WO2009043264A1 - A control method and device for image code rate - Google Patents

A control method and device for image code rate Download PDF

Info

Publication number
WO2009043264A1
WO2009043264A1 PCT/CN2008/072415 CN2008072415W WO2009043264A1 WO 2009043264 A1 WO2009043264 A1 WO 2009043264A1 CN 2008072415 W CN2008072415 W CN 2008072415W WO 2009043264 A1 WO2009043264 A1 WO 2009043264A1
Authority
WO
WIPO (PCT)
Prior art keywords
bits
image
quantization parameter
image block
current
Prior art date
Application number
PCT/CN2008/072415
Other languages
English (en)
French (fr)
Inventor
Fangyao Chen
Original Assignee
Huawei Technologies Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Publication of WO2009043264A1 publication Critical patent/WO2009043264A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/124Quantisation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/146Data rate or code amount at the encoder output
    • H04N19/147Data rate or code amount at the encoder output according to rate distortion criteria
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/176Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to an image rate control method and apparatus. Background technique
  • Existing image signals such as mpeg2, mepg4, and H.264 image signals, are often transformed using corresponding image transformation techniques depending on the application and transmission channel requirements.
  • the code rate conversion method is used to adjust the code rate of the image
  • the image format conversion method and the frame rate conversion method are used to adjust the image size and frame rate. All of these image transformation processes require rate control of the output code rate to ultimately meet the requirements of adapting to the channel and reducing bandwidth.
  • the output code stream needs to be subjected to rate control so that the output code rate reaches the specified requirement. Therefore, the control of the image bit rate is very important.
  • the approximate working principle of the image rate control apparatus in the existing rate conversion technique for controlling the image code rate is as follows: Assume that Ratel is the input image code rate, Rate2 is the image rate required to be output, and the compressed image data is decompressed. After outputting uncompressed DCT (Discrete Cosine Transformation) i or image data, the DCT domain image data is inverse quantized, re-quantized, and finally compressed, and finally the original image rate is from Ratel to Rate2. Transformation process.
  • DCT Discrete Cosine Transformation
  • the raw input image rate Rate of the received feedback, the image rate Rate2, which is required to be output, and the quantization parameter (represented as Qpi) of each macroblock of the original image and the original number of bits of each macroblock (represented as mb-biti) Calculating the complexity X of the current frame image and the complexity xi of each macroblock of the current frame image according to the feedback Qpi and mb_bitsi, Then calculating the conversion quantization parameter reference value Qt: calculating the target use bit number R_target_using of the transformed macroblock according to the complexity of the transformed macroblock in the current frame image and the complexity of the current one frame image;
  • the target bit number R_target_ used of the transformed macroblock is compared with the actual used bit number R_act_ used of the transformed macroblock, when the difference between the two is greater than the defined error range threshold, ie, R_target_used-R- When act- used > threshold, the quantization parameter is increased to obtain the quantization parameter OUT_Qp required for re-quantization
  • the values of X_org and R_act_used are updated, and the quantization parameters required for requantization are determined according to the above method, and the subsequent macroblocks are further transformed according to the quantization parameters determined by the rate control module.
  • the existing image rate control method needs to use a predefined error range when determining the quantization parameter of the current macroblock: when the difference between the target bit number of the converted macroblock and the actual used bit number is greater than the error range
  • the quantization parameter reduces the quantization parameter when the difference between the target bit number of the transformed macroblock and the actually used bit number is smaller than the error range, and the obtained quantization parameter is only a predicted value, which is not completely consistent with the quantization of the current macroblock. Actual requirements, therefore, using the quantization parameter to quantize the current bit rate after the current macroblock may not completely meet the requirements. If the output is too high or too low, the average bit rate can only be achieved by controlling the subsequent macroblock rate. Rate control has hysteresis.
  • Embodiments of the present invention provide an image rate control method and apparatus thereof, which are used to select appropriate quantization parameters for each image block, not only the actual output code rate and requirements of each image block after rate conversion.
  • the difference in output bit rate is small, and the cumulative error can be reduced.
  • An embodiment of the present invention provides an image rate control method, including:
  • the quantization parameter that makes the actual number of bits of the current image block closest to the target number of bits is selected, wherein the bit number error of each image block The difference between the actual number of bits of the image block and the corresponding target number of bits;
