CN102130897A - Cloud computing-based video acquisition and analysis system and method - Google Patents
Cloud computing-based video acquisition and analysis system and method Download PDFInfo
- Publication number
- CN102130897A CN102130897A CN2010102308911A CN201010230891A CN102130897A CN 102130897 A CN102130897 A CN 102130897A CN 2010102308911 A CN2010102308911 A CN 2010102308911A CN 201010230891 A CN201010230891 A CN 201010230891A CN 102130897 A CN102130897 A CN 102130897A
- Authority
- CN
- China
- Prior art keywords
- pixels
- video
- block
- pixel
- video image
- 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
Images
Abstract
The invention relates to a video acquisition and analysis system and a video acquisition and analysis method. The method comprises the following steps of: uploading a video image, wherein the video image is a part (140) which is cut from a video stream; dividing the video image into one or more first pixels or one or more first pixel blocks (412) comprising a plurality of pixels; respectively mapping the divided one or more first pixels or first pixel blocks into one or more corresponding second pixels or pixel blocks (413) according to a certain mapping relationship; obtaining information (414) of the one or more second pixels or pixel blocks according to a position relationship between the first pixels or pixel blocks and the corresponding one or more second pixels or pixel blocks; and merging the video image consisting of the one or more first pixels or one or more second pixels or pixel blocks with other video images except the cut partial video image in the video stream to form a continuous video stream (417). Corresponding to the method, the invention provides a corresponding system for implementing the method.
Description
Technical field
What the present invention relates to is image processing field, particularly relates to a kind of video acquisition analytical system and method based on cloud computing.
Background technology
In the current life, various occasions all can have the application of video system, for example carry out the public safety system, satellite imagery system, medical image system etc. of video monitoring with camera.In these are used, might be of poor quality, CCD (charge coupled device) and reasons such as cmos sensor resolution is not high, camera motion, light variation because of camera, cause the image blurring of collection, be difficult to recognize.At present, the superresolution restoration/reconstruction technique of video sequence is the hot technology that addresses the above problem, and it is meant and constructs the one or more resolution enhance technology of high-definition picture clearly from a series of fuzzy low resolution observed images.
Some super-resolution blurred picture treatment systems are arranged on the market now.For example, " knowing intelligent " blurred picture treatment system of He Lan " the shadow doctor " and the U.S..These systems not only involve great expense, and also have a lot of defectives.For example, because these systems are single board computer, it is easy to do the video file analysis, but is difficult to carry out the real-time processing of video file.In addition, when the fuzzy video of gathering is handled, need convert the analog video data of gathering to digital signal, operand is big, under the situation of single-chip microcomputer, is not easy to realize the requirement of real time high-speed system.And, when the fuzzy video that gathers is rebuild amount of calculation when very big, if the computing meeting on unit of software architecture routinely causes operation time long even cause crashing.Moreover if there is the data source of a large amount of acquired original in the video analytic system of handling based on the server device, how so huge data being passed to server, to be unlikely to cause network blockage also be a new difficult problem.
Summary of the invention
The object of the present invention is to provide a kind of effectively and handle video information at high speed, extract effective content, recover the real-time acquisition analysis system of the high-speed video based on cloud computing and the method for clear video image.
This overall system adopts the star framework of centralized control unit and a plurality of terminals.Centralized control unit is based on the virtual server of cloud computing, and terminal is made of a computer and a video frequency collection card, and the interface between computer and the video frequency collection card can be PCIE or USB.Wherein, a plurality of terminals are converted to digital signal transfers to centralized control unit with the analog video signal of gathering, and then, centralized control unit utilizes super-resolution technique of the present invention to analyze a plurality of vision signals, recover or reconstruct vision signal clearly.
The present invention provides a kind of video acquisition analytical method in one aspect, may further comprise the steps: upload video image, this video image is to shear a part of coming out from a video flowing, described video image is cut into one or more first pixels or is cut into one or more first block of pixels that contain a plurality of pixels, one or more first pixels after the described cutting or first block of pixels are mapped to corresponding one or more second pixels or block of pixels respectively according to certain mapping relations, and according to described first pixel or block of pixels and and described second pixel or block of pixels between the position relation draw the information of described one or more second pixel or block of pixels.
In addition, this video acquisition analytical method also comprises in video image that described one or more first block of pixels and corresponding one or more second pixels or block of pixels are formed and the described video flowing merging the continuous video flowing of formation except other video images of shearing the partial video image that comes out.
Wherein aforesaid first pixel or block of pixels are to be judged to be fuzzy image according to a criterion.
Moreover, one or more first pixels after the cutting of aforesaid video image, the cutting or first block of pixels to the mapping of corresponding one or more second block of pixels and according to described first pixel or block of pixels and and described mapping after one or more second pixels or the relation of the position between the block of pixels step that draws the information of described one or more second pixel or block of pixels be respectively to realize by the SPLIT function in the MapReduce programming model, Map function and Reduce function.
