CN105979218A - High-definition real-time monitoring video efficient transmission method - Google Patents
High-definition real-time monitoring video efficient transmission method Download PDFInfo
- Publication number
- CN105979218A CN105979218A CN201610420589.XA CN201610420589A CN105979218A CN 105979218 A CN105979218 A CN 105979218A CN 201610420589 A CN201610420589 A CN 201610420589A CN 105979218 A CN105979218 A CN 105979218A
- Authority
- CN
- China
- Prior art keywords
- grades
- cachings
- data
- cache
- request
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a high-definition real-time monitoring video efficient transmission method. The method is realized on the basis of an ARM video transmission board by adopting a multi-input multi-output system of technologies of multi-thread processing, multi-level data caching, high-concurrent mutual exclusion access and the like. The hardware mainly comprises the ARM video transmission board, a USB high-definition camera and a power supply module; and the software treating process mainly comprises a multi-camera input part, an image data warehouse and a multi-request output part. The method occupies a few resources, is efficient in operation, can guarantee the quality and the instantaneity of high-definition real-time video, and is applicable in the field of real-time video monitoring, particularly in the fields having high efficiency requirement, limited hardware resources and efficient software method requirement.
Description
Technical field
The present invention relates to real-time video monitoring field, efficiently pass particularly to a kind of real-time monitor video of high definition
Transmission method.
Background technology
Real-time video monitoring system is as the important component part of industrial monitoring Yu safety monitoring, directly perceived with it
The abundant form of traversal is widely used in the occasions such as factory, bank, road and community, real-time video monitoring
System generally by video acquisition, transmit, the part such as display forms.
The occasion serious in the compression of some costs, efficiency requirements is harsh, such as the area such as mountain forest, desert, one
Aspect is wide due to area, and monitor node requires more, and the configuration of equipment therefor is too high is easily caused device
Internal power source quickly consumes, and too increases holistic cost.Therefore it is required that the cost requirement of single monitor node
As far as possible low, individual node monitoring range require as far as possible high.On the other hand, for existing to monitoring
The requirement of field monitoring in real time, needs to be supervised by this video camera within a short period of time at photographic head institute monitoring range
The video data controlled quick, safe and efficient is transferred to destination.
Affect the factor of video monitoring system high efficiency of transmission, mainly uploaded videos data big, and lower line number
Smaller according to such as controlling data stream, upstream bandwidth size also determines the access quantity of photographic head simultaneously.By
The network bandwidth upper limit needed for actual transmissions, can only often much larger than the network bandwidth required by this agreement
The mode using height configuration hardware to exchange the network bandwidth for improves transmission quality.Real-time video monitoring method at present
Use higher hardware configuration more, be compressed to after video acquisition as H.264, the data of the agreement such as RTP lead to
Cross network transmission.While the method using height configuration hardware solves high efficiency of transmission, the system that also brings is opened
Send out that cycle length, software recycling rate of waterused be low, the overall expensive hardware of system, the defects such as system failure rate is high.
In sum, existing real-time video monitoring method cannot meet hardware complexity low, run effect
Rate is high, the high requirement of network bandwidth utilization factor, needs a kind of realization height under the conditions of relatively low hardware configuration
The method of clear real-time video high efficiency of transmission.
Summary of the invention
In view of the defect on above-mentioned real-time video monitoring method framework and technical deficiency, the present invention is main
Purpose is a kind of high definition real-time monitor video efficient transmission system of design, it is achieved low in hardware platform complexity
Platform on carry out the high efficiency of transmission of video data.
A kind of high definition real-time monitor video high-efficiency transmission method, comprises the following steps:
Gather the video data of all photographic head, the video data of each photographic head is decomposed into continuous print list
Two field picture, and every two field picture is stored in the level cache individually distributed for each photographic head；Then will be each
View data in individual level cache moves to two grades of the view data warehouse as all level caches to be delayed
In depositing；After receiving the request extracting corresponding camera video content and being started by acceptance to perform, create one
Individual corresponding three grades of cachings, and the view data in L2 cache corresponding for request is moved to three grades of cachings,
Persistently send three grades caching in view data until request terminate, finally terminate these three grades caching with
Discharge the resource shared by these three grades cachings.
Described method, described level cache is the single frames latest image data depositing corresponding photographic head
Extra buffer, after single-frame images is copied to L2 cache, empties the data in level cache to connect
Receive next frame latest image.
