CN103561270A - Coding control method and device for HEVC - Google Patents

Abstract

The invention discloses a coding control method and device for HEVC. The method includes the steps that frames of a current image to be compressed are read, corresponding auxiliary information of the image to be compressed is read at the same time, and a quantization coefficient value of the frames of the current image to be compressed is used as a reference quantization coefficient of a current maximum coding unit to be coded in the frames of the current image to be compressed; according to the fact of whether bit rate control exists or not and the type of the auxiliary information of the image, a quantization parameter modification value of the current maximum coding unit to be coded is acquired based on the reference quantization coefficient and an average value of the corresponding auxiliary information of the image to be compressed of the current maximum coding unit to be coded; according to the fact of whether the quantization parameter modification value exceeds a threshold or not and adjustment of the quantization parameter modification value, a final quantization parameter modification value is acquired, and then coding is carried out. An image higher in quality can be acquired as far as possible under the condition of the same bit rate under a bit rate control mode configuration, and the bit rate can be decreased to a lower value on the premise that image quality is not seriously affected as far as possible under a non-bit rate control mode configuration.

Description

A kind of coding control method for HEVC and device

Technical field

The invention belongs to high-performance video coding (High Efficient Video Coding, HEVC) technical field, particularly relate to a kind of coding control method for HEVC and the device of realizing encoder bit rate and video quality optimization.

Background technology

In recent years, video content is gradually towards high image quality and high-res development, in order to answer this demand, the ISO/IEC of International Standards Organization and ITU-T have worked out video encoding and decoding standard High Efficient Video Coding of new generation (HEVC), than several generations coding techniques H.264/AVC and before, provide higher compression efficiency.HEVC can be enough provides the video of identical definition, bandwidth and the reduction maintenance cost that can save a large amount of memory spaces, transmission of video than few half code check H.264.

HEVC standard code process is roughly as follows: first frame is cut into a plurality of boxed area, and HEVC has proposed based on large scale quadtree coding structure, uses coding unit, predicting unit and converter unit to describe whole cataloged procedure, and variable range is 64x64 to 8x8.First image be take maximum coding unit and is encoded as unit, carries out partition, until become minimum code unit in maximum coding unit inside according to quad-tree structure.For each coding unit, HEVC realizes the forecasting process of this coding unit by predicting unit, predicting unit size-constrained in affiliated coding unit, and it can be that square can be also rectangle.HEVC supports the transcoding, coding transform of 4x4 to 32x32, and the converter unit of take carries out transform and quantization as elementary cell.For improving the code efficiency of large scale coding unit, dct transform adopts the mapped structure of quaternary tree equally.In the frame of HEVC, inter prediction adopts adjacent block reconstructed pixel to carry out infra-frame prediction to current block, from the motion vector of adjacent block, selects motion vectors, supports multiple reference frame prediction etc.Meanwhile, HEVC has adopted as multi-angle prediction, and the multiple technologies such as high-precision motion compensating, improve precision of prediction greatly.The infra-frame prediction of HEVC expands to 33 kinds by prediction direction, has increased the accuracy of infra-frame prediction.The encoder inside of HEVC has increased bit depths of pixels, and maximum can be supported the decoded picture output of 12bit, has improved the precision of information of decoded picture, in bi directional motion compensation process, all will use the precision of 14bit to carry out correlation computations.HEVC adopts context adaptive binary arithmetic coding in addition, for high efficient coding.In frame or interframe predict the outcome and real screen between residual error data after conversion, sampling, quantification, entropy coding, be transferred to decoder together with information of forecasting.Encoder is by communicating predicted pattern information and motion vector (MV), compute motion compensated, and then prediction data between reconstruction frames.

In existing HEVC intraframe coding process, input picture can be cut into a plurality of boxed area before coding.HEVC makes encoder more efficient to high-resolution video coding based on large scale quadtree coding structure.Code tree unit (CTU) can include a coding unit (Coding Unit, CU) or cut into a plurality of less coding units, as shown in Figure 1,32 * 32 CU can further be cut into 16 * 16 CU, 8 * 8 CU ... the coding unit of high efficiency video coding HEVC can utilize easily four minutes tree (Quadtree) mode present, as shown in Figure 2.Within four minutes, tree is a kind of recursive structure, and whether this node of the numeral on four marks continues to do to cut apart, if this node can continue to cut apart, is 1, if do not continue, cutting apart is 0.Numeral on node is 0 o'clock as seen from Figure 2, and this node is coding unit.In other words, in code tree, leaf node (Leaf Node) is coding unit.The order of code tree, as shown in the dotted line of Fig. 1, adopts the mode of Z-scan on coding unit, is depth-first traversal concerning code tree.The size of coding unit is supported 2Nx2N, and the high depth (Depth) of wherein N=4,8,16 or 32, so the Si Fenshu of HEVC is 4.

