CN104113761A - Code rate control method for video encoding and encoder - Google Patents

Abstract

The invention relates to the field of data and video processing, and discloses a code rate control method and an encoder in video coding, which are used to improve the accuracy of the code rate control method and reduce complexity. The method is as follows: first determine the current video image The first gradient value and the second gradient value of the previous video image, and then calculate the target bit rate of the current video image according to the preset average bit rate, then determine the specified performance parameters of at least one encoded video image, and then, in the first When the gradient value and/or the second gradient value are not 0, determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and performance parameters; The QP of the current video image is determined according to the rate and performance parameters. Finally, the current video image is encoded according to the QP. In this way, the accuracy of the code rate control method is improved, and the complexity is also reduced.

Description

A code rate control method and encoder in video encoding

technical field

The invention relates to the field of data and video processing, in particular to a code rate control method and an encoder in video encoding.

Background technique

In order to save transmission bandwidth and storage space, and support applications such as digital TV broadcasting, remote monitoring, digital video on demand, and wireless multimedia communication, video coding has become one of the hot spots in research and industrial applications at home and abroad. The coding technologies involved in the current video coding standards mainly include: intra prediction, inter prediction, integer transformation, reconstructed image, bit rate control, quantization and entropy coding, inverse quantization and inverse transformation, deblocking filtering, etc., as shown in Figure 1 shown.

Among them, bit rate control is an indispensable technology in video coding. Any standard without bit rate control will have its application limited. For example, in the transmission process under bandwidth constraints, if there is no suitable bit rate control method, the client The buffer may overflow, that is, the purpose of rate control is to ensure that the optimal image quality can be achieved under the condition of generating the specified target bit rate, and to ensure that the codec buffer will not overflow. The main output of rate control is QP (Quantization Parameter, quantization parameter), which is used in the encoding process to obtain a good quantization step size for image quantization.

At present, single I-frame coding technology is widely used due to its following advantages. For example, single I-frame coding technology only utilizes the correlation within the frame, and avoids the accumulation of GOP-based coding while ensuring compression efficiency. error, effectively improving the quality of video coding; another example, each frame in a single I-frame coding technology is coded and decoded separately, this feature makes it easy for the encoder to adopt acceleration technologies such as parallel codec and distributed codec , effectively improving the encoding rate.

The existing H.264/AVC standard cannot control the output code rate in a single I frame encoding mode well due to the following problems:

1. In the H.264/AVC standard, a macroblock-level code rate control algorithm is adopted. The method is as follows: firstly, linear prediction is used to obtain the MAD (Mean Absolute Difference) of the current macroblock, and then two Power model, calculate the corresponding QP. This method is suitable for inter-frame coding process, not for intra-frame coding. For a single I-frame mode, since QP is also related to intra-frame complexity, a new statistic is needed to measure the complexity of the image frame degree, thus increasing the computational load to the encoder and increasing the computational complexity;

Two, in the existing H.264/AVC standard, the code rate control is carried out at the macroblock level, and the MAD of the macroblock at the corresponding position of the current video image is predicted by the MAD of the macroblock at the corresponding position of the previous video image. This method is suitable for video images with less motion and small scene changes. For motion areas or when the lens moves or zooms, there may be larger Such prediction will inevitably lead to huge errors, which will lead to unreasonable bit rate allocation and deteriorate video quality. Since the single I frame mode deals with video images with large motion or frequent camera movement, it will cause obvious errors, reduce the accuracy of the determined QP, and also reduce the accuracy of bit rate control.

In the prior art, in order to improve the accuracy of bit rate control, the transformed and quantized DCT (DiscreteCosine Transformation, discrete cosine transform) coefficients are analyzed to establish the relationship between the DCT coefficients and the output bit rate, and then determined by the DCT coefficients Output code rate, but this method not only has higher complexity, but also takes a long time, so there is no bit rate control with higher accuracy and lower complexity for single I-frame video coding at present. method.

Contents of the invention

An embodiment of the present invention provides a code rate control method in video coding, which is used to solve the problem of poor accuracy and high complexity of the code rate control method for single I-frame video coding in the prior art.

A bit rate control method in video coding, used to carry out bit rate control to the video image of single I frame, comprising:

determining the first gradient value of the current video image and the second gradient value of the previous video image;

Calculate the target bit rate of the current video image according to the preset average bit rate;

determining a performance parameter of the specified at least one encoded video image;

Judging whether the first gradient value and/or the second gradient value is 0, if so, determine the quantization parameter QP of the current video image according to the target code rate and the performance parameter, otherwise, determine the quantization parameter QP according to the target bit rate The code rate, the first gradient value, the second gradient value and the performance parameter determine the QP of the current video image;

Encode the current video image according to the QP.