  • the current image block is quantized according to the determined quantization parameter.
  • An embodiment of the present invention provides an image rate control apparatus, including:
  • a rate control module configured to determine a target number of bits of the current image block, where the target number of bits is a number of bits when the current image block is output at a required output rate
  • a quantization parameter determining module configured to: in all of the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, select a quantization parameter that approximates the actual number of bits of the current image block to the target number of bits, where each The bit number error of an image block is the difference between the actual number of bits of the image block and the corresponding target number of bits;
  • a quantization module configured to quantize the current image block according to the quantization parameter determined by the quantization parameter determination module.
  • An embodiment of the present invention provides an image rate control apparatus applied to a rate conversion technology, including:
  • variable length decoding module configured to send, to the rate control module, the quantization parameter and the original number of bits of each image block of the current frame original image, and the total number of image blocks of the current frame original image;
  • a rate control module configured to determine a destination of the current image block according to the information provided by the variable length decoding module, the required output code rate, and the original input code rate of the current frame original image. a number of target bits, the number of target bits being the number of bits when the current image block is output at a desired output rate;
  • a quantization parameter determining module configured to: in all of the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, select a quantization parameter that approximates the actual number of bits of the current image block to the target number of bits, where each The bit number error of an image block is the difference between the actual number of bits of the image block and the corresponding target number of bits;
  • a quantization module configured to quantize the current image block according to the quantization parameter determined by the quantization parameter determination module.
  • An embodiment of the present invention provides an image rate control apparatus applied to an image coding technology, including:
  • An image analysis module configured to calculate the complexity of each image block of the current frame original image, the total number of bits of the current frame original image, the original input code rate of the current frame original image, and the total number of image blocks of the current frame original image.
  • the rate control module is configured to determine a target number of bits of the current image block according to the information provided by the image analysis module and the required output code rate, where the target number of bits is the current image The number of bits when the block outputs the bit rate as required;
  • a quantization parameter determining module configured to: in all of the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, select a quantization parameter that approximates the actual number of bits of the current image block to the target number of bits, where each The bit number error of an image block is the difference between the actual number of bits of the image block and the corresponding target number of bits;
  • a quantization module configured to quantize the current image block according to the quantization parameter determined by the quantization parameter determination module.
  • the image rate control method and apparatus thereof provided by the embodiments of the present invention select the actual number of bits of the current image block to be closest to the target by selecting all quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks.
  • the quantization parameter of the bit number quantizes the current image block according to the quantization parameter, so that not only the actual output code rate of each image block output and the required output code rate error are small, but also does not increase the actuality of all the transformed image blocks.
  • the cumulative error of the number of bits and the corresponding number of target bits causes the cumulative error of the image blocks of the plurality of image blocks or the entire frame to approach zero, so that the quality of the output image is maintained at a high quality.
  • FIG. 1 is a schematic flowchart of an image bit rate control method according to an embodiment of the present invention
  • FIG. 2 is another schematic flowchart of an image bit rate control method according to an embodiment of the present invention
  • FIG. 4 is a schematic flowchart of a method for determining a quantization parameter required for quantizing a current macroblock according to an embodiment of the present invention
  • FIG. 5 is a schematic flow chart of an image bit rate control method according to another embodiment of the present invention.
  • FIG. 6 is a schematic structural flow chart of an image bit rate control device according to an embodiment of the present invention.
  • FIG. 7 is a schematic structural flow chart of a quantization parameter determining module according to an embodiment of the present invention;
  • FIG. 8 is