Wherein, described video acquisition analytical method also comprises the step that the information of described one or more second pixels or block of pixels is excavated.
Foregoing video acquisition analytical method, when merging except other video images of shearing the partial video image that comes out in the video image that described one or more first pixels or block of pixels and second pixel corresponding with it or block of pixels are formed and described video flowing, adjusting resolution makes the resolution unanimity of the two.
Aforesaid video acquisition analytical method wherein, can be carried out from the time the cutting of image, cuts frame of video fuzzy in the video image, or spatially carries out, and cuts pixel fuzzy in the video image.
Pixel Information described in the aforesaid video acquisition analytical method comprises color of pixel and/or gray scale.
Another aspect of the present invention comprises a kind of video acquisition analytical system, and this system comprises: the data load module, be used to load video image, and this video image is a part that cuts out in the middle of a video flowing; The multiple programming environment module, at first described video image is cut into one or more first pixels or is cut into one or more first block of pixels that contain a plurality of pixels, again with one or more first pixels after the described cutting or first block of pixels according to certain mapping relations be mapped to respectively corresponding one or more second pixels or or block of pixels, at last according to described one or more first pixels or block of pixels and and corresponding one or more second pixel or block of pixels between the position relation draw the information of described one or more second pixel or block of pixels.
In aforesaid video acquisition analytical system, described video image is to be blurred picture according to a criterion.
Aforesaid video acquisition analytical system also comprises a parallel data excavation module, and the information of described one or more second pixels or block of pixels is excavated.
Aforesaid video acquisition analytical system, also comprise a terminal, be used for providing described video image, and merge the continuous video flowing of formation except other video images of shearing the partial video image that comes out in the video image that described one or more first block of pixels and one or more second pixel or block of pixels are formed and the described video flowing to described data load module.
Because system architecture of the present invention adopts the star framework, not only can handle a plurality of data sources, and because the special concurrency design of inside center control unit, processing terminal can infinitely enlarge, optimum combination the characteristics of cloud computing platform, for example low-cost, highly reliable, high expansion, feature such as virtual, the while has partly been considered high speed processing at video acquisition, makes whole system become the high-speed real-time system.
Description of drawings
Figure 1 shows that the overall structure schematic diagram of the video acquisition analytical system that the present invention is based on cloud computing;
Figure 2 shows that the schematic diagram of the super-resolution video analysis device in the video acquisition analytical system that the present invention is based on cloud computing;
Fig. 3 (a) is depicted as video acquisition analytical system and the method fuzzy video image that comprises a plurality of pixels or block of pixels to be processed that the present invention is based on cloud computing;
Fig. 3 (b) is depicted as and comprises a plurality of fuzzy video image pixel or a plurality of clear video image pixel of block of pixels mapping or clear video images of block of pixels from Fig. 3 (a);
Figure 4 shows that the handling process of the super-resolution video analysis device in the video acquisition analytical system that the present invention is based on cloud computing;
Figure 5 shows that first example of the terminal installation in the video acquisition analytical system that the present invention is based on cloud computing;
Figure 6 shows that the handling process of first example of the terminal installation in the video acquisition analytical system that the present invention is based on cloud computing;
Figure 7 shows that the handling process of second example of the terminal installation in the video acquisition analytical system that the present invention is based on cloud computing;
Figure 8 shows that the handling process of the 3rd example of the terminal installation in the video acquisition analytical system that the present invention is based on cloud computing;
Figure 9 shows that the handling process of the 4th example of the terminal installation in the video acquisition analytical system that the present invention is based on cloud computing; And
Figure 10 shows that the 5th example of the terminal installation in the video acquisition analytical system that the present invention is based on cloud computing.
Embodiment
Cloud computing (Cloud Computing) is a kind of method of emerging shared architecture, huge system pool can be linked together so that various IT services to be provided.Several factors has promoted the demand to this class environment, the sharp increase of using comprising the employing of connection device, real time data stream, SOA and search, such Web2.0 such as open cooperation, community network and Mobile business.The invention provides real-time acquisition analysis system of a kind of high-speed video and method based on cloud computing, improved the super-resolution software design framework in the blurred picture treatment technology dexterously, make it at utmost to utilize the resource of cloud computing platform, make the real-time processing of super-resolution algorithms come true.
Fig. 1 is the overall structure schematic diagram that the present invention is based on the real-time acquisition analysis system of high-speed video of cloud computing.This system 10 comprises a super-resolution video analysis device 20 as centralized control unit, and is attached thereto the one or more terminals 30 that form the star framework.Described super-resolution video analysis device 20 for example is the one or more computers with Internet connection.Terminal 30 for example is, the computer of video flowing upload device, band camera and the data acquisition board of video flowing upload device, band camera, can upload video flowing computer, can upload video flowing and with the equipment of video acquisition plate etc.Super-resolution video analysis device 20 is used for the fuzzy video image of processing terminal 30, obtains video clearly.
