CN103634613A - Method and system for automatically synchronizing mobile terminal and broadcast television channels - Google Patents

Abstract

The invention provides a method and a system for automatically synchronizing a mobile terminal and broadcast television channels. The method comprises the following steps: a server side respectively acquires channel audio files of the broadcast television channels in real time and carries out feature extraction to obtain channel audio feature data packets; the mobile terminal records recording audio signals of programs; the mobile terminal extracts audio features of the recording audio signals to obtain recording audio feature data and sends the recording audio feature data to the server side; the server side judges whether channel audio feature data matched with the recording audio feature data exists in a channel audio feature library and outputs a matching channel identifier if channel audio feature data matched with the recording audio feature data exists in the channel audio feature library. As long as a user simply operates the mobile terminal, the server side can identify the television channel which the user is currently watching, thus enhancing the user experience.

Description

The autosynchronous method and system of mobile terminal and broadcast TV channel

Technical field

The invention belongs to broadcasting television technology field, be specifically related to the autosynchronous method and system of a kind of mobile terminal and broadcast TV channel.

Background technology

Along with the development of broadcast television technique and mobile terminal technology, mobile terminal starts by a large number for radio and television interactive application, and mobile terminal is regarded as the second screen of radio and television.Spectators, when watching radio and television, usually can participate in the current interactive application of releasing of radio and television, such as: the interactive application such as spectators' scoring, guess.

In prior art, spectators, when participating in some interactive application, usually need the television channel name that participates in to the manual input of mobile terminal, on the one hand, complex operation, thus make troubles for spectators; On the other hand, owing to being manual input, there is the problem of input error, thereby reduced user's experience.

Summary of the invention

The defect existing for prior art, the invention provides the autosynchronous method and system of a kind of mobile terminal and broadcast TV channel, user only needs simple operations mobile terminal, and server end can recognize the current television channel of watching of user, thereby has promoted user's experience.

The technical solution used in the present invention is as follows:

The invention provides a kind of autosynchronous method of mobile terminal and broadcast TV channel, comprise the following steps:

S1, server end is P broadcast TV channel { M of Real-time Collection respectively ₁, M ₂m ₂p channel audio file { W of current playing program ₁, W ₂w _p, to broadcast TV channel M _ithe channel audio file W corresponding with it _icarry out S1.1-S1.3, obtain P channel audio characteristic bag { Z ₁, Z ₂z _p; Specifically comprise the following steps:

S1.1, arranges and reads for unique identification broadcast TV channel M _ichannel identication ID _i;

S1.2, to described channel audio file W _icarry out audio feature extraction, obtain channel audio characteristic F _i;

S1.3, by channel identication ID _iwith channel audio characteristic F _imerge, obtain channel audio characteristic bag Z _i;

S2, server end is by P channel audio characteristic bag { Z ₁, Z ₂z _pstore in the local channel audio feature database of setting up;

S3, for the in progress appointed program of television set, when described mobile terminal need to be inquired about the channel identication corresponding with described appointed program, described mobile terminal is recorded appointed program at the current recorded audio signal constantly of playing, and obtains recorded audio file;

S4, described mobile terminal carries out audio feature extraction to described recorded audio file, obtains recorded audio characteristic;

S5, described mobile terminal sends to described server end by described recorded audio characteristic;

S6; described server end be take described recorded audio characteristic as inquiry critical data; inquire about described channel audio feature database; judge and in described channel audio feature database, whether have the channel audio characteristic matching with described recorded audio characteristic; if had; obtain the channel audio characteristic bag at described channel audio characteristic place, and then obtain the channel identication of described channel audio characteristic bag storage;

S7, described server end is transferred to described mobile terminal by the described channel identication obtaining.

Preferably, after S2, also comprise:

Server end is deleted the expired channel audio characteristic bag of storing in local channel audio feature database automatically.

Preferably, in S1, for channel audio file W _i, also comprise execution following steps:

By channel audio file W _ibeing divided into duration is a the channel audio section { W of TW _i1, W _i2w _ia, between adjacent two channel audio sections, have overlappingly, overlapping duration is TS; It is initial point that the channel of take is play initial time, calculates the initial time tp of each channel audio section, obtains { tp _i1, tp _i2tp _ia, wherein, initial time tp _i2for channel audio section W _i1in the time interval of playing initial time apart from channel, j ∈ (1,2 ... a); Each channel audio section is carried out to channel audio feature extraction, obtain a channel audio feature subdata { C _i1, C _i2c _ia; By { tp _i1, tp _i2tp _iaand { C _i1, C _i2c _iabe designated as channel audio characteristic F _i;

By channel identication ID _iand channel audio characteristic F _imerge and store channel audio characteristic bag Z into _i;

S4 specifically comprises the following steps:

By recorded audio file division, be that duration is b the recorded audio section { B of TW ₁, B ₂b _b, between adjacent two recorded audio sections, have overlappingly, overlapping duration is TS; The recording initial time of take is initial point, calculates the initial time tr of each recorded audio section, obtains { tr ₁, tr ₂tr _b, wherein, initial time tr _jfor recorded audio section B _japart from the time interval of recording initial time, wherein, j ∈ (1,2 ... b); Each recorded audio section is carried out to recorded audio feature extraction, obtain audio frequency characteristics subdata { E between b record ₁, E ₂e _b;