HEVC allows the quantization parameter different to the employing of each CU.By giving the absolute value of quantization parameter corresponding to parameter c u_qp_delta_abs definition CU monochrome information with difference with reference to CU, and define its corresponding sign bit by parameter c u_qp_delta_sign.If cu_qp_delta_sign=0, the difference CU that it is corresponding and quantization parameter with reference to CU be on the occasion of, otherwise be negative value, if simultaneously cu_qp_delta_sign to be defined explanation difference its corresponding CU and quantization parameter with reference to CU be 0.But HEVC standard does not define how to determine cu_qp_delta_abs and cu_qp_delta_sign.What the art lacked that self adaptation adjusts that quantization parameter improves compression ratio can implementation.

Summary of the invention

The present invention overcomes the deficiencies in the prior art, provides a kind of according to image auxiliary information, and self adaptation is adjusted quantization parameter, improves the compression coding technology scheme based on HEVC framework of compression ratio.

Technical scheme provided by the invention comprises a kind of coding control method for HEVC, comprises the following steps:

Step 1 is read in successively a picture frame from video, and the picture frame current to be compressed reading in is read in to corresponding image auxiliary information to be compressed simultaneously;

Step 2 is extracted successively a maximum coding unit to be encoded from current picture frame to be compressed, to current maximum coding unit to be encoded, usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀, the coordinate information of current maximum coding unit to be encoded is designated as (i, j);

Step 3, adds up the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed, adds up mean value y(i, the j of the current corresponding image auxiliary information to be compressed of maximum coding unit to be encoded according to coordinate information (i, j));

Step 4, detects and has or not Rate Control;

Step 5, according to the monitoring result of step 4 and image auxiliary information type, adopts corresponding pattern function to draw the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded;

Step 6, whether the quantization parameter correction value QP (i, j) of the current maximum coding unit to be encoded of determining step 5 gained crosses the border and the quantization parameter correction value QP (i, j) crossing the border is adjusted, and obtains final quantization parameter correction value QP _z(i, j);

Step 7, according to the final quantization parameter correction value of step 6 gained QP _z(i, j) encodes to current maximum coding unit to be encoded;

Step 8, return to step 2, from current picture frame to be compressed, extract maximum coding unit and coordinate information (i next to be encoded, j), the maximum coding unit that the next one is to be encoded is processed as new maximum coding unit current to be encoded, until current picture frame end-of-encode to be compressed enters step 9;

Step 9, return to step 1, from video, read in successively next picture frame as new picture frame current to be compressed and corresponding image auxiliary information to be compressed, new picture frame current to be compressed is processed, until all images frame end-of-encode in video, process ends.

And image auxiliary information type is 1 or 2, be that the larger expression of 1 o'clock image auxiliary information value to be compressed is more important, be the less expression of 2 o'clock image auxiliary information values to be compressed is more important; In step 5, described pattern function is as follows,

While having Rate Control,

Image auxiliary information type is 1 o'clock, and the function A that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0-m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0-m+\frac{m+n}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function B that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m+\frac{m+n}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0-n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

During without Rate Control,

Image auxiliary information type is 1 o'clock, and the function C that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+\frac{m}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function D that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m+\frac{m}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Wherein, m, n is preset parameter value, α, β is multiplied by a corresponding default coefficient value by the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed and obtains.

And, while having Rate Control, put m=2, n=2, α=2x * 0.5, β=2x * 0.8, puts m=4 during without Rate Control, α=2 * x * 0.5, β=2x * 0.8.

And maximum coding unit is of a size of 64 * 64.

