme that the present invention solves the problems of the technologies described above employing is as follows.
A kind of intelligent dressing system based on cloud computing, it is characterized in that, comprise: Cloud Server and intelligent terminal, intelligent terminal is common minute surface embedded processor, high-definition camera, display screen, wireless network card, reader, high-definition camera gathers the pictorial information of client, reader identification clothing labels information, embedded processor stores local file and deletes, and show and check the result returned from server end, wireless network card by the transmitting data file that collects to Cloud Server, the information received is stored in distributed file system HDFS by cloud server end invokes application design interface API, the information of storage is divided into the collection split of a series of block by cloud server end, copy in the cluster and produce a series of file fragment, file fragment is given idle machine resource and go process, and produce a series of Map task and Reduce task, cloud server end distributes a Map task or Reduce task to idle customer side, block is mapped to multiple customer side by idle customer side, and each customer side is called Map function and produced key assignments key in the middle of block, stored in memory, the position of buffer memory key is sent to abbreviation work station reduce worker by cloud server end, the key/value collection corresponding with unique middle key assignments key is passed to Reduce function, and the output of Reduce function is stored in a series of output file.
When multiple user sends request, collaborative work between each server of Cloud Server, be respective segments by parallel task division, and be assigned on different idling-resources and process, for each fragment produced, Cloud Server creates Map(automatically and maps) function is used for processing fragment, then utilizes Reduce(abbreviation) function carries out Macro or mass analysis, obtains clothes effect and collocation suggestion, color and price return to intelligent terminal.
The present invention devises a kind of Intelligent fitting mirror based on cloud computing, integrated use intelligent terminal technology, image processing techniques, the network communications technology, cloud computing technology.Each intelligent terminal collects fileinfo by wireless network transmissions to cloud server end, and cloud server end adopts distributed storage and parallel computation, and result of calculation is returned to terminal.
The invention has the beneficial effects as follows: the Intelligent fitting mirror being different from prior art substantially depends on unit operation system and simple network service and can not carry out mass data processing, and can not meet the defect of the different computation requests of multiple user.
Accompanying drawing explanation
Fig. 1 overall system design figure;
Fig. 2 mobile terminal integral module design drawing;
Fig. 3 Photo Viewer systematic schematic diagram;
Fig. 4 cloud server end stores schematic diagram;
Fig. 5 cloud server end task process schematic diagram.
Embodiment
Cloud computing is a kind of emerging technology, it has a powerful resource pool, it is put together the storage of multiple stage computer and computing capability, and storage and computing capability are distributed to uniformly on the multiple nodes in cluster, achieves the huge storage to super large data set and computing capability.Cloud computing take parallel computation as core, scheduling on demand distribution of computation tasks and computational resource, and cloud storage achieves storage Full-virtualization, and provide stronger storage and sharing functionality, Information Security is high, can mass memory and magnanimity calculating.And very low to the equipment requirement of client, only required simple configuration, client does not need to buy large number quipments yet, it is that the mode of serving uses calculating and storage resources, takes as required, pays as required, like this for enterprise has saved a large amount of costs.Therefore, the Intelligent fitting mirror based on cloud computing technology will have very large application prospect.
Below in conjunction with accompanying drawing, the present invention is further described.
Consult Fig. 1, it is the global design figure of system.In this Intelligent fitting mirror, comprise intelligent object, data acquisition unit, display device and Cloud Server.Wherein data acquisition unit comprises high-definition camera, RFID reader.High-definition camera shooting user is from the fitting picture of all angles, and RFID reader reads the information of clothes self, as manufacturer, price, color, material by clothes drop label.These information collected are sent to the intelligent object of fitting mirror, intelligent object upload the data to cloud server end by wireless network.Utilize cloud computing data processing centre to process data, and according to the order received, result of calculation is returned to user.