S6 specifically comprises the following steps:

By b recorded audio section { B ₁, B ₂b _brespectively with P channel audio file { W ₁, W ₂w _pmate, to each channel audio file, all set up a statistic histogram, vertical P the statistic histogram of building together; When the vertical variable number value of any one statistic histogram reaches L, stop the cycling of following S6.2-S6.3;

Wherein, to any channel audio file W _i, the statistic histogram of setting up is HIST-ID _i(dt, number); HIST-ID _ithe process of establishing of (dt, number) comprises the following steps:

S6.1, makes number=0;

S6.2, makes j=1;

S6.3, by recorded audio section B _jrecord between audio frequency characteristics subdata respectively with each channel audio section { W _i1, W _i2w _iadistinguish corresponding channel audio feature subdata { C _i1, C _i2c _iamate, whether the Hamming distance between judgement record between audio frequency characteristics subdata and each channel audio feature subdata is greater than threshold value TH, if be greater than, the two does not mate; Otherwise, the two coupling; If with recorded audio section B _jthe channel audio feature subdata that the match is successful is C _ij, with recorded audio section B _jthe channel audio section that the match is successful is W _ij, channel audio section is W _ijinitial time be tp _ij; Recorded audio section B _jrecording initial time be tr _j; : at dt=tp _ij-tr _jabscissa position, make number=number+1; I=i+1; Returning to circulation and carrying out S6.2 and S6.3, until in P statistic histogram, while existing the number value of the vertical variable of a statistic histogram to reach L, termination cyclic process;

S6.4, the corresponding channel identication of statistic histogram that the vertical variable number value of order reaches L is at first ID _end: channel identication is ID _endchannel be the channel with recorded audio Signal Matching.

The present invention also provides a kind of mobile terminal and the autosynchronous system of broadcast TV channel, comprises server end and mobile terminal;

Described server end comprises:

Real-time Collection module, for Real-time Collection P broadcast TV channel { M ₁, M ₂m _pp channel audio file { W of current playing program ₁, W ₂w _p;

Module is set, the broadcast TV channel M collecting for described Real-time Collection module is set _ichannel identication ID _i;

Channel audio characteristic extracting module,, to described channel audio file W _icarry out audio feature extraction, obtain channel audio characteristic F _i;

Merge module, for by channel identication ID _iwith channel audio characteristic F _imerge, obtain channel audio characteristic bag Z _i;

Channel audio feature database, P the channel audio characteristic bag { Z obtaining for storing described merging block merging ₁, Z ₂z _p;

Described mobile terminal comprises:

Recording module, for recording appointed program at the current recorded audio signal constantly of playing, obtains recorded audio file;

Recorded audio characteristic extracting module, carries out audio feature extraction for described recording module being recorded to the recorded audio file obtaining, and obtains recorded audio characteristic;

Sending module, for sending to described server end by described recorded audio characteristic;

Described server end also comprises:

Enquiry module, take described recorded audio characteristic as inquiry critical data, inquires about described channel audio feature database;

Judge module, judges in described channel audio feature database, whether there is the channel audio characteristic matching with described recorded audio characteristic;

Output module, while for judgment result is that when described judge module being, obtain the channel audio characteristic bag at described channel audio characteristic place, and then obtain the channel identication that described channel audio characteristic bag is stored, then the described channel identication obtaining is transferred to described mobile terminal.

Preferably, described server end also comprises removing module;

Described removing module is used for: automatically delete the expired channel audio characteristic bag that local channel audio feature database is stored.

Preferably, described server end also arranges computing module, channel audio section characteristic extracting module and the channel audio section characteristic merging module of channel segmentation module, channel audio section initial time;

Wherein, described channel segmentation module is used for: by channel audio file W _ibeing divided into duration is a the channel audio section { W of TW _i1, W _i2w _ia, between adjacent two channel audio sections, have overlappingly, overlapping duration is TS;

The computing module of described channel audio section initial time is used for: it is initial point that the channel of take is play initial time, calculates the initial time tp of each channel audio section, obtains { tp _i1, tp _i2tp _ia, wherein, initial time tp _i2for channel audio section W _i1in the time interval of playing initial time apart from channel, j ∈ (1,2 ... a);

Described channel audio section characteristic extracting module, for each channel audio section is carried out to channel audio feature extraction, obtains a channel audio feature subdata { C _i1, C _i2c _ia; By { tp _i1, tp _i2tp _iaand { C _i1, C _i2c _iabe designated as channel audio characteristic F _i;

Described channel audio section characteristic merges module and is used for: by channel identication ID _iand channel audio characteristic F _imerge and store channel audio characteristic bag Z into _i;

Described mobile terminal comprises: recording segmentation module, the computing module of recorded audio section initial time, recorded audio section characteristic extracting module;

Described recording segmentation module is used for: by recorded audio file division, be that duration is b the recorded audio section { B of TW ₁, B ₂b _b, between adjacent two recorded audio sections, have overlappingly, overlapping duration is TS;

