CN107169480A - A kind of distributed character identification system of live video stream - Google Patents
A kind of distributed character identification system of live video stream Download PDFInfo
- Publication number
- CN107169480A CN107169480A CN201710504745.5A CN201710504745A CN107169480A CN 107169480 A CN107169480 A CN 107169480A CN 201710504745 A CN201710504745 A CN 201710504745A CN 107169480 A CN107169480 A CN 107169480A
- Authority
- CN
- China
- Prior art keywords
- video
- task
- decoding
- unit
- text region
- 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
- 238000005111 flow chemistry technique Methods 0.000 claims abstract description 38
- 238000012545 processing Methods 0.000 claims abstract description 37
- 238000000034 method Methods 0.000 claims abstract description 13
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000012805 post-processing Methods 0.000 claims description 10
- 230000005540 biological transmission Effects 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 abstract description 7
- 230000004048 modification Effects 0.000 abstract description 3
- 238000012986 modification Methods 0.000 abstract description 3
- 238000000926 separation method Methods 0.000 abstract description 3
- 238000006467 substitution reaction Methods 0.000 abstract description 2
- 238000012015 optical character recognition Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/21—Server components or server architectures
- H04N21/218—Source of audio or video content, e.g. local disk arrays
- H04N21/2181—Source of audio or video content, e.g. local disk arrays comprising remotely distributed storage units, e.g. when movies are replicated over a plurality of video servers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
- H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
- H04N21/23406—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving management of server-side video buffer
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
- H04N21/4402—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving reformatting operations of video signals for household redistribution, storage or real-time display
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Character Discrimination (AREA)
Abstract
The invention discloses a kind of distributed character identification system of live video stream, belong to video streaming content identification technology field.Present system includes video stream management module, task scheduling modules, decoding video stream module and Text region module, and present system realizes the processing of many concurrent live video streams while video flowing Text region is handled using distributed task dispatching framework；The separation of decoding video stream and Text region makes system have more preferable adaptability, Text region module can not interfere with the decoder module of video flowing with modifications or substitutions, and mismatch problem of two modules in processing speed can be solved using the multiple Text region modules of decoder module correspondence；Real-time video stream process that is many and giving is realized using distributed system architecture simultaneously.Present invention is mainly used in the high video flow processing of requirement of real-time, and the processing requirement of many concurrent live video streams is disclosure satisfy that, present system has good practical value.
Description
Technical field
The invention belongs to video streaming content identification technology field, more particularly, to a kind of distribution of live video stream
Character identification system.
Background technology
In recent years with the fast development of image recognition technology, the identification demand to picture material in video is also constantly increasing
Plus.In traditional safety monitoring video, the character image in video is the main target of identification, and in desktop office system and
In the video of the concern text information such as online education, the target of identification is exactly the text information in video.Obtained in real time to reach
The requirement of the text information in identification video is taken, it is necessary to carry out timely identifying processing to the live video stream transmitted in network.
Generally required simultaneously in actual use environment to multiple video flowings while Text region processing is carried out, so solution is more simultaneously again
The problem of sending out the real-time text identifying processing of video flowing.
OCR (Optical Character Recognition, optical character identification) is that image information is carried out at analysis
Reason obtains the technology of text information.The application relative maturity of OCR technique, application field at home and abroad widely, including is handed over
Logical, finance, insurance, education etc..The error rate of identification of the OCR technique in English character and numerically is very low, is primarily due to English
Word mother's number and numeral only have 36 altogether, and the quantity of Chinese character is tens thousand of, and Chinese characters in common use also have thousands of, so
It is more much larger than English and numeral in the difficulty of identification.OCR Chinese identification also experienced the growth change of long period, each
The Chinese recognition accuracy for planting OCR under new technology and method is also being improved constantly.The OCR technique of current main-stream is all based on machine
Study or deep learning realize that wherein Google OCR Open Frameworks Tesseract is a Typical Representative.English and
In numeral identification, the Tesseract being good to English and numeral identification is used；In Chinese identification, then depth can be utilized
Practise realizing.
When in face of how concurrent Video processing, typically all handled using distributed system scheme.Distributed system
One very big feature be that it has good scalability and fault-tolerance.When in face of high concurrent and big data processing, generally
Single server is that task can not be handled in time.And appropriate clothes can be increased newly at any time as needed under distributed system
Device is engaged in realize the timely processing of task.Distributed system has multiple species, is distributed formula scheduling, Distributed Calculation and distribution
Storage.Distributed Calculation is currently used most distributed system, under MapReduce and Spark frameworks are Distributed Calculation
Typical Representative.In the processing of video file, many existing systems use MapReduce frameworks.
Processing system for video based on MapReduce frameworks is relatively more suitable for the processing to video file, and to video text
Unlike the processing of part, the processing of live video stream is with very high requirement of real-time, in video decoding, general single line
Journey decoding is difficult to meet live video stream decoding request.For the processing of live video stream, be not suitable for using MapReduce frames
Frame.On the one hand it is because video data frame has dependence in decoding in video flowing.Frame after video compress is divided into three kinds：I
Frame, P frames and B frames.I frames are directly can to carry out decoding obtaining original image number, and P frames and B frames are to need the I frames before and after relying on
Or P frames could be decoded.Realize and correctly decode, one has the I frames, P frames and B frames of dependence preferably same
Decoded at decoding.On the other hand, under the certain real-time video stream of frame per second, received in the unit interval of video flow processing end
Pending the video data volume and not less than its disposal ability.For video file, multitude of video can be read in the unit interval
Data, can be distributed to other machines for the video data beyond machine its own processing capabilities and handle；And real-time video
Stream is different, because video frame rate is certain, the video data volume that can be read in the unit interval from network is general not over machine
The disposal ability of device itself.
The content of the invention
For the disadvantages described above or Improvement requirement of prior art, the invention provides a kind of distribution text of live video stream
Word recognition system, its object is to video flowing is carried out while multithreading decoding using the scheduling pair of distributed task dispatching framework
The Text region processing of image is decoded, the Text region of many concurrent live video streams is thus solved the problems, such as.
To achieve the above object, according to one aspect of the present invention, there is provided a kind of distributed word of live video stream
Identifying system, the system includes：
Video stream management module, for managing being turned on and off for video flow processing task, video flow processing task is opened
When to task scheduling modules submit video flow processing task unlatching request and video flowing address；When video flow processing task is closed
The turn-off request of video flow processing task is submitted to task scheduling modules；
Task scheduling modules, for storing video flowing address and view data, the unlatching for receiving video flow processing task please
An idle decoding video stream module is dispatched after asking and performs video decoding task；After the Text region request for receiving view data
The idle Text region module of scheduling one performs Text region task；Receive and storage is deleted after the turn-off request of video flow processing task
The video flowing address deposited；
Decoding video stream module, appoints for receiving
Video flowing address is read after the video decoding task for scheduler module of being engaged in from task scheduling modules, video data is obtained
And decoding process is carried out, decode obtained view data deposit task scheduling modules and submit the image to task scheduling modules
Text region is asked；Video decoding module inquires about the video flowing address of its processing to task scheduling modules at regular intervals simultaneously
Whether it has been deleted, has stopped the decoding process of video flowing if being deleted, otherwise continue executing with decoding operate；
Read after Text region module, the Text region task for receiving task scheduling modules from task scheduling modules
View data carries out Text region, deletes identified view data after the completion of identification from task scheduling modules.
Further, the task scheduling modules include：
Buffer unit, for storing video flowing address and view data；
Distributed task dispatching server unit, receives scheduling one after the unlatching request of video flow processing task idle
Decoding video stream module performs video decoding task；An idle word is dispatched after the Text region request for receiving view data
Identification module performs Text region task；Receive the video deleted and stored in buffer unit after the turn-off request of video flow processing task
Flow address.
Further, the video stream management module includes：
The opening unit of video flow processing task, submits video flow processing to appoint for distributed task dispatching server unit
The unlatching request of business, and video flowing address is sent to buffer unit storage；
The closing unit of video flow processing task, to distributed task dispatching service when being closed for video flow processing task
Device unit submits the turn-off request of video flow processing task.
Further, the decoding video stream module includes：
Stop decoding unit, for whether being spaced video flowing address that query caching unit using according to set time
It is deleted, stops the video decoding task of video flowing reading and decoding unit if deleting；Otherwise do not process；
Video flowing is read and decoding unit, and it is single that caching is read in the scheduling for receiving distributed task dispatching server unit
Video flowing address in member performs video decoding task, obtains decoding image, and decoding image is transmitted into image recognition request list
Member；
In image recognition request unit, the greyscale image data deposit buffer unit for extracting decoding image and to distribution
Formula task scheduling server unit sends the Text region request of view data.
Further, the Text region module includes：
Images to be recognized reading unit, the Text region for receiving the transmission of distributed task dispatching server unit please
Ask, read view data from buffer unit and send to image character recognition unit；
Character recognition unit, for judging whether view data needs to carry out Text region processing, if desired then to image
Data carry out Text region and send the Text region result of acquisition to post-processing unit；
Post-processing unit, for receiving Text region result, and deletes what character recognition unit in buffer unit was treated
View data.
In general, by the contemplated above technical scheme of the present invention compared with prior art, it is special with following technology
Levy and beneficial effect：
(1) present invention in user can by video flowing address live video stream be linked into system carry out decoding and
Text region processing, while the processing of many concurrent live video streams can be realized using distributed task dispatching framework.
(2) decoding and the separation of identification module not only cause system to have good adaptability, are only needed under different demands
Corresponding modification is carried out to identification module can just realize the processing of video flowing, but also can be many by a decoder module correspondence
Individual Text region module solves the mismatch problem of the decoding and Text region of video flowing in speed.
(3) whole system is the distributed system realized based on distributed task dispatching framework；With autgmentability, Ke Yisui
When to system add decoding or Text region module；With fault-tolerance, can be when decoding or abnormal identification module processing generation
System can resend to task available decoding or identification module is continued with.
Brief description of the drawings
Fig. 1 is the structural representation of present system；
Fig. 2 is the overall workflow figure of present system；
Fig. 3 is the schematic layout pattern of present system embodiment.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, not
For limiting the present invention.As long as in addition, technical characteristic involved in each embodiment of invention described below that
Not constituting conflict between this can just be mutually combined.
As shown in figure 1, present system includes video stream management module 100, decoding video stream module 200, Text region
Module 300 and task scheduling modules 400, are described in further detail to each several part below：
Task scheduling modules 400 are used for schedule video stream decoder module 200 and perform video decoding task and Text region mould
Block 300 performs Text region task, while stopping video flow processing signal, video flowing address date and pending text for transmitting
The view data of word identification.
Task scheduling modules 400 include buffer unit 401 and distributed task dispatching server unit 402：
Buffer unit 401 is used to store decodes obtained view data depending on video flowing address and decoding video stream module 200.
Distributed task dispatching server unit 402 is used for the decoding that schedule video stream decoder module 200 performs video flowing
Task and Text region module 300 perform Text region task to the decoded view data of decoding video stream module 200.Distribution
If the unlatching video flowing Text region processing that formula task scheduling server unit 402 receives the transmission of video stream management module 100 please
Ask, then schedule video stream decoder module 200 performs decoding video stream task；If distributed task dispatching server unit 402 connects
The request that obtained image progress Text region is decoded to decoding video stream module 200 that video flowing decoder module 200 is sent is received,
Then scheduling Text region module 300 performs Text region task.
Video stream management module 100 is used for the execution for being turned on and off a video flowing Text region task, specifically
It is to be realized by controlling decoding video stream module 200 to whole video flowing Text region task start and closing.The present invention's
Video stream management module 100 includes the opening unit 101 of video flow processing task and the closing unit of video flow processing task
102。
The opening unit 101 of video flow processing task is used for the execution for opening a video flowing Text region task.Open
When to distributed task dispatching server unit 402 send a decoding video stream task requests, one is carried in request and is regarded
Frequency stream address, distributed task dispatching server unit 402 receives the video flowing of request and the carrying of decoding video stream task
Gone with performing video flowing using the video flowing address scheduling one free time available decoding video stream module 200 received behind address
The decoding task of the corresponding video flowing in location；The video flowing address is stored into buffer unit 401 simultaneously as video flowing word to know
Manage the open signal of task in other places.
The closing unit 102 of video flow processing task is used for closing one and is carrying out video flowing Text region processing task
Execution.The closing unit 102 of video flow processing task deletes the video flowing address to be closed, video from buffer unit 401
The video flowing pair that stream decoder module 200 is decoded in buffer unit 401 is detected without current video stream decoder module 200
Stop the execution of current video stream decoding task behind the video flowing address answered, realize that the Text region task of the video flowing is stopped
Only.
Decoding video stream module 200 is with receiving the video flowing of the transmission of distributed task dispatching server unit 402
The video flowing address received from distributed task dispatching server unit 402 is performed behind location and decoding video stream task corresponding to regard
Frequency stream decoding task, every time decoding obtains that all one can be sent to distributed task dispatching server unit 402 after an image
The Text region that image is obtained to decoding handles request, while checking current video stream in buffer unit 401 at regular intervals
The corresponding video flowing address of video flowing that decoder module is decoded whether there is to determine whether to continue executing with the solution of video flowing
Code task.The decoding video stream module 200 of the present invention includes stopping decoding unit 201, video flowing reading and the and of decoding unit 202
Image recognition request unit 203.
Stop decoding unit 201 and inquire about current just in decoding video stream module from buffer unit 401 at regular intervals
The video flowing address decoded on 200 whether there is, and decoding is continued executing with if there is then current video stream decoder module 200
Task, otherwise terminates the execution for the decoding video stream task that current video stream decoder module 200 is carried out.
Video flowing is read and decoding unit 202 is receiving what the scheduling of distributed task dispatching server unit 402 was performed
The decoding of video flowing is carried out after decoding video stream task and corresponding video flowing address.Decoding from general video file is different
It is that decoding is carried out to video flowing with very high requirement of real-time；If the decoding speed of video flowing is than video data on network
Transmission speed will result in video LOF slowly, due to having front and rear dependence between frame of video in decoding, after influenceing
The picture quality that the decoding of continuous frame of video is obtained is so as to influence the Text region effect of image.Read and decoding unit in video flowing
Multithreading pipelining is employed in 202 to avoid video LOF and realize the real-time processing of video flowing：
What multithreading pipelining was decoded, which implements, is：
1) frame of video thread Reader is read from video flowing；
2) frame of video to reading carries out decoding thread Decoder；
3) post processing thread Postprocesser is carried out to the decoding data obtained after decoding.Reader threads are from network
It is middle to read the video data transmitted, in the queue for the video data bag of reading being put into a packet_queue；
Decoder threads are successively read video data bag from packet_queue queues, and the video data bag to reading is decoded
Decoding data is obtained, in the queue for obtained decoding data being put into a frame_queue；Postprocesser threads
The greyscale image data in decoding data, separation decoding data is obtained from frame_queue, and greyscale image data information
It is sent to image recognition request unit 203.
Image recognition request unit 203 receives greyscale image data information deposit buffer unit 401 and to distribution
Task scheduling server unit 402 sends the Text region request of the gray level image received.Wherein to distributed task dispatching
The parameter carried in the request that server unit 402 is sent is that the gray level image number of deposit can be obtained from buffer unit 401
According to corresponding key.Text region module 300 obtains corresponding greyscale image data to carry out by the key from buffer unit 401
Text region processing.
Text region module 300 is receiving the Text region that the scheduling of distributed task dispatching server unit 402 is performed
Corresponding gray level image being read from buffer unit 401 after task and carrying out Text region processing, finally recognition result is protected
The post processing such as deposit.The Text region module 300 of the present invention includes images to be recognized reading unit 301, image character recognition unit
302 and post-processing unit 303.
Images to be recognized reading unit 301 is to receive the word of the scheduling of distributed task dispatching server unit 402
Identification mission and for from buffer unit 401 obtain greyscale image data key after, from buffer unit 401 read key institute it is right
The greyscale image data answered is sent to image character recognition unit 302.
Image character recognition unit 302 is that the gray level image sended over to image fetching unit 301 to be identified enters style of writing
Word is recognized.The judgement to be identified first is made whether to gray level image before identification, if image is not belonging to images to be recognized
Scope then directly abandon image, otherwise carrying out related Text region to gray level image is handled.Text region result
It is sent to post-processing unit 303.
Post-processing unit 303 is tied after the Text region result of the transmission of image character recognition unit 302 is received to identification
Fruit carries out the processing such as preserving, while deleting the greyscale image data information that completion has been recognized in buffer unit 401.
As shown in Fig. 2 flow of the present invention is：The opening unit 101 of video flow processing task takes to distributed task dispatching
Business device unit 402 sends the processing request of video flowing and video flowing address is write buffer unit 401, and distributed task dispatching takes
Business device unit 402 receives one decoding video stream module 200 of scheduling after video flow processing request and goes the decoding for performing video flowing to appoint
Business, the reading of video flowing and decoding unit 202 carry out decoding process to video flowing, and obtaining please by image recognition after an image
Unit 203 is asked to preserve the word knowledge that image sends image to buffer unit 401 and to distributed task dispatching server unit 402
Do not invite and ask.Distributed task dispatching server unit 402 receives one Text region module 300 of scheduling after Text region request and entered
The Text region processing of row image, images to be recognized reading unit 301 is read after image data from buffer unit 401, character
Recognition unit 302 carries out Text region to image, and the text information after 303 pairs of identifications of post-processing unit carries out the processing such as preserving.
Stop decoding unit 201 to detect whether to stop the signal of decoding process.Execution if then terminating current decoding task,
After the decoding of video flowing stops, Text region module will not also receive the pictograph identifying processing request of correspondence video flowing, from
And realize the end of whole video flow processing task；Otherwise the decoding of video flowing and the Text region processing of image are repeated.
The system that Fig. 3 is the present invention implements schematic diagram.Decoding video stream and Text region are all to have used multiserver collection
Group, these servers all can use is scheduled execution corresponding video stream solution by distributed task dispatching server unit 402
Code or Text region task.The server of buffer unit 401 is to pass to Text region server as decoded view data
Caching use and close the transmission pond of decoding video stream signal.Client is used to manage video flowing, to distributed task dispatching
Server unit 402 sends decoding video stream request or closes decoding video stream process signal to the transmission of buffer unit 401.
In order to as it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not
To limit the present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention etc. all should
Within protection scope of the present invention.
Claims (5)
1. the distributed character identification system of a kind of live video stream, it is characterised in that the system includes：
Video stream management module (100), for managing being turned on and off for video flow processing task, video flow processing task is opened
When to task scheduling modules (400) submit video flow processing task unlatching request and video flowing address；Video flow processing task
The turn-off request of video flow processing task is submitted during closing to task scheduling modules (400)；
Task scheduling modules (400), for storing video flowing address and view data, the unlatching for receiving video flow processing task please
An idle decoding video stream module (200) is dispatched after asking and performs video decoding task；Receive the Text region of view data
An idle Text region module (300) is dispatched after request and performs Text region task；Receive the pass of video flow processing task
Close the video flowing address that storage is deleted after request；
From task scheduling mould after decoding video stream module (200), the video decoding task for receiving task scheduling modules (400)
Video flowing address is read in block (400), video data is obtained and carries out decoding process, obtained view data deposit task is decoded
Scheduler module (400) simultaneously submits the Text region of the view data to ask to task scheduling modules (400)；The mould of video decoding simultaneously
Whether the video flowing address that block (200) inquires about its processing to task scheduling modules (400) at regular intervals has been deleted, if by
The decoding process for then stopping video flowing being deleted, decoding operate is otherwise continued executing with；
From task scheduling modules after Text region module (300), the Text region task for receiving task scheduling modules (400)
(400) view data is read in and carries out Text region, identified figure is deleted after the completion of identification from task scheduling modules (400)
As data.
2. the distributed character identification system of a kind of live video stream according to claim 1, it is characterised in that described
Business scheduler module (400) includes：
Buffer unit (401), for storing video flowing address and view data；
Distributed task dispatching server unit (402), a free time is dispatched after the unlatching request for receiving video flow processing task
Decoding video stream module (200) perform video decoding task；A sky is dispatched after the Text region request for receiving view data
Not busy Text region module (300) performs Text region task；Receive and caching list is deleted after the turn-off request of video flow processing task
The video flowing address of storage in first (401).
3. a kind of distributed character identification system of live video stream according to claim 1 or 2, it is characterised in that institute
Stating video stream management module (100) includes：
The opening unit (101) of video flow processing task, video flowing is submitted for distributed task dispatching server unit (402)
The unlatching request of processing task, and video flowing address is sent to buffer unit (401) storage；
The closing unit (102) of video flow processing task, takes when being closed for video flow processing task to distributed task dispatching
The turn-off request for device unit (402) submission video flow processing task of being engaged in.
4. a kind of distributed character identification system of live video stream according to claim 1 or 2, it is characterised in that institute
Stating decoding video stream module (200) includes：
Stop decoding unit (201), for being spaced the video flowing that query caching unit (401) is being used according to set time
Whether location is deleted, and stops the video decoding task of video flowing reading and decoding unit (202) if deleting；Otherwise do not do and locate
Reason；
Video flowing is read and decoding unit (202), and the scheduling for receiving distributed task dispatching server unit (402) is read
Video flowing address in buffer unit (401) performs video decoding task, obtains decoding image, and decoding image is transmitted into image
Recognize request unit (203)；
In image recognition request unit (203), the greyscale image data deposit buffer unit (401) for extracting decoding image simultaneously
The Text region request of view data is sent to distributed task dispatching server unit (402).
5. a kind of distributed character identification system of live video stream according to claim 1 or 2, it is characterised in that institute
Stating Text region module (300) includes：
Images to be recognized reading unit (301), the word for receiving distributed task dispatching server unit (402) transmission is known
Do not invite and ask, read view data from buffer unit (401) and send to image character recognition unit (302)；
Character recognition unit (302), for judging whether view data needs to carry out Text region processing, if desired then to image
Data carry out Text region and send the Text region result of acquisition to post-processing unit (303)；
Post-processing unit (303), for receiving Text region result, and deletes character recognition unit in buffer unit (401)
(302) treated view data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710504745.5A CN107169480B (en) | 2017-06-28 | 2017-06-28 | Distributed character recognition system of real-time video stream |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710504745.5A CN107169480B (en) | 2017-06-28 | 2017-06-28 | Distributed character recognition system of real-time video stream |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107169480A true CN107169480A (en) | 2017-09-15 |
| CN107169480B CN107169480B (en) | 2020-05-19 |
Family
ID=59827944
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710504745.5A Active CN107169480B (en) | 2017-06-28 | 2017-06-28 | Distributed character recognition system of real-time video stream |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107169480B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114466227A (en) * | 2021-12-22 | 2022-05-10 | 天翼云科技有限公司 | Video analysis method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102404615A (en) * | 2011-11-29 | 2012-04-04 | 广东威创视讯科技股份有限公司 | Video processing system based on cloud computing |
| CN104410838A (en) * | 2014-12-15 | 2015-03-11 | 成都鼎智汇科技有限公司 | Distributed video monitoring system |
| CN106131185A (en) * | 2016-07-13 | 2016-11-16 | 腾讯科技（深圳）有限公司 | The processing method of a kind of video data, Apparatus and system |
-
2017
- 2017-06-28 CN CN201710504745.5A patent/CN107169480B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102404615A (en) * | 2011-11-29 | 2012-04-04 | 广东威创视讯科技股份有限公司 | Video processing system based on cloud computing |
| CN104410838A (en) * | 2014-12-15 | 2015-03-11 | 成都鼎智汇科技有限公司 | Distributed video monitoring system |
| CN106131185A (en) * | 2016-07-13 | 2016-11-16 | 腾讯科技（深圳）有限公司 | The processing method of a kind of video data, Apparatus and system |
Non-Patent Citations (5)
| Title |
|---|
| 刘吉刚: "视频文字识别技术的研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 姚志鹏: "基于Hadoop平台的印刷体蒙古文字识别系统的研究与实现", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 孟帅: "Hadoop平台下的并行文字识别技术研究", 《2014全国计算机网络与通信学术会议优秀论文》 * |
| 杨帆: "分布式系统Hadoop平台的视频转码", 《计算机系统应用》 * |
| 颜建强: "图像视频复杂场景中文字检测识别方法研究", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114466227A (en) * | 2021-12-22 | 2022-05-10 | 天翼云科技有限公司 | Video analysis method and device, electronic equipment and storage medium |
| CN114466227B (en) * | 2021-12-22 | 2023-08-04 | 天翼云科技有限公司 | Video analysis method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107169480B (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20190075178A1 (en) | Push Notification Delivery System with Feedback Analysis | |
| CN108985293A (en) | A kind of image automation mask method and system based on deep learning | |
| CN110971939B (en) | Illegal picture identification method and related device | |
| CN109918184A (en) | Picture processing system, method and relevant apparatus and equipment | |
| CN110806954A (en) | Method, device and equipment for evaluating cloud host resources and storage medium | |
| CN109902195A (en) | Monitoring image querying method, device, equipment and medium | |
| US8112450B2 (en) | Priority messaging and priority scheduling | |
| CN108959337A (en) | Big data acquisition methods, device, equipment and storage medium | |
| US11494409B2 (en) | Asynchronously processing sequential data blocks | |
| CN109766435A (en) | The recognition methods of barrage classification, device, equipment and storage medium | |
| CN106033438B (en) | Public sentiment data storage method and server | |
| CN110647544A (en) | Account checking method based on stream data | |
| CN110275779B (en) | Resource acquisition method, device, equipment and storage medium | |
| CN102129619A (en) | Service data processing method and system based on cloud storage | |
| CN107169480A (en) | A kind of distributed character identification system of live video stream | |
| CN113723926A (en) | Bank pipelining processing method and device combining RPA and AI and electronic equipment | |
| CN112231344A (en) | Real-time stream data query method and device | |
| Wang et al. | Equilibrium customer strategies in the Geo/Geo/1 queue with single working vacation | |
| CN115062676B (en) | Data processing method, device and computer readable storage medium | |
| CN113269577B (en) | Data acquisition method, device, server and storage medium | |
| CN107730422A (en) | One kind is based on face recognition online testing analysis system | |
| US20210055976A1 (en) | Processing of a message stream | |
| CN111611523A (en) | Resource management system, resource management method, device, and storage medium | |
| CN111143688A (en) | Evaluation method and system based on mobile news client | |
| CN110968434A (en) | Method for managing machine and related device |
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 |