Described method, described L2 cache is to preserve each level cache with queue form to move
The view data warehouse of single-frame images data, after corresponding capacity of queue exceedes default threshold value, removes
A oldest two field picture is to receive up-to-date image.
Described method, three grades of described cachings replicate for storage single frames responds request from L2 cache
The extra buffer of the view data come.
Described method, when the request extracting corresponding camera image content is single frames picture, first creates
Three grades of cachings, then take out view data corresponding in L2 cache and move to three grades of cachings, delay from three grades
Deposit transmission view data, finally terminate three grades of cachings to discharge the space that this three grades of cachings take.
Described method, when the request extracting corresponding camera image content is live video stream, first creates
Build three grades of cachings, then a oldest frame in image queue data corresponding in L2 cache is moved to three
Level caching, sends view data from three grades of cachings, then empties the view data in three grades of cachings, continue
From L2 cache, mobile corresponding view data to three grades of cachings and circulates execution above-mentioned steps until request
Terminate, finally terminate three grades of cachings to discharge the space that this three grades of cachings take.
Described method, produces conflict when being provided with in described L2 cache for preventing write and reading
Mutual exclusion lock, certain image queue data of L2 cache being write or during read operation, this image
Mutual exclusion lock corresponding to queuing data carries out locking only to allow uniquely to write these image queue data
Enter or read, unlocking again after writing or read accordingly.
The method have technical effect that: the use of level cache makes the read-write to L2 cache not by each
The impact of camera collection speed and more efficient, reduce multiple photographic head to L2 cache simultaneously simultaneously
The competitive relation of write, not by each when the use of three grades of cachings makes to reply multiple requests respectively
The impact of network environment and more efficient, also reduce multiple output function thread and L2 cache read simultaneously
Competitive relation when taking.The picture exported owing to have employed photographic head has carried out frame data compression and has not carried out frame
Between compress, single frames picture is taken up space little, is reduced cpu load, exchanged for CPU time with the network bandwidth
Mode carry out the transmission of HD video data.Whole transmitting procedure is not related to video compression operation, therefore
The hardware platform of run the method is configured extremely low, and operational efficiency is high.
Accompanying drawing explanation
Fig. 1 is the data stream of the present invention and performs flow chart；
Fig. 2 is the System Initialization Procedure figure of the present invention；
Fig. 3 is that the system of the present invention closes cleaning work flow chart；
Fig. 4 is the input initialization module frame figure of the present invention；
Fig. 5 is the input operation thread frame diagram of the present invention；
Fig. 6 is the global image data warehouse content of the present invention；
Fig. 7 is the output initialization module frame diagram of the present invention；
Fig. 8 is the output function thread frame diagram of the present invention.
Detailed description of the invention
Below in conjunction with the accompanying drawings and be embodied as the present invention is described in further detail again.
The present embodiment comprises the following steps:
Gather the video data of all photographic head, the video data of each photographic head is decomposed into continuous print list
Two field picture, and every two field picture is stored in the level cache individually distributed for each photographic head；Then will be each
View data in individual level cache moves to two grades of the view data warehouse as all level caches to be delayed
In depositing；After receiving the request extracting corresponding camera video content and being started by acceptance to perform, create one
Individual corresponding three grades of cachings, and the view data in L2 cache corresponding for request is moved to three grades of cachings,
Persistently send three grades caching in view data until request terminate, finally terminate these three grades caching with
Discharge the resource shared by these three grades cachings.
Level cache is the extra buffer of the single frames latest image data depositing corresponding photographic head, by single
After two field picture copies to L2 cache, empty the data in level cache to receive next frame latest image.
L2 cache is the image preserving the single-frame images data that each level cache moves with queue form
Data warehouse, after corresponding capacity of queue exceedes default threshold value, removes a oldest two field picture to connect
Receiving up-to-date image, threshold value can set according to specific needs, and the threshold value in the present embodiment is set to 160 frames.
Three grades of cachings responds from L2 cache for storage single frames asks and the view data that replicates interim
Relief area.
When the request extracting corresponding camera image content is single frames picture, first create three grades of cachings, so
View data corresponding in rear taking-up L2 cache moves to three grades of cachings, sends picture number from three grades of cachings
According to, finally terminate three grades of cachings to discharge the space that this three grades of cachings take.