The computing module of described recorded audio section initial time is used for: take and record initial time as initial point, calculate the initial time tr of each recorded audio section, obtain { tr ₁, tr ₂tr _b, wherein, initial time tr _jfor recorded audio section B _japart from the time interval of recording initial time, wherein, j ∈ (1,2 ... b);

Described recorded audio section characteristic extracting module is used for: each recorded audio section is carried out to recorded audio feature extraction, obtain audio frequency characteristics subdata { E between b record ₁, E ₂e _b;

Described server end also arranges matching module, and described matching module is used for: by b recorded audio section { B ₁, B ₂b _brespectively with P channel audio file { W ₁, W ₂w ₂mate, to each channel audio file, all set up a statistic histogram, vertical P the statistic histogram of building together; When the vertical variable number value of any one statistic histogram reaches L, stop the cycling of following S6.2-S6.3;

Wherein, to any channel audio file W _i, the statistic histogram of setting up is HIST-ID _i(dt, number); HIST-ID _ithe process of establishing of (dt, number) comprises the following steps:

S6.1, makes number=0;

S6.2, makes j=1;

S6.3, by recorded audio section B _jrecord between audio frequency characteristics subdata respectively with each channel audio section { W _i1, W _i2w _iadistinguish corresponding channel audio feature subdata { C _i1, C _i2c _iamate, whether the Hamming distance between judgement record between audio frequency characteristics subdata and each channel audio feature subdata is greater than threshold value TH, if be greater than, the two does not mate; Otherwise, the two coupling; If with recorded audio section B _jthe channel audio feature subdata that the match is successful is C _ij, with recorded audio section B _jthe channel audio section that the match is successful is W _ij, channel audio section is W _ijinitial time be tp _ij; Recorded audio section B _jrecording initial time be tr _j; : at dt=tp _ij-tr _jabscissa position, make number=number+1; I=i+1; Returning to circulation and carrying out S6.2 and S6.3, until in P statistic histogram, while existing the number value of the vertical variable of a statistic histogram to reach L, termination cyclic process;

S6.4, the corresponding channel identication of statistic histogram that the vertical variable number value of order reaches L is at first ID _end: channel identication is ID _endchannel be the channel with recorded audio Signal Matching.

Beneficial effect of the present invention is as follows:

The autosynchronous method and system of mobile terminal provided by the invention and broadcast TV channel, user only needs simple operations mobile terminal, and server end can recognize the current television channel of watching of user, thereby has promoted user's experience.And server end is identified the channel identication under the program that TV broadcasting automatically based on recorded audio feature, has advantages of that recognition accuracy is high.

Accompanying drawing explanation

Fig. 1 is the autosynchronous method flow schematic diagram of mobile terminal provided by the invention and broadcast TV channel.

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in detail:

Embodiment mono-

As shown in Figure 1, the invention provides a kind of autosynchronous method of mobile terminal and broadcast TV channel, it is characterized in that, comprise the following steps:

S1, server end is P broadcast TV channel { M of Real-time Collection respectively ₁, M ₂m _pp channel audio file { W of current playing program ₁, W ₂w _p, to broadcast TV channel M _ithe channel audio file W corresponding with it _icarry out S1.1-S1.3, obtain P channel audio characteristic bag { Z ₁, Z ₂z _p.

Because the present invention is used for the just channel identication under broadcast program of Real time identification television set, therefore, server end needs the channel audio file that all television channels of Real-time Collection are corresponding, and the channel query request that could effectively send for any one mobile terminal is served.

This step specifically comprises the following steps:

S1.1, arranges and reads for unique identification broadcast TV channel M _ichannel identication ID _i;

S1.2, to described channel audio file W _icarry out audio feature extraction, obtain channel audio characteristic F _i;

S1.3, by channel identication ID _iwith channel audio characteristic F _imerge, obtain channel audio characteristic bag Z _i;

S2, server end is by P channel audio characteristic bag { Z ₁, Z ₂z _pstore in the local channel audio feature database of setting up.

Due to channel audio file corresponding to all television channels of server end Real-time Collection, thereby obtain the channel audio characteristic bag that all television channels are corresponding, for alleviating server end storage pressure, for the program time-interval having broadcasted, server end deletes channel audio characteristic bag corresponding to the period of broadcast items of storing automatically.For example: the program data that server end Real-time Collection BTV plays, the program data collecting is only stored 0.1-0.2 second in channel audio feature database, then deletes.

S3, for the in progress appointed program of television set, when described mobile terminal need to be inquired about the channel identication corresponding with described appointed program, described mobile terminal is recorded appointed program at the current recorded audio signal constantly of playing, and obtains recorded audio file;

S4, described mobile terminal carries out audio feature extraction to described recorded audio file, obtains recorded audio characteristic;

S5, described mobile terminal sends to described server end by described recorded audio characteristic;

S6; described server end be take described recorded audio characteristic as inquiry critical data; inquire about described channel audio feature database; judge and in described channel audio feature database, whether have the channel audio characteristic matching with described recorded audio characteristic; if had; obtain the channel audio characteristic bag at described channel audio characteristic place, and then obtain the channel identication of described channel audio characteristic bag storage;

S7, described server end is transferred to described mobile terminal by the described channel identication obtaining.