Fig. 2 is intelligent terminal intelligent object design drawing.Processor is then a kind of intelligent mobile terminal processor, can adopt ARM or Intel atom processor etc., to realize the display of picture piecemeal, data upload download, the functions such as user operation control.Memory be used for interim storage of processor computing need data, comprise No. ID of clothing labels, color, price, material, the place of production, the pictorial information tried on.Network interface card is used for realizing the interface of terminal and cloud server end network service, realize physical connection between Network Transfer Media and signal of telecommunication coupling, also relate to the transmission of frame and reception, the encapsulation of frame and opening, medium access control, the encoding and decoding of data and the function etc. of data buffer storage.Data input/output interface is used for external I/O equipment.
Whole intelligent terminal software system can adopt Windows CE embedded OS, and this system is non-PC field operating system, carries powerful communication function and DLL (dynamic link library), has very strong real-time and robustness.The operation interface main menu of fitting mirror is mainly checks and uploads two functions.The data of wherein checking are divided into local pictorial information and server info two class, and local picture is fitting effects figure, can individual slide display or the display of multiple breviary comparison diagrams.When process checked by picture, this intelligent terminal adopts the Photo Viewer based on Windows CE system.When designing Photo Viewer, the CEFileManager module carried by operating system enumerates the picture file on storage medium to user, and provide preview with thumbnail form to user, facilitate user to select interested picture of trying on to appreciate in detail, operate.In the process enumerating picture file, a new thread can be created in main window thread and carry out locating file, facilitate user when picture file is too many like this, the picture that arbitrarily can operate a certain preview carries out amplification to be appreciated, and alleviates the pressure that processing speed is brought to user.With CEPicViewer module to user provide actual size is amplified to certain pictures check pattern, and support simple operations function: select a upper pictures, select next pictures, lantern slide browse mode, be rotated counterclockwise, turn clockwise, delete, be switched to preview mode, exit.The application programming interfaces api function that this reader utilizes Windows to provide and the programming of the Component Object Model com interface, the format decoder JPG Decoder carried by Windows CE system carries out read-write to camera file and decodes and show.
Picture checks the model of system as shown in Figure 3.Server info be download return collocation suggestion and model or other people upload try picture on.The content uploaded comprises to be uploaded oneself and satisfied tries picture on, upload voluntarily and download reference for others, or oneself advise the collocation of this part clothes, or picture is tried in deletion on.
Fig. 4 is that cloud server end stores schematic diagram.Cloud server end is the cloud computing service platform built with HADOOP distributed computing framework, adopts HADOOP distributed file system HDFS (Hadoop Distributed Files System) to store.The a large amount of message file collected by multiple intelligent terminal is by wireless network transmissions to cloud server end, and file is stored in HDFS by the application programming interface API that directly calling HADOOP provides.Be made up of the back end Datanode of a management node Namenode and some in HDFS cluster, general cloud server end only runs a Namenode, and other intelligent terminals in cluster run Datanode respectively, therefore the number of back end Datanode is determined by the intelligent terminal number in cluster.Management node is in charge of the name space (namespace) of file system and client to the access of file.Back end is in charge of the storage on its place node.When file writes, client Client initiates file write request to management node Namenode, management node Namenode is according to file size and file configuration situation, return to the information of its institute administrative section Datanode of Client, super large file division is become multiple blocks of files, and the most basic storage cell of HDFS acquiescence is the data block of 64M, therefore super large file is divided into the data block of 64M mono-piece, except last, all data blocks are all onesize.Each data block is assigned with and is stored on back end, in order to ensure the fault-tolerance of system, and to each deposit data copy (being defaulted as 3 copies), when certain back end Datanode break down cause picture file to read time, system can obtain data from other back end Datanode.Distributed file system HDFS communications protocol is based upon on ICP/IP protocol.In HDFS, deploy a set of remote procedure call (RPC) mechanism, it carries out RPC by one group of agreement to call the communication that just can realize between server, HADOOP defines oneself communication protocol, client is connected to Namenode by a configurable tcp port, when client process is wanted to communicate with Namenode process time, personal code work realizes reciprocal process by ClientProtocol agreement.Datanode process usage data node protocol DatanodeProtocol and Namenode process interaction, use InterDatanodeProtoclol agreement to communicate between Datanode process.