When the request extracting corresponding camera image content is live video stream, first create three grades of cachings,
Then a oldest frame in image queue data corresponding in L2 cache is moved to three grades of cachings, from three
Level caching sends view data, then empties the view data in three grades of cachings, continues from L2 cache
Mobile corresponding view data is to three grades of cachings and circulates execution above-mentioned steps until request terminates, and finally ties
Restraint three grades of cachings to discharge the space that this three grades of cachings take.
Produce the mutual exclusion lock of conflict when L2 cache being provided with for preventing write and reading, two grades are being delayed
Certain the image queue data deposited carry out writing or during read operation, mutual corresponding to these image queue data
Lock is scolded to carry out locking only to allow these image queue data are uniquely write or read, accordingly
Unlock again after writing or having read.
The present embodiment in terms of hardware configuration by ARM transmission of video plate, USB high-definition camera and power supply
Module form, the work on software processing method mainly have global image data warehouse, system initialization,
Input initialization module, input operation thread, output initialization module, each output function thread and system
Close cleaning work.
Wherein ARM transmission of video plate is made up of core board and base plate；ARM transmission of video core board is by core
The IPEX interface of heart processor AR9331, internal memory 64MB DDR2,16MB SPI Flash and extraction
Composition.Wherein, AR9331 core is MIPS24K monokaryon, and dominant frequency 400MHz can be on a single chip
Realize IEEE 802.11n 1x1AP or routing function.Base plate is by video acquisition module, debugging module, net
Network transport module, power module and signal amplification module composition.Video acquisition module has drawn 2 USB 2.0
Host interface, for controlling USB high-definition camera output HD video data, and passes through USB2.0
Agreement by digital independent to internal memory；Debugging module is drawn 1 UART serial ports and is changed by CH340G
After, USB MINI interface draw, for program burn writing, run debugging, data transmission；Network passes
Defeated module draws 2 WAN/LAN 10/100M self adaptation network interfaces, for by wired mode and the external world
Carry out network service；Power module introduces 3.3V DC source, for providing for each module of transmission of video plate
Power supply；The wireless signal of the IEEE802.11b/g/n agreement that IPEX interface is drawn is put by signal amplification module
After big and launch or receive into, transmission rate is 150Mbps to the maximum, and transmission range is up to 500
Rice.
The complete of in management method all key datas is specifically included that on the software processing method of this system
Office view data warehouse；For starting system and carrying out the of overall importance initialized system unrelated with concrete operations
System initialization module；For initializing each photographic head, distribute the input initialization mould of each level cache
Block；For performing gathering and the input operation thread of storage work of each photographic head；For receiving the external world
HTTP request, and distribute the output initialization module of thread process work；For performing the content of request,
I.e. transmit each output function thread of real-time video；End is pressed for performing system user when front stage operation
Only after key CTRL+C, the system of cleaned system work closes cleaning module.
Described global image data warehouse includes: running mark position, depot data address, warehouse are respectively taken the photograph
As head size of data, the mutual exclusion lock array of offer multithreading secure access, it is provided that whether have latest data
Conditional-variable array.The photographic head one_to_one corresponding that wherein mutual exclusion lock array element sorts with each, described
The photographic head one_to_one corresponding that conditional-variable array sorts with each.
Described system initialization module includes: initializes global image data, opens system log,
Ignoring SIGPIPE signal, arrange SIGINT signal process function, write-in program version is to system record
Device, performs input initialization module, performs to export initialization module, create and perform each output function line
Journey, enters sleep state wait and is waken up.Wherein cleaned system work includes: write terminates daily record to system
Recorder, the running mark arranging global image data warehouse is stopping, and the handle opened also is closed in time delay,
Destroy mutual exclusion lock and the conditional-variable of global image data warehouse, close system log.
Described input initialization module contents includes: open each camera device file, test shooting
Whether head supports video acquisition, supports that then arranging photographic head is output, arranges resolution and output format,
Application caching, puts into buffer queue the caching of request, distributes level cache.Wherein distribute each one-level
Cache contents is that each photographic head distributes a data field that can preserve single-frame images.
Described input operation thread content includes: arranges thread cleaning operation collection, starts and read correspondence
The image of photographic head, request take global data mutual exclusion lock, the image of reading is put into warehouse and update right
Should the warehouse size of photographic head, notify that the conditional-variable of thread of all this photographic head of wait is for there being new number
According to can read, request releases global data mutual exclusion lock.Its thread cleaning operation collection includes: close correspondence
Photographic head exports, the level cache of release distribution, discharges photographic head handle, discharges photographic head status information,
Release photographic head filename.