A kind of encoder, is used for carrying out code rate control to the video image of single I frame, comprises:

The first determination unit is used to determine the first gradient value of the current video image and the second gradient value of the previous video image;

A calculation unit, configured to calculate the target bit rate of the current video image according to a preset average bit rate;

A second determining unit, configured to determine a specified performance parameter of at least one coded video image;

A third determination unit, configured to determine whether the first gradient value and/or the second gradient value is 0, and if so, determine the quantization parameter QP of the current video image according to the target code rate and the performance parameter , otherwise, determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and the performance parameter;

A coding unit, configured to code the current video image according to the QP.

In the embodiment of the present invention, the encoder first determines the first gradient value of the current video image and the second gradient value of the previous video image when performing bit rate control on each video image during the encoding process, and then according to the preset Calculate the target bit rate of the current video image at the average bit rate, then determine the specified performance parameter of at least one encoded video image, and finally, when the first gradient value and/or the second gradient value are 0, according to The target code rate and performance parameters determine the QP of the current video image, and if the first gradient value and/or the second gradient value are not 0, it is determined according to the target code rate, the first gradient value, the second gradient value and the performance parameters The QP of the current video image. Finally, the current video image is encoded according to the QP. Since the gradient value is one of the inherent attributes of the video image, it is used to quantify the complexity of the video image, and the QP measures the complexity of the video image. By The accuracy of the QP calculated by the gradient value is relatively high, but in the present invention, the specific method used to calculate the QP is determined by the first gradient value and the second gradient value, and the first gradient value and the second gradient value are not equal to In the case of 0, the QP of the current video image is determined according to the target code rate, the first gradient value, the second gradient value and performance parameters. Therefore, the scheme provided by the embodiments of the present invention is used to calculate the QP, which can improve the calculated QP accuracy, and then improve the accuracy of the bit rate control method for single I-frame video coding, therefore, solve the problem that the accuracy of the bit rate control method in the single I-frame video coding is poor at present, at the same time, due to In the process of determining the QP, there is no need to set a new statistic to measure the complexity of the image frame, so the complexity of the rate control method for single I-frame video coding is reduced.

Description of drawings

FIG. 1 is a schematic diagram of the functional structure of an encoder in the prior art;

Fig. 2 is the first detailed embodiment of code rate control in the embodiment of the present invention;

Fig. 3 is the second detailed embodiment of code rate control in the embodiment of the present invention;

Fig. 4 is the third detailed embodiment of code rate control in the embodiment of the present invention;

FIG. 5 is a schematic diagram of a logic function structure of an encoder in an embodiment of the present invention;

FIG. 6A is a schematic diagram of a logic function structure of a computing unit in an embodiment of the present invention;

Fig. 6B is a schematic diagram of the logic function structure of the second determination unit in the embodiment of the present invention.

Detailed ways

In order to solve the problem of poor accuracy and high complexity of the bit rate control method in the single I-frame video encoding in the prior art, in the embodiment of the present invention, the encoder performs an encoding process on each video image during the encoding process. During code rate control, first determine the first gradient value of the current video image and the second gradient value of the previous video image, then calculate the target code rate of the current video image according to the preset average code rate, and then determine the specified At least one performance parameter of the coded video image, finally, in the case of the first gradient value and/or the second gradient value of 0, determine the QP of the current video image according to the target bit rate and the performance parameter, in the first gradient value and /or when the second gradient value is not 0, determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and performance parameters, and finally, encode the current video image according to the QP, due to the gradient Value is one of the intrinsic properties of video images, which is used to quantify the complexity of video images, and what QP measures is also the complexity of video images. The accuracy of QP calculated by the gradient value is higher, and in the present invention, through the first A gradient value and a second gradient value are used to determine the specific method used to calculate QP, and when the first gradient value and the second gradient value are not 0, according to the target code rate, the first gradient value, and the second gradient value and performance parameters to determine the QP of the current video image, therefore, adopting the scheme provided by the embodiment of the present invention to calculate QP can improve the accuracy of the calculated QP, and then improve the efficiency of the bit rate control method for single I-frame video encoding Accuracy, therefore, solves the problem that the bit rate control method in the single I-frame video coding is poor in accuracy at present, at the same time, because there is no need to set a new statistic to measure the image frame in the process of determining the QP Therefore, the complexity of the rate control method for single I-frame video coding is reduced.

Preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

In the embodiment of the present invention, the specific implementation process is as follows:

Referring to Fig. 2, in an embodiment of the present invention, the detailed flow of the code rate control method in video coding is as follows:

Step 200: Determine the first gradient value of the current video image and the second gradient value of the previous video image.

In the embodiment of the present invention, there are multiple ways to determine the first gradient value of the current video image. Preferably, a gradient function is used to determine the first gradient value of the current video image, wherein the gradient function includes at least a pixel parameter and a gradient parameter, As shown in Formula 1, and the gradient parameter is 0 or 1.

$\mathrm{Gradient}=\underset{i=0}{\overset{m}{\mathrm{\Σ}}}\underset{j=0}{\overset{N}{\mathrm{\Σ}}}\left(\right|I_{x,\mathrm{the\; y}}-I_{x+1,\mathrm{the\; y}}|+|I_{x,\mathrm{the\; y}}{-I}_{x,\mathrm{the\; y}+1}\left|\right)\&Center\; Dot;\mathrm{Sgn}(x,\mathrm{the\; y})$ (Formula 1)

Wherein, Gradient is gradient value; M is the length of current video image; N is the width of current video image; I _{x, y} is the pixel parameter of (x, y) position, and x is the abscissa of pixel point, and y is pixel point The ordinate; Sgn(x,y) is the gradient parameter.

The value of Sgn(x, y) in Formula 1 is 0 or 1, and it is determined according to the preset threshold value when taking the value. The specific determination method is shown in Formula 2:

$\mathrm{Sgn}(x,\mathrm{the\; y})=\left\{\begin{array}{cc}1& \left(\right|I_{x,\mathrm{the\; y}}-I_{x+1,\mathrm{the\; y}}|+|I_{x,\mathrm{the\; y}}-I_{x,\mathrm{the\; y}+1}\left|\right)>\mathrm{Threshold}\\ 0& \left(\right|I_{x,\mathrm{the\; y}}-I_{x+1,\mathrm{the\; y}}|+|I_{x,\mathrm{the\; y}}-I_{x,\mathrm{the\; y}+1}\left|\right)\≤\mathrm{Threshold}\end{array}\right.$ (Formula 2)

Wherein, Threshold is a preset threshold value.

In the embodiment of the present invention, in order to obtain a suitable preset threshold value and further improve the accuracy of the determined gradient value, Threshold can be calculated according to formula three:

$\mathrm{Threshold}=\mathrm{QStep}/a$ (Formula 3)

Wherein, a can be any value, preferably, the value of a is 3.1; QStep has a corresponding relationship with the QP of the previous video image, therefore, the specific value of QStep is determined according to the QP of the previous video image.

In the embodiment of the present invention, when determining the second gradient value of the previous video image, there are many ways to be adopted. Preferably, the gradient function used when determining the first gradient value of the current video image is adopted, specifically as Formula 1 shows.

Step 210: Calculate the target bit rate of the current video image according to the preset average bit rate.

In the embodiment of the present invention, there are many ways to calculate the target code rate of the current video image. Preferably, the target code rate of the current video image is calculated according to the preset average code rate, wherein, all video images used in the entire encoding process The preset average bit rate is the same, does not change as the video image changes, and the target bit rate is the expected bit rate before encoding the video image, not the actual output bit rate after the video image is encoded .

In the embodiment of the present invention, there are many ways to calculate the target bit rate of the current video image according to the preset average bit rate. For example, the target average data size of the encoded video image can be determined according to the preset average bit rate, and based on the target average The data size calculates the total target data size of N video images including the current video image, and compares the total target data size with the rest of the N video images not including the current video image. The difference between the total actual data size of the video image is used as the target code rate of the current video image, wherein the data size refers to the memory size occupied by the code stream output after the video image is encoded, and all video images are encoded in the encoding process The average data size of the subsequent targets is the same.

For example: the preset average bit rate a in the process of video image encoding is 500kbps, the preset frame rate b is 25 frames/s, and the data sizes of the first three encoded video images are respectively: 22kbit, 21kbit, and 20kbit, then When determining the target bit rate of the fourth frame of video image (current video image) to be encoded, first determine the target average data size c after encoding each frame of video image according to the preset average bit rate a: adopt (Formula 4) to calculate, Then determine the total target data size d of the first four video images: d=c×n (Formula 5) where n is the number of specific video images, that is, d=20kbit/frame*4 frames=80kbit, then the fourth The target data size e of the video image is calculated according to Formula 6: e=d-specify the total size of the video image (Formula 6), that is Therefore, the target bit rate of the fourth video image is 17kbit/frame, that is, 680bps.

