CN103945246A - Video transmission method and video transmission device - Google Patents

Abstract

The invention provides a video transmission method. The method comprises the step that each layer of a video sequence generated by scalable video coding (SVC) is distributed to a sub-channel of an antenna of an MIMO system for transmission, wherein important layers are distributed to high-quality sub-channels. The invention also provides a video transmission device. The device comprises a distributing module, wherein the distributing module is used for distributing each layer of the video sequence generated by SVC to the sub-channel of the antenna of the MIMO system for transmission, and the important layers are distributed to the high-quality sub-channels. The video transmission method and the video transmission device guarantee the quality of the video at a receiving terminal.

Description

Video transmission method and device

Technical field

The present invention relates to the communications field, in particular to video transmission method and device.

Background technology

In correlation technique, utilize scalable video (SVC) technology to produce the video flowing with sandwich construction, by the video flowing of different layers is dispatched on the subchannel of the each antenna of upper many antennas (MIMO) system randomly, transmit simultaneously, thereby support large video code rate.

The time-varying characteristics of wireless channel have determined that not in the same time, the quality of each sub-channels is incomplete same.The importance of each layer data in SVC video flowing is not exclusively the same.But in correlation technique, base layer data may be scheduled on the very poor subchannel of quality condition, thereby have a strong impact on the quality of receiving terminal video.

Summary of the invention

The present invention aims to provide video transmission method and device, to address the above problem.

In an embodiment of the present invention, a kind of video transmission method is provided, comprise: will on each layer in the video sequence that utilize SVC to produce subchannel that is assigned to respectively the each antenna of multiaerial system, send, wherein, important layer is assigned on the measured subchannel of matter.

In an embodiment of the present invention, a kind of video frequency transmitter is provided, has comprised: distribution module, has sent for each layer of video sequence that utilizes SVC to produce being assigned to respectively on the subchannel of the each antenna of multiaerial system, wherein, important layer is assigned on the measured subchannel of matter.

The video transmission method that the present invention is above-mentioned and device are assigned to important layer on the measured subchannel of matter, have therefore guaranteed the quality of receiving terminal video.

Brief description of the drawings

Accompanying drawing described herein is used to provide a further understanding of the present invention, forms the application's a part, and schematic description and description of the present invention is used for explaining the present invention, does not form inappropriate limitation of the present invention.In the accompanying drawings:

Fig. 1 shows the flow chart of video transmission method according to the preferred embodiment of the invention.

Embodiment

Below with reference to the accompanying drawings and in conjunction with the embodiments, describe the present invention in detail.

One embodiment of the present of invention provide a kind of video transmission method, comprise: will on each layer in the video sequence that utilize SVC to produce subchannel that is assigned to respectively the each antenna of multiaerial system, send, wherein, important layer is assigned on the measured subchannel of matter.

At receiving terminal, user accepts all video layers, and carries out digital demodulation, channel-decoding with this, and video decode obtains decoded video flowing.In correlation technique, distribute randomly each layer in video sequence, and in the present embodiment, important layer is assigned on the measured subchannel of matter, this can improve the communication quality of importance, thereby improves the communication quality of whole video sequence, thereby guarantees the quality of receiving terminal video.

In correlation technique, all video layers are put on an equal footing, and in the present embodiment, according to the quality condition of each sub-channels and the importance of different layers video flowing, reasonably carry out data dispatch, have very large effect to improving receiving terminal video quality.

Preferably, important layer is assigned on the measured subchannel of matter and is comprised: basic layer is assigned on top-quality subchannel.The present invention states in the process of transmission of video on the implementation, find that in correlation technique, at least there are the following problems: the importance of each layer data in SVC video flowing is not exclusively the same, put on an equal footing without distinction, in the time that base layer data is scheduled for the very poor subchannel of quality condition, can have a strong impact on the quality of receiving terminal video.This preferred embodiment thinks that basic layer is most important layer, by basic layer is assigned on top-quality subchannel, thereby has improved communication quality.