Super-resolution video analysis device 20 of the present invention comprises cloud computing platform, and based on the super-resolution application software of cloud computing platform.This super-resolution application software adopts software multithreading and multi-process parallel at the characteristics of cloud computing platform Distributed Calculation dexterously, and the resource of at utmost dispatching cloud computing platform improves the speed that video intelligent is analyzed.
From system logic structure, described super-resolution video analysis device 20 comprises three levels, is distributed cloud computing platform layer, video data digging podium level and video analysis service application layer according to going up down.
More specifically, distributed cloud computing platform layer comprises following three partial function modules: distributed file system (DFS) module 21, multiple programming environment module 22, distributed system management module 23.
Distributed file system (DFS) module 21 is used to store the view data of self terminal 30, and a parallel programmed environment is provided.Terminal 30 of the present invention is when delivering to video described super-resolution video analysis device 20 and handle, be not to send all frame of video, but at first the frame in 30 pairs of video flowings of terminal carries out ambiguity and judges, only fuzzy video is sheared out then to be sent to super-resolution video analysis device 20 and to handle.The shearing of described fuzzy video can be from time intercepting video flowing, is sequence of pictures 1,2 such as video flowing ... N, intercepting for example is picture etc. between the sequence 5 ~ 20 so, is the temporal part of original video stream; The shearing of described video flowing can also be to intercept from the space, such as only submitting fuzzy pixel to or comprising the block of pixels of a plurality of fuzzy pixels.So, described DFS21 storage be blurred picture data after terminal 30 transmits the shearing that comes up, because terminal 30 provides the source of video data, the present invention creatively combines the distributed data file memory function of cloud computing platform with the extensive distributivity of terminal 30, provide possess high reliability, the storage platform of high stability.
In multiple programming environment module 22, original fuzzy vedio data collection after the shearing that will handle be decomposed into a plurality of little/the subdata collection, (or several) little/subdata collection after these decompose is respectively by a plurality of nodes in the cloud computing platform cluster (generally being common computer) parallel processing, for example, a little/subdata collection is by a node processing, and generation intermediate treatment result.Described intermediate object program is merged by a large amount of nodes again, forms final result.In this module, the video analysis software that described original fuzzy vedio data through shearing is carried out parallel processing for example is based on the programming model of the MapReduce of google company.
More specifically, as shown in Figure 3, described video analysis software is mapped to the fuzzy video pictures (a) of the low resolution mode with grid in the high-resolution video pictures (b) of an establishment.At first, low-resolution video sequence of pictures (a) original among the DFS21 (original fuzzy sets of video data) is divided into independently grid, each grid can be one independently pixel also can be a block of pixels (seeing Fig. 3 (a), pixel/block of pixels 1,2,3 and 4) that contains a plurality of pixels.The process of this division is for example finished by the SPLIT function in the MapReduce programming model.Then, in the low-resolution video sequence of pictures (a) after will dividing with the MAP function in the MapReduce programming model independently pixel or block of pixels be mapped to the sub-pixel or the sub-pixel piece of one or more estimations in the high resolution video image sequence (b) according to certain rule, and pixel/block of pixels that (b) in the described high resolution video image sequence estimated with respect to the relative position relation information mapping of original pixels/block of pixels to different REDUCE functions.The REDUCE function can calculate pixel or block of pixels information, for example color and/or the half-tone information etc. estimating or estimate according to these concrete relative position relations.The Map of described pixel/block of pixels, Reduce carry out on each node (computer) in the cloud computing system.
More specifically, for example, by the MAP function, for example, pixel 1 or block of pixels 1 can be mapped as estimator pixel or sub-pixel piece 1-1,1-2 and the 1-3 in the high-resolution video picture (b) in the low-resolution video picture (a); Pixel 2 in the low-resolution video picture (a) or block of pixels 2 can be mapped as sub-pixel or sub-pixel piece 2-1,2-2 and the 2-3 etc. that estimate in the high-resolution video picture (b); Pixel/block of pixels 3 in the low-resolution video picture (a) can be mapped as pixel/block of pixels 3-1,3-2, the 3-3 of the estimation in the high-resolution video picture (b); Pixel 4 in the low-resolution video picture (a) can be mapped as pixel/block of pixels 4-1,4-2 and the 4-3 that estimates in the high-resolution video picture (b).And according to the known pixels/block of pixels (1 in described high-resolution pictures (b), 2,3,4) with estimation pixel or block of pixels (1-1,1-2,1-3,2-1,2-2,2-3,3-1,3-2,3-3,4-1,4-2, relative position relation 4-3) is mapped to different REDUCE functions, the REDUCE function can be mapped to different REDUCE tasks according to these concrete relative position relations and finish, such as sub-pixel/block of pixels 1-2,3-2 can be mapped to same REDUCE task, for example, because all be the pixel (1-2 that belongs to estimation, 3-2) be in marginal position and the same day picture known pixels (1,3) at described estimation pixel (1-2, straight line phase ortho position 3-2), for this special position, no matter from which pictures or block of pixels, the algorithm of processing also is consistent.Through drawing the pixel/block of pixels data message of all estimations, for example gray scale or colouring information after the REDUCE processing.At last, result is delivered to described video data digging platform and carried out data mining.Through such processing, original low-resolution video image data (pixel) can not be modified, and the sub-pixel or the sub-pixel piece of just original pixels or block of pixels being derived are out independently operated, improved the highly-parallel pattern effectively, and each operation complexity is not high, promptly can successfully apply to the software architecture of MapReduce.Because the demand of terminal 30 has nothing in common with each other, the distributed system management module 23 of cloud computing platform of the present invention just in time helps the relatively independent management of each terminal.
