CN1452331A - Digitized real time multi-channel video and audio differential mode detecting method - Google Patents

Digitized real time multi-channel video and audio differential mode detecting method Download PDF

Info

Publication number
CN1452331A
CN1452331A CN03128702A CN03128702A CN1452331A CN 1452331 A CN1452331 A CN 1452331A CN 03128702 A CN03128702 A CN 03128702A CN 03128702 A CN03128702 A CN 03128702A CN 1452331 A CN1452331 A CN 1452331A
Authority
CN
China
Prior art keywords
differential mode
frame image
image
audio frequency
classification
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN03128702A
Other languages
Chinese (zh)
Other versions
CN1206827C (en
Inventor
洪钧
肖子辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
YONGXIN TONGFANG INFORMATION ENGRG CO Ltd BEIJING
Original Assignee
YONGXIN TONGFANG INFORMATION ENGRG CO Ltd BEIJING
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by YONGXIN TONGFANG INFORMATION ENGRG CO Ltd BEIJING filed Critical YONGXIN TONGFANG INFORMATION ENGRG CO Ltd BEIJING
Priority to CNB031287026A priority Critical patent/CN1206827C/en
Publication of CN1452331A publication Critical patent/CN1452331A/en
Application granted granted Critical
Publication of CN1206827C publication Critical patent/CN1206827C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Abstract

A digitalizing method for real-time detecting the multi-channel video-audio differential mode is characterized by use of the digital image processing technique for video differential mode and the S/N ratio detection for audio differential mode, and includes such steps as digitalizing video and audio signals, data processing, and comparing the result with differential mode threshold.