Further, in order to improve the stability of the determined target bit rate of the video image, after determining the target average data size of the encoded video image according to the preset average bit rate, calculate the current video image based on the target average data size The total target data size of the N video images, and the difference between the total target data size and the total actual data size of the remaining video images not including the current video image in the N video images As the preprocessing target code rate of the current video image, the target code rate of the current video image is calculated according to the preprocessing target code rate and the specified weight value of the video image.

For example: in the process of video image encoding, the preset average bit rate a is 500kbps, the preset frame rate b is 25 frames/s, and the total actual data sizes of the first three encoded video images are: 22kbit, 21kbit, 20kbit , the weight value ratio of the third video image and the fourth video image is 2:1, according to the above-mentioned embodiment, it can be known that the target data size of the fourth video image is 17kbit, then the target code rate of the fourth video image is That is 720bps.

Since the gradient parameter is introduced when calculating the gradient value, the gradient parameter can remove some pixels and improve the accuracy of the calculated gradient value. The gradient value represents the complexity of the video image, and the QP can be further calculated. Therefore, The error of the determined QP is avoided, and the accuracy of the QP is improved, thereby improving the accuracy of the code rate control method for single I-frame video encoding.

Step 220: Determine the performance parameter of the specified at least one coded video image.

In the embodiment of the present invention, the designated performance parameters of at least one coded video image include at least the following parameters: actual average code rate and actual average QP.

In the embodiment of the present invention, when determining the actual average bit rate of the specified at least one encoded video image, first obtain the actual bit rate of the specified at least one encoded video image, and then determine the actual average bit rate according to the actual bit rate , as shown in Equation 7:

${\mathrm{Rate}}_{\mathrm{last}}=\underset{k=1}{\overset{\mathrm{no}}{\mathrm{\Σ}}}{\mathrm{Rate}}_k/\mathrm{no}$ (Formula 7)

Wherein, Rate _last is the actual average code rate of at least one coded video image specified; Rate _k is the actual code rate of at least one coded video image specified; n is the number of coded video images specified.

For example: the actual bit rates of the first three encoded video images are respectively: 12kbps, 20kbps, and 24kbps, then the actual average bit rate of the first three encoded video images is the actual bit rate of the first three encoded video images The average is 18.67kbps.

In the embodiment of the present invention, when determining the actual average QP of the specified at least one encoded video image, first obtain the actual QP of the specified at least one encoded video image, and then determine the actual average QP according to the actual QP, as shown in formula 8 Shown:

${\mathrm{QP}}_{\mathrm{last}}=\underset{k=1}{\overset{\mathrm{no}}{\mathrm{\Σ}}}{\mathrm{QP}}_k/\mathrm{no}$ (Formula Eight)

Wherein, QP _last is the actual average QP of the specified at least one coded video image; QP _k is the actual QP of the specified at least one coded video image; n is the number of specified coded video images.

For example: the actual QP of the first three encoded video images are: 12, 20, 24 respectively, then the actual average QP of the first three encoded video images is the average value of the actual QP of the first three encoded video images 18.67 , since the QP can only be an integer, the calculated average QP is 19.

In this embodiment of the present invention, the specified encoded video image may be one encoded video image, or N encoded video images, where N is a positive integer greater than 1. Whether it is a coded video image or N coded video images, it is the previous or the previous N coded video images closest to the current video image, that is, it is not any coded video image obtained. video image.

For example: the current video image to be encoded is the twelfth video image, that is, there are eleven encoded video images in the front, and the determined performance parameters are the performance parameters of three encoded video images, then the three encoded video images The images are the eleventh video image, the tenth video image, and the ninth video image, instead of randomly selecting three video images from the encoded eleven video images.

Step 230: Determine whether the first gradient value and/or the second gradient value is 0, if yes, perform step 240; otherwise, perform step 250.

In the embodiment of the present invention, formula 1 is used to calculate the first gradient value and the second gradient value. Therefore, if the gradient parameter is 0, the calculated first gradient value and the second gradient value are 0.

Step 240: Determine the QP of the current video image according to the target code rate and performance parameters.

