CN105574035A - Intelligent recognition and retrieval system and method for mass graphic images - Google Patents
Intelligent recognition and retrieval system and method for mass graphic images Download PDFInfo
- Publication number
- CN105574035A CN105574035A CN201410547776.5A CN201410547776A CN105574035A CN 105574035 A CN105574035 A CN 105574035A CN 201410547776 A CN201410547776 A CN 201410547776A CN 105574035 A CN105574035 A CN 105574035A
- Authority
- CN
- China
- Prior art keywords
- image
- target object
- module
- ice device
- intelligent recognition
- 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.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides an intelligent recognition and retrieval system and method for mass graphic images. The system comprises an ICE device, wherein the ICE device is connected with a source image database and a target non-graphical feature model library; the ICE device is connected with a result retrieval and recognition device; and the method comprises the following recognition and retrieval steps: (1) an intelligent modeling and classifying module for the graphic images; (2) a graphic image mode matching evaluation module; (3) a distributed high-reliability storage module for mass data; and (4) a distributed parallel processing module for computing tasks. The intelligent recognition and retrieval system and method have the characteristics of being quick to retrieve, time-saving and efficient.
Description
Technical field
The invention belongs to image recognition retrieval technique field, be specifically related to a kind of magnanimity graph image Intelligent Recognition searching system and search method thereof.
Background technology
Much having the place of high-definition video monitoring (as hotel, traffic sluice gate, customs, army), preserve a large amount of high-definition video monitoring data, and in certain special cases, need retrieve in multitude of video data and identify data slot (the such as location of traffic hazard vehicle comprising specific objective object, secret storehouse enters personnel's supervision, customs designated person follows the tracks of), in addition, the more external private TV media of picture and Internet video media, also there is retrieval in multitude of video data and comprise the demand (such as retrieving the data of football match in a large amount of image data) of intended target object.Above-mentioned application demand often because the restriction of the aspects such as technology, hardware resource, the expense budget that once drops into, can not be well solved in long-time, has to rely on artificial or that treatment effeciency is very low system, becomes industry pain spot.
Traditional graph image identification processing system, most dependence possesses the equipment such as the minicomputer of superpower processing power, calculation process is carried out to the image/video data of centralized stores, due to resource restriction and the expansion restriction of hardware device, make its processing power and efficiency after data reach certain magnitude, cannot support further again, become the bottleneck of system for restricting development.And the distributed storage of traditional graphics and image indexing recognition technology and current large data fields and parallel computing creatively organically combine by the present invention program, by actual for this polytechnic Integrated research and development are landed, effectively improve efficiency and the processing power of graphics and image indexing identification, its horizontal extension ability supported breaches the resource cap of separate unit high-end devices, some application scenarioss restricted because of hardware capabilities bottleneck are in the past made to become possibility, adopt the solution of the present invention, only need to dispose abundant cheap blade server to store for sharing and calculate, the demand of " impossible " in the past scheme can be realized.When follow-up data amount and portfolio continue to increase, the blade server quantity that can be configured equally by parallel expansion realizes the upgrading of storage and computing power.For the application pain spot perplexing multiple industry provides a solution.
Summary of the invention
In order to overcome above-mentioned the deficiencies in the prior art, the object of this invention is to provide magnanimity graph image Intelligent Recognition searching system and search method thereof, there is retrieval rapidly, the feature of time-saving and efficiency.
To achieve these goals, the technical solution used in the present invention is: a kind of magnanimity graph image Intelligent Recognition searching system, and include ICE device, ICE device is connected without figure feature model library with target with source image database; ICE device is connected with retrieval result device.
Magnanimity graph image Intelligent Recognition search method, comprises the following steps:
1) graph image intelligent modeling and sort module
By target object image, after being loaded into software, starting modeling learning process, set up the characteristic model for this target object image in software; If multiple target object, then automatically set up the classification image feature model of target object;
2) graph image pattern matching score module
After target object image feature model is set up, source image data to be analyzed is loaded into software, calculates the matching degree of these image datas and characteristic model, provide evaluation score value, thus as the foundation identified and retrieve;
3) the distributed highly reliable memory module of mass data
Utilize the HDFS technology in Hadoop system, realize storage administration and the storage security of massive image data;
4) calculation task distributed variable-frequencypump module
Utilize the Map-Reduce technology in Hadoop system, realize distributed treatment and parallel computation, thus make full use of hardware computing resource, promote the efficiency of integral operation.
Target of the present invention is to retrieve efficiently, quickly and accurately and identifying, reply magnanimity source data, there is provided can smoothly expand, the upgrading ability of low basic configuration, reduce the requirement of one-time investment, really carry out flexible planning according to the data volume demand of practical application.
The present invention program adopts Images Classification and feature modeling recognition technology, content-based modeling analysis is carried out to large nuber of images, whether the metadata not relying on view data complete, thus solve present stage industry can not the problem of retrieval because metadata is imperfect to the retrieval of the view data in irregular source.
Accompanying drawing explanation
Accompanying drawing 1 is structural representation of the present invention.
Embodiment
Below in conjunction with accompanying drawing, structural principle of the present invention and principle of work are described in further detail.
See Fig. 1, a kind of magnanimity graph image Intelligent Recognition searching system, includes ICE device 1, ICE device 1 and is connected without figure feature model library 4 with target with source image database 2; ICE device 1 is connected with retrieval result device 3.
To the magnanimity set of source data of retrieval be needed to inject ICE device 1, start up system analysis be retrieved, and system will identify to retrieve and include target data to be retrieved in source image database 2.