Description

The method that digitized real-time multichannel is looked, audio frequency differential mode detects
Technical field the invention belongs to the method for video, the detection of audio frequency differential mode, is applied in and detects video, the alarm of audio frequency differential mode in the broadcast television monitoring network.
Background technology is in China, and along with the fast development of cable TV nerve of a covering, the broadcast television monitoring network that adapts with it is with very fast formation.It is to improving broadcast television transmissions and broadcast quality, verify the broadcast television coverage effect, provide scientific basis for drafting, revise the nerve of a covering technical plan, being the judging basis of carrying out program making, transmission and broadcast system technical quality comparation and assessment contest, is the state-of-the-art technology means that radio and television administration authorities at different levels and TV station at different levels, wired TV station carry out scientific management.
The main task of wined tv monitoring network is to monitor video quality and the content of TV program that broadcasts in the whole cable TV network in real time, and video, audio frequency differential mode comprise that mainly no video, absence of audio and video image are static.At present, the video in the cable TV network, the differential mode of audio frequency detect and mainly finish by manual observation, are difficult to realize long-time, multichannel detection.
Summary of the invention the objective of the invention is to provide a kind of digitized real-time multi-channel video, audio frequency differential mode detection method for overcoming existing weak point to the wined tv monitoring means.Realize that long-time, real-time, multichannel video, audio frequency differential mode detect automatically.
The present invention includes behind multi-channel video, the audio signal digitizing, data are handled in real time, with result and relevant differential mode threshold ratio, exceed threshold value and will be judged to differential mode, this method specifically comprises:
Vision signal differential mode detects, and the steps include:
1) utilizes video frequency collection card to gather multi-channel video signal in real time, be stored in the frame image buffering area;
2) extract present frame picture inversion becoming digitlization gray scale image in the frame image buffering area after, to classifying of this each row gray value of current frame image with nC;
3) distribute whether come evenly to determine whether this classification is non-noise classification with the nC that differentiates in each classification;
4) if accounting for the ratio of whole columns, the columns that belongs to non-noise classification reaches more than the even distribution threshold value SR, then think the color end or color bar signal, further, be the color end if not the noise classification number is counted threshold value CBMin less than minimum colour bar, counting threshold value CBMax greater than CBMin and less than maximum colour bar is colour bar, wherein, 0%＜SR＜100%, SR are big more, and condition is harsh more, 1＜CBMin＜5,8＜CBMax＜12;
5), calculate being used for and the correlation A of present frame image certain patterns vector relatively of present frame image and storage in advance if belong to ratio that the columns of non-noise classification accounts for whole columns less than SR;
6) if correlation A greater than similar threshold value Amin then think that present frame image is similar to certain patterns, then current frame image differential mode adopts different certain patterns can judge the different differential mode of present frame image, wherein, 90%＜Amin＜100%, Amin are big more, and then condition is harsh more;
Audio signal differential mode detects, and the steps include:
1) gathers multipath audio signal and carry out therefrom extracting the audio frequency signal amplitude value after analog-to-digital conversion becomes digitized signal, deduct the average noise amplitude and obtain signal to noise ratio snr;
2) comparing audio Signal-to-Noise snr and audio frequency threshold value AT, if snr is greater than AT then think that audio frequency is normal, otherwise think voiceless sound, and continue to judge whether audio frequency is normal, then judge absence of audio differential mode if the voiceless sound time surpasses quiet time interval threshold value MT; Wherein, 3＜AT＜30db, 1 second＜MT＜10 minutes.
Whether said sorting technique can be judges distance that nC and last class respectively be listed as the mean value of nC (being center of gravity nCg) greater than classification thresholds S, greater than then producing a new class, otherwise is included into last class, wherein, 100＜S＜500, S is big more, and condition is loose more.
Each interior nC that classifies of said differentiation distributes, and whether uniform method can be the mean square deviation N that calculates nC in the classification, compares with linear threshold nL, and mean square deviation then is even the distribution less than nL's, promptly be defined as non-noise classification, 5000＜nL＜15000, nL is big more, and condition is loose more.
The method of calculating the vector of said certain patterns and present frame image vector can be: pattern is divided into the fritter of a plurality of area (for example 8*8), calculates the mean value of each piece gray scale, from arranging left to bottom right, be the vector of pattern then.The method of calculating the correlation A of present frame image vector and certain patterns vector can be: establishing present frame image vector is X (x1, x2, ... .xn), x1 wherein, x2......xn is the average gray of each fritter in the image, the certain patterns vector is Y (y1, y2 ... .yn), y1 wherein, y2......yn is the average gray of each fritter in the certain patterns, then correlation A = X · Y = x 1 y 1 + x 2 y 2 + . . . + xnyn ( x 1 2 + x 2 2 + . . . + x n 2 ) ( y 1 2 + y 2 2 + . . . + yn 2 ) ;
Said certain patterns can comprise former frame image, noisy image and self-defined differential mode image;
If present frame image and former frame image similarity judge that then prior image frame is static differential mode; If present frame image image is similar to the specific noise pattern, judge that then prior image frame is no video differential mode; If present frame image and self-defined differential mode image similarity judge that then prior image frame is self-defined differential mode.