Cloud computing server end data handling principle schematic diagram of the present invention shown in Fig. 5.The present invention uses MapReduce framework.MapReduce by abstract for complicated parallel procedure operated on large-scale cluster be two functions: map Map and abbreviation Reduce.Cloud server end is called MapReduce(and is mapped abbreviation) data message of storage is divided into the collection split of M block by storehouse, copy in the cluster and produce a series of file fragment, file fragment is given idle machine resource and go process, and produce a series of Map task and Reduce task; Cloud server end distributes a Map task or Reduce task to idle customer side; Idle customer side is by the data-mapping of segmentation to multiple customer side, and each customer side is called Map function and produced key assignments key in the middle of block, stored in memory; The position of buffer memory key is sent to abbreviation work station reduce worker by cloud server end; The key/value collection corresponding with unique middle key assignments key is passed to Reduce function, and the output of Reduce function is stored in a series of (R) output file.
Map function call is distributed in multiple stage client, the Data Segmentation inputted in memory becomes the collection split of M block, can be processed in parallel in different clients, by the middle key assignments key of segmentation defined with segmentation function, form R data file output, Reduce calls and is distributed to the enterprising row relax of multiple stage client, and the value of dividing number M is determined by block size, and output file number R is specified in advance by user.Concrete implementation step is as follows:
1. call MapReduce storehouse, according to the sheet size preset, the pictorial information file division of fitting effects is become multiple (M) sheet, form block cluster, then carry out copy in the cluster and produce file fragment, for each file fragment produced, file fragment to be given idle machine resource and is gone process by the calculation task distributed according to load-balancing algorithm, and produces M Map task and R Reduce task will be assigned with.The size of each is generally redefined for 16 to 64MB, also can be other size.
2. Cloud Server inquires about the state of all customer sides, and Cloud Server distributes a Map task or Reduce task to idle customer side.
3. the content of input block split is read in the customer side being assigned with Map task, and be mapped to multiple customer side, key assignments/operand attribute in the middle of key/value(is therefrom extracted in each customer side) right, then key/value to passing to user-defined Map function.Produce key assignments key in the middle of block by Map function to be cached in internal memory.
4. the middle key assignments be buffered in internal memory is periodically written on the local disk of customer side, by segmentation function they write R regions.On local disk, the position of buffer memory key is transmitted to cloud server end primary control program block, and primary control program block is responsible for these positions to send reduce worker(abbreviation processing workstation to).
5., when a reduce worker receives the position informing of primary control program block, the key that inquiring position is corresponding, make that there is the together content-aggregated of identical key by sequence.
6. reduce worker iteration is drained through the intermediate data of sequence, for the middle key assignments key that each is unique, the key/value collection corresponding with key is passed to user-defined Reduce function, and the output of Reduce function is added in the final output file of Reduce segmentation.
On successful completion, the output performing MapReduce operation is left in R output file (each Reduce task produces a file by user-assigned name word).Generally, do not need to merge this R file and become a file, often these files are used as the MapReduce that an input passes to other and call, or use them in the Distributed Application that can process multiple divided file.
The present invention adopts the advanced technologies such as cloud computing, Internet of Things and intelligent terminal, common visual information is converted into computer vision information, data store and data calculating is all carry out in cloud server end, intelligent terminal only needs to carry out shirtsleeve operation, just can obtain powerful storage capacity and computing capability, so not only can bring more convenience to client, simultaneously also for businessman has saved a large amount of costs.
Claims (1)
1. the intelligent dressing system based on cloud computing, it is characterized in that, comprise: Cloud Server and intelligent terminal, intelligent terminal comprises processor, high-definition camera, display screen, wireless network card, reader, high-definition camera gathers the pictorial information of client, reader identification clothing labels information, embedded processor stores processor pictorial information and label information, cloud server end is transferred to by wireless network card, the information received is stored in distributed file system HDFS by cloud server end invokes application design interface API, the information of storage is divided into the collection split of a series of block by cloud server end, copy in the cluster and produce a series of file fragment, file fragment is given idle machine resource and go process, and produce a series of Map task and Reduce task, cloud server end distributes a Map task or Reduce task to idle customer side, block is mapped to multiple customer side by idle customer side, and each customer side is called Map function and produced key assignments key in the middle of block, stored in memory, the position of buffer memory key is sent to abbreviation work station reduce worker by cloud server end, the key/value collection corresponding with unique middle key assignments key is passed to Reduce function, and the output of Reduce function is stored in a series of output file, information stores and comprises further, be made up of a management node Namenode and multiple back end Datanode in distributed file system HDFS cluster, cloud server end runs a Namenode, client runs Datanode respectively, management node is in charge of the name space of file system and client to the access of file, and back end is in charge of the storage on the node of place, when file writes, client initiates data write request to management node, management node is according to data file size and configuration, return to the information of its institute administrative section Datanode of client, super large file division is become the data block of multiple 64M, each data block is assigned with and is stored on back end, to each deposit data copy, when certain back end Datanode break down cause picture file to read time, data are obtained from other back end Datanode, client is connected to Namenode by a tcp port, when client is wanted to communicate with Namenode time, reciprocal process is realized by client protocol, Datanode process usage data node protocol and Namenode process interaction, in HDFS, client is connected to Namenode by a configurable tcp port, when client process is wanted to communicate with Namenode process time, personal code work realizes reciprocal process by ClientProtocol agreement, Datanode process usage data node protocol DatanodeProtocol and Namenode process interaction, use InterDatanodeProtoclol agreement to communicate between Datanode process.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210320051.3A CN102857565B (en) | 2012-09-03 | 2012-09-03 | Intelligent clothes trying-on system based on cloud computing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210320051.3A CN102857565B (en) | 2012-09-03 | 2012-09-03 | Intelligent clothes trying-on system based on cloud computing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102857565A CN102857565A (en) | 2013-01-02 |
CN102857565B true CN102857565B (en) | 2015-05-27 |
Family
ID=47403753
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210320051.3A Active CN102857565B (en) | 2012-09-03 | 2012-09-03 | Intelligent clothes trying-on system based on cloud computing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102857565B (en) |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104065685B (en) * | 2013-03-22 | 2018-04-27 | 中国银联股份有限公司 | Data migration method towards in the bedding storage system of cloud computing environment |
CN103440316A (en) * | 2013-08-27 | 2013-12-11 | 重庆邮电大学 | Intelligent wardrobe based on cloud calculations and control and processing method thereof |
CN104133942A (en) * | 2014-07-08 | 2014-11-05 | 上海置森数码技术有限公司 | Intelligent 3D life data cloud service system |
CN104881581A (en) * | 2015-05-28 | 2015-09-02 | 成都艺辰德迅科技有限公司 | IoT (Internet of Things) data high-efficiency analysis method |
CN105608584A (en) * | 2015-07-16 | 2016-05-25 | 宇龙计算机通信科技(深圳)有限公司 | Clothing collocation recommendation method, household terminal and server |
CN107515860A (en) * | 2017-08-07 | 2017-12-26 | 中译语通科技(青岛)有限公司 | A kind of machine translation method based on neuron |
CN107467987A (en) * | 2017-08-28 | 2017-12-15 | 厦门臻鑫豪科技有限公司 | Intelligent fitting mirror |
CN109101188B (en) * | 2017-11-21 | 2022-03-01 | 新华三大数据技术有限公司 | Data processing method and device |
CN107968786A (en) * | 2017-12-05 | 2018-04-27 | 北京奇艺世纪科技有限公司 | Distributed RPC systems and data processing method and processing device |
CN110415062A (en) * | 2018-04-28 | 2019-11-05 | 香港乐蜜有限公司 | The information processing method and device tried on based on dress ornament |
CN109120663A (en) * | 2018-06-07 | 2019-01-01 | 滨州学院 | A kind of long-range processing method of panoramic imagery based on cloud computing and application system |
CN112214466A (en) * | 2019-07-12 | 2021-01-12 | 海能达通信股份有限公司 | Distributed cluster system, data writing method, electronic equipment and storage device |
CN111245958A (en) * | 2020-03-09 | 2020-06-05 | 中国能源建设集团陕西省电力设计院有限公司 | Unified cloud computing system for ubiquitous power Internet of things |
CN113406894A (en) * | 2021-07-22 | 2021-09-17 | 深圳市伟峰科技有限公司 | Intelligent household control system, method, equipment and storage medium based on cloud computing |
CN113656176B (en) * | 2021-08-19 | 2022-08-23 | 北京百度网讯科技有限公司 | Cloud equipment distribution method, device and system, electronic equipment, medium and product |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101425166A (en) * | 2008-10-17 | 2009-05-06 | 幻境湾网络科技(北京)有限公司 | Intelligent fitting system in store, intelligent fitting system and method |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4604525B2 (en) * | 2004-03-17 | 2011-01-05 | セイコーエプソン株式会社 | Network system, program, and printing apparatus control method |
-
2012
- 2012-09-03 CN CN201210320051.3A patent/CN102857565B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101425166A (en) * | 2008-10-17 | 2009-05-06 | 幻境湾网络科技(北京)有限公司 | Intelligent fitting system in store, intelligent fitting system and method |
Non-Patent Citations (2)
Title |
---|
Hadoop官网.《Hadoop分布式文件系统》.《http://hadoop.apache.org/docs/r0.19.1/cn/hdfs_design.html,Hadoop分布式文件系统》.2008, * |
百度百科.《MapReduce》.《http://baike.baidu.com/history/33351241,MapReduce》.2012, * |
Also Published As
Publication number | Publication date |
---|---|
CN102857565A (en) | 2013-01-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102857565B (en) | Intelligent clothes trying-on system based on cloud computing | |
US20220223035A1 (en) | Ubiquitous visual computing witness | |
CN102713848B (en) | For using lightweight client to calculate, with virtualization, the method that service is docked by network | |
CN106453572B (en) | Method and system based on Cloud Server synchronous images | |
CN106033476A (en) | Incremental graphic computing method in distributed computing mode under cloud computing environment | |
US20200159764A1 (en) | Method for Processing and Displaying Real-Time Social Data on Map | |
CN106407463A (en) | Hadoop-based image processing method and system | |
CN103780685A (en) | Method for adaptive data sharing among diversified intelligent equipment | |
CN103905833A (en) | Distributed network video data mining and collecting system based on cloud calculation | |
CN115022342B (en) | Data processing method, device, electronic equipment and computer readable storage medium | |
CN103440316A (en) | Intelligent wardrobe based on cloud calculations and control and processing method thereof | |
CN111935663B (en) | Sensor data stream processing method, device, medium and electronic equipment | |
CN108536617A (en) | Buffer memory management method, medium, system and electronic equipment | |
CN115238015A (en) | Space-time big data platform based on micro-service | |
CN103442056B (en) | A kind of intelligent shoe cabinet control system based on cloud platform | |
Zuo et al. | Bandwidth and energy efficient image sharing for situation awareness in disasters | |
CN105872731A (en) | Data processing method and device | |
CN108898267A (en) | Electric energy substitution extension system and method based on image recognition | |
CN102752884B (en) | The Internet of things system of sing on web technology and the data processing method of this system | |
CN112181678A (en) | Service data processing method, device and system, storage medium and electronic device | |
CN102902593A (en) | Protocol distribution processing system based on cache mechanism | |
CN112087632A (en) | Video processing system, method, storage medium and computer device | |
CN212541377U (en) | Electric power AI training platform based on containerization management | |
CN207304622U (en) | Industrial Internet of things system based on data cloud platform | |
CN110430464A (en) | A kind of electron album, electronic whiteboard display systems and method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |