CN102708903B - Automatic video playback system - Google Patents
Automatic video playback system Download PDFInfo
- Publication number
- CN102708903B CN102708903B CN201210196785.5A CN201210196785A CN102708903B CN 102708903 B CN102708903 B CN 102708903B CN 201210196785 A CN201210196785 A CN 201210196785A CN 102708903 B CN102708903 B CN 102708903B
- Authority
- CN
- China
- Prior art keywords
- video
- playback
- image
- carry out
- match
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an automatic video playback system, which comprises a shooting unit, a video allocator, a video playback machine, a switching unit and a terminal, wherein the shooting unit is used for collecting N-channel video signal; the video allocator is connected with the shooting unit and is used for allocating the N-channel video signals collected by the shooting unit, the video playback machine is connected with the video allocator and is used for storing the N-channel video signals transmitted by the video allocator, matching the read image and judging whether the read image is used as a splendid pictures to be automatically played; the switching unit is respectively connected with the video allocator and the video playback machine and is used for directly broadcasting or playing back the video signal transmitted by the video allocator or the video signal transmitted by the video playback machine and for selecting which signal is directly broadcasted or played back; and the terminal is connected with the switching unit and is used for displaying the video signal transmitted by the switching unit and for transmitting a switching order to the switching unit. The system simultaneously supports a hand-operated mode and a model recognition mode, realizes the automatic playback and recognizes a playback point.
Description
Technical field
The present invention relates to a kind of full-automatic video playback system.
Background technology
Traditional DVR is mainly Embedded, is not to be applied to field of broadcast televisions, but uses as monitor.The product that belongs to low side class of trade.Hard disk corruptions, does not have early warning completely, and system program is still in operation, and the data of in fact recording a video cannot be write in the middle of hard disk, thereby cause loss of data; Even if pinpoint the problems, scene also cannot be keeped in repair.All hardware all integrate, as long as wherein there is a module to break down, have substantially just caused whole system normally to move, and must more renew machine or beam back producer's maintenance.Owing to being subject to, the effect of restriction playback of chip is extremely undesirable, and when carrying out LAN (Local Area Network) and extranet access, effect can not allow people accept especially, and display effect is undesirable, and time delay is larger.Video processing part is integrated on a plank, if need to expand 1 tunnel, only way can only be the device of changing planes, so poor expandability.
The software that need to provide according to manufacturer for the DVR user of existing PC framework and card provide PC voluntarily, and then configuration is installed.Be the system that a set of user assembles by various boards, reliability is not high, and poor stability is also only applicable to general commercial environment, such as monitoring.
For some, can in the time of need to carrying out playback, also can only carry out playback by the mode of artificial identification for the video playback machine of competitive sports, artificial playback point of selection, re-starts broadcasting by a certain section of video.For the competitive sports that playback number of times is few, can also deal with, but control completely artificially playback for game of golf, workload is quite large, need to have a plurality of people to be responsible in real time video playback specially.Therefore can realize automatic playback is pattern-recognition, to whole video playback machine field, is a very large breakthrough, especially, for golf race, is extremely necessary.
Summary of the invention
The present invention is directed to the proposition of above problem, and develop a kind of full-automatic video playback system.The technological means that the present invention adopts is as follows:
A full-automatic video back method, is characterized in that comprising the steps:
S1, reading a two field picture, judge whether this image is gray level image, is not to be converted into gray level image;
S2, the gray level image that step S1 is obtained are treated as histogram, and then setting threshold, finds bright spot and abandon dim spot simultaneously;
S3, the image after step S2 is processed carry out binary Images Processing; Read Template Information, i.e. the monochrome information of ball simultaneously;
S4, according to template domain division method, cut apart image, then by characteristic matching and two kinds of methods of Gray-scale Matching, carry out images match;
S5, judge whether that the match is successful, if the match is successful, return to the beginning judgement of a frame again;
If S6 does not have that the match is successful, return to previous frame image and carry out above-mentioned computing, carry out images match, and judge whether that the match is successful, if the match is successful for previous frame, illustrate that ball has entered hole, system can select now picture forward N for starting broadcasting pictures, as excellent picture, automatically play second.
Described step S4 also comprises: whether judge templet information reads successfully, as reads and successfully carry out images match operation, exits the step that recognition mode carries out manual mode as unsuccessful.