Every in the present invention threshold value can set up on their own according to actual conditions, and certain patterns also can be self-defined according to actual conditions.
Characteristics of the present invention are: the present invention utilizes digital image processing techniques to carry out real time automatic detection to the present frame image, determines video differential mode image (do not have that image, image are static, colour bar, the color end and any self-defined image); Utilize detecting audio frequency signal to noise ratio method detects automatically audio frequency and determines its whether differential mode; And can realize long-time, multichannel detection; Also can be according to the present invention detected differential mode situation, be easy to realize the function of automatic alarm report and record.
Description of drawings
Fig. 1 is a video differential mode detection method embodiment flow chart of the present invention;
Fig. 2 is an audio frequency differential mode detection method embodiment flow chart of the present invention.
Embodiment further specifies below in conjunction with drawings and Examples.The embodiment of the inventive method is: 5 videos, audio collection card are installed in video, Audio Processing work station, and every Ka4Lu is used for detecting in real time 20 road videos, audio frequency.Giving the video, the detection of audio frequency differential mode that are solidificated in the work station of adopting the inventive method to work out by real-time 20 road videos, the voice data of gathering of this card alarms application program and handles, so just can be in real time detected various videos, the alarm of audio frequency differential mode be sent to monitoring (branch) center by network, and can work in this locality stand in store video, audio frequency differential mode alarm log information.
Video, audio collection card can adopt the AVE2000 of Thakral card in an embodiment of the present invention, and work station adopts the Pentium 4 industrial computer, and operating system is WINDOWS2000.
The video differential mode detection method of present embodiment is after extracting 20 road video images that video frequency collection card gathers in real time and carrying out analog-to-digital conversion, to become digitized image, to the handling procedure flow process of current frame image as shown in Figure 1, may further comprise the steps:
1) classifying to each row gray value of current frame image with nC, sorting technique can be judge distance that nC and last class respectively be listed as the mean value of nC (being center of gravity nCg) whether greater than classification thresholds S greater than then producing a new class, otherwise be included into last class, the classification thresholds S=300 of present embodiment;
2) differentiate each classification then and whether evenly distribute, determine whether this classification is non-noise classification, and method of discrimination is: calculate the mean square deviation N of nC in each classification, with linear threshold nL (present embodiment line taking threshold value nL=10000; ) relatively, mean square deviation then is evenly to distribute less than nL's, promptly is defined as non-noise classification;
3) calculate the ratio that equally distributed classification midrange accounts for whole columns, present embodiment is got even distribution threshold value SR=80%, if this ratio is greater than SR, then think the color end or color bar signal, further, differentiation is the colour bar or the color end, get minimum colour bar and count threshold value CBMin=4, maximum colour bar is counted threshold value CBMax=10, if non-noise classification number is the color end less than CBMin, is colour bar greater than CBMin and less than CBMax;
4) if sample number accounts for whole sample proportions less than SR in the equally distributed classification, the correlation A of computational picture and certain patterns vector; The computational methods of the vector of certain patterns and present frame image vector are: pattern is divided into the fritter of 8*8, calculates the mean value of each piece gray scale, from arranging left to bottom right, be the vector of pattern then;
The method of calculating the correlation A of this current frame image vector and certain patterns vector is: establishing present frame image vector is X (x1, x2, ... .xn), x1 wherein, x2......xn is the average gray of each fritter in the image, the certain patterns vector is Y (y1, y2 ... .yn), y1 wherein, y2......yn is the average gray of each fritter in the certain patterns, then correlation A = X · Y = x 1 y 1 + x 2 y 2 + . . . + xnyn ( x 1 2 + x 2 2 + . . . + xn 2 ) ( y 1 2 + y 2 2 + . . . + yn 2 ) ;
5) if correlation A is greater than similar threshold value Amin (present embodiment gets 0.98), then the decidable image is similar to certain patterns, adopts different certain patterns can judge that image is static, not have image (noise) differential mode;
The certain patterns of present embodiment comprises former frame image, noisy image and self-defined differential mode image;
If present frame image and former frame image similarity judge that then prior image frame is static differential mode; If present frame image image is similar to the specific noise pattern, judge that then prior image frame is no video differential mode (not having image); If present frame image image and self-defined differential mode image similarity judge that then prior image frame is self-defined differential mode.
It is to gather 20 tunnel audio signals to carry out after analog-to-digital conversion becomes digitized signal from capture card that the audio frequency differential mode of present embodiment detects, and to the handling process of this digital signal as shown in Figure 2, may further comprise the steps:
1) extracts the audio frequency signal amplitude value, deduct the average noise amplitude and obtain signal to noise ratio snr;
2) compare snr and audio frequency threshold value AT, if snr＞AT thinks that then audio frequency is normal, otherwise think absence of audio, if, then judge absence of audio differential mode at the MT that continues absence of audio all the time in the time, its sound intermediate frequency threshold value AT gets 20db, and quiet time interval threshold value MT got 20 seconds.