The second layer is the video data digging podium level, five modules that this layer mainly comprises: workflow module 24, data load module 25, parallel ETL module 26, parallel data are excavated module 27, parallel display module 28 as a result.
Parallel ETL (extract, conversion, load) module 26 is used for the initial data of the blurred picture of the low resolution through shearing that loads is carried out preliminary treatment to obtain mining data.The ETL process is: in order to obtain high-quality data, must do a series of complicated conversions (Transform) to the initial data that extraction (Extract) is come out and handle, load (Load) at last in data warehouse.This from the initial data to the data warehouse between, the operation that data are carried out is called the ETL process.As previously mentioned, the present invention does not have simply original video flowing directly to be submitted to cloud computing platform, but initial data is passed through certain processing in terminal, ETL (extract, change, the load) function that this processing neither utilize cloud computing platform itself to provide fully mainly relies on the hardware of system terminal 30 to realize.The video flowing that terminal 30 only will be judged as fuzzy space or temporal shearing is submitted to cloud computing platform, submits pending ETL (Extract, transform, and load) task to.The present invention can be certain frame or a few frame from time intercepting video flowing in the shearing of the described video flowing at described terminal 30 places, also can intercept from the space, such as for the sequence of video images (a) of figure in (3), only submit wherein pixel or block of pixels 1 and 3 to, doing like this may be because pixel or block of pixels 2 and 4 itself may be exactly clearly, does not need to handle.
Parallel data is excavated module 27, according to the difference of terminal hardware, realizes satisfying the data mining algorithm of service needed.Because the described data of collecting often have noise, these data of possibility are also inconsistent in addition, address these problems and will carry out data mining.Excavate in the module 27 in this parallel data of the present invention, can use the technology of cloud computing platform.Described parallel data is excavated module and can be collected described a plurality of according to the vedio data after the MapReduce models treated, a process has a MapReduce processing procedure, a plurality of processes are excavated module in parallel data and are gathered, pass through data mining algorithm, reclassify, sending into the video analysis service application layer in an orderly manner, is high resolution video image from the image of video traffic application layer process.Data mining algorithm C4.5, SLIQ commonly used, SPRINT, correlation rule, K-mean value, K-arest neighbors, Bayesian network, artificial neural net, genetic algorithm etc.
Parallel data is excavated module 27 and is submitted pending clustering algorithm task to cloud computing platform, one group of video data is classified according to similitude and otherness, make that the similitude between the data belong to same classification is big as far as possible, the similitude between the data in different classes of is as far as possible little.Calculate platform by cloud and carry out and feedback result, deposit in DFS module 21;
The 3rd layer is the video analysis service application layer, mainly comprises user's gui interface module 29 and algorithms library API module 31.The GUI user interface is used to make the user can carry out the loading of original video flow data by user interface, initiatively extracts a part of video (ETL operation) that needs processing from original video stream.In addition for example, can also provide the several data mining algorithm to the user, the user can be configured by described gui interface, and selection will be used the algorithm of data mining.Algorithm in the algorithms library API module 31 is mainly the super-resolution optimized Algorithm, by several different methods such as interpolation, reconstruction, estimation, study, with information known in the high-resolution pictures in the video data digging layer (a) such as pixel 1,2,3,4, and calculate by the REDUCE function and to estimate pixel 1-1,1-2, information such as 2-1 are further optimized, and obtain picture more clearly.
Based on above-mentioned description to super-resolution video analysis device 20 of the present invention, can obtain, the workflow of super-resolution video analysis device 20 of the present invention is seen accompanying drawing 4.