There are multiple ways to determine the QP of the current video image according to the target code rate and performance parameters. Preferably, the first Q calculation P function is used to determine the QP of the current video image. It is negatively correlated, as shown in Formula 9:

${\mathrm{QP}}_{\mathrm{this}}=-\frac{1}{\mathrm{\λ}}\&Center\; Dot;\ln (e^{{-\mathrm{\λQP}}_{\mathrm{last}}}\&Center\; Dot;\frac{{\mathrm{Rate}}_{\mathrm{this}}}{{\mathrm{Rate}}_{\mathrm{last}}})$ (formula nine)

Wherein, QP _this is the QP value of the current video image; Rate _this is the target bit rate of the current video image; Rate _last is the actual average bit rate of at least one encoded video image specified; QP _last is at least one encoded video specified The actual average QP of the image; the value range of λ is [0.1,0.2].

Step 250: Determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and performance parameters.

In the embodiment of the present invention, there are many ways to determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value, and performance parameters. Preferably, the second QP calculation function is used to determine the QP of the current video image. QP, where QP is positively correlated with the target code rate and the second gradient value, and negatively correlated with the first gradient value and performance parameters, specifically as shown in Formula 10:

${\mathrm{QP}}_{\mathrm{this}}=-\frac{1}{\mathrm{\λ}}\&Center\; Dot;\ln \left(e^{{-\mathrm{\λQP}}_{\mathrm{last}}}\&Center\; Dot;\frac{{\mathrm{Rate}}_{\mathrm{this}}\&Center\; Dot;{\mathrm{Gradient}}_{\mathrm{last}}}{{\mathrm{Rate}}_{\mathrm{last}}\·{\mathrm{Gradient}}_{\mathrm{this}}}\right)$ (Formula ten)

Wherein, QP _this is the QP value of the current video image; Gradient _this is the first gradient value of the current video image; Rate _this is the target code rate of the current video image; Rate _last is the actual average code of at least one coded video image specified rate; Gradient _last is the second gradient value of the previous video image; QP _last is the actual average QP of at least one coded video image specified; the value range of λ is [0.1,0.2].

Step 260: Encode the current video image according to the QP.

In the embodiment of the present invention, there are multiple ways to encode the current video image according to the QP. Preferably, the MEPG-2 standard is used to encode the current video image according to the QP; or, the JVT-G012 standard is used to encode the current video image according to the QP. The image is encoded.

In order to better understand the embodiment of the present invention, a specific application scenario is given below, and a further detailed description is made for the process of bit rate control in video coding:

The following takes determining the QP of the second video image as an example, and the specific execution process is shown in Figure 3 (wherein, the preset average bit rate is 2500kbps, and the actual data size of the first video image is 98824bits (that is, the actual bit rate is 3952.96 bps), the QP of the first video image is 40, a is 3.1, and λ is 0.12):

Step 300: Determine the first gradient value of the second video image and the second gradient value of the first video image;

In this step, when calculating the first gradient value, the Threshold is first determined according to Formula 3. The QP of the first video image is 40. By consulting the standard document, the corresponding QStep is 64 when the QP is 40. Therefore, it can be seen that the Threshold is 20.6, and then, bring the Threshold of 20.6 into Formula 2 to determine the gradient parameter Sgn(x, y), and finally, bring the determined gradient parameter Sgn(x, y) into Formula 1 to obtain the second video image The first gradient value is 566596.

When calculating the second gradient value, the process of calculating the first gradient value is the same as the process of calculating the first gradient value. Referring to the above process, the second gradient value is finally obtained as 567917.

Step 310: Calculate the target bit rate of the second video image (ie the current video image) according to the preset average bit rate of 2500kbps;

In this embodiment, the preset average code rate is 2500kbps, and the video frame rate is 25 frames/s. According to formula 4, the target average data size of each video image is 100kbit, that is, 100000bit. Since the encoded first video image The actual data size of is 98824bits, then according to Formula 6, the target bit rate of the second video image (that is, the current video image) is 101176bit/frame, which is 4047.04bps.

Step 320: Determine the performance parameters of the first video image, wherein the performance parameters include: actual average code rate, actual average QP;

In this step, owing to only determining the performance parameter of an encoded video image, therefore, the actual average code rate and the actual average QP are the code rate and the QP of the first video image, that is, the actual average code rate is 3952.96bps, the actual average QP for 40.