The present invention is also corresponding provides a kind of coding-control device for HEVC, comprises with lower unit:

Picture frame reading unit for read in successively a picture frame from video, reads in corresponding image auxiliary information to be compressed to the picture frame current to be compressed reading in simultaneously;

Maximum coding unit reading unit, for extract successively a maximum coding unit to be encoded from current picture frame to be compressed, to current maximum coding unit to be encoded, usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀, the coordinate information of current maximum coding unit to be encoded is designated as (i, j);

Supplementary mean value statistic unit, for adding up the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed, according to coordinate information (i, j), add up mean value y(i, the j of the current corresponding image auxiliary information to be compressed of maximum coding unit to be encoded);

Rate Control monitoring means, for detection of having or not Rate Control;

Quantization parameter amending unit, for according to the monitoring result of Rate Control monitoring means and image auxiliary information type, adopts corresponding pattern function to draw quantization parameter correction value QP(i, the j of current maximum coding unit to be encoded);

Quantization parameter correction value modulating unit, for judging the quantization parameter correction value QP(i of the current maximum coding unit to be encoded of quantization parameter amending unit gained, j) whether cross the border and to the quantization parameter correction value QP(i, the j that cross the border) adjust, obtain final quantization parameter correction value QP _z(i, j);

Correction value coding unit, for according to the final quantization parameter correction value of quantization parameter correction value modulating unit gained QP _z(i, j) encodes to current maximum coding unit to be encoded;

Picture frame end-of-encode judging unit, for controlling maximum coding unit reading unit, from current picture frame to be compressed, extract maximum coding unit and coordinate information (i next to be encoded, j), the maximum coding unit that the next one is to be encoded is processed as new maximum coding unit current to be encoded, until current picture frame end-of-encode to be compressed;

Video coding finishes judging unit, for control chart picture frame reading unit, from video, read in successively next picture frame as new picture frame current to be compressed and corresponding image auxiliary information to be compressed, new picture frame current to be compressed is processed, until all images frame end-of-encode in video.

And image auxiliary information type is 1 or 2, be that the larger expression of 1 o'clock image auxiliary information value to be compressed is more important, be the less expression of 2 o'clock image auxiliary information values to be compressed is more important; In quantization parameter amending unit, described pattern function is as follows,

While having Rate Control,

Image auxiliary information type is 1 o'clock, and the function A that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0-m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0-m+\frac{m+n}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function B that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m+\frac{m+n}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0-n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

During without Rate Control,

Image auxiliary information type is 1 o'clock, and the function C that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+\frac{m}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function D that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m+\frac{m}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Wherein, m, n is preset parameter value, α, β is multiplied by a corresponding default coefficient value by the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed and obtains.

And, while having Rate Control, put m=2, n=2, α=2x * 0.5, β=2x * 0.8, puts m=4 during without Rate Control, α=2 * x * 0.5, β=2x * 0.8.

And maximum coding unit is of a size of 64 * 64.

According to the image auxiliary information to be compressed of material information that the CU of compressed image is provided, the quantization parameter that the present invention has proposed image C U to be compressed according to this information is adjusted scheme, and the impact based on each CU on video quality determines the quantization parameter that CU adopts.For the unessential CU of video quality, the present invention adopts larger quantization parameter, reduces the code check of its corresponding code stream that produces.For the important CU of video quality, the present invention adopts less quantization parameter, to guarantee the quality after its compression.Without Rate Control in the situation that, CU quantization parameter in relative anchor-frame, i.e. in the situation of the identical quantization parameter of the unified use of all CU, the present invention's code check that effectively minimizing produces.In the situation that having Rate Control, by the CU quantization parameter of the video quality based on CU, control, in same frame code check situation, the present invention can effectively improve video quality.More reasonable grade of image quantization through image auxiliary information optimization, can further improve compression efficiency.

Accompanying drawing explanation

Fig. 1 is the coding unit cutting schematic diagram of HEVC in prior art.

Fig. 2 is the four minutes tree schematic diagrames of coding unit of HEVC in prior art.

Fig. 3 is image 64x64CU to be compressed and the corresponding reference quantization coefficient QP of the embodiment of the present invention ₀schematic diagram.

Fig. 4 is the image auxiliary information schematic diagram to be compressed of the embodiment of the present invention.

Fig. 5 is image 64x64CU to be compressed and the final quantization parameter correction value QP of the embodiment of the present invention _zthe schematic diagram of (i, j).