Fig. 4 is the workflow schematic diagram of the present invention as the super-resolution video analysis device 20 of centralized control unit.As shown in Figure 4, in step 410, described parallel ETL module 26 is uploaded the vedio data of fuzzy/low resolution through terminal 30 shear treatment to be processed from client 30.In step 411, fuzzy/low-resolution video view data that storage is uploaded by described parallel ETL module 26 is to DFS module 21.In step 412, the low-resolution video view data in the DFS module 21 is divided into a plurality of grids, the block of pixels (pixel/block of pixels 1,2,3,4 among Fig. 3 (a)) that each grid comprises a pixel or is made of a plurality of pixels.Then step 413 with described step 412 in a plurality of low-resolution pixel of cutting or block of pixels be mapped to one or more sub-pixel/block of pixels (seeing pixel/block of pixels 1-1,1-2,1-3,2-1,2-2,2-3,3-1,3-2,3-3,3-4,4-1,4-2,4-3 and 4-4 among Fig. 3 (b)) in the high resolution video image respectively.Comprise described low-resolution pixel or block of pixels (seeing the pixel/block of pixels 1,2,3 and 4 among Fig. 3 (b)) in the described high resolution video image, and the sub-pixel/block of pixels (pixel/block of pixels 1-1, the 1-2 among Fig. 3 (b), 1-3,2-1 etc.) after the mapping.Then, in step 414, by calculating the mapping relations in the described step 413, and the relative position relation between the sub-pixel/block of pixels of the low-resolution pixel in the described high resolution video image and their correspondences, obtain the sub-pixel/block of pixels after the described mapping.Described step 412,413,414 can be realized by the SPLIT in the MapReduce programming model, MAP, REDUCE function respectively.Sub-pixel/block of pixels data that the described step of in step 415 collection being handled through MapReduce 414 draws are carried out data mining, in step 416, the data of excavating are carried out the super-resolution analysis again, obtain video image clearly by methods such as interpolation, reconstruction, estimation, study.Think in video image clearly that in step 417, will obtain again and the client 30 that unambiguous video merges, and forms continuous video image clearly.Before merging, can make the video image clearly of formation consistent with the video image resolution of client.Some intermediate object programs in the described treatment step can be deposited among the described DFS21.
Described terminal 30 for example can be five types of equipment combination in any hereinafter mentioning.Certainly, these five kinds of equipment are being for the purpose of illustration only property explanation principle of the present invention herein, and the present invention is not limited thereto.Those skilled in the art can change according to the example in principle of the present invention and this specification, and these variations still belong to category of the present invention.
Figure 5 shows that first example of terminal 30.As shown in Figure 5, terminal 30 is the common computer 33 of band camera 31 with data acquisition board 32.Common computer 33 links to each other with data acquisition board 32 by the interface of PCIE or USB, and data acquisition board 32 is connected with camera or CCD31.Camera 31 can be gathered analog video signal in real time, then the analog video signal of gathering is converted to digital signal by A/D (mould/number), then digital signal is passed to the FPGA (field programmable gate array) on the data acquisition board 32, carry out complicated compressed encoding by collection plate inside and handle, pass to centralized control unit 20 by network interface again with packing.
Shown in Figure 6 is the workflow of above-mentioned first exemplary terminal.
In step 601, terminal 30 (common computer 33 of band camera 31 and data acquisition board 32) is connected with centralized control unit 20 by network.In step 602, gather analog video data by camera or CCD31.In step 603, the analog video data of gathering is carried out digital sample then, carry out analog-to-digital conversion.In the video acquisition process, the pixel/block of pixels in the marking video sequence of pictures frame by frame.In step 604, carry out the analysis of video flowing ambiguity again, and in step 605, video flowing is carried out fuzzy Judgment frame by frame, judge whether frame of video blurs.Analysis of video ambiguity and judgement can be carried out with reference to prior art, for the purpose of the statement of the principle of the invention is clear, do not give unnecessary details herein.If not fuzzy frame of video, then flow process proceeds to step 606 stores video frames.Otherwise, if in step 605, judge it is fuzzy frame of video, then in step 607, fuzzy frame of video to be sheared, described shearing can be to go up the time to shear frame of video, for example, shears fuzzy picture frame, or shears on the space, for example, shears fuzzy pixel.Then video flowing is compressed in step 608 and pack.Then, in step 609 by computer 31 with transfer of data to cloud computing platform 22, i.e. centralized control unit.After cloud computing platform is handled (referring to Fig. 4 and related description thereof), in step 610, receive the video flowing clearly that described super-resolution video analysis device feeds back to.In step 611, for example, improve the pixel or the block of pixels resolution of storage in step 606 then, itself and the resolution of video flowing clearly that obtains in step 610 are consistent.Then, the clear video flowing that will obtain in step 612 merges with video stream stored in step 606, forms continuous video image.In step 613, should carry out digital-to-analogue (D/A) conversion by the clear video flowing of numeral, produce analog video stream clearly, in step 614, to show.