Embodiment bis-

On the basis of embodiment mono-, as a kind of preferential execution mode, the present invention also provides a kind of method of segmentation coupling recorded audio file and channel audio file, thereby improves the probability that recorded audio file matches its affiliated channel audio file, concrete, comprise the following steps:

(1) for channel audio file W _i, carry out following steps:

By channel audio file W _ibeing divided into duration is a the channel audio section { W of TW _i1, W _i2w _ia, between adjacent two channel audio sections, have overlappingly, overlapping duration is TS; It is initial point that the channel of take is play initial time, calculates the initial time tp of each channel audio section, obtains { tp _i1, tp _i2tp _ia, wherein, initial time tp _i2for channel audio section W _i1in the time interval of playing initial time apart from channel, j ∈ (1,2 ... a); Each channel audio section is carried out to channel audio feature extraction, obtain a channel audio feature subdata { C _i1, C _i2c _ia; By { tp _ia, tp _i2tp _iaand { C _i1, C _i2c _iabe designated as channel audio characteristic F _i;

By channel identication ID _iand channel audio characteristic F _imerge and store channel audio characteristic bag Z into _i.

Wherein, channel audio feature subdata can be the spectrum energy of channel audio section, and the computational methods of the spectrum energy of channel audio section are:

Channel audio section is carried out to discrete Fourier transform, calculate the amplitude of Fourier coefficient, in spectrum space, select N frequency range, calculate respectively the mean-square value of Fourier coefficient amplitude in each frequency range as the spectrum energy of this frequency range; Spectrum energy is designated as ENERGY (m, n), m ∈ 0,1 ..., M-1 }, n ∈ 0,1 ..., N-1}; Parameter K, M and N meet following condition:

K＝(M-1)*N

According to formula below, calculate binary audio feature, in this formula m, n span be m ∈ 0,1 ..., M-2}, n ∈ 0,1 ..., N-1

$\mathrm{AF}(m,n)=\left\{\begin{array}{cc}1,& \mathrm{IENERGY}(m+1,n)>\mathrm{ENERGY}(m,n)I\\ 0,& \mathrm{II}\end{array}\right..$

The leaching process of channel audio feature is described with an example below:

Channel audio file duration is 1 second, and TW value is 0.06 second, and TS value is 0.04 second, and M value is that 48, N value is that 4, K value is that 188,4 frequency ranges are respectively 500Hz-707Hz, 707Hz-1000Hz, 1000Hz-1414Hz and 1414Hz-2000Hz.The channel audio file division that is 1 second by duration is that 48 durations are the channel audio section of 0.06 second, and between adjacent two channel audio sections, overlapping duration is 0.04 second; Channel audio section is carried out to discrete Fourier transform, calculate the amplitude of Fourier coefficient, in spectrum space, select 4 frequency ranges, calculate respectively the mean-square value of Fourier coefficient amplitude in each frequency range as the spectrum energy of this frequency range; Spectrum energy is designated as ENERGY (m, n), m ∈ 0,1 ..., 47 }, n ∈ 0,1 ..., 3}; According to formula below, calculate channel audio feature, in this formula m, n span be m ∈ 0,1 ..., 46, n ∈ 0,1 ..., 3; Finally obtain 188 binary features.

$\mathrm{AF}(m,n)=\left\{\begin{array}{cc}1,& \mathrm{IENERGY}(m+1,n)>\mathrm{ENERGY}(m,n)I\\ 0,& \mathrm{II}\end{array}\right..$

(2), for recorded audio file, carry out following steps:

By recorded audio file division, be that duration is b the recorded audio section { B of TW ₁, B ₂b _b, between adjacent two recorded audio sections, have overlappingly, overlapping duration is TS; The recording initial time of take is initial point, calculates the initial time tr of each recorded audio section, obtains { tr ₁, tr ₂tr _b, wherein, initial time tr _jfor recorded audio section B _japart from the time interval of recording initial time, wherein, j ∈ (1,2...b); Each recorded audio section is carried out to recorded audio feature extraction, obtain audio frequency characteristics subdata { E between b record ₁, E ₂... E _b.

In the present invention, recorded audio section is carried out to feature extraction, the process of audio frequency characteristics subdata between being recorded, identical with channel audio characteristic extraction procedure, between record, audio frequency characteristics subdata is similarly the spectrum energy of recorded audio section, does not repeat them here.

(3) matching process of recorded audio file and channel audio file is:

By b recorded audio section { B ₁, B ₂b _brespectively with P channel audio file { W ₁, W ₂w _pmate, to each channel audio file, all set up a statistic histogram, vertical P the statistic histogram of building together; When the vertical variable number value of any one statistic histogram reaches L, stop the cycling of following S6.2-S6.3;

Wherein, to any channel audio file W _i, the statistic histogram of setting up is HIST-ID _i(dt, number); HIST-ID _ithe process of establishing of (dt, number) comprises the following steps:

S6.1, makes number=0;

S6.2, makes j=1;

S6.3, by recorded audio section B _jrecord between audio frequency characteristics subdata respectively with each channel audio section { W _i1, W _i2w _iadistinguish corresponding channel audio feature subdata { C _i1, C _i2c _iamate, whether the Hamming distance between judgement record between audio frequency characteristics subdata and each channel audio feature subdata is greater than threshold value TH, if be greater than, the two does not mate; Otherwise, the two coupling; If with recorded audio section B _jthe channel audio feature subdata that the match is successful is C _ij, with recorded audio section B _jthe channel audio section that the match is successful is W _ij, channel audio section is W _ijinitial time be tp _ij; Recorded audio section B _jrecording initial time be tr _j; : at dt=tp _ij-tr _jabscissa position, make number=number+1; I=i+1; Returning to circulation and carrying out S6.2 and S6.3, until in P statistic histogram, while existing the number value of the vertical variable of a statistic histogram to reach L, termination cyclic process;

S6.4, the corresponding channel identication of statistic histogram that the vertical variable number value of order reaches L is at first ID _end: channel identication is ID _endchannel be the channel with recorded audio Signal Matching.

Understand for convenience above-mentioned principle, row cite a plain example and describe:

If L=3, establishes recorded audio file and is split into 6 recorded audio section { B ₁, B ₂b ₆, as itself and channel audio file W ₁during coupling, establish channel audio file W ₁be divided into 9 channel audio section { W ₁₁, W ₁₂w ₁₉; The initial time that each channel audio section is play initial point with respect to channel is respectively { 0.4s, 0.6s ... 2.0s; Recorded audio section { B ₁, B ₂b ₆respectively corresponding recording initial time be { 0s, 0.2s ... 1.0s; Set up statistic histogram, the vertical variable of statistic histogram is number, and horizontal variable is dt; First by recorded audio section B ₁mate respectively 9 channel audio sections, suppose W ₁₂with B ₁the match is successful, and in dt=0.6s position, number value is set to 1; Then by recorded audio section B ₂mate respectively 9 channel audio sections, suppose W ₁₃with B ₂the match is successful, and in dt=0.6s position, number value is set to 2; Then by recorded audio section B ₃mate respectively 9 channel audio sections, suppose W ₁₄with B ₃the match is successful, and in dt=0.6s position, number value is set to 3, now, number value has equaled L value, that is to say, the match is successful for three sections of continuous recorded audio sections and continuous 3 sections of channel audio sections, thinks the recording that this recorded audio file is this channel audio file section content.

Otherwise, due to recorded audio section and channel audio section duration all very short, so, there is following situation: first by recorded audio section B ₁mate respectively 9 channel audio sections, suppose W ₁₂with B ₁the match is successful, and in dt=0.6s position, number value is set to 1; Then by recorded audio section B ₂mate respectively 9 channel audio sections, suppose W ₁₄with B ₂the match is successful, and in dt=0.8s position, number value is set to 1; Then by recorded audio section B ₃mate respectively 9 channel audio sections, suppose not find and B ₃the channel audio section that the match is successful, statistic histogram does not change, and the rest may be inferred, if do not occur that number value is 3 situation, thinks that this recorded audio file is not the recording of this channel audio file section content.

By audio feature segmentation matching process provided by the invention, improved the accuracy of coupling, prevent the situation that a bit of recorded audio section is accidentally identical with channel audio section and occur.

Embodiment tri-

The present embodiment provides a kind of mobile terminal and the autosynchronous system of broadcast TV channel, comprises server end and mobile terminal.

(1) mobile terminal

Mobile terminal comprises:

Recording module, for recording appointed program at the current recorded audio signal constantly of playing, obtains recorded audio file;

Recorded audio characteristic extracting module, carries out audio feature extraction for described recording module being recorded to the recorded audio file obtaining, and obtains recorded audio characteristic;

Sending module, for sending to described server end by described recorded audio characteristic

(2) server end

Server end comprises:

Real-time Collection module, for Real-time Collection P broadcast TV channel { M ₁, M ₂m _pp channel audio file { W of current playing program ₁, W ₂w _p;

Removing module, described removing module is used for: automatically delete the expired channel audio characteristic bag that local channel audio feature database is stored.

Module is set, the broadcast TV channel M collecting for described Real-time Collection module is set _ichannel identication ID _i;

Channel audio characteristic extracting module,, to described channel audio file W _icarry out audio feature extraction, obtain channel audio characteristic F _i;

Merge module, for by channel identication ID _iwith channel audio characteristic F _imerge, obtain channel audio characteristic bag Z _i;

Channel audio feature database, P the channel audio characteristic bag { Z obtaining for storing described merging block merging ₁, Z ₂z _p;

Server end also comprises:

Enquiry module, take described recorded audio characteristic as inquiry critical data, inquires about described channel audio feature database;

Judge module, judges in described channel audio feature database, whether there is the channel audio characteristic matching with described recorded audio characteristic;

Output module, while for judgment result is that when described judge module being, obtain the channel audio characteristic bag at described channel audio characteristic place, and then obtain the channel identication that described channel audio characteristic bag is stored, then the described channel identication obtaining is transferred to described mobile terminal.

Embodiment tetra-

On the basis of embodiment tri-, the present invention also provides a kind of mobile terminal and the autosynchronous system of broadcast TV channel, comprises server end and mobile terminal.

(1) mobile terminal

Mobile terminal comprises: recording segmentation module, the computing module of recorded audio section initial time, recorded audio section characteristic extracting module;

Described recording segmentation module is used for: by recorded audio file division, be that duration is b the recorded audio section { B of TW ₁, B ₂b _b, between adjacent two recorded audio sections, have overlappingly, overlapping duration is TS;

The computing module of described recorded audio section initial time is used for: take and record initial time as initial point, calculate the initial time tr of each recorded audio section, obtain { tr ₁, tr ₂tr _b, wherein, initial time tr _jfor recorded audio section B _japart from the time interval of recording initial time, wherein, j ∈ (1,2 ... b);

Described recorded audio section characteristic extracting module is used for: each recorded audio section is carried out to recorded audio feature extraction, obtain audio frequency characteristics subdata { E between b record ₁, E ₂e _b.

(2) server end

Wherein, server end also arranges computing module, channel audio section characteristic extracting module and the channel audio section characteristic merging module of channel segmentation module, channel audio section initial time;

Wherein, described channel segmentation module is used for: by channel audio file W _ibeing divided into duration is a the channel audio section { W of TW _i1, W _i2w _ia, between adjacent two channel audio sections, have overlappingly, overlapping duration is TS;

The computing module of described channel audio section initial time is used for: it is initial point that the channel of take is play initial time, calculates the initial time tp of each channel audio section, obtains { tp _i1, tp _i2tp _ia, wherein, initial time tp _i2for channel audio section W _i1in the time interval of playing initial time apart from channel, j ∈ (1,2 ... a);

Described channel audio section characteristic extracting module, for each channel audio section is carried out to channel audio feature extraction, obtains a channel audio feature subdata { C _i1, C _i2c _ia; By { tp _i1, tp _i2tp _iaand { C _i1, C _i2c _iabe designated as channel audio characteristic F _i;

Described channel audio section characteristic merges module and is used for: by channel identication ID _iand channel audio characteristic F _imerge and store channel audio characteristic bag Z into _i;

Server end also arranges matching module, and described matching module is used for: by b recorded audio section { B ₁, B ₂b _brespectively with P channel audio file { W ₁, W ₂w _pmate, to each channel audio file, all set up a statistic histogram, vertical P the statistic histogram of building together; When the vertical variable number value of any one statistic histogram reaches L, stop the cycling of following S6.2-S6.3;

Wherein, to any channel audio file W _i, the statistic histogram of setting up is HIST-ID _i(dt, number); HIST-ID _ithe process of establishing of (dt, number) comprises the following steps:

S6.1, makes number=0;

S6.2, makes j=1;

S6.3, by recorded audio section B _jrecord between audio frequency characteristics subdata respectively with each channel audio section { W _i1, W _i2w _iadistinguish corresponding channel audio feature subdata { C _i1, C _i2c _iamate, whether the Hamming distance between judgement record between audio frequency characteristics subdata and each channel audio feature subdata is greater than threshold value TH, if be greater than, the two does not mate; Otherwise, the two coupling; If with recorded audio section B _jthe channel audio feature subdata that the match is successful is C _ij, with recorded audio section B _jthe channel audio section that the match is successful is W _ij, channel audio section is W _ijinitial time be tp _ij; Recorded audio section B _jrecording initial time be tr _j; : at dt=tp _ij-tr _jabscissa position, make number=number+1; I=i+1; Returning to circulation and carrying out S6.2 and S6.3, until in P statistic histogram, while existing the number value of the vertical variable of a statistic histogram to reach L, termination cyclic process;

S6.4, the corresponding channel identication of statistic histogram that the vertical variable number value of order reaches L is at first ID _end: channel identication is ID _endchannel be the channel with recorded audio Signal Matching.

In sum, broadcast TV program sign and recognition methods and system based on audio frequency characteristics provided by the present invention, have the following advantages: user only needs simple operations mobile terminal, server end can recognize the current television channel of watching of user, thereby has promoted user's experience.And server end is identified the channel identication under the program that TV broadcasting automatically based on recorded audio feature, has advantages of that recognition accuracy is high.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be looked protection scope of the present invention.

Claims (6)

1. the autosynchronous method of mobile terminal and broadcast TV channel, is characterized in that, comprises the following steps:

S1, server end is P broadcast TV channel { M of Real-time Collection respectively ₁, M ₂m _pp channel audio file { W of current playing program ₁, W ₂w _p, to broadcast TV channel M _ithe channel audio file W corresponding with it _icarry out S1.1-S1.3, obtain P channel audio characteristic bag { Z ₁, Z ₂z _p; Specifically comprise the following steps:

S1.1, arranges and reads for unique identification broadcast TV channel M _ichannel identication ID _i;

S1.2, to described channel audio file W _icarry out audio feature extraction, obtain channel audio characteristic F _i;

S1.3, by channel identication ID _iwith channel audio characteristic F _imerge, obtain channel audio characteristic bag Z _i;

S2, server end is by P channel audio characteristic bag { Z ₁, Z ₂z _pstore in the local channel audio feature database of setting up;

S3, for the in progress appointed program of television set, when described mobile terminal need to be inquired about the channel identication corresponding with described appointed program, described mobile terminal is recorded appointed program at the current recorded audio signal constantly of playing, and obtains recorded audio file;

S4, described mobile terminal carries out audio feature extraction to described recorded audio file, obtains recorded audio characteristic;

S5, described mobile terminal sends to described server end by described recorded audio characteristic;

S6; described server end be take described recorded audio characteristic as inquiry critical data; inquire about described channel audio feature database; judge and in described channel audio feature database, whether have the channel audio characteristic matching with described recorded audio characteristic; if had; obtain the channel audio characteristic bag at described channel audio characteristic place, and then obtain the channel identication of described channel audio characteristic bag storage;

S7, described server end is transferred to described mobile terminal by the described channel identication obtaining.

2. the autosynchronous method of mobile terminal according to claim 1 and broadcast TV channel, is characterized in that, after S2, also comprises:

Server end is deleted the expired channel audio characteristic bag of storing in local channel audio feature database automatically.

3. the autosynchronous method of mobile terminal according to claim 1 and broadcast TV channel, is characterized in that, in S1, for channel audio file W _i, also comprise execution following steps:

By channel audio file W _ibeing divided into duration is a the channel audio section { W of TW _i1, W _i2w _ia, between adjacent two channel audio sections, have overlappingly, overlapping duration is TS; It is initial point that the channel of take is play initial time, calculates the initial time tp of each channel audio section, obtains { tp _i1, tp _i2tp _ia, wherein, initial time tp _i2for channel audio section W _i1in the time interval of playing initial time apart from channel, j ∈ (1,2 ... a); Each channel audio section is carried out to channel audio feature extraction, obtain a channel audio feature subdata { C _i1, C _i2c _ia; By { tp _i1, tp _i2tp _iaand { C _i1, C _i2c _iabe designated as channel audio characteristic F _i;

By channel identication ID _iand channel audio characteristic F _imerge and store channel audio characteristic bag Z into _i;

S4 specifically comprises the following steps:

By recorded audio file division, be that duration is b the recorded audio section { B of TW ₁, B ₂b _b, between adjacent two recorded audio sections, have overlappingly, overlapping duration is TS; The recording initial time of take is initial point, calculates the initial time tr of each recorded audio section, obtains { tr ₁, tr ₂tr _b, wherein, initial time tr _jfor recorded audio section B _japart from the time interval of recording initial time, wherein, j ∈ (1,2 ... b); Each recorded audio section is carried out to recorded audio feature extraction, obtain audio frequency characteristics subdata { E1, E2 between b record ... Eb };

S6 specifically comprises the following steps:

By b recorded audio section { B ₁, B ₂b _brespectively with P channel audio file { W ₁, W ₂w _pmate, to each channel audio file, all set up a statistic histogram, vertical P the statistic histogram of building together; When the vertical variable number value of any one statistic histogram reaches L, stop the cycling of following S6.2-S6.3;

Wherein, to any channel audio file W _i, the statistic histogram of setting up is HIST-ID _i(dt, number); HIST-ID _ithe process of establishing of (dt, number) comprises the following steps:

S6.1, makes number=0;

S6.2, makes j=1;

S6.3, by recorded audio section B _jrecord between audio frequency characteristics subdata respectively with each channel audio section { W _i1, W _i2w _iadistinguish corresponding channel audio feature subdata { C _i1, C _i2c _iamate, whether the Hamming distance between judgement record between audio frequency characteristics subdata and each channel audio feature subdata is greater than threshold value TH, if be greater than, the two does not mate; Otherwise, the two coupling; If with recorded audio section B _jthe channel audio feature subdata that the match is successful is C _ij, with recorded audio section B _jthe channel audio section that the match is successful is W _ij, channel audio section is W _ijinitial time be tp _ij; Recorded audio section B _jrecording initial time be tr _j; : at dt=tp _ij-tr _jabscissa position, make number=number+1; I=i+1; Returning to circulation and carrying out S6.2 and S6.3, until in P statistic histogram, while existing the number value of the vertical variable of a statistic histogram to reach L, termination cyclic process;

S6.4, the corresponding channel identication of statistic histogram that the vertical variable number value of order reaches L is at first ID _end: channel identication is ID _endchannel be the channel with recorded audio Signal Matching.

4. the autosynchronous system of mobile terminal and broadcast TV channel, is characterized in that, comprises server end and mobile terminal;

Described server end comprises:

Real-time Collection module, for Real-time Collection P broadcast TV channel { M ₁, M ₂m _pp channel audio file { W of current playing program ₁, W ₂w _p;

Module is set, the broadcast TV channel M collecting for described Real-time Collection module is set _ichannel identication ID _i;

Channel audio characteristic extracting module,, to described channel audio file W _icarry out audio feature extraction, obtain channel audio characteristic F _i;

Merge module, for by channel identication ID _iwith channel audio characteristic F _imerge, obtain channel audio characteristic bag Z _i;

Channel audio feature database, P the channel audio characteristic bag { Z obtaining for storing described merging block merging ₁, Z ₂z _p;

Described mobile terminal comprises:

Recording module, for recording appointed program at the current recorded audio signal constantly of playing, obtains recorded audio file;

Recorded audio characteristic extracting module, carries out audio feature extraction for described recording module being recorded to the recorded audio file obtaining, and obtains recorded audio characteristic;

Sending module, for sending to described server end by described recorded audio characteristic;

Described server end also comprises:

Enquiry module, take described recorded audio characteristic as inquiry critical data, inquires about described channel audio feature database;

Judge module, judges in described channel audio feature database, whether there is the channel audio characteristic matching with described recorded audio characteristic;

Output module, while for judgment result is that when described judge module being, obtain the channel audio characteristic bag at described channel audio characteristic place, and then obtain the channel identication that described channel audio characteristic bag is stored, then the described channel identication obtaining is transferred to described mobile terminal.

5. the autosynchronous system of mobile terminal according to claim 4 and broadcast TV channel, is characterized in that, described server end also comprises removing module;

Described removing module is used for: automatically delete the expired channel audio characteristic bag that local channel audio feature database is stored.

6. the autosynchronous system of mobile terminal according to claim 4 and broadcast TV channel; it is characterized in that, computing module, channel audio section characteristic extracting module and channel audio section characteristic that described server end also arranges channel segmentation module, channel audio section initial time merge module;

Wherein, described channel segmentation module is used for: by channel audio file W _ibeing divided into duration is a the channel audio section { W of TW _i1, W _i2w _ia, between adjacent two channel audio sections, have overlappingly, overlapping duration is TS;

The computing module of described channel audio section initial time is used for: it is initial point that the channel of take is play initial time, calculates the initial time tp of each channel audio section, obtains { tp _i1, tp _i2tp _ia, wherein, initial time tp _i2for channel audio section W _i1in the time interval of playing initial time apart from channel, j ∈ (1,2 ... a);

Described channel audio section characteristic extracting module, for each channel audio section is carried out to channel audio feature extraction, obtains a channel audio feature subdata { C _i1, C _i2c _ia; By { tp _i1, tp _i2tp _iaand { C _i1, C _i2c _iabe designated as channel audio characteristic F _i;

Described channel audio section characteristic merges module and is used for: by channel identication ID _iand channel audio characteristic F _imerge and store channel audio characteristic bag Z into _i;

Described mobile terminal comprises: recording segmentation module, the computing module of recorded audio section initial time, recorded audio section characteristic extracting module;

Described recording segmentation module is used for: by recorded audio file division, be that duration is b the recorded audio section { B of TW ₁, B ₂b _b, between adjacent two recorded audio sections, have overlappingly, overlapping duration is TS;

The computing module of described recorded audio section initial time is used for: take and record initial time as initial point, calculate the initial time tr of each recorded audio section, obtain { tr ₁, tr ₂tr _b, wherein, initial time tr _jfor recorded audio section B _japart from the time interval of recording initial time, wherein, j ∈ (1,2 ... b);

Described recorded audio section characteristic extracting module is used for: each recorded audio section is carried out to recorded audio feature extraction, obtain audio frequency characteristics subdata { E between b record ₁, E ₂e _b;

Described server end also arranges matching module, and described matching module is used for: by b recorded audio section { B ₁, B ₂b _brespectively with P channel audio file { W ₁, W ₂w _pmate, to each channel audio file, all set up a statistic histogram, vertical P the statistic histogram of building together; When the vertical variable number value of any one statistic histogram reaches L, stop the cycling of following S6.2-S6.3;

Wherein, to any channel audio file W _i, the statistic histogram of setting up is HIST-ID _i(dt, number); HIST-ID _ithe process of establishing of (dt, number) comprises the following steps:

S6.1, makes number=0;

S6.2, makes j=1;

S6.3, by recorded audio section B _jrecord between audio frequency characteristics subdata respectively with each channel audio section { W _i1, W _i2w _iadistinguish corresponding channel audio feature subdata { C _i1, C _i2c _iamate, whether the Hamming distance between judgement record between audio frequency characteristics subdata and each channel audio feature subdata is greater than threshold value TH, if be greater than, the two does not mate; Otherwise, the two coupling; If with recorded audio section B _jthe channel audio feature subdata that the match is successful is C _ij, with recorded audio section B _jthe channel audio section that the match is successful is W _ij, channel audio section is W _ijinitial time be tp _ij; Recorded audio section B _jrecording initial time be tr _j; : at dt=tp _ij-tr _jabscissa position, make number=number+1; I=i+1; Returning to circulation and carrying out S6.2 and S6.3, until in P statistic histogram, while existing the number value of the vertical variable of a statistic histogram to reach L, termination cyclic process;

S6.4, the corresponding channel identication of statistic histogram that the vertical variable number value of order reaches L is at first ID _end: channel identication is ID _endchannel be the channel with recorded audio Signal Matching.

Images