Claims (6)

1, the method that a kind of digitized real-time multichannel is looked, audio frequency differential mode detects is characterized in that, comprises behind multi-channel video, the audio signal digitizing, data are handled in real time, with result and relevant differential mode threshold ratio, exceed threshold value and will be judged to differential mode, this method specifically comprises:
Vision signal differential mode detects, and the steps include:
1) utilizes the real-time multi pass acquisition vision signal of video frequency collection card, be stored in the frame image buffering area;
2) from the frame image buffering area, extract present frame picture inversion becoming digitlization gray scale image after, to classifying of this each row gray value of current frame image with nC;
3) distribute whether come evenly to determine whether this classification is non-noise classification with the nC that differentiates in each classification;
4) if accounting for the ratio of whole columns, the columns that belongs to non-noise classification reaches more than the even distribution threshold value SR, then think the color end or color bar signal, further, if not counting threshold value CBMin less than minimum colour bar, the noise classification number is the color end, greater than CBMin and to count threshold value CBMax less than maximum colour bar be colour bar, wherein, 0%＜SR＜100%, 1＜CBMin＜5,8＜CBMax＜12;
5), calculate being used for and the correlation A of present frame image certain patterns vector relatively of present frame image and storage in advance if belong to ratio that the columns of non-noise classification accounts for whole columns less than SR;
6) if correlation A greater than similar threshold value Amin then think that present frame image is similar to certain patterns, then current frame image differential mode adopts different certain patterns can judge the different differential mode of present frame image, wherein, 90%＜Amin＜100%;
Audio signal differential mode detects, and the steps include:
1) gathers multipath audio signal and carry out therefrom extracting the audio frequency signal amplitude value after analog-to-digital conversion becomes digitized signal, deduct the average noise amplitude and obtain signal to noise ratio snr;
2) comparing audio Signal-to-Noise snr and audio frequency threshold value AT, if snr is greater than AT then think that audio frequency is normal, otherwise think voiceless sound, and continue to judge whether audio frequency is normal, then judge absence of audio differential mode if the voiceless sound time surpasses quiet time interval threshold value MT; Wherein, 3＜AT＜30db, 1 second＜MT＜10 minutes.
2, the method that digitized real-time multichannel as claimed in claim 1 is looked, audio frequency differential mode detects, it is characterized in that, whether said sorting technique respectively is listed as the distance of mean value of nC greater than classification thresholds S for judging nC and last class, greater than then producing a new class, otherwise be included into last class, wherein, 100＜S＜500.
3, the method that digitized real-time multichannel as claimed in claim 1 is looked, audio frequency differential mode detects, it is characterized in that, nC in each classification of said differentiation distributes, and whether uniform method is for calculating the mean square deviation N of nC in the classification, compare with linear threshold nL, mean square deviation then is evenly to distribute less than nL's, promptly be defined as non-noise classification, 5000＜nL＜15000.
4, the method that digitized real-time multichannel as claimed in claim 1 is looked, audio frequency differential mode detects, it is characterized in that, the computational methods of the vector of said certain patterns and present frame image vector are: the fritter that pattern is divided into a plurality of area, calculate the mean value of each piece gray scale, from arranging left to bottom right, be the vector of pattern then.
5, the method that digitized real-time multichannel as claimed in claim 4 is looked, audio frequency differential mode detects is characterized in that, the computational methods of the correlation A of said present frame image vector and certain patterns vector are: establishing present frame image vector is X (x1, x2, ... .xn), x1 wherein, x2......xn is the average gray of each fritter in the image, the certain patterns vector is Y (y1, y2 ... .yn), y1 wherein, y2......yn is the average gray of each fritter in the certain patterns, then correlation A = X · Y = x 1 y 1 + x 2 y 2 + . . . + xnyn ( x 1 2 + x 2 2 + . . . + xn 2 ) ( y 1 2 + y 2 2 + . . . + yn 2 ) .
6, the method that digitized real-time multichannel as claimed in claim 1 is looked, audio frequency differential mode detects is characterized in that said certain patterns comprises one or more of former frame image, noisy image or self-defined differential mode image; If present frame image and former frame image similarity judge that then prior image frame is static differential mode; If present frame image image is similar to the specific noise pattern, judge that then prior image frame is no video differential mode; If present frame image and self-defined differential mode image similarity judge that then prior image frame is self-defined differential mode.
CNB031287026A 2003-04-24 2003-04-24 Digitized real time multi-channel video and audio differential mode detecting method Expired - Fee Related CN1206827C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB031287026A CN1206827C (en) 2003-04-24 2003-04-24 Digitized real time multi-channel video and audio differential mode detecting method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB031287026A CN1206827C (en) 2003-04-24 2003-04-24 Digitized real time multi-channel video and audio differential mode detecting method

Publications (2)

Publication Number Publication Date
CN1452331A true CN1452331A (en) 2003-10-29
CN1206827C CN1206827C (en) 2005-06-15

Family

ID=29222920

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB031287026A Expired - Fee Related CN1206827C (en) 2003-04-24 2003-04-24 Digitized real time multi-channel video and audio differential mode detecting method