Described output initialization module content includes: arranges thread cleaning function, opens socket, if
It is reusable for putting socket, bound socket, monitors socket and circulates acceptance request and create output behaviour
Process as thread.
Described output function thread content includes: read a line HTTP request, it is judged that request content is
No completely, resolve user name, password whether in data base and correct, the most then perform request-reply behaviour
Make.Wherein response operation includes: wait the correspondence indicating whether new data of global image data warehouse
Conditional-variable also pins mutual exclusion lock, distributes three grades of cachings and resets, replicating warehouse content to three grades caching,
Unlock global image data warehouse corresponding mutual exclusion lock, send HTTP reply, three grades of cache contents and
boundary。
Described system request and response content are simply efficient, and after once asking to be received, content judges to close
Each frame after method is replied data and is only included content type, content-length, single frames image data and boundary.
The data stream of system of the present invention and perform flow chart as it is shown in figure 1, following to data stream and holding
Each step of row flow process is described in detail:
Step one: system initialization, it performs frame diagram as shown in Figure 2.
First beginning to part when system initialization is system start-up, concrete execution is as follows:
(1) Global Repository Array for structural body is initialized；
(2) open a programming system recorder, and preserve its link；
(3) receiving SIGPIPE signal, default-action is the execution terminating native system, ignores in native system
This signal, to prevent from receiving this signal and making program by system closure when buttoned-up socket sends data；
(4) receiving SIGINT signal, default-action is also the execution terminating native system, and arranging at this should
Signal process function, receives this signal to prevent native system to be transfused to CTRL+C when performing on foreground straight
Connect and exit and do not do cleaning work；
Wherein, the system of signal process function closes cleaning work flow process as it is shown on figure 3, concrete execution is:
1. print and terminate daily record to system log；2. arranging running mark position, warehouse is to stop, so that all inspections
The thread looking into this flag bit normally exits loop body and then exits the execution of thread；3. time delay 1s waits all
Exiting of sub-line journey；4. each handle opened is closed in circulation, prevents the waste to system handle count；⑤
Destroy mutual exclusion lock and the conditional-variable in warehouse；6. system log is closed.
(5) version information of write-in program is to system log；
(6) reading parameter configuration files and initialize the parameter of input-output operation structure, native system is permitted
Static initialization parameter configuration files is to determine some parameter when this method performs the most before execution；
(7) the required L2 cache of each photographic head in warehouse, each photographic head required memory it are sequentially allocated
Being determined by its resolution, the biggest resolution makes single frames picture the biggest, it is therefore desirable to the most two grades delay
Deposit；
(8) perform input initialization module carry out each photographic head initialization and distribute each one-level delay
Deposit；
(9) perform output initialization module and receive extraneous HTTP request, and distribute thread process work；
(10) create and perform fast threads distribution when each output function thread is provided with request arrival；
(11) system initialization has so far worked and has entered sleep state, waits and being waken up up by signal, as
SIGINT signal.
Step 2: photographic head input channel initializes, and its role is to initialize each photographic head so that it is energy
The video data of specified format in enough output profiles, but and be not turned on the output function of each photographic head,
Shown in program flow diagram such as Fig. 4 (a)
(1) from system initialization part arrange input operation structure extract photographic head implementor name,
The information of resolution, frame frequency, video output formats etc.；
(2) analyze and camera information structure is set；
(3) circulation carries out the initialization of photographic head, shown in its initialization procedure such as Fig. 4 (b):
1. the device file of corresponding photographic head is opened；
2. judge whether photographic head supports video acquisition, support that then carrying out photographic head is set to output；
3. photographic head parameter is set, including resolution, formats, driving form is set；
4. the frame frequency of photographic head is set, asks 4 to be in kernel state for photographic head number in MMAP mode
The relief area preserved according to circulation；
5. buffer queue is put in the relief area of request, complete the initialization of photographic head.
(4) it is that native system distribution is for depositing the extra buffer of single frames up-to-date camera image data, i.e.
Level cache, for preserving the up-to-date photographic head image data collected, prevents from being directly placed into L2 cache
The performance consumption that other operation L2 caches are brought；
(5) create and perform output function thread.
Step 3: photographic head input operation thread, it performs shown in frame diagram such as Fig. 5 (a).