What Fig. 7 described is the flow chart of the second example work of described terminal 30.In this second example, terminal 30 is a video flowing uploading device.Can be made of a device with network interface and file input port, for example be a set-top box with last network interface.This terminal will need the video flowing handled to pass to centralized control unit in the mode of file or Streaming Media by network interface to handle.Idiographic flow at first, in step 701, is connected by network terminal 30 (device with network interface and file input port) as shown in Figure 7 with centralized control unit 20.The video flowing that will handle in step 702 uploads to terminal then.In step 703, carry out the analysis of video flowing ambiguity again, and in step 704, video flowing is carried out fuzzy Judgment frame by frame, to isolate fuzzy video image.If not fuzzy video, then flow process proceeds to step 705 stores video frames.Otherwise, if fuzzy video then carries out video and shears in step 706.Described shearing for example can be to shear fuzzy frame of video from the time, perhaps for example is the fuzzy pixel of spatially shearing in the fuzzy video.And the fuzzy video stream data of described shearing compressed in step 707 and pack.Then, in step 708 by computer with the transfer of data of described compression and packing to centralized control unit 20.After cloud computing platform is handled (referring to Fig. 4 and related description thereof), in step 709, obtain video flowing clearly.In step 710, improve then original in step 705 video stream stored the resolution of the video image of other non-fuzzies, itself and the resolution of video flowing clearly that obtains in step 709 are consistent, make the user observe the video flowing of reconstruction more intuitively.Then, the clear video flowing that will obtain in step 711 merges with the original video stream of storing in step 705, forms continuous video image.In step 712, should carry out digital-to-analogue (D/A) conversion by the clear video flowing of numeral, produce analog video stream clearly, in step 713, to show.
Fig. 8 is the workflow diagram of the 3rd example of terminal installation of the present invention.Described terminal 30 is for uploading the common computer of video flowing.This Daepori energising brain need have network interface, and the computer video file that will need to handle passes to centralized control unit by network interface like this.Video file can just be stored on the hard disk originally, perhaps by external tapping such as USB, equipment such as DVD drive copy into file.The treatment step 701 ~ 712 that occurred in described treatment step 801 ~ 812 and described second exemplary terminal is similar, repeats no more herein.Just in this case at the video flowing of step 813 clear display.
Fig. 9 is the workflow diagram of the 4th example of terminal installation of the present invention, and terminal 30 is for uploading video flowing and with the terminal of video acquisition.This terminal can be thought a synthesizer of the first exemplary terminal type and the second exemplary terminal type, just need make a control switch in inside, have operating personnel select current to be processed be the video image of gathering, still the video file that itself generates, or directly by the outside copy into video file.Above-mentioned deterministic process is carried out in step 905.The processing of step 901 ~ 903 and 601 ~ 603 similar repeats no more.Step 904,906 ~ 913 with the Fig. 6 that occurs previously and Fig. 7 in corresponding treatment step similar.
Figure 10 is the 5th example of terminal installation of the present invention.Described terminal 30 is that the video of a band camera is uploaded box.This a kind of portable unit that can think first kind of terminal example of the present invention no longer needs computer, directly is with network interface, can communicate by letter with centralized control unit.The analog video that camera is filmed is converted to digital signal by the A/D device, sends into FPGA then and compresses processing, again by microprocessor (MCU) conversion of packing, from network interface output and centralized control unit.Workflow is with the workflow (see figure 6) of described first kind of terminal example.
The present invention for example can be used for video detection system, and camera can be distributed in any one corner, the video information of collecting is carried out real time high-speed handle, and video comes (720P) thereby recover clearly.The present invention also can be used for consuming industry, and video image comes to wish to recover clearly anywhere by the fuzzy video that will oneself take by this system such as anyone.Also having a kind of application is can use in intelligent community, Smart Home.
The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and any those of ordinary skill in the art are in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.
Claims (13)
1. video acquisition analytical method may further comprise the steps:
Upload video image, this video image is to shear a part (410) of coming out from a video flowing,
Described video image is cut into one or more first pixels or is cut into one or more first block of pixels (412) that contain a plurality of pixels,
One or more first pixels after the described cutting or first block of pixels are mapped to corresponding one or more second pixels or block of pixels (413) respectively according to certain mapping relations,
According to described first pixel or block of pixels and and corresponding one or more second pixel or block of pixels between the position relation draw the information (414) of described one or more second pixel or block of pixels.
2. video acquisition analytical method according to claim 1 is characterized in that it also comprises in video image that described one or more first block of pixels and one or more second pixel or block of pixels are formed and the described video flowing and merges the continuous video flowing (417) of formation except other video images of shearing the partial video image that comes out.
3. video acquisition analytical method according to claim 1 is characterized in that wherein said first pixel or block of pixels are to be judged to be fuzzy image according to a criterion.
4. according to claim 2 or 3 described video acquisition analytical methods, it is characterized in that wherein said step (412), (413), (414) are respectively to realize by the SPLIT function in the MapReduce programming model, Map function and Reduce function.
5. video acquisition analytical method according to claim 4 is characterized in that wherein after handling through step (414) information of described one or more second pixels or block of pixels being excavated (415).
6. video acquisition analytical method according to claim 5 is characterized in that wherein described step (413) and (414) of described a plurality of second pixels or block of pixels execution being carried out on one or more computing platforms.
