CN100542298C - A kind of block size computing method of cylinder panorama video code and transmission method - Google Patents

Abstract

The invention provides a kind of block size computing method and transmission method of cylinder panorama video code, comprise that cylinder panorama video does not carry out the direct precoding of piecemeal, a code check of the panorama video code that obtains; This video carries out the piecemeal precoding according to an optional piece size, obtains the encoder bit rate under this piece size; Calculate the increment of above two code checks; Factor of influence according to this this video of incremental computations; Calculate the optimum code piece size of this video at last according to this factor.The piece size that the present invention calculates according to said method provides a kind of panoramic video to carry out the method for block encoding transmission again.The present invention can calculate optimum encoding block size by twice precoding, uses this optimum piece size to carry out cylinder panorama video code, can be under the identical situation of user's visual quality transmitted data amount minimum.

Description

A kind of block size computing method of cylinder panorama video code and transmission method

Technical field

The present invention relates to video and image coding and decoding field, particularly the block size computing method of cylinder panorama video code and based on the transmission technology of the cylinder panorama video code of piecemeal.

Background technology

Panoramic video is a kind of emerging medium, compares with traditional video, and it is big to have a visual field, resolution height, characteristics such as data volume is big.Utilize panoramic video, observer's viewpoint is constant, change direction of observation and can observe whole scenes on every side, and common two-dimensional video has only reacted certain part of panoramic video.

Cylinder panorama video is a kind of common panoramic video, and it is equivalent to a virtual video camera, and the three-dimensional body in the space is projected on the cylinder.The generation of cylinder panorama video can utilize multi-cam or the collection of single camera acquisition system to form.Because projection surface is a cylinder, so above-below direction has some blind spots.

Because the field range of panoramic video is 5～6 times of ordinary video, providing to the user under the situation of identical visual quality, the data volume of panoramic video is 5～6 times of ordinary video.If according to traditional video transmission scheme, the use of panoramic video under network environment becomes difficult.But, because at synchronization, certain part that the required content of seeing of user is a panoramic video is so block encoding and transmission become the common scheme of panoramic video Network Transmission.

In the block encoding of panoramic video, the size of piecemeal is for the code efficiency and the transmission region important influence of panoramic video, and these two factors need are directly determining data quantity transmitted.If the encoding block size is little, then transmission region is less, but code efficiency can be lower; If the encoding block size is big, then code efficiency is higher, but transmission region is also bigger.So under identical visual quality, different encoding block sizes, it is different needing data quantity transmitted.But in systems more both domestic and external, do not consider the size of piecemeal basically now, generally according to the experience decision, this has just caused the increase of transmitted data amount to block size.

Summary of the invention

The objective of the invention is provides a kind of computational methods fast and effectively for the selection of cylinder panorama video code piece size, makes under local visual situation identical in quality required data quantity transmitted minimum.

To achieve these goals, the invention provides a kind of cylinder panorama video code block size computing method, utilize the width of coding panoramic video, the panoramic video peak width that relates to, and the information calculations such as factor of influence of current video obtain optimum code piece size.Comprise following steps:

Cylinder panorama video does not carry out the direct precoding of piecemeal, the code check R of the panorama video code that obtains ₀

Cylinder panorama video carries out the piecemeal precoding according to an optional block size a, obtains the encoder bit rate R under block size a _a

According to formula ω _a=(R _a-R ₀)/R ₀, calculate code check R _aThan code check R ₀Increment ω _a

According to formula C=ω _aAW/ (W-a) calculates the factor of influence C of this video, and wherein W represents that the width of this video is a W macro block;

According to formula $a_{\mathrm{me}}=\left[\sqrt{\mathrm{CM}/(1-\frac{C}{W})}\right],$ Calculate the optimum code piece size a of this video _Me, wherein M represents that the target area width is a M macro block.

Based on this block size computing method, the present invention also provides a kind of cylinder panorama video code transmission method, comprises the steps:

Calculate the piece size a that gets according to said method _MeThis video is carried out piecemeal;

Behind the piecemeal, the sequence of each piece is independently encoded;

Select the data behind the required coding to transmit.

By cylinder panorama video code block size computing method provided by the invention, can access a piece size that compression efficiency is higher and transmission region is more appropriate that makes panoramic video.Need the data quantity transmitted minimum thereby make.

Description of drawings

Below, describe embodiments of the invention in conjunction with the accompanying drawings in detail, wherein:

Fig. 1 is a cylinder panorama video piecemeal schematic diagram.

Fig. 2 is ω _aWith encoding block number and ω _aGraph of a relation with the encoding block size.

Fig. 3 is the operational flow figure of cylinder panorama video optimum code block size computing method.

Fig. 4 is compression of cylinder panorama video piecemeal and transmission schematic diagram.

Embodiment

The partial view visual quality is determined by panoramic video original coding quality.In the process of coding, control the quality of panorama video code with quantization parameter (QP).Need under the certain condition of QP, calculate block size a and make the data quantity transmitted minimum, even transmission code rate r _aMinimum block size a.So at first calculate transmission code rate r _aFunctional relation with block size a.