Data acquisition for each photographic head in system correspond to an input operation thread, and it is held
Row flow process is as follows:
(1) thread cleaning function is set, carries out clearing up and discharging work when function exits, perform flow process
As shown in Fig. 5 (b):
1. the video frequency output of corresponding photographic head is closed；
2. interim single frames image cache i.e. level cache is discharged；
3. release photographic head handle is in order to avoid meaningless must consume system file handle count；
4. release photographic head status information is in order to avoid causing memory overflow；
5. the internal memory of photographic head filename is discharged；
6. circulation performs the collecting work of photographic head；
7. the cleaning work of thread is performed when program exits.
(2) camera collection, it performs shown in frame diagram such as Fig. 5 (c).
1. obtain a two field picture of corresponding photographic head, its operating process such as Fig. 5 (d) so: judgement if
Perform then to start photographic head for the first time to start to export video data；Data are obtained from video buffer queue；
Judge that image data size, whether less than threshold value, if less than then giving up, waits next frame image；If
Be not less than, then image reproduction data is in the i.e. level cache in photographic head single-frame images relief area；Again should
Data field is put into video buffer queue and is waited that webcam driver is filled in data；
2. record image size is to journal file；
3. pin the mutual exclusion lock of this photographic head of data warehouse, will sleep without obtaining mutual exclusion lock, as
Fruit obtains mutual exclusion lock resource and other threads will be made cannot to obtain this mutual exclusion lock resource；
4. monitoring view data integrity, if there being damage, repairs, and then replicates level cache number
According in corresponding warehouse photographic head L2 cache and update the value of the current size of this caching；
5. trigger all threads performing to wait the conditional-variable of corresponding photographic head, i.e. wake up up in output channel
Execution thread etc. stream to be written or file；
6. unlock the mutual exclusion lock of this photographic head of data warehouse so that other etc. this mutual exclusion lock resource to be obtained
Thread again have an opportunity obtain mutual exclusion lock resource；
7. output journal is to system log.
Step 4: video data enters global image data warehouse, it performs frame diagram as shown in Figure 6.
Particular content and acting as: 1. running mark position, it is provided that run and the mark stopped in corresponding native system
Will is pointed out.It is set to running mark when running and starting, is set to when needs terminate and run stop mark,
The thread of each circulation circulates every time all can check whether this mark is off, if it is, thread will be terminated
Circulation, terminates the execution of thread；
2. depot data address array, it is provided that the storage ground of each photographic head L2 cache in corresponding native system
Space, location.The L2 cache regional address of the fixing corresponding photographic head of each item in array, L2 cache is not
Being same as one-level and three grades of cachings, this space can store the image more than a frame；
3. warehouse photographic head size of data array, it is provided that to each photographic head L2 cache space in this method
The size of current latest data amount, in array, each item fixes to obtain the L2 cache of a corresponding photographic head
District's current data amount size；
4. mutual exclusion lock array, it is provided that the mutual exclusion of each photographic head L2 cache digital independent in corresponding native system
Property access.The L2 cache mutual exclusion lock of the fixing corresponding photographic head of each item in array, this mutual exclusion lock makes
Multiple threads to be unlikely to read reading photographic head L2 cache when be not the number of desired locations
According to, the position also making multiple thread be unlikely to be written to writing photographic head L2 cache when is not
Desired locations.
5. conditional-variable array, it is provided that in this method for certain photographic head latest data write two grades
To waiting that this photographic head latest data reads the notice of thread after caching successfully, read at input operation thread
Taking photographic head to after level cache success writing L2 cache, notice waited has slept
Output function thread is converted to run mode, and is read out L2 cache and carries out next step operation.
Step 5: photographic head output channel initializes.It performs frame diagram as it is shown in fig. 7, it is specifically grasped
Make execution as follows:
(1) from input operation structure, the information such as port numbers, voucher is extracted；
(2) the network connection i.e. structure of thread information is allocated and set up for network connecting part；
(3) the cleaning function of thread is set；
(4) socket is opened；
(5) arrange socket to make it possible to reuse；
(6) bound socket；
(7) socket is monitored；
(8) circulation performs the process of request, processes content and includes: 1. create connection structural bodies to unify
Management；2. request is accepted；3. write request information is to system log；4. create and perform output function
Thread；5. thread is separated so that output initialization thread need not wait the end of output function thread, defeated
Go out to operate can discharge immediately after thread terminates the system resource oneself taken.