The making step of described Template Information comprises: cut-away view picture, determine whether gray level image, not to be converted into gray level image, add up the gray level probability of this frame gray level image, generate histogram, with maximum variance between clusters, definite threshold, passing threshold is cut apart generation bianry image, carries out profile extraction, and obtain characteristic information by SOBEL operator edge detection method.
Utilize the above-mentioned device that full-automatic video back method carries out automatic video frequency playback of stating, it is characterized in that comprising: for gathering the image unit of N road vision signal; Connect image unit, for the N road vision signal to image unit collection, carry out the video distributor of allocation process; Connect video distributor, the N road vision signal that video distributor is sended over is stored, and the video playback machine that carries out video playback process according to above-mentioned video playback method; Connect respectively video distributor and video playback machine, the vision signal that video distributor is sended over or the vision signal that video playback machine is sended over, carry out live or carry out video playback according to above-mentioned video playback method, and selecting switch unit live or playback any road signal; Connection switching unit, the vision signal that switch unit is transmitted shows, and the control terminal that sends switching command to switch unit.Described switch unit and control terminal carry out data connection by network; Described network is television network or Ethernet; In addition, described network is also same for watching the user terminal of video frequency program to be connected.
Compared with the existing technology advantage of the present invention is apparent, specific as follows: this system is supported manual mode and pattern-recognition pattern simultaneously, playback automatically, identification playback point; Mode identification technology can realize full-automatic video playback, without manual operation; Support multichannel playback; Support remote network control; Support closely bluetooth, infraredly shake control; And can many concurrent workings etc. advantage.Because this system architecture is simple, easy to usely, and with low cost be in addition suitable for extensive popularization.
Accompanying drawing explanation
Fig. 1 is net cast playback work schematic diagram of the present invention;
Fig. 2 full-automatic video playback system workflow diagram;
Fig. 3 template training process flow diagram.
Embodiment
The method and system of full-automatic video playback, in the live of each race and relay playback, especially golf race, realizes full-automatic playback as shown in Figure 1 to Figure 3.Native system is based on PC formula framework, and extension storage hard disk, can expand to 6 tunnels easily, from different perspectives with gathering multiple different pictures, thereby can select more suitably picture to carry out multichannel playback.Can be because of gathering several roads video due to playback engine, and miss excellent picture.Except being undertaken, outside playback, providing the function of automatic playback by artificial identification playback point, be not also applied to this type of auto back playing function of sports event live broadcast and relay at present.For game of golf, very large to the number of times demand of playback, each player's takes turns match at least needs the playback of 18 times, so game of golf realizes automatic playback, is necessary very much.The automatic playback of this case utilizes mode identification technology, by mathematical model algorithm combining image, processes to judge whether ball enters hole.Ball enters behind hole, starts to carry out automatic playback, enters this process of hole carry out playback tens seconds from push rod to ball.So do not need to take special human resources and be responsible in real time playback, and people arranges for playback needs the regular hour, find playback point, be relatively not easy assurance, lose time.If cannot identify by implementation pattern because go wrong, system can transfer artificial recognition mode automatically to, i.e. manual mode.Two kinds of playback modes are also deposited actual effect and the playback quality that has guaranteed playback.Simple video clipping function is also provided in addition, and infrared, Bluetooth remote control function, can many concurrent workings.
Full-automatic video back method as shown in Figure 1, comprises the steps: S1, reads a two field picture, judges whether this image is gray level image, is not to be converted into gray level image (object is to reduce quantity of information); S2, the gray level image that step S1 is obtained are treated as histogram, and then setting threshold, finds bright spot and abandon dim spot simultaneously; Conventionally Threshold is 192, is greater than 192 bright spot and stays, and is less than 192 dim spot and abandons; S3, the image after step S2 is processed carry out binary Images Processing; Read Template Information, i.e. the monochrome information of ball (comprising: the size of gray-scale value, luminance frequency) simultaneously; S4, according to template domain division method, cut apart image, then by characteristic matching and two kinds of methods of Gray-scale Matching, carry out images match (, eigenwert is extracted, wherein matching process adopts staggered scanning, mates); S5, judge whether that the match is successful (, whether capture ball), if the match is successful, return to the beginning judgement of a frame again; If S6 does not have that the match is successful, return to previous frame image and carry out above-mentioned computing, carry out images match, and judge whether that the match is successful, if the match is successful for previous frame, illustrate that ball has entered hole, system can select now picture forward N for starting broadcasting pictures, as excellent picture, automatically play second.Described step S4 also comprises: whether judge templet information reads successfully, as reads and successfully carry out images match operation, exits the step that recognition mode carries out manual mode as unsuccessful.The making step of described Template Information comprises: cut-away view picture, determine whether gray level image, not to be converted into gray level image, add up the gray level probability of this frame gray level image, generate histogram, with maximum variance between clusters, definite threshold, passing threshold is cut apart generation bianry image, carries out profile extraction, and obtain characteristic information by SOBEL operator edge detection method.
Fig. 2 is video playback system workflow diagram, in the information of obtaining, reads a two field picture from video distributor, by file reading head, judge whether this two field picture is gray level image, if not gray level image, by extracting G component, makes it be converted into gray level image.Add up in this frame gray level image, gray level probability, generate histogram, with maximum variance between clusters, definite threshold, passing threshold is cut apart generation bianry image, read the Template Information of XML form, judge whether information reads successfully, if read Template Information failure, the mode of secession identification, enters into manual mode.If read Template Information success, according to template domain division method, cut apart image, then by characteristic matching and two kinds of methods of Gray-scale Matching, carry out images match, if captured ball, the match is successful, automatically starts playback after ball enters hole, if this two field picture coupling is unsuccessful, do not capture ball, whether the match is successful to take a fancy to a two field picture, if the match is successful for previous frame, illustrates that ball has entered hole, system enters auto back mode playback, as previous frame, the match is successful, reads the next frame of present frame, repeats aforesaid operations.
Fig. 3 template training process flow diagram, intercepting one two field picture, judges by file reading head whether this two field picture is gray level image, if not gray level image, by extracting G component, makes it be converted into gray level image.Add up the gray level probability of this frame gray level image, generate histogram.With maximum variance between clusters, definite threshold, passing threshold is cut apart generation bianry image, by SOBEL operator edge detection method, carry out profile extraction, and obtain characteristic information, obtain the characteristic information of golf, comprise: form parameter: the area that the girth A that F=C2/4 π A(C is ball is ball), color characteristic (regional average value), generates Template Information, and preservation (can save as XML file).With reference to masterplate, by 2 kinds of modes, obtain: the golf (or other ball) of an existing resolution below 100*100, also can obtain by scene, on-the-spot acquisition refers to that the video returning by video camera intercepts the mode of picture.After obtaining reference template, be changed to gray level image, mode is: Y=G(wherein Y is the rear gray-scale value of conversion, G is green component), the number of adding up afterwards each gray scale generates grey level histogram, sport balls substantially all be take a kind of color as main body, very large at main background difference, gray-scale map presents unimodal state, just in time with maximum variance between clusters, carrys out definite threshold; This threshold value also can be determined by actual use experience.In feature extraction, this programme mainly extracts from gray feature and shape facility.Gray scale belongs to the extraction of color character, before to mention spheroid be all to take a kind of color as main body, color distribution is regular and very even, and spheroid is the impact that not changed by image rotation and translation, therefore use simple half-tone information to extract and can use this programme, the feature of extracting comprises gray average, adjacent 2 gray scale differences.The extraction of shape facility adopts the method for Sobel rim detection to realize the extraction of contour feature, coupling adopts centered by SDA method source images match point, define a window W, add up its window gray-scale value (this value be binaryzation after) and, then in source images, progressively calculate itself and masterplate gray scale difference value and, final search to the center pixel in region of difference minimum be coupling initial point.The normalizing eliminate indigestion Gray-scale Matching of being correlated with:
Wherein:
for the average gray of the corresponding former figure of template (subgraph), the average gray of template; Each point of the former figure of i and j is each point
the average gray that represents former figure, S (m, n) in former picture with the gray scale in the region of template Corresponding matching position and, the summation of T (m, n) template gray level image; With the region of template Corresponding matching position in m template or former picture, n
M and N are template image height and width, and m ∈ (0, M), n ∈ (0, N);
the average gray of template, the explanation to method: 1. define absolute error; 2. get a constant threshold value; 3. template subgraph corresponding to each pixel of the former figure of surface sweeping point to be matched,, stops adding up when accumulated value surpasses threshold value according to calculating absolute error according to the formula in 1, stops the scanning of this template subgraph, the position of recording strip match point and accumulative frequency; 4. until scanning complete graph; 5. the minimum pixel of accumulative frequency is just optimal match point.