Step 330: Determine that both the first gradient value 566596 and the second gradient value 567917 are not 0, therefore, determine the QP of the second video image according to the target code rate, the first gradient value, the second gradient value and performance parameters;

In this step, when determining the QP of the second video image according to the target code rate, the first gradient value, the second gradient value and the performance parameter, the second QP function can be used for determination, wherein the second QP function is as shown in formula ten , put the above specific values into Formula 10 to get:

${\mathrm{<span\; class="notranslate">\; QP</span>}}_{\mathrm{<span\; class="notranslate">\; this</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 0.12</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; ln</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.12</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; 40</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\frac{\mathrm{<span\; class="notranslate">\; 4047.04</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; 567917</span>}}{\mathrm{<span\; class="notranslate">\; 3952.96</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; 566596</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 39.785</span>}$

Step 340: Encode the second video image according to the QP, and output a code stream.

The following takes determining the QP of the third video image as an example, and the specific execution process is shown in Figure 4 (wherein, the preset average code rate is 2500kbps, the actual data size of the first video image is 98824bits (that is, the actual code rate is 3952.96 bps), the QP of the first video image is 40, a is 3.1, and λ is 0.12):

Step 400: Determine the first gradient value of the third video image and the second gradient value of the second video image;

In this step, when calculating the first gradient value, first determine Threshold according to formula three, a is 3.1, because the QStep of the current video image is related to the actual average QP of the specified encoded video image, in this embodiment, the specified The coded video images are the first video image and the second video image, the QP of the first video image is 40, and the QP of the second video image is 40 (the actual QP of the second video image is 39.785, because QP can only be an integer , therefore, the QP of the second video image is optimized to 40), then the actual average QP is 40. According to the standard document, when the QP is 40, the corresponding QStep is 64. Therefore, it can be seen that the Threshold is 20.6, and then Threshold Put the gradient parameter Sgn(x,y) into formula 2 for 20.6 to determine the gradient parameter Sgn(x,y), and finally, bring the determined gradient parameter Sgn(x,y) into formula 1 to get the first gradient value of the third video image 565791 .

When calculating the second gradient value, the process of calculating the first gradient value is the same as the process of calculating the first gradient value, referring to the above process to finally obtain the second gradient value of 566596.

Step 410: Calculate the target bit rate of the third video image (ie the current video image) according to the preset average bit rate of 2500kbps;

In this embodiment, the preset average code rate is 2500kbps, and the video frame rate is 25 frames/s. According to formula 4, the target average data size of each video image is 100kbit, that is, 100000bit. Since the encoded first video image The actual data size of the encoded second video image is 98824bits, and the actual data size of the encoded second video image is 98176bits. According to formula 6, the target bit rate of the third video image (ie, the current video image) is 103000bit/frame, that is, 4120bps.

Step 420: Determine the performance parameters of the first video image and the second video image, wherein the performance parameters include: actual average code rate, actual average QP;

In this step, since it is determined the performance parameters of two encoded video images, the actual average code rate is the average value of the actual code rates of the first video image and the second video image, and the actual average QP is the first video image The average value of the actual QP of the image and the second video image, that is, the actual average QP is 40, and the actual average bit rate is 100912bit/frame, that is, 4036.48bps.

Step 430: Determine that both the first gradient value 565791 and the second gradient value 566596 are not 0, therefore, determine the QP of the third video image according to the target code rate, the first gradient value, the second gradient value and performance parameters;

In this step, when determining the QP of the third video image according to the target code rate, the first gradient value, the second gradient parameter and the performance parameter, the second QP function can be used for determination, wherein the second QP function is as shown in formula ten , λ is 0.12, and the above specific value is brought into Formula 10 to get:

${\mathrm{<span\; class="notranslate">\; QP</span>}}_{\mathrm{<span\; class="notranslate">\; this</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 0.12</span>}}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; ln</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.12</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; 40</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\frac{\mathrm{<span\; class="notranslate">\; 4120</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; 566596</span>}}{\mathrm{<span\; class="notranslate">\; 4036.48</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; 565791</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 39.615</span>}$

Since the QP can only be an integer, the actual QP of the third video image is 40. After encoding, the actual data size of the third video image is 99656 bits.

Step 440: Encode the third video image according to the QP, and output a code stream.

Based on the above technical solution, as shown in FIG. 5, in the embodiment of the present invention, the encoder includes a first determination unit 50, a calculation unit 51, a second determination unit 52, a third determination unit 53, and an encoding unit 54, wherein

The first determining unit 50 is used to determine the first gradient value of the current video image and the second gradient value of the previous video image;

Calculation unit 51, for calculating the target code rate of current video image according to preset average code rate;

The second determination unit 52 is configured to determine the specified performance parameter of at least one coded video image;

The third determining unit 53 is used to judge whether the first gradient value and/or the second gradient value are 0, if so, determine the quantization parameter QP of the current video image according to the target code rate and performance parameters, otherwise, according to the target code rate, the second gradient value A gradient value, a second gradient value and performance parameters determine the QP of the current video image;

The encoding unit 54 is configured to encode the current video image according to the QP.

The first determining unit 50 is specifically configured to determine a first gradient value of the current video image according to a gradient function, wherein the gradient function includes at least a pixel parameter and a gradient parameter, and the gradient parameter is 0 or 1.

In the embodiment of the present invention, the calculation unit 51 specifically determines the target average data size unit 510, the calculation total target data size unit 5100, and the calculation target code rate unit 51000, as shown in FIG. 6A, wherein:

Determine the target average data size unit 510, which is used to determine the target average data size after video image encoding according to the preset average code rate;

Calculate the total target data size unit 5100, for calculating the total target data size of N video images including the current video image based on the target average data size;

Calculating the target code rate unit 51000, which is used to use the difference between the total target data size and the total actual data size of the remaining video images not including the current video image in the N video images as the target code rate of the current video image; or , which is used to use the difference between the total target data size and the total actual data size of the remaining video images excluding the current video image in the N video images as the preprocessing target code rate of the current video image, and then according to the preprocessing Process the target bit rate and the weight value of the specified video image to calculate the target bit rate of the current video image.

In the embodiment of the present invention, the performance parameters of the specified at least one coded video image determined by the second determination unit 52 at least include: an actual average code rate and an actual average QP.

Further, the second determination unit 52 specifically includes an acquisition unit 520 and a performance parameter determination unit 5200, as shown in FIG. 6B, wherein the acquisition unit 520 is used to acquire the actual code rate and the actual QP of at least one specified encoded video image ;

The performance parameter determining unit 5200 is configured to respectively determine an actual average code rate and an actual average QP according to the actual code rate and the actual QP.

To sum up, in the embodiment of the present invention, the encoder first determines the first gradient value of the current video image and the second gradient value of the previous video image when performing rate control on each video image during the encoding process. , and then calculate the target bit rate of the current video image according to the preset average bit rate, then determine the specified performance parameter of at least one encoded video image, and finally, when the first gradient value and/or the second gradient value is 0 In the case of , determine the QP of the current video image according to the target code rate and performance parameters, and if the first gradient value and/or the second gradient value are not 0, determine The value and performance parameters determine the QP of the current video image. Finally, the current video image is encoded according to the QP. Since the gradient value is one of the inherent attributes of the video image, it is used to quantify the complexity of the video image, and QP also measures the video image. complexity, the accuracy of the QP calculated by the gradient value is relatively high, and in the present invention, the specific method used to calculate the QP is determined by the first gradient value and the second gradient value, and the first gradient value and the second gradient value When the second gradient value is not 0, determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and performance parameters, therefore, adopt the scheme provided by the embodiments of the present invention to calculate QP, which can improve The accuracy of the calculated QP improves the accuracy of the bit rate control method for single I-frame video coding, and therefore solves the problem that the current accuracy of the bit rate control method for single I-frame video coding is poor. At the same time, since there is no need to set a new statistic to measure the complexity of the image frame in the process of determining the QP, the complexity of the rate control method for single I-frame video coding is reduced.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and combinations of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a A device for realizing the functions in one or more procedures of the flow chart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device implements the functions in one or more procedures of the flow chart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions in the flow chart or flow charts and/or the block diagram block or blocks.

While preferred embodiments of the invention have been described, additional changes and modifications to these embodiments can be made by those skilled in the art once the basic inventive concept is appreciated. Therefore, it is intended that the appended claims be construed to cover the preferred embodiment as well as all changes and modifications which fall within the scope of the invention.

Apparently, those skilled in the art can make various changes and modifications to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. In this way, if the modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and equivalent technologies, the present invention also intends to include these modifications and variations.

Claims (12)

1. a bit rate control method in video coding, is used for carrying out bit rate control to the video image of single I frame, is characterized in that, comprises: determining the first gradient value of the current video image and the second gradient value of the previous video image; Calculate the target bit rate of the current video image according to the preset average bit rate; determining a performance parameter of the specified at least one encoded video picture; Judging whether the first gradient value and/or the second gradient value is 0, if so, determine the quantization parameter QP of the current video image according to the target code rate and the performance parameter, otherwise, determine the quantization parameter QP according to the target bit rate The code rate, the first gradient value, the second gradient value and the performance parameter determine the QP of the current video image; Encode the current video image according to the QP. 2. The method according to claim 1, wherein determining the first gradient value of the current video image comprises: Determine a first gradient value of the current video image according to a gradient function, wherein the gradient function includes at least a pixel parameter and a gradient parameter, and the gradient parameter is 0 or 1. 3. The method according to claim 1 or 2, wherein calculating the target bit rate of the current video image according to a preset average bit rate comprises: Determine the target average data size after video image encoding according to the preset average code rate, and calculate the total target data size of N video images including the current video image based on the target average data size, and convert the total target data size The difference between the data size and the total actual data size of the remaining video images not including the current video image in the N video images is used as the target code rate of the current video image; or, Determine the target average data size after video image encoding according to the preset average code rate, and calculate the total target data size of N video images including the current video image based on the target average data size, and convert the total target data size The difference between the data size and the total actual data size of the remaining video images not including the current video image in the N video images is used as the preprocessing target code rate of the current video image, according to the preprocessing target code rate Calculate the target bit rate of the current video image using the rate and the specified weight value of the video image. 4. The method according to claim 1, wherein the determined specified performance parameters of at least one coded video image at least include: an actual average code rate and an actual average QP. 5. The method as claimed in claim 4, wherein determining the actual average code rate and the actual average QP of at least one coded video image specified comprises: Acquire the actual code rate and actual QP of at least one specified coded video image, and respectively determine the actual average code rate and the actual average QP according to the actual code rate and the actual QP. 6. The method according to claim 1, wherein determining the quantization parameter QP of the current video image according to the target code rate and the performance parameter comprises: Using a first QP calculation function to determine the QP of the current video image according to the target code rate and the performance parameter, wherein the QP is positively correlated with the target code rate and negatively correlated with the performance parameter. 7. The method according to claim 1, wherein the quantization parameter QP of the current video image is determined according to the target code rate, the first gradient value, the second gradient value and the performance parameter ,include: Using the second QP calculation function to determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and the performance parameter, wherein the QP and the target code Both the rate and the second gradient value are positively correlated, and are negatively correlated with the first gradient value and the performance parameter. 8. The method according to claim 1, wherein encoding the current video image according to the QP comprises: Encoding the current video image according to the QP using the MEPG-2 standard; or, The current video image is encoded according to the QP by using the JVT-G012 standard. 9. A kind of coder, is used for carrying out bit rate control to the video image of single I frame, is characterized in that, comprises: The first determination unit is used to determine the first gradient value of the current video image and the second gradient value of the previous video image; A calculation unit, configured to calculate the target bit rate of the current video image according to a preset average bit rate; A second determining unit, configured to determine a specified performance parameter of at least one coded video image; A third determination unit, configured to determine whether the first gradient value and/or the second gradient value is 0, and if so, determine the quantization parameter QP of the current video image according to the target code rate and the performance parameter , otherwise, determine the QP of the current video image according to the target code rate, the first gradient value, the second gradient value and the performance parameter; A coding unit, configured to code the current video image according to the QP. 10. The encoder according to claim 9, wherein the calculation unit specifically includes a unit for determining a target average data size, a unit for calculating a total target data size, and a unit for calculating a target code rate, wherein: The unit for determining the target average data size is used to determine the target average data size of the encoded video image according to the preset average code rate; The unit for calculating the total target data size is used to calculate the total target data size of N video images including the current video image based on the target average data size; The unit for calculating the target code rate is used to use the difference between the total target data size and the total actual data size of the remaining video images not including the current video image in the N video images as the The target code rate of the current video image; or, for first calculating the difference between the total target data size and the total actual data size of the remaining video images not including the current video image in the N video images The value is used as the preprocessing target code rate of the current video image, and then the target code rate of the current video image is calculated according to the preprocessing target code rate and the specified weight value of the video image. 11. The encoder according to claim 9, wherein the performance parameters of the specified at least one coded video image determined by the second determination unit at least include: an actual average code rate and an actual average QP. 12. The encoder according to claim 11, wherein the second determination unit specifically includes an acquisition unit and a performance parameter determination unit, wherein: The acquiring unit is configured to acquire the actual code rate and actual QP of at least one specified encoded video image; The performance parameter determining unit is configured to respectively determine the actual average code rate and the actual average QP according to the actual code rate and the actual QP.