Country Status (1)

Country Link
CN (1) CN1206827C (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101136233B (en) * 2006-07-14 2010-06-23 索尼株式会社 Playback apparatus, playback method, system and recording medium
CN103581663A (en) * 2012-08-03 2014-02-12 展讯通信（上海）有限公司 Method and system for testing circuit board
CN104754324A (en) * 2015-01-09 2015-07-01 北京正奇联讯科技有限公司 Video color bar detecting method and video color bar detecting device
CN106910823A (en) * 2017-04-05 2017-06-30 上海天马微电子有限公司 A kind of flexible display panels and flexible display apparatus
CN110677725A (en) * 2019-10-31 2020-01-10 飞思达技术（北京）有限公司 Audio and video anomaly detection method and system based on Internet television service
CN112261406A (en) * 2020-09-18 2021-01-22 北京中视广信科技有限公司 Ultrahigh-definition video color bar anomaly real-time detection method based on filtering

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101136233B (en) * 2006-07-14 2010-06-23 索尼株式会社 Playback apparatus, playback method, system and recording medium
CN103581663A (en) * 2012-08-03 2014-02-12 展讯通信（上海）有限公司 Method and system for testing circuit board
CN103581663B (en) * 2012-08-03 2015-07-29 展讯通信（上海）有限公司 The method of testing of circuit board and system
CN104754324A (en) * 2015-01-09 2015-07-01 北京正奇联讯科技有限公司 Video color bar detecting method and video color bar detecting device
CN106910823A (en) * 2017-04-05 2017-06-30 上海天马微电子有限公司 A kind of flexible display panels and flexible display apparatus
CN110677725A (en) * 2019-10-31 2020-01-10 飞思达技术（北京）有限公司 Audio and video anomaly detection method and system based on Internet television service
CN112261406A (en) * 2020-09-18 2021-01-22 北京中视广信科技有限公司 Ultrahigh-definition video color bar anomaly real-time detection method based on filtering
CN112261406B (en) * 2020-09-18 2023-08-22 北京中视广信科技有限公司 Filtering-based ultra-high definition video color bar anomaly real-time detection method

Also Published As

Publication number Publication date
CN1206827C (en) 2005-06-15

Similar Documents

Publication Publication Date Title
CN101103246B (en) Automatic judging system
CN107915102B (en) Elevator blocking door behavior detection system and detection method based on video analysis
CN1206827C (en) Digitized real time multi-channel video and audio differential mode detecting method
EP1696396A3 (en) Image pickup apparatus and image distributing method
CN107610099A (en) A kind of emotion state of mind detecting system and method based on pupil information
CN105812618B (en) A kind of method for testing motion and motion detection apparatus
CH695526A5 (en) A method and installation for the measurement of audience ratings.
US20010022852A1 (en) Method of monitoring the quality of distributed digital images by detecting false contours
DE102016207705A1 (en) Smoke detection device, method for detecting smoke of a fire and computer program
DE112004000873T5 (en) Automatic discrimination system
CN101807297B (en) Medical ultrasonic image line detection method
CN113114750B (en) Multimedia digital exhibition hall intelligent regulation and control cloud platform based on big data and cloud computing
CN111355949B (en) Method for constructing audio and video multimedia database and evaluating subjective quality of multimedia
CN2471036Y (en) Real-time monitoring apparatus for quality of TV signal
CN117079219B (en) Vehicle running condition monitoring method and device applied to trailer service
CN101984451B (en) Video-based shielded flame detecting method and device
CN112183406A (en) Video blockage detection method and detection system
CN1130074C (en) Video signal character converting device and method of same
CN1595957A (en) Method for determining automatic detection threshold of bad pixel of medical image
CN110019060B (en) Method and device for automatically synchronizing locomotive video file and operation record file
CN113382208B (en) Information service cloud platform integrated management system based on internet
EP1973351B1 (en) Monitor
CN107203993B (en) A kind of downhole monitoring system based on computer vision
CN110059679A (en) The non-targeted detection method of pollutant in a kind of feedstuff based on near-infrared image
CN114627420A (en) City management violation event information acquisition method and system

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
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20050615

Termination date: 20170424