Step 6: photographic head output channel operation thread, outside output function thread effect is that circulation receives
Boundary asks, it is judged that give a response according to request type after legitimacy.It performs shown in frame diagram such as Fig. 8 (a),
Concrete execution is as follows:
(1) thread cleaning function is set；
(2) opening socket, it is reusable for arranging it, binds and monitors socket, wait connection request；
(3) when stopping mark and being set, stop circulation, exit thread；
(4) when have request arrive, then create new thread process request, record request time, request type,
The information such as request source IP and port numbers, request to create processes thread and thread carries out lock out operation so that
New thread can Resource recovery independently after being finished.Hereafter, output function thread continues cycling through and connects
Receive new request.
Wherein, processing shown in process thread flow such as Fig. 8 (b) of request, request processing thread is for respectively
The legitimacy verification of individual request and the response to difference request.Will read HTTP for each request please
Seek content, it is judged that request type and account number cipher, just responsed by rear, not by being shut off socket
And remove distributed resource.
If request type is single frames picture, shown in flow chart such as Fig. 8 (c), first attempt to pin correspondence in warehouse
Mutual exclusion lock, distribute single-frame images internal memory, i.e. three grades cachings, taking-up warehouse correspondence memory after pinning successfully
Place's view data copies to three grades of cachings, then release warehouse correspondence mutual exclusion lock, the reply made requests on.
Recovery Process is first to send HTTP to reply, and retransmits view data.The finally space of release three grades caching,
Complete request processing thread.
If request type is live video stream, shown in flow chart such as Fig. 8 (d), send HTTP and reply,
Cycle through single-frame images data and form video flowing.The stopping mark of data warehouse is set, then circulate end
Only, three grades of caching backed off after random request processing threads are discharged.Stop mark not being set, then loop body is always
Circulation performs.Continue waiting for updating respective conditions variable after input operation thread reads data, do not update then
This thread is absorbed in sleep；If updated, this thread will be waken up up by conditional-variable, right in replicating L2 cache
Answer view data to three grades caching, adjust three grades of cache size the most in right amount, unlock corresponding mutual exclusion lock,
Send the reply of request.Recovery Process is first to send to include that the HTTP of content type and content-length replys,
Retransmit the view data in three grades of cachings, finally send frame boundaries.
Claims (7)
1. a high definition real-time monitor video high-efficiency transmission method, it is characterised in that comprise the following steps:
Gather the video data of all photographic head, the video data of each photographic head is decomposed into continuous print list
Two field picture, and every two field picture is stored in the level cache individually distributed for each photographic head；Then will be each
View data in individual level cache moves to two grades of the view data warehouse as all level caches to be delayed
In depositing；After receiving the request extracting corresponding camera video content and being started by acceptance to perform, create one
Individual corresponding three grades of cachings, and the view data in L2 cache corresponding for request is moved to three grades of cachings,
Persistently send three grades caching in view data until request terminate, finally terminate these three grades caching with
Discharge the resource shared by these three grades cachings.
Method the most according to claim 1, it is characterised in that described level cache is right for depositing
Answer the extra buffer of the single frames latest image data of photographic head, single-frame images is being copied to L2 cache
After, empty the data in level cache to receive next frame latest image.
Method the most according to claim 1, it is characterised in that described L2 cache is with queue
Form preserves the view data warehouse of the single-frame images data that each level cache moves, when corresponding team
After row capacity exceedes default threshold value, remove a oldest two field picture to receive up-to-date image.
Method the most according to claim 1, it is characterised in that three grades of described cachings are single for storage
Frame responds request from L2 cache and replicates the extra buffer of the view data come.
Method the most according to claim 1, it is characterised in that in extracting corresponding camera image
When the request held is single frames picture, first creates three grades of cachings, then take out image corresponding in L2 cache
Data move to three grades of cachings, send view data from three grades of cachings, finally terminate three grades of cachings with release
The space that this three grades of cachings take.
Method the most according to claim 1, it is characterised in that in extracting corresponding camera image
When the request held is live video stream, first create three grades of cachings, then by image corresponding in L2 cache
A frame the oldest in queuing data moves to three grades of cachings, sends view data from three grades of cachings, the most clearly
View data in empty three grades of cachings, continues mobile corresponding view data from L2 cache and delays to three grades
Deposit and circulate execution above-mentioned steps until request terminates, finally terminate three grades of cachings to discharge these three grades cachings
The space taken.