7. video acquisition analytical method according to claim 6, when it is characterized in that wherein in video image that step (417) forms described second pixel or block of pixels and described video flowing merging except other video images of shearing the partial video image that comes out, adjusting resolution makes the resolution unanimity of the two.
8. video acquisition analytical method according to claim 7, it is characterized in that wherein can carrying out from the time the cutting of image in step (413), cut frame of video fuzzy in the video image, or spatially carry out, cut pixel fuzzy in the video image.
9. according to claim 1 or 8 described video acquisition analytical methods, it is characterized in that wherein said Pixel Information comprises color of pixel and/or gray scale.
10. video acquisition analytical system is characterized in that it comprises:
Data load module (25) is used to load video image, and this video image is a part that cuts out in the middle of a video flowing,
Multiple programming environment module (22), at first described video image is cut into one or more first pixels or is cut into one or more first block of pixels that contain a plurality of pixels, again with one or more first pixels after the described cutting or first block of pixels according to certain mapping relations be mapped to respectively corresponding one or more second pixels or or block of pixels, at last according to described one or more first pixels or block of pixels and and corresponding one or more second pixel or block of pixels between the position relation draw the information of described one or more second pixel or block of pixels.
11. video acquisition analytical system according to claim 10 is characterized in that wherein said video image is is blurred picture according to a criterion.
12. video acquisition analytical system according to claim 11 is characterized in that it comprises that also parallel data excavates module (27) information of described one or more second pixels or block of pixels is excavated.
13. video acquisition analytical system according to claim 12, it is characterized in that it also comprises a terminal (30), be used for providing described video image, and merge the continuous video flowing of formation except other video images of shearing the partial video image that comes out in the video image that described one or more first block of pixels and one or more second pixel or block of pixels are formed and the described video flowing to described data load module (25).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102308911A CN102130897A (en) | 2010-04-26 | 2010-07-07 | Cloud computing-based video acquisition and analysis system and method |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010156652.6 | 2010-04-26 | ||
| CN201010156652 | 2010-04-26 | ||
| CN2010102308911A CN102130897A (en) | 2010-04-26 | 2010-07-07 | Cloud computing-based video acquisition and analysis system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102130897A true CN102130897A (en) | 2011-07-20 |
Family
ID=44268789
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010102308911A Pending CN102130897A (en) | 2010-04-26 | 2010-07-07 | Cloud computing-based video acquisition and analysis system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102130897A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102571991A (en) * | 2012-03-27 | 2012-07-11 | 电子科技大学 | Multistage-mapping-based large-scale multi-copy distributed storage system and application method thereof |
| CN102625145A (en) * | 2011-11-28 | 2012-08-01 | 苏州奇可思信息科技有限公司 | Method for processing multi-server-node video on basis of multiple processing algorithms |
| CN103207662A (en) * | 2012-01-11 | 2013-07-17 | 联想(北京)有限公司 | Method and device for obtaining physiological characteristic information |
| CN103379336A (en) * | 2012-04-25 | 2013-10-30 | 全视科技有限公司 | Method, apparatus and system for exchanging video data in parallel |
| CN103631568A (en) * | 2013-12-20 | 2014-03-12 | 厦门大学 | Medical-image-oriented multi-thread parallel computing method |
| CN103905833A (en) * | 2013-07-12 | 2014-07-02 | 吉首大学 | Distributed network video data mining and collecting system based on cloud calculation |
| CN104281991A (en) * | 2013-09-06 | 2015-01-14 | 北京宇图天下软件有限公司 | Smart community three-dimensional monitoring platform and method |
| CN106254458A (en) * | 2016-08-04 | 2016-12-21 | 山东大学 | A kind of image processing method based on cloud robot vision, platform and system |
| CN108122246A (en) * | 2017-12-07 | 2018-06-05 | 中国石油大学(华东) | Video monitoring intelligent identifying system |
| CN109886396A (en) * | 2019-03-18 | 2019-06-14 | 国家电网有限公司 | A kind of transmission line galloping on-line prediction system and method |
| CN111314741A (en) * | 2020-05-15 | 2020-06-19 | 腾讯科技（深圳）有限公司 | Video super-resolution processing method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5446806A (en) * | 1993-11-15 | 1995-08-29 | National Semiconductor Corporation | Quadtree-structured Walsh transform video/image coding |
| CN1578449A (en) * | 2003-06-27 | 2005-02-09 | 成津C&C株式会社 | Method for recovering and reconsisting super-resolution image from low-resolution compression image |
| CN101226634A (en) * | 2008-01-30 | 2008-07-23 | 哈尔滨工业大学 | Method for reestablishment of single frame image quick super-resolution based on nucleus regression |