Preferably, important layer is assigned on the measured subchannel of matter and is also comprised: the data of n enhancement layer are put on the subchannel that quality n+1 is good.

Preferably, expect video distortion by assessing minimized user, important layer is assigned on the measured subchannel of matter.

Preferably, assess minimized user and expect that video distortion comprises:

$A*=\underset{\left\{A\right\}}{\arg \min }\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-P_k(A,R,r)))$

s.t.

$a_{\mathrm{ij}}\∈\left\{\mathrm{0,1}\right\},\∀1\≤i,j\≤L$

$\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}=1,\∀1\≤j\≤L$

$\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}=1,\∀1\≤i\≤L$

Wherein, L represents video layer quantity, A={a _ijl × L matrix, represent video layer-antenna mapping relation, a _ij=1 represents that i layer video is scheduled for j sub-channels and transmits, otherwise a _ij=0, R=[R ₁, R ₂..., R _l] expression subchannel bandwidth vector, R _irepresent the bandwidth of i sub-channels, r=[r ₁, r ₂..., r _l] expression video layer code check vector, r _irepresent the code check of i layer video, w _irepresent the weight of i layer video, P _k(A, R, r) represents the packet loss of the k layer video calculating according to the code check r of A, bandwidth R and k layer video.

A* is best video layer-antenna mapping relational matrix.Also in the time that mapping relations are A*, can obtain minimum video distortion.The implication of argmin, exactly in the time that distortion obtains minimum value, is returned to the value of independent variable below.

Preferably, assess minimized user and expect that video distortion comprises:

$u*=\underset{\left\{u\right\}}{\arg \max }I\left(u\right)$

Wherein, u is that length is 0/1 character string, $I\left(u\right)=-\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-P_k(u,R,r)));$

L represents video layer quantity, R=[R ₁, R ₂..., R _l] expression subchannel bandwidth vector, R _irepresent the bandwidth of i sub-channels, r=[r ₁, r ₂..., r _l] expression video layer code check vector, r _irepresent the code check of i layer video, w _irepresent the weight of i layer video, P _k(A, R, r) represents the packet loss of the k layer video calculating according to the code check r of A, bandwidth R and k layer video.

Preferably, assess minimized user and expect that video distortion comprises: be right $u*=\underset{\left\{u\right\}}{\arg \max }I\left(u\right)$ Solve as follows:

Definition event functions: h ₊(X), when event X is true time, this function returns to 1, otherwise returns to 0, definition p _ifor u _i=1 probability, as given Probability p=[p ₁, p ₂..., p _k] time, the probability of character string u is: $f(u,p)=\underset{i=1}{\overset{K}{\mathrm{\Π}}}{p}_i^{h_{+}(u_i=1)}{(1-p_i)}^{1-h_{+}(u_i=1)};$

For the set of one group of u, its cross entropy is defined as: $Q\left(p\right)=-\frac{1}{N_u}\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)\ln f\left(u\right[n],p);$

Probability when wherein p has represented that last iteration finishes, N _urepresent the quantity in set, u[n] be that d is the desired value of expecting, in iterative process, constantly updates, and finally approaches target optimal solution according to the random character string generating of Probability p;

Replacement criteria is $p_i=\frac{\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)h_{+}\left(u_i\right[n]=1)}{\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)},$ Obtain the p after upgrading, then generate random sequence u according to p, calculate cross entropy Q (p), make cross entropy minimum again upgrade p, and then regenerate random sequence u, iterate, until meet certain condition, exit iteration, obtain target solution.

Preferably, receiving terminal is estimated the signal to noise ratio of each sub-channels, and feeds back to transmitting terminal.

Fig. 1 shows the flow chart of video transmission method according to the preferred embodiment of the invention, comprising:

Step S10, utilize SVC to produce the video sequence with sandwich construction, as video source, according to signal to noise ratio, the code check of video layer and the importance of video layer of different subchannels, the video layer of different layers is mapped on the subchannel of different quality and transmits, the principle of video layer-channel mapping is that minimum user is expected video distortion, wherein, expect that video quality can be by the channel signal to noise ratio of feedback, importance and video layer-antenna channel relation of video layer, obtain by numerical computations, as a kind of desired value;

Step S20, at receiving terminal, user accepts all video layers, and carries out digital demodulation, channel-decoding with this, and video decode obtains decoded video flowing;

Step S30, receiving terminal is responsible for estimating the signal to noise ratio of each sub-channels, and feeds back the signal to noise ratio of each sub-channels by feedback channel;

So far, whole transmitting procedure finishes.

Below further describe the preferred embodiments of the present invention and how to estimate to expect video distortion, and the matching relationship of How to choose video layer-interchannel.Symbol definition:

L: represent video layer quantity, also represent the quantity of transmitting antenna (subchannel), video layer quantity equals the quantity of transmitting antenna;

A _{l × L}={ a _ij}: video layer-antenna mapping relational matrix, a _ij=1 represents that i layer video is scheduled for j sub-channels and transmits, otherwise a _ij=0;

R=[R ₁, R ₂..., R _l]: subchannel bandwidth vector, R _irepresent the bandwidth of i sub-channels;

R=[r ₁, r ₂..., r _l]: video layer code check vector, r _irepresent the code check of i layer video;

W=[w ₁, w ₂..., w _l]: video layer weight vectors, w _irepresent the weight of i layer video;

transmission error probability after FEC chnnel coding, P _ij ^crepresent that i layer video is mapped to the probability making a mistake while transmission on j root;

P=[P ₁, P ₂..., P _l]: video layer transmission error probability, P _ithe probability making a mistake while representing the transmission of i layer video;

Video distortion:

In above-mentioned definition, be subject to the impact of video layer code check, channel signal to noise ratio and channel width.And P=[P ₁, P ₂..., P _l] again by and video layer-antenna match relational matrix A _{l × L}={ a _ijinstitute determine,

$p_i={\left[A_{L\×L}\right]}_i\×{\left[P_{L\×L}^c\right]}_i^T$

Wherein represent [] _ithe i of matrix is capable, [] ^tthe transposition of representing matrix.

Finally, the expectation video distortion in (3) can be expressed as follows:

$D=\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-p_k))$

Problem formalized description:

Target is to select best video layer-channel matched relational matrix, makes to expect video distortion minimum,

$A*=\underset{\left\{A\right\}}{\arg \min }\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-P_k(A,R,r)))$

s.t.

$a_{\mathrm{ij}}\∈\left\{\mathrm{0,1}\right\},\∀1\≤i,j\≤L$

$\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}=1,\∀1\≤j\≤L$

$\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}=1,\∀1\≤i\≤L$

Problem redefines:

Due to feasible video layer-channel matched matrix one Kind.Wherein, to show and only have an element be 1 to the different rows difference of each matrix.Therefore, the solution space of matrix can be with a length 0/1 character string u be mapped one by one, i.e. a unique corresponding video layer-channel matched matrix of character string, equally also unique corresponding one expect video distortion.The set that defines all feasible u is U, now, problem can redefine into:

$u*=\underset{\left\{u\right\}}{\arg \max }I\left(u\right)$

Wherein $I\left(u\right)=-\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-P_k(u,R,r))).$

Problem solving:

In order to find best video layer-channel matched relational matrix, can adopt Cross-Entropy Algorithm.

First, event functions a: h of definition ₊(X), when event X is true time, this function returns to 1, otherwise returns to 0, definition p _ifor u _i=1 probability.As given Probability p=[p ₁, p ₂..., p _k] time, the probability of character string u is:

$f(u,p)=\underset{i=1}{\overset{K}{\mathrm{\Π}}}{p}_i^{h_{+}(u_i=1)}{(1-p_i)}^{1-h_{+}(u_i=1)}$

For the set of one group of u, their cross entropy is defined as:

$Q\left(p\right)=-\frac{1}{N_u}\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)\ln f\left(u\right[n],p)$

Probability when wherein p has represented that last iteration finishes, N _urepresent the quantity in set, u[n] be that d is the desired value of expecting, in iterative process, constantly updates, and finally approaches target optimal solution according to the random character string generating of Probability p.

According to cross entropy principle, the replacement criteria of Probability p is for minimizing cross entropy,

$p^{*}=\underset{p}{\arg \min }Q\left(p\right)$

And then the replacement criteria obtaining is

$p_i=\frac{\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)h_{+}\left(u_i\right[n]=1)}{\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)}$

Obtain the later p after upgrading, then generate random sequence u according to p, calculate cross entropy Q (p), make cross entropy minimum again upgrade p, and then regenerate random sequence u.Iterate, until meet certain condition, exit iteration, obtain target solution.

Suppose to have now 2 antennas (note position ab) and 3 layer videos (being designated as 12), the possible corresponding relation one between them has in 2 so, i.e. (a-1, b-2); (a-2, b-1), further, supposes to have L root antenna and L layer video, and so possible mapping relations have the factorial kind of L, and essence is exactly all solutions, can be any as for the order of label.By the factorial kind combination label of this L, then use binary number representation phase label, as described in " problem redefines " that chapter.Like this, each combination is just mapped one by one with a label, and label is binary sequence, and by character string, u represents.For example, if there are 3 antennas, the length of u is just 3 so, for example u=[001] represent the first matching relationship, u=[100] represent the 4th kind of matching relationship.

Therefore, above-mentioned formula is mainly that iterative computation goes out each probability that is 1 in u sequence, be pi, once pi has restrained, each in u is also just fixed so, and then the whole sequence of u is also just fixed, further obtain u corresponding be which matching relationship, then just can be according to distortion formula calculated distortion.

The mapping relations of antenna and video are transformed, connect with a binary sequence u, or 2 antennas are example:

Possible mapping relations: basic layer a is mapped to first antenna, and the first enhancement layer is mapped to second antenna, is designated as the first solution [01]; Basic layer a is mapped to second antenna, and the first enhancement layer is mapped to first antenna, is designated as the second solution [10]; 2 factorial equals 2, so travel through also with regard to both of these case, left calculating u, if last u=[01], so with regard to corresponding the first solution, if u=[10], so with regard to corresponding the second solution.

In an embodiment of the present invention, a kind of video frequency transmitter is provided, has comprised: distribution module, has sent for each layer of video sequence that utilizes SVC to produce being assigned to respectively on the subchannel of the each antenna of multiaerial system, wherein, important layer is assigned on the measured subchannel of matter.

Preferably, distribution module, for expecting video distortion by assessing minimized user, is assigned to important layer on the measured subchannel of matter.

Obviously, those skilled in the art should be understood that, above-mentioned of the present invention each module or each step can realize with general calculation element, they can concentrate on single calculation element, or be distributed on the network that multiple calculation elements form, alternatively, they can be realized with the executable program code of calculation element, thereby, they can be stored in storage device and be carried out by calculation element, or they are made into respectively to each integrated circuit modules, or the multiple modules in them or step are made into single integrated circuit module to be realized.Like this, the present invention is not restricted to any specific hardware and software combination.

The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any amendment of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (10)

1. a video transmission method, is characterized in that, comprising:

To on each layer in the video sequence that utilize SVC to produce subchannel that is assigned to respectively the each antenna of multiaerial system, send, wherein, important layer is assigned on the measured subchannel of matter.

2. method according to claim 1, is characterized in that, important layer is assigned on the measured subchannel of matter and is comprised:

Basic layer is assigned on top-quality subchannel.

3. method according to claim 2, is characterized in that, important layer is assigned on the measured subchannel of matter and is also comprised:

The data of n enhancement layer are put on the subchannel that quality n+1 is good.

4. method according to claim 1, is characterized in that, expects video distortion by assessing minimized user, and important layer is assigned on the measured subchannel of matter.

5. method according to claim 1, is characterized in that, assesses minimized user and expects that video distortion comprises:

$A*=\underset{\left\{A\right\}}{\arg \min }\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-P_k(A,R,r)))$

s.t.

$a_{\mathrm{ij}}\∈\left\{\mathrm{0,1}\right\},\∀1\≤i,j\≤L$

$\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}=1,\∀1\≤j\≤L$

$\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{a}_{\mathrm{ij}}=1,\∀1\≤i\≤L$

Wherein, L represents video layer quantity, A={a _ijl × L matrix, represent video layer-antenna mapping relation, a _ij=1 represents that i layer video is scheduled for j sub-channels and transmits, otherwise a _ij=0, R=[R ₁, R ₂..., R _l] expression subchannel bandwidth vector, R _irepresent the bandwidth of i sub-channels, r=[r ₁, r ₂..., r _l] expression video layer code check vector, r _irepresent the code check of i layer video, w _irepresent the weight of i layer video, P _k(A, R, r) represents the packet loss of the k layer video calculating according to the code check r of A, bandwidth R and k layer video.

6. method according to claim 5, is characterized in that, assesses minimized user and expects that video distortion comprises:

$u*=\underset{\left\{u\right\}}{\arg \max }I\left(u\right)$

Wherein, u is that length is 0/1 character string, $I\left(u\right)=-\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{w}_i(1-\underset{k=1}{\overset{i}{\mathrm{\Π}}}(1-P_k(u,R,r)));$

L represents video layer quantity, R=[R ₁, R ₂..., R _l] expression subchannel bandwidth vector, R _irepresent the bandwidth of i sub-channels, r=[r ₁, r ₂..., r _l] expression video layer code check vector, r _irepresent the code check of i layer video, w _irepresent the weight of i layer video, P _k(A, R, r) represents the packet loss of the k layer video calculating according to the code check r of A, bandwidth R and k layer video.

7. method according to claim 6, is characterized in that, assesses minimized user and expects that video distortion comprises: be right solve as follows:

Definition event functions: h ₊(X), when event X is true time, this function returns to 1, otherwise returns to 0, definition p _ifor u _i=1 probability, as given Probability p=[p ₁, p ₂..., p _k] time, the probability of character string u is: $f(u,p)=\underset{i=1}{\overset{K}{\mathrm{\Π}}}{p}_i^{h_{+}(u_i=1)}{(1-p_i)}^{1-h_{+}(u_i=1)}$

For the set of one group of u, its cross entropy is defined as: $Q\left(p\right)=-\frac{1}{N_u}\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)\ln f\left(u\right[n],p)$

Probability when wherein p has represented that last iteration finishes, N _urepresent the quantity in set, u[n] be that d is the desired value of expecting, in iterative process, constantly updates, and finally approaches target optimal solution according to the random character string generating of Probability p;

Replacement criteria is $p_i=\frac{\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)h_{+}\left(u_i\right[n]=1)}{\underset{n=1}{\overset{N_u}{\mathrm{\Σ}}}{h}_{+}(I\left(u\right[n\left]\right)\≥d)},$ Obtain the p after upgrading, then generate random sequence u according to p, calculate cross entropy Q (p), make cross entropy minimum again upgrade p, and then regenerate random sequence u, iterate, until meet certain condition, exit iteration, obtain target solution.

8. method according to claim 1, is characterized in that, receiving terminal is estimated the described signal to noise ratio of each sub-channels, and feeds back to transmitting terminal.

9. a video frequency transmitter, is characterized in that, comprising:

Distribution module, sends for each layer of video sequence that utilizes SVC to produce being assigned to respectively on the subchannel of the each antenna of multiaerial system, wherein, important layer is assigned on the measured subchannel of matter.

10. device according to claim 9, is characterized in that, described distribution module, for expecting video distortion by assessing minimized user, is assigned to important layer on the measured subchannel of matter.