Fig. 1 is net cast playback work schematic diagram of the present invention, and camera acquisition to N road (for example: the 6 tunnels) image unit 1 of vision signal; Connect image unit 1, for N road that image unit 1 is gathered (for example: 6 tunnels) vision signal is carried out the video distributor 2 of allocation process; Connect video distributor 2, the N road vision signal that video distributor 2 is sended over is stored, and the video playback machine 3 that carries out video playback process according to above-mentioned video playback method; Connect respectively video distributor 2 and video playback machine 3, the vision signal that video distributor 2 is sended over or the vision signal that video playback machine 3 is sended over, carry out live or playback, and carry out switch unit 4 live or which road signal of playback is selected; Connection switching unit 4, the vision signal that switch unit 4 is transmitted shows, and the control terminal 6 that sends switching command to switch unit 4.Described switch unit 4 and control terminal 6 carry out data connection by network 5; Described network 5 is television network or Ethernet; In addition, described network is also same for watching the user terminal of video frequency program to be connected.The N road vision signal that image unit 1 collects (for example: 6 tunnels, concrete how many roads can be adjusted according to actual needs), directly send video distributor to, by video distributor 2, send respectively switch unit 4 and video playback machine 3 each road vision signals to again, the signal that sends switch unit 4 to can directly be play.By network 5, send user terminal 7(television terminal or network user terminals to) play.Vision signal in the middle of video playback machine 3 after treatment, automatically select to need the video information part of playback, then the N road signal of handling well is sent to switch unit 4, by cut bank, determines any road playback, finally by TV signal or network signal playback.
The above; it is only preferably embodiment of the present invention; but protection scope of the present invention is not limited to this; anyly be familiar with those skilled in the art in the technical scope that the present invention discloses; according to technical scheme of the present invention and inventive concept thereof, be equal to replacement or changed, within all should being encompassed in protection scope of the present invention.
Claims (5)
1. a full-automatic video back method, is characterized in that comprising the steps:
S1, reading a two field picture, judge whether this image is gray level image, is not to be converted into gray level image;
S2, the gray level image that step S1 is obtained are treated as histogram, and then setting threshold, finds bright spot and abandon dim spot simultaneously;
S3, the image after step S2 is processed carry out binary Images Processing; Read Template Information, i.e. the monochrome information of ball simultaneously;
S4, according to template, with domain division method, cut apart image, then by characteristic matching and two kinds of methods of Gray-scale Matching, carry out images match;
S5, judge whether that the match is successful, if the match is successful, return to the beginning judgement of a frame again;
If S6 does not have that the match is successful, return to previous frame image and carry out above-mentioned computing, carry out images match, and judge whether that the match is successful, if the match is successful for previous frame, illustrate that ball has entered hole, system can select now picture forward N for starting broadcasting pictures, as excellent picture, automatically play second.
2. a kind of full-automatic video back method according to claim 1, is characterized in that described step S4 also comprises: whether judge templet information reads successfully, as read, successfully carries out images match operation, as unsuccessful, exits the step that full-automatic video playback system carries out manual mode.
3. a kind of full-automatic video back method according to claim 1, the making step that it is characterized in that described Template Information comprises: cut-away view picture, determine whether gray level image, not to be converted into gray level image, add up the gray level probability of this frame gray level image, generate histogram, use maximum variance between clusters definite threshold, passing threshold is cut apart generation bianry image, by SOBEL operator edge detection method, carries out profile extraction, and obtains the characteristic information of golf.
4. utilize the device that full-automatic video back method carries out automatic video frequency playback described in claim 1, it is characterized in that comprising:
For gathering the image unit (1) of N road vision signal;
Connect image unit (1), for the N road vision signal that image unit (1) is gathered, carry out the video distributor (2) of allocation process;
The video playback machine (3) that the N road vision signal that connect video distributor (2), video distributor (2) is sended over is stored and carried out video playback process according to video playback method described in claim 1;
Connect respectively video distributor (2) and video playback machine (3), the vision signal that video distributor (2) is sended over or the vision signal that video playback machine (3) is sended over, carry out live or playback and carry out switch unit (4) live or which road signal of playback is selected;
The control terminal (6) that connection switching unit (4), the vision signal that switch unit (4) is transmitted show and send switching command to switch unit (4).