If the width of target area is a M macro block, dividing block size is a macro block, as shown in Figure 1.Because the position that the target area is dropped in the panoramic video can change, thereby the number of relevant piecemeal also can change thereupon.A kind of situation be the right margin of target area in the scope of a ', relevant piecemeal number is N ', another kind of situation is that the right margin of target area is at a " scope in, relevant piecemeal number is N ".Then:

Formula 1

The probability that makes the target area drop in a ' scope is P ', and a " probability in the scope is P " is dropped in the target area.Thereby, when minute block size is a, the relevant range width W _aDesired value can calculate by following formula:

$W_a=N^{\′}\·a\·P^{\′}+N^{\′\′}\·a\·P^{\′\′}$

Formula 2

$=M+a.$

Transmission code rate is the code check sum of all relevant piecemeals that need transmit, and calculates with following formula.

$r_a=\frac{N_a}{N}\·R_a=\frac{W_a}{W}\·R_a$ Formula 3

Wherein, r _aTransmission code rate when the expression block size is a, R _aThe encoder bit rate of expression block size panoramic video when being a, the code check sum of all block encodings just, N _aComprise the number of the macro block of target area when representing the panoramic video block size for a, N represents panoramic video macro block sum, W _aThe width sum of the macro block that comprises the target area when expression panoramic video block size is a, W represents the width of panoramic video.

In order to calculate under the certain situation of original coding quality the optimum code piece size a of transmission code rate minimum _Me, the most direct method utilizes above-mentioned formula to calculate the optimum code size by enumerating mode exactly.Embodiment obtains panorama video code code check R at first each possible block size is all encoded _a, calculate the width W of relevant piecemeal then according to formula 2, calculate transmission code rate r according to formula 3 then again _a, by the transmission code rate of more different block sizes, the block size that obtains the transmission code rate minimum is exactly the optimum code size at last.Example forest, village, bridge and four sequences of assembly hall as shown in table 2 enumerate result of calculation.

Adopt each possible block size to encode to panoramic video, through relatively finding the optimum code piece, such method is unpractical in actual applications.The present invention proposes a kind of piece of optimum code fast size Selection method based on the inverse ratio model.

At first define λ _aBe R _aAnd R ₀Ratio.

${\mathrm{\&lambda;}}_a=\frac{R_a}{{\mathrm{<figure-callout\; id="0"\; label="R"\; filenames="C20071017544300122.png,C20071017544300131.png"\; state="\{\{state\}\}">R</figure-callout>}}_0}$ Formula 4

R wherein _aPanorama video code code check when being of a size of a for encoding block, R ₀Be the panoramic video encoder bit rate during piecemeal not.According to the relation between panorama video code block size and the code efficiency, λ _aAll the time be a real number more than or equal to 1.

By formula 3, formula 4 can obtain:

$r_a=\frac{W_a}{W}\&CenterDot;R_a={\mathrm{\&lambda;}}_a\&CenterDot;{\mathrm{<figure-callout\; id="0"\; label="R"\; filenames="C20071017544300122.png,C20071017544300131.png"\; state="\{\{state\}\}">R</figure-callout>}}_0\&CenterDot;\frac{W_a}{W}\&CenterDot;$ Formula 5

For same encoding condition and panoramic video, different encoding block size a, R ₀With W is identical, so definition:

P _a=λ _aW _aFormula 6

Make P as long as calculate _aMinimum encoding block size promptly gets optimum code piece size.

Further, definition

${\mathrm{\&omega;}}_a=\frac{R_a-R_0}{{\mathrm{<figure-callout\; id="0"\; label="R"\; filenames="C20071017544300122.png,C20071017544300131.png"\; state="\{\{state\}\}">R</figure-callout>}}_0}={\mathrm{\&lambda;}}_a-1.$ Formula 7

Table 1 has been listed the bridge sequence under different encoding block sizes, different ω _a,, used 4 different quantization parameters (QP) and encoded for each encoding block size.

Table Bridge 1 beam sequence ω _aWith the encoding block size relationship

The data of table in the arrangement, ω _aAnd the relation between encoding block number and the encoding block size is shown in Fig. 2 a.ω _aAnd proportional relation between the encoding block number, owing to be inverse relation between encoding block number and the encoding block size, so ω _aAnd be inverse relation between the encoding block size, shown in Fig. 2 b.

Because ω _aBe inverse relation with the encoding block size, the present invention sets up a model between the two, describes between the two to concern.

${\mathrm{\&omega;}}_a=\frac{C}{a}-\frac{C}{W}.$ Formula 8

For same video C is constant, when the panoramic video block encoding, and a=W, at this time ω _a=0, meet us for ω _aDefinition.

Can obtain the computing formula of C by formula 8:

$C=\frac{{\mathrm{\&omega;}}_a\&CenterDot;a\&CenterDot;W}{W-a}.$ Formula 11

Thus, video is encoded by any one encoding block size a, calculate ω _a, then can computational constant C.

By formula 6,7,8 calculate:

$P_a=W_a\&CenterDot;{\mathrm{\&lambda;}}_a=(M+a)\&CenterDot;(1+{\mathrm{\&omega;}}_a)=(M+a)\&CenterDot;(1+\frac{C}{a}-\frac{C}{W}).$ Formula 9

Calculate P _aTo the derivative of a is 0 point, is P _aExtreme point, this moment a value be optimum code piece size.

${P_a}^{\&prime;}=0\&DoubleRightArrow;a_{\mathrm{me}}=\left[\sqrt{\frac{\mathrm{CM}}{1-\frac{C}{W}}}\right].$ Formula 10

Can calculate optimum code piece size a in view of the above _Me

According to above-mentioned analysis, according to the width of panoramic video, information calculations encoding block sizes such as target area width.This computational methods embodiment is as follows as shown in Figure 3:

At first, cylinder panorama video to be encoded is not carried out piecemeal, integral body is carried out precoding, the code check R of the panorama video code that obtains ₀Then, select a piece size a arbitrarily, carry out the piecemeal precoding to panoramic video, obtain the encoder bit rate R under piece size a according to this piece size for cylinder panorama video to be encoded _aAgain according to formula ω _a=(R _a-R ₀)/R ₀Calculate precoding code check R _aThan code check R ₀Increment ω _aAgain then, according to formula C=ω _aAW/ (W-a) calculates the factor of influence C of current cylinder panorama video, and wherein W represents the width of panoramic video.At last, according to formula $a_{\mathrm{me}}=\left[\sqrt{\mathrm{CM}/(1-\frac{C}{W})}\right],$ Calculate the optimum code piece size a of cylinder panorama video _Me, wherein M is the target area width.

According to the block size computing method of above-mentioned cylinder panorama video code, the transmission method of cylinder panorama video provided by the invention, as shown in Figure 4, concrete steps are as follows:

At first calculate the piece size that gets panoramic video is carried out piecemeal according to the block size computing method of above-mentioned cylinder panorama video code.

According to above piecemeal, the sequence of each piece is independently encoded then.

Select the data behind the needed coding to transmit again then.Can select data according to the current visual angle of user at this.Transmission medium can be the appropriate combination of internet, wireless network, local area network (LAN), optic network, other suitable transmission medium or these transmission mediums.

Last decoding end receives after the data, and these piece sequences are independently decoded and projective transformation, obtains required image.

Table 2 encoding block size and transmission code rate corresponding relation

Table 2 has illustrated in the transmission of cylinder panorama video image the encoding block size to the influence of actual transmissions code check, its with resolution be 4 image sequences of 1920x352 at quantization parameter QP=28, be transmitted as example under the condition of target area M=20 and describe.For being target image with forest, village, bridge and assembly hall, carry out the coding transmission test respectively with the different masses size, can observe different images optimum block size separately from each piece size corresponding code rate.For example, when target image was forest, the piece size equaled at 2 o'clock, and code check is 194.89, and this value is the minimum value of all different masses size encoder bit rates.The piece size that adopts the inventive method to calculate also is 2, and is consistent with the optimal value of enumerating test.

Enumerate calculating optimal result and the inventive method comparison sheet as a result during table 3 different Q P

It is that 4 sequences of target image are utilized the actual test of piece size coding full blast piece size that obtains and the optimum piece size comparative result that obtains with formula of the present invention enumerated that table 3 has illustrated under different quantization parameter QP respectively with forest, village, bridge and assembly hall.

Above by analysis form learns, through enumerating method to calculate full blast block size and the block size basically identical that utilizes the present invention to calculate, what have only band asterisk in the table 3 is not quite identical, but the two is also very approaching.

Should be noted that and understand, under the situation that does not break away from the desired the spirit and scope of the present invention of accompanying Claim, can make various modifications and improvement the present invention of foregoing detailed description.Therefore, the scope of claimed technical scheme is not subjected to the restriction of given any specific exemplary teachings.

Claims (4)

1. the block size computing method of a cylinder panorama video code comprises the following steps:

Cylinder panorama video does not carry out the direct precoding of piecemeal, the code check R of the panorama video code that obtains ₀

This video carries out the piecemeal precoding according to an optional block size a, obtains the encoder bit rate R under block size a _a

According to formula ω _a=(R _a-R ₀)/R ₀, calculate described code check R _aThan described code check R ₀Increment ω _a

According to formula C=ω _aAW/ (W-a) calculates the factor of influence C of this video, and wherein W represents that the width of panoramic video is a W macro block;

According to formula $a_{\mathrm{me}}=\left[\sqrt{\mathrm{CM}/(1-\frac{C}{W})}\right],$ Calculate the optimum code piece size a of this video _Me, wherein M represents that the target area width is a M macro block.

2. a cylinder panorama video code transmission method comprises the following steps:

Calculate the piece size a that gets according to the method for claim 1 _MeDescribed panoramic video is carried out piecemeal;

After the described piecemeal, the sequence of each piece is independently encoded;

Data are transmitted after selecting required coding.

3. method as claimed in claim 2, it is characterized in that described transfer of data after, decoding end receives this data, and described data are independently decoded and projective transformation.

4. method as claimed in claim 2 is characterized in that the data based user perspective behind the described required coding is selected.

Images