Citing illustrates:
Such as to capture in the video got off from network a large amount of, retrieve the video of all Basketball Match, so, first user can provide the picture of several basketballs for system self study to system, after system completes self study, specify video set position to be analyzed to system, start up system analyzes search function, and all video datas including basketball object are retrieved by system.If arrange some rules (accounting such as including frame all frames in this video of basketball object reaches predetermined threshold value) before analysis, then can filter out result for retrieval more accurately.
Magnanimity graph image Intelligent Recognition search method, comprises the following steps:
1) graph image intelligent modeling and sort module
By target object image, after being loaded into software, starting modeling learning process, set up the characteristic model for this target object image in software; If multiple target object, then automatically set up the classification image feature model of target object;
2) graph image pattern matching score module
After target object image feature model is set up, source image data to be analyzed is loaded into software, calculates the matching degree of these image datas and characteristic model, provide evaluation score value, thus as the foundation identified and retrieve;
3) the distributed highly reliable memory module of mass data
Utilize the HDFS technology in Hadoop system, realize storage administration and the storage security of massive image data;
4) calculation task distributed variable-frequencypump module
Utilize the Map-Reduce technology in Hadoop system, realize distributed treatment and parallel computation, thus make full use of hardware computing resource, promote the efficiency of integral operation.
Claims (2)
1. a magnanimity graph image Intelligent Recognition searching system, is characterized in that, includes ICE device (1), and ICE device (1) is connected without figure feature model library (4) with target with source image database (2); ICE device (1) is connected with retrieval result device (3).
2. a magnanimity graph image Intelligent Recognition search method, is characterized in that, comprise the following steps:
1) graph image intelligent modeling and sort module
By target object image, after being loaded into software, starting modeling learning process, set up the characteristic model for this target object image in software; If multiple target object, then automatically set up the classification image feature model of target object;
2) graph image pattern matching score module
After target object image feature model is set up, source image data to be analyzed is loaded into software, calculates the matching degree of these image datas and characteristic model, provide evaluation score value, thus as the foundation identified and retrieve;
3) the distributed highly reliable memory module of mass data
Utilize the HDFS technology in Hadoop system, realize storage administration and the storage security of massive image data;
4) calculation task distributed variable-frequencypump module
Utilize the Map-Reduce technology in Hadoop system, realize distributed treatment and parallel computation, thus make full use of hardware computing resource, promote the efficiency of integral operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410547776.5A CN105574035A (en) | 2014-10-16 | 2014-10-16 | Intelligent recognition and retrieval system and method for mass graphic images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410547776.5A CN105574035A (en) | 2014-10-16 | 2014-10-16 | Intelligent recognition and retrieval system and method for mass graphic images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105574035A true CN105574035A (en) | 2016-05-11 |
Family
ID=55884178
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410547776.5A Pending CN105574035A (en) | 2014-10-16 | 2014-10-16 | Intelligent recognition and retrieval system and method for mass graphic images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105574035A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106686108A (en) * | 2017-01-13 | 2017-05-17 | 中电科新型智慧城市研究院有限公司 | Video monitoring method based on distributed detection technology |
| CN111125410A (en) * | 2019-12-19 | 2020-05-08 | 湖北南楚网络传媒有限公司 | Intelligent identification and retrieval system for massive graphic images |
-
2014
- 2014-10-16 CN CN201410547776.5A patent/CN105574035A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106686108A (en) * | 2017-01-13 | 2017-05-17 | 中电科新型智慧城市研究院有限公司 | Video monitoring method based on distributed detection technology |
| CN111125410A (en) * | 2019-12-19 | 2020-05-08 | 湖北南楚网络传媒有限公司 | Intelligent identification and retrieval system for massive graphic images |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11837027B2 (en) | Activity recognition systems and methods | |
| Xu et al. | Video structured description technology based intelligence analysis of surveillance videos for public security applications | |
| US20200250435A1 (en) | Activity recognition method and system | |
| WO2021217934A1 (en) | Method and apparatus for monitoring number of livestock, and computer device and storage medium | |
| Ding et al. | Violence detection in video by using 3D convolutional neural networks | |
| Kim et al. | RGB color model based the fire detection algorithm in video sequences on wireless sensor network | |
| CN103841367A (en) | Monitoring system | |
| CN109241325A (en) | A kind of extensive face retrieval method and apparatus based on depth characteristic | |
| CN107786848A (en) | The method, apparatus of moving object detection and action recognition, terminal and storage medium | |
| De Marsico et al. | FACE: Face analysis for commercial entities | |
| Gao et al. | Research on food recognition of smart refrigerator based on SSD target detection algorithm | |
| CN110119730A (en) | A kind of monitor video processing method, system, terminal and storage medium | |
| CN113239792A (en) | Big data analysis processing system and method | |
| CN104834911A (en) | Portrait dynamic comparison system based on distributed cloud transformation platform | |
| CN105468741A (en) | Police affair big data processing system | |
| CN202120402U (en) | Mass video analysis based on cloud computing and rapid retrieval system | |
| CN104820711A (en) | Video retrieval method for figure target in complex scene | |
| CN106844654A (en) | Towards the massive video distributed search method of police service practical | |
| Sun et al. | Unsupervised fast anomaly detection in crowds | |
| CN105574035A (en) | Intelligent recognition and retrieval system and method for mass graphic images | |
| CN104484651A (en) | Dynamic portrait comparing method and system | |
| Xia et al. | Recognition of suspicious behavior using case-based reasoning | |
| Loukil et al. | Design and test of smart IP-camera within reconfigurable platform | |
| Zhou et al. | Social network construction of the role relation in unstructured data based on multi-view | |
| CN112306985A (en) | Digital retina multi-modal feature combined accurate retrieval method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160511 |