Method the most according to claim 1, it is characterised in that be provided with use in described L2 cache
The mutual exclusion lock of conflict is produced, in certain the image queue data to L2 cache in time preventing write and read
Carrying out writing or during read operation, this mutual exclusion lock corresponding to image queue data carries out locking only to allow
These image queue data are uniquely write or read, solves again after writing or read accordingly
Lock.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610420589.XA CN105979218A (en) | 2016-06-13 | 2016-06-13 | High-definition real-time monitoring video efficient transmission method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610420589.XA CN105979218A (en) | 2016-06-13 | 2016-06-13 | High-definition real-time monitoring video efficient transmission method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105979218A true CN105979218A (en) | 2016-09-28 |
Family
ID=57021231
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610420589.XA Pending CN105979218A (en) | 2016-06-13 | 2016-06-13 | High-definition real-time monitoring video efficient transmission method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105979218A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108038462A (en) * | 2017-12-22 | 2018-05-15 | 深圳英飞拓科技股份有限公司 | A kind of method and device that Face datection is carried out to real-time video and is identified |
| CN108090008A (en) * | 2016-11-22 | 2018-05-29 | 谷歌有限责任公司 | For the system and method for parallel power monitoring |
| CN108415857A (en) * | 2018-03-01 | 2018-08-17 | 合肥威艾尔智能技术有限公司 | A kind of universal process method of serial data |
| CN108551490A (en) * | 2018-05-14 | 2018-09-18 | 西京学院 | A kind of industry flow data coding/decoding system and method |
| CN110072082A (en) * | 2019-03-22 | 2019-07-30 | 视联动力信息技术股份有限公司 | A kind of video broadcasting method and transfer server based on view networking |
| WO2020135306A1 (en) * | 2018-12-29 | 2020-07-02 | 华为技术有限公司 | Method for controlling camera, control device, network apparatus and camera |
| CN112003992A (en) * | 2020-08-14 | 2020-11-27 | 迅镭智能（广州）科技有限公司 | Transmission system and method based on scanning gun |
| CN113660426A (en) * | 2021-09-08 | 2021-11-16 | 英望科技(山东)有限公司 | Method and system for preventing photographing jitter of user and computer readable storage medium |
| CN114339349A (en) * | 2021-12-10 | 2022-04-12 | 海信视像科技股份有限公司 | Display device, data transmission method and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101562635A (en) * | 2009-05-15 | 2009-10-21 | 中兴通讯股份有限公司 | Method and player for mobile streaming media on demand |
| CN101715145A (en) * | 2008-10-06 | 2010-05-26 | 辉达公司 | Equipment, method and computer program product for evaluating processing capacity by utilizing cascade memories |
| CN102387409A (en) * | 2011-12-07 | 2012-03-21 | 江苏远为科技有限公司 | Method for video playing and based on stream media system |
| CN103098410A (en) * | 2010-09-14 | 2013-05-08 | 英特尔公司 | Increasing hybrid automatic repeat request (HARQ) throughput |
| US8707282B2 (en) * | 2009-12-14 | 2014-04-22 | Advanced Micro Devices, Inc. | Meta-data based data prefetching |
| CN104935680A (en) * | 2015-06-18 | 2015-09-23 | 中国互联网络信息中心 | Recursive domain name service system and method of multi-level shared cache |
| CN105045532A (en) * | 2015-07-09 | 2015-11-11 | 首都师范大学 | Three-level cache storage apparatus and method for dynamic reconfigurable bus monitoring system |
-
2016
- 2016-06-13 CN CN201610420589.XA patent/CN105979218A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101715145A (en) * | 2008-10-06 | 2010-05-26 | 辉达公司 | Equipment, method and computer program product for evaluating processing capacity by utilizing cascade memories |
| CN101562635A (en) * | 2009-05-15 | 2009-10-21 | 中兴通讯股份有限公司 | Method and player for mobile streaming media on demand |
| US8707282B2 (en) * | 2009-12-14 | 2014-04-22 | Advanced Micro Devices, Inc. | Meta-data based data prefetching |
| CN103098410A (en) * | 2010-09-14 | 2013-05-08 | 英特尔公司 | Increasing hybrid automatic repeat request (HARQ) throughput |
| CN102387409A (en) * | 2011-12-07 | 2012-03-21 | 江苏远为科技有限公司 | Method for video playing and based on stream media system |