-
2010
- 2010-07-07 CN CN2010102308911A patent/CN102130897A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5446806A (en) * | 1993-11-15 | 1995-08-29 | National Semiconductor Corporation | Quadtree-structured Walsh transform video/image coding |
| CN1578449A (en) * | 2003-06-27 | 2005-02-09 | 成津C&C株式会社 | Method for recovering and reconsisting super-resolution image from low-resolution compression image |
| CN101226634A (en) * | 2008-01-30 | 2008-07-23 | 哈尔滨工业大学 | Method for reestablishment of single frame image quick super-resolution based on nucleus regression |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102625145A (en) * | 2011-11-28 | 2012-08-01 | 苏州奇可思信息科技有限公司 | Method for processing multi-server-node video on basis of multiple processing algorithms |
| CN103207662A (en) * | 2012-01-11 | 2013-07-17 | 联想(北京)有限公司 | Method and device for obtaining physiological characteristic information |
| CN102571991B (en) * | 2012-03-27 | 2014-06-11 | 电子科技大学 | Multistage-mapping-based large-scale multi-copy distributed storage system and application method thereof |
| CN102571991A (en) * | 2012-03-27 | 2012-07-11 | 电子科技大学 | Multistage-mapping-based large-scale multi-copy distributed storage system and application method thereof |
| CN103379336B (en) * | 2012-04-25 | 2017-04-12 | 豪威科技股份有限公司 | Method, apparatus and system for exchanging video data in parallel |
| CN103379336A (en) * | 2012-04-25 | 2013-10-30 | 全视科技有限公司 | Method, apparatus and system for exchanging video data in parallel |
| CN103905833A (en) * | 2013-07-12 | 2014-07-02 | 吉首大学 | Distributed network video data mining and collecting system based on cloud calculation |
| CN103905833B (en) * | 2013-07-12 | 2017-06-27 | 上海易钧信息技术有限公司 | Based on cloud computing distributed network video data digging acquisition system |
| CN104281991A (en) * | 2013-09-06 | 2015-01-14 | 北京宇图天下软件有限公司 | Smart community three-dimensional monitoring platform and method |
| CN103631568A (en) * | 2013-12-20 | 2014-03-12 | 厦门大学 | Medical-image-oriented multi-thread parallel computing method |
| CN106254458A (en) * | 2016-08-04 | 2016-12-21 | 山东大学 | A kind of image processing method based on cloud robot vision, platform and system |
| CN106254458B (en) * | 2016-08-04 | 2019-11-15 | 山东大学 | A kind of image processing method based on cloud robot vision, platform and system |
| CN108122246A (en) * | 2017-12-07 | 2018-06-05 | 中国石油大学(华东) | Video monitoring intelligent identifying system |
| CN109886396A (en) * | 2019-03-18 | 2019-06-14 | 国家电网有限公司 | A kind of transmission line galloping on-line prediction system and method |
| CN111314741A (en) * | 2020-05-15 | 2020-06-19 | 腾讯科技（深圳）有限公司 | Video super-resolution processing method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102130897A (en) | Cloud computing-based video acquisition and analysis system and method | |
| US9830736B2 (en) | Segmenting objects in multimedia data | |
| Gargees et al. | Incident-supporting visual cloud computing utilizing software-defined networking | |
| Ahmad et al. | An efficient divide-and-conquer approach for big data analytics in machine-to-machine communication | |
| WO2017024975A1 (en) | Unmanned aerial vehicle portable ground station processing method and system | |
| US8953674B2 (en) | Recording a sequence of images using two recording procedures | |
| KR101559838B1 (en) | Visualizaion method and system, and integrated data file generating method and apparatus for 4d data | |
| WO2012095867A4 (en) | An integrated intelligent server based system and method/systems adapted to facilitate fail-safe integration and /or optimized utilization of various sensory inputs | |
| CN109739820A (en) | A kind of E-government information service system based on big data analysis | |
| CN103905833A (en) | Distributed network video data mining and collecting system based on cloud calculation | |
| Distefano et al. | Device-centric sensing: an alternative to data-centric approaches | |
| Xing et al. | The challenges of image segmentation in big remotely sensed imagery data | |
| CN103517035A (en) | Intelligent park security panorama monitoring system and method | |
| Balchandani et al. | A deep learning framework for smart street cleaning | |
| CN111782745B (en) | Space-time big data grid coding efficient visualization method and system | |
| CN106155773A (en) | High resolution ratio satellite remote-sensing product large-scale production | |
| CN108228900B (en) | Power equipment multispectral data center model building method based on hierarchical structure | |
| Hu et al. | FEVA: A federated video analytics architecture for networked smart cameras | |
| CN103136739A (en) | Registering method of controllable vidicon surveillance video and three-dimensional model in complex scene | |
| CN207589029U (en) | A kind of intelligent mobile full-view video image monitoring system | |
| CN110096042A (en) | Artificial swamp monitors methods of exhibiting and system on-line | |
| CN114819852A (en) | Logistics planning system and method for smart city | |
| CN202374374U (en) | Shopping mall crowd flow dynamic management system based on remote images | |
| Di Staso et al. | Heterogeneous-Resolution and multi-source terrain builder for CesiumJS WebGL virtual globe | |
| US10747812B1 (en) | Video analytics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110720 |