Embodiment

Below in conjunction with drawings and Examples explanation technical solution of the present invention.

The CU piece that the quantization parameter that the present invention utilizes image auxiliary information to help to revise current C U obtains each 64x64 more suitably quantizes coefficient QP value.Consider in a two field picture, it not the image that everywhere all needs identical fine degree, the present invention adopts a kind of image mask to match as supplementary sequence and video to be compressed, the quantization parameter at every place all image auxiliary information is from here revised, and obtains final quantization parameter correction value.And detect and have or not Rate Control, under Rate Control pattern configurations, object is under same code rate, to obtain as far as possible higher-quality image, and under non-Rate Control pattern configurations, object is not lose under too much picture quality prerequisite and reducing more multi code Rate of Chinese character as far as possible.Finally, judge that quantization parameter correction value whether within the quantization parameter correction value thresholding of HEVC, obtains final quantization parameter correction value.

The embodiment of the present invention realizes on HM11.0 experiment porch, and those skilled in the art can adopt software engineering to realize automatic operational process, and described flow process comprises that step is as follows:

Step 1 is read in successively a picture frame from video, to the picture frame current to be compressed reading in, reads in corresponding image auxiliary information to be compressed.The present invention improves in the framework of the existing cataloged procedure of HEVC, reads in image auxiliary information to be compressed when reading in image to be compressed simultaneously.During concrete enforcement, can in chronological sequence extract picture frame.

In the picture frame of embodiment, each point has corresponding image auxiliary information to be compressed, represent the importance of this point, in current picture frame to be compressed, the somewhat corresponding image auxiliary information to be compressed of institute forms the corresponding image auxiliary information to be compressed of current picture frame to be compressed, and the corresponding image auxiliary information to be compressed of current picture frame to be compressed can be considered an image auxiliary information frame to be compressed.During concrete enforcement, the image auxiliary information to be compressed of each point can adopt 8 bit data to represent, value is 0,1 ... between 255.Can define the larger expression of value more important, or the less expression of value is more important, these two kinds of modes are all feasible.These two kinds of modes of definable image auxiliary information type identification, when image auxiliary information type is 1, image auxiliary information to be compressed in picture frame represents compared with vital point with smaller value, when image auxiliary information type 2, image auxiliary information to be compressed in picture frame represents compared with vital point with higher value, according to image auxiliary information to be compressed can the current picture frame to be compressed of significant reaction in which CU more important.

Step 2 is extracted successively a maximum coding unit to be encoded from current picture frame to be compressed, to current maximum coding unit to be encoded, usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀.During concrete enforcement, generally from picture frame, extract from top to bottom, left to right maximum coding unit to be encoded.

In embodiment, the CU piece that maximum coding unit to be encoded is 64x64 is got one by one the CU piece of 64x64 and is processed equally from current maximum coding unit to be encoded.CU piece to the 64x64 of current taking-up, starts to carry out the calculating of coding unit quantization parameter from this step.Embodiment reads the coordinate information (i, j) of CU piece of the 64x64 of current taking-up, and usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀.That is to say, using the reference quantization coefficient QP of the CU piece of the quantization parameter value of current picture frame to be compressed all 64x64 in frame ₀.

Step 3, add up the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed and the mean value y(i of the current corresponding image auxiliary information to be compressed of maximum coding unit to be encoded, j), wherein (i, j) is the coordinate of relevant block in current maximum coding unit to be encoded and image auxiliary information frame to be compressed.

Embodiment, according to the coordinate (i, j) of the CU piece of 64x64 to be encoded, adds up the mean value y (i, j) of the image auxiliary information frame corresponding coordinate 64x64 of place piece to be compressed, and the mean value x of image auxiliary information frame to be compressed.During concrete enforcement, first the maximum coding unit to be encoded extracting in current picture frame to be compressed is carried out to the mean value x that step 3 obtains, during other maximum coding units of the current picture frame to be compressed of subsequent treatment, can directly use this value, without repeating statistics.

As shown in Figure 3, it is the CU of 64x64 that current picture frame to be compressed has 13x7 block size, and the piece that black box marks is the current CU encoding, the reference quantization coefficient QP of the CU piece of all 64x64 in frame ₀be 32.Shown in Fig. 4, corresponding image auxiliary information frame to be compressed has an image auxiliary information matching of formed objects, the mean value x that adds up image auxiliary information frame to be compressed with and the mean value y(i of the corresponding 64x64 piece of the CU piece of corresponding each 64x64 of image to be compressed, j), wherein (i, j) is the coordinate of 64x64 size block in frame.For example, in Fig. 4, mean value y(i, the j of each 64x64 piece of image auxiliary information frame to be compressed) there are values such as 0,123,255.

Step 4, detects and has or not Rate Control.

The control function of opening Rate Control is provided in the framework of the existing cataloged procedure of HEVC, and embodiment reads this control function value can learn that cataloged procedure is is also under non-Rate Control pattern configurations in Rate Control pattern configurations.

Step 5, reads the assistant images type of use, draws the quantization parameter correction value of current maximum coding unit to be encoded according to corresponding pattern function.

In Rate Control pattern configurations, if image auxiliary information type is 1, smaller value represents more important CU, and to value, the less CU that is importance is higher adopts less quantization parameter, and the concrete grammar that quantization parameter is controlled is: make quantization parameter correction value

and by preset value, control the thresholding of quantization parameter correction value QP (i, j).If image auxiliary information type is 2, higher value represents more important CU, to value, is more greatly that the CU that importance is higher adopts less quantization parameter, and the concrete grammar that quantization parameter is controlled is: make quantization parameter correction value (y (i, j)-α), and by preset value, control the thresholding of quantization parameter correction value QP (i, j).

In non-Rate Control pattern configurations, if image auxiliary information type is 1, smaller value represents more important CU, and to value, the less CU that is importance is higher adopts less quantization parameter, and the concrete grammar that quantization parameter is controlled is: make quantization parameter correction value

and by preset value, control the thresholding of quantization parameter correction value QP (i, j).If image auxiliary information type is 2, higher value represents more important CU, to value, is more greatly that the CU that importance is higher adopts less quantization parameter, and the concrete grammar that quantization parameter is controlled is: make quantization parameter correction value

α), and by preset value, control the thresholding of quantization parameter correction value QP (i, j).

Be summarized as follows:

While having Rate Control,

Image auxiliary information type is 1 o'clock, and the function A that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0-m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0-m+\frac{m+n}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function B that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m+\frac{m+n}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0-n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

During without Rate Control,

Image auxiliary information type is 1 o'clock, and the function C that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+\frac{m}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function D that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+m+\frac{m}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Wherein, m, n is preset parameter value, controls the thresholding of quantization parameter correction value QP (i, j), guarantees that quantization parameter correction value QP (i, j) meets required effect.Parameter alpha, β is multiplied by a corresponding default coefficient value by the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed and obtains.Can by those skilled in the art, set in advance parameter value and coefficient value as the case may be, for example whether Rate Control opens.

Embodiment judges whether Rate Control opens, and in this way, enters operation a, as no, enters operation b.

Operation a, the video compression coding based under Rate Control: embodiment arranges m=2, n=2, α=2x * 0.5, β=2x * 0.8.Class1 if, obtains the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0-2,& y(i,j)\&le;x\\ {\mathrm{QP}}_0-2+\frac{4}{2x\&times;0.3}\&times;(y(i,j)-\mathrm{x\&alpha;}),& x<y(i,j)\&le;2x\&times;0.8\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+2,& y(i,j)>2x\&times;0.8\end{array}\right.$

Type 2 if, obtain the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+2,& y(i,j)\&le;\mathrm{<figure-callout\; id="0"\; label="x\; QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">x</figure-callout>}& \\ {\mathrm{QP}}_{}{}_0+2-\frac{4}{2x\&times;0.3}\&times;(y(i,j)-\mathrm{x\&alpha;}),& x<y(i,j)\&le;2x\&times;0.8\\ {\mathrm{QP}}_0-2,& y(i,j)>2x\&times;0.8\end{array}\right.$

Operation b, the video compression coding based under non-Rate Control: embodiment m=4 here, α=2 * x * 0.5, β=2x * 0.8.Class1 if, obtains the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)\&le;\mathrm{<figure-callout\; id="0"\; label="x\; QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">x</figure-callout>}& \\ {\mathrm{QP}}_{}{}_0+\frac{4}{2x\&times;0.3}\&times;(y(i,j)-x),& x<y(i,j)\&le;2x\&times;0.8\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+4& y(i,j)>2x\&times;0.8\end{array}\right.$

Type 2 if, obtain the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0+4,& y(i,j)\&le;\mathrm{<figure-callout\; id="0"\; label="x\; QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">x</figure-callout>}& \\ {\mathrm{QP}}_{}{}_0+4-\frac{4}{2x\&times;0.3}\&times;(y(i,j)-x),& x<y(i,j)\&le;2x\&times;0.8\\ {\mathrm{<figure-callout\; id="0"\; label="QP"\; filenames="1311081628082.png"\; state="\{\{state\}\}">QP</figure-callout>}}_0,& y(i,j)>2x\&times;0.8\end{array}\right.$

Step 6, whether the quantization parameter correction value QP (i, j) of the current maximum coding unit to be encoded of determining step 5 gained crosses the border and the quantization parameter correction value QP (i, j) crossing the border is adjusted, and obtains final quantization parameter correction value QP _z(i, j).

Whether, during concrete enforcement, can utilize HEVC available frame judgement quantization parameter correction value QP (i, j) to cross the border, boundary is [0-51].When quantization parameter correction value QP (i, j) value is 0,1 ..., when arbitrary in 50,51, do not cross the border, when quantization parameter correction value QP (i, j) value is less than 0 or cross the border while being greater than 51.When quantization parameter correction value QP (i, j) value is less than 0, is adjusted into and equals 0, when quantization parameter correction value QP (i, j) value is greater than 51, is adjusted into and equals 51, final quantization parameter correction value QP after making to adjust _z(i, j) meets within the scope of 0-51.

Step 7, according to the final quantization parameter correction value of step 6 gained QP _z(i, j) encodes to current maximum coding unit to be encoded.

During concrete enforcement, according to the framework of the existing cataloged procedure of HEVC, by the final quantization parameter correction value QP of the CU of current 64x64 _z(i, j) brings HEVC framework into and encodes, for example, according to QP _z(i, j)=33 obtain cu_qp_delta_abs=33-32=1 and cu_qp_delta_sign for just, and specific implementation is prior art.

Step 8, return to step 2, from current picture frame to be compressed, extract maximum coding unit and coordinate information (i next to be encoded, j), using the next one maximum coding unit to be encoded as new maximum coding unit current to be encoded, 2 to 7 pairs of new maximum coding units current to be encoded of repeating step are processed like this, until current picture frame end-of-encode to be compressed enters step 9.

Step 9, return to step 1, from video, read in successively next picture frame as new picture frame current to be compressed, and read in corresponding image auxiliary information to be compressed, 1 to 8 pair of new picture frame current to be compressed of repeating step is processed, until all images frame end-of-encode in video, process ends.

The flow scheme design of embodiment is, after step 7 completes, to judge the coding unit of whether not encoded in addition in present frame, in this way, returning to step 2 extracts the CU of next 64x64 and upgrades coordinate information (i, j), as no, present frame has been encoded, and further judges whether to arrive video tail, as arrived, all end-of-encodes, as do not arrived, return to step 1, read in next frame image to be compressed and read in corresponding next frame image auxiliary information to be compressed simultaneously.Last Fig. 5 has been for to have compressed the final quantization parameter correction value of rear image distribution example according to Fig. 4 image auxiliary information, and the final quantization parameter correction value of the CU of each 64x64 has the values such as 32,33,34.

During concrete enforcement, can also adopt that software modularity technology is corresponding provides a kind of coding-control device for HEVC, comprise with lower unit:

Picture frame reading unit for read in successively a picture frame from video, reads in corresponding image auxiliary information to be compressed to the picture frame current to be compressed reading in simultaneously;

Maximum coding unit reading unit, for extract successively a maximum coding unit to be encoded from current picture frame to be compressed, to current maximum coding unit to be encoded, usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀, the coordinate information of current maximum coding unit to be encoded is designated as (i, j);

Supplementary mean value statistic unit, for adding up the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed, according to coordinate information (i, j), add up the mean value of the current corresponding image auxiliary information to be compressed of maximum coding unit to be encoded _y(i, j);

Rate Control monitoring means, for detection of having or not Rate Control;

Quantization parameter amending unit, for according to the monitoring result of Rate Control monitoring means and image auxiliary information type, draws the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded according to corresponding pattern function;

Quantization parameter correction value modulating unit, for judging the quantization parameter correction value QP (i of the current maximum coding unit to be encoded of quantization parameter amending unit gained, j) whether cross the border and to the quantization parameter correction value QP (i crossing the border, j) adjust, obtain final quantization parameter correction value QP _z(i, j);

Correction value coding unit, for according to the final quantization parameter correction value of quantization parameter correction value modulating unit gained QP _z(i, j) encodes to current maximum coding unit to be encoded;

Picture frame end-of-encode judging unit, for controlling maximum coding unit reading unit, from current picture frame to be compressed, extract maximum coding unit and coordinate information (i next to be encoded, j), the maximum coding unit that the next one is to be encoded is processed as new maximum coding unit current to be encoded, until current picture frame end-of-encode to be compressed;

Video coding finishes judging unit, for control chart picture frame reading unit, from video, read in successively next picture frame as new picture frame current to be compressed and corresponding image auxiliary information to be compressed, new picture frame current to be compressed is processed, until all images frame end-of-encode in video.

Each unit specific implementation is corresponding with process step, and it will not go into details.

Below be only that specific embodiment of the invention case is described, not in order to limit of the present invention can practical range.All equivalent deformations, replacement or modification that those of ordinary skill in the art complete under the spirit indicated without prejudice to the present invention and principle, be still included in the scope of the claims in the present invention.

Claims (8)

1. for a coding control method of HEVC, it is characterized in that, comprise the following steps:

Step 1 is read in successively a picture frame from video, and the picture frame current to be compressed reading in is read in to corresponding image auxiliary information to be compressed simultaneously;

Step 2 is extracted successively a maximum coding unit to be encoded from current picture frame to be compressed, to current maximum coding unit to be encoded, usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀, the coordinate information of current maximum coding unit to be encoded is designated as (i, j);

Step 3, adds up the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed, adds up mean value y(i, the j of the current corresponding image auxiliary information to be compressed of maximum coding unit to be encoded according to coordinate information (i, j));

Step 4, detects and has or not Rate Control;

Step 5, according to the monitoring result of step 4 and image auxiliary information type, adopts corresponding pattern function to draw the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded;

Step 6, whether the quantization parameter correction value QP (i, j) of the current maximum coding unit to be encoded of determining step 5 gained crosses the border and the quantization parameter correction value QP (i, j) crossing the border is adjusted, and obtains final quantization parameter correction value QP _z(i, j);

Step 7, according to the final quantization parameter correction value of step 6 gained QP _z(i, j) encodes to current maximum coding unit to be encoded;

Step 8, return to step 2, from current picture frame to be compressed, extract maximum coding unit and coordinate information (i next to be encoded, j), the maximum coding unit that the next one is to be encoded is processed as new maximum coding unit current to be encoded, until current picture frame end-of-encode to be compressed enters step 9;

Step 9, return to step 1, from video, read in successively next picture frame as new picture frame current to be compressed and corresponding image auxiliary information to be compressed, new picture frame current to be compressed is processed, until all images frame end-of-encode in video, process ends.

2. according to claim 1 for the coding control method of HEVC, it is characterized in that: image auxiliary information type is 1 or 2, being that the larger expression of 1 o'clock image auxiliary information value to be compressed is more important, is that the less expression of 2 o'clock image auxiliary information values to be compressed is more important; In step 5, described pattern function is as follows,

While having Rate Control,

Image auxiliary information type is 1 o'clock, and the function A that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0-m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0-m+\frac{m+n}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0+n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function B that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0+m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0+m+\frac{m+n}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0-n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

During without Rate Control,

Image auxiliary information type is 1 o'clock, and the function C that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0+\frac{m}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0+m,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function D that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0+m& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0+m+\frac{m}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Wherein, m, n is preset parameter value, α, β is multiplied by a corresponding default coefficient value by the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed and obtains.

3. according to claim 2 for the coding control method of HEVC, it is characterized in that: while having Rate Control, put m=2, n=2, α=2x * 0.5, β=2x * 0.8, puts m=4 during without Rate Control, α=2 * x * 0.5, β=2x * 0.8.

According to described in claim 1 or 2 or 3 for the coding control method of HEVC, it is characterized in that: maximum coding unit is of a size of 64 * 64.

5. for a coding-control device of HEVC, it is characterized in that, comprise with lower unit:

Picture frame reading unit for read in successively a picture frame from video, reads in corresponding image auxiliary information to be compressed to the picture frame current to be compressed reading in simultaneously;

Maximum coding unit reading unit, for extract successively a maximum coding unit to be encoded from current picture frame to be compressed, to current maximum coding unit to be encoded, usings the quantization parameter value of current picture frame to be compressed as with reference to quantization parameter QP ₀, the coordinate information of current maximum coding unit to be encoded is designated as (i, j);

Supplementary mean value statistic unit, for adding up the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed, according to coordinate information (i, j), add up mean value y(i, the j of the current corresponding image auxiliary information to be compressed of maximum coding unit to be encoded);

Rate Control monitoring means, for detection of having or not Rate Control;

Quantization parameter amending unit, for according to the monitoring result of Rate Control monitoring means and image auxiliary information type, adopts corresponding pattern function to draw the quantization parameter correction value QP (i, j) of current maximum coding unit to be encoded;

Quantization parameter correction value modulating unit, for judging the quantization parameter correction value QP (i of the current maximum coding unit to be encoded of quantization parameter amending unit gained, j) whether cross the border and to the quantization parameter correction value QP (i crossing the border, j) adjust, obtain final quantization parameter correction value QP _z(i, j);

Correction value coding unit, for according to the final quantization parameter correction value of quantization parameter correction value modulating unit gained QP _z(i, j) encodes to current maximum coding unit to be encoded;

Picture frame end-of-encode judging unit, for controlling maximum coding unit reading unit, from current picture frame to be compressed, extract maximum coding unit and coordinate information (i next to be encoded, j), the maximum coding unit that the next one is to be encoded is processed as new maximum coding unit current to be encoded, until current picture frame end-of-encode to be compressed;

Video coding finishes judging unit, for control chart picture frame reading unit, from video, read in successively next picture frame as new picture frame current to be compressed and corresponding image auxiliary information to be compressed, new picture frame current to be compressed is processed, until all images frame end-of-encode in video.

6. according to claim 5 for the coding-control device of HEVC, it is characterized in that: image auxiliary information type is 1 or 2, being that the larger expression of 1 o'clock image auxiliary information value to be compressed is more important, is that the less expression of 2 o'clock image auxiliary information values to be compressed is more important; In quantization parameter amending unit, described pattern function is as follows,

While having Rate Control,

Image auxiliary information type is 1 o'clock, and the function A that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0-m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0-m+\frac{m+n}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0+n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function B that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0+m,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0+m+\frac{m+n}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0-n,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

During without Rate Control,

Image auxiliary information type is 1 o'clock, and the function C that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0,& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0+\frac{m}{\mathrm{\&beta;}-\mathrm{\&alpha;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0+m,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Image auxiliary information type is 2 o'clock, and the function D that uses a model is as follows,

$\mathrm{QP}(i,j)=\left\{\begin{array}{cc}{\mathrm{QP}}_0+m& y(i,j)\&le;\mathrm{\&alpha;}\\ {\mathrm{QP}}_0+m+\frac{m}{\mathrm{\&alpha;}-\mathrm{\&beta;}}\&times;(y(i,j)-\mathrm{\&alpha;}),& \mathrm{\&alpha;}<y(i,j)\&le;\mathrm{\&beta;}\\ {\mathrm{QP}}_0,& y(i,j)>\mathrm{\&beta;}\end{array}\right.$

Wherein, m, n is preset parameter value, α, β is multiplied by a corresponding default coefficient value by the mean value x of the corresponding image auxiliary information to be compressed of current picture frame to be compressed and obtains.

7. according to claim 6 for the coding-control device of HEVC, it is characterized in that: while having Rate Control, put m=2, n=2, α=2x * 0.5, β=2x * 0.8, puts m=4 during without Rate Control, α=2 * x * 0.5, β=2x * 0.8.

According to described in claim 5 or 6 or 7 for the coding-control device of HEVC, it is characterized in that: maximum coding unit is of a size of 64 * 64.

Images