| CN104935680A (en) * | 2015-06-18 | 2015-09-23 | 中国互联网络信息中心 | Recursive domain name service system and method of multi-level shared cache |
| CN105045532A (en) * | 2015-07-09 | 2015-11-11 | 首都师范大学 | Three-level cache storage apparatus and method for dynamic reconfigurable bus monitoring system |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108090008A (en) * | 2016-11-22 | 2018-05-29 | 谷歌有限责任公司 | For the system and method for parallel power monitoring |
| CN108038462A (en) * | 2017-12-22 | 2018-05-15 | 深圳英飞拓科技股份有限公司 | A kind of method and device that Face datection is carried out to real-time video and is identified |
| CN108415857A (en) * | 2018-03-01 | 2018-08-17 | 合肥威艾尔智能技术有限公司 | A kind of universal process method of serial data |
| CN108551490A (en) * | 2018-05-14 | 2018-09-18 | 西京学院 | A kind of industry flow data coding/decoding system and method |
| CN108551490B (en) * | 2018-05-14 | 2021-06-18 | 西京学院 | Industrial stream data coding and decoding system and method |
| WO2020135306A1 (en) * | 2018-12-29 | 2020-07-02 | 华为技术有限公司 | Method for controlling camera, control device, network apparatus and camera |
| CN110072082A (en) * | 2019-03-22 | 2019-07-30 | 视联动力信息技术股份有限公司 | A kind of video broadcasting method and transfer server based on view networking |
| CN112003992A (en) * | 2020-08-14 | 2020-11-27 | 迅镭智能（广州）科技有限公司 | Transmission system and method based on scanning gun |
| CN113660426A (en) * | 2021-09-08 | 2021-11-16 | 英望科技(山东)有限公司 | Method and system for preventing photographing jitter of user and computer readable storage medium |
| CN114339349A (en) * | 2021-12-10 | 2022-04-12 | 海信视像科技股份有限公司 | Display device, data transmission method and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105979218A (en) | High-definition real-time monitoring video efficient transmission method | |
| DE102020132078A1 (en) | RESOURCE ALLOCATION BASED ON APPLICABLE SERVICE LEVEL AGREEMENT | |
| DE112011101321B4 (en) | Querying performance data on a parallel computer system having compute nodes | |
| WO2017173919A1 (en) | Concurrent data caching method and structure | |
| DE102019104394A1 (en) | COMMAND SET ARCHITECTURE TO ENABLE ENERGY EFFICIENT CALCULATION FOR EXASCALE ARCHITECTURES | |
| CN104506379B (en) | Network Data Capturing method and system | |
| US20160321776A1 (en) | Model Parallel Processing Method and Apparatus Based on Multiple Graphic Processing Units | |
| DE102020133738A1 (en) | FIRMWARE UPDATE TECHNIQUES | |
| CN105893083A (en) | Container-based mobile code unloading support system under cloud environment and unloading method thereof | |
| CN108874588A (en) | A kind of database instance restoration methods and device | |
| CN104954795B (en) | A kind of image acquisition transmission system based on JPEG2000 | |
| US20090110326A1 (en) | High bandwidth image processing system | |
| CN104683860A (en) | Multipath audio and video concurrent decoding acceleration card and decoding acceleration method for same | |
| DE112012004551B4 (en) | Multi-core interconnection in a network processor | |
| US11698863B1 (en) | Data set and node cache-based scheduling method and device | |
| DE112004001133T5 (en) | Queue locks with monitor memory wait | |
| CN108920153A (en) | A kind of Docker container dynamic dispatching method based on load estimation | |
| DE102022104207A1 (en) | Pooling of network processing resources | |
| CN107077390A (en) | A kind of task processing method and network interface card | |
| CN107179878A (en) | The method and apparatus of data storage based on optimizing application | |
| CN102457578B (en) | A kind of distributed network monitoring method based on case mechanism | |
| CN107430554A (en) | Storage cache performance is improved as the standard that cache is inserted by using the compressibility of data | |
| CN107153643A (en) | Tables of data connection method and device | |
| CN110362617A (en) | Batch data method and system is quickly exported from database based on more concurrent technologies | |
| CN107615746A (en) | For the Integrated Solution being intelligently imaged |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160928 |