5. a kind of device that carries out automatic video frequency playback according to claim 4, is characterized in that: described switch unit (4) and control terminal (6) carry out data connection by network (5); Described network (5) is television network or Ethernet; In addition, described network (5) is also same for watching the user terminal (7) of video frequency program to be connected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210196785.5A CN102708903B (en) | 2012-06-14 | 2012-06-14 | Automatic video playback system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210196785.5A CN102708903B (en) | 2012-06-14 | 2012-06-14 | Automatic video playback system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102708903A CN102708903A (en) | 2012-10-03 |
| CN102708903B true CN102708903B (en) | 2014-12-10 |
Family
ID=46901598
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210196785.5A Active CN102708903B (en) | 2012-06-14 | 2012-06-14 | Automatic video playback system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102708903B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106659927B (en) * | 2014-07-16 | 2020-05-22 | 杰森·拉瑟 | Method and apparatus for predicting likelihood of success of a golf swing |
| CN105916049A (en) * | 2016-04-18 | 2016-08-31 | 深圳Tcl数字技术有限公司 | Method and device for automatically switching watching mode |
| CN107835397B (en) * | 2017-12-22 | 2019-12-24 | 成都华栖云科技有限公司 | Multi-lens video synchronization method |
| CN108259864A (en) * | 2018-03-28 | 2018-07-06 | 周毅 | A kind of vision signal optical cable transmission system and the video camera with the Transmission system |
| CN111246126A (en) * | 2020-03-11 | 2020-06-05 | 广州虎牙科技有限公司 | Direct broadcasting switching method, system, device, equipment and medium based on live broadcasting platform |
| CN112000315A (en) * | 2020-08-24 | 2020-11-27 | 浪潮云信息技术股份公司 | Automatic generation method of user-defined parameter panel based on operator |
| CN112492341B (en) * | 2020-11-30 | 2024-02-23 | 深圳前海微众银行股份有限公司 | Live broadcast time shift caching method, device and system and computer readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1510907A (en) * | 2002-12-24 | 2004-07-07 | 中国科学院计算技术研究所 | Method for detecting exciting part in sports game video frequency |
| CN201523428U (en) * | 2009-06-25 | 2010-07-07 | 上海市杨浦区控江二村小学 | Monitoring playback system |
-
2012
- 2012-06-14 CN CN201210196785.5A patent/CN102708903B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1510907A (en) * | 2002-12-24 | 2004-07-07 | 中国科学院计算技术研究所 | Method for detecting exciting part in sports game video frequency |
| CN201523428U (en) * | 2009-06-25 | 2010-07-07 | 上海市杨浦区控江二村小学 | Monitoring playback system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102708903A (en) | 2012-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102708903B (en) | Automatic video playback system | |
| US11790504B2 (en) | Monitoring method and apparatus | |
| US10515471B2 (en) | Apparatus and method for generating best-view image centered on object of interest in multiple camera images | |
| US9437012B2 (en) | Multiple-object tracking and team identification for game strategy analysis | |
| US8805007B2 (en) | Integrated background and foreground tracking | |
| CN103716594B (en) | Panorama splicing linkage method and device based on moving target detecting | |
| CN101383910B (en) | Apparatus and method for rendering a 3d scene | |
| US20110102678A1 (en) | Key Generation Through Spatial Detection of Dynamic Objects | |
| US10163256B2 (en) | Method and system for generating a three-dimensional model | |
| US20180137363A1 (en) | System for the automated analisys of a sporting match | |
| CN111013150B (en) | Game video editing method, device, equipment and storage medium | |
| WO2005099423A2 (en) | Automatic event videoing, tracking and content generation system | |
| WO2020062053A1 (en) | Methods and apparatus to generate photo-realistic three-dimensional models of photographed environment | |
| CN103106404A (en) | Apparatus, method and system | |
| RU2387011C2 (en) | Movement tracking based on image analysis | |
| US20170201723A1 (en) | Method of providing object image based on object tracking | |
| CN101753852A (en) | Sports event dynamic mini- map based on target detection and tracking | |
| CN101715070A (en) | Method for automatically updating background in specifically monitored video | |
| CN111970434A (en) | Multi-camera multi-target athlete tracking shooting video generation system and method | |
| JP6456244B2 (en) | Camera calibration method and apparatus | |
| CN116668481A (en) | Intelligent identification system of live broadcast education APP | |
| CN103100204A (en) | Method, apparatus and program | |
| CA2393803C (en) | Method and apparatus for real time insertion of images into video | |
| KR20010025404A (en) | System and Method for Virtual Advertisement Insertion Using Camera Motion Analysis | |
| CN114025183A (en) | Live broadcast method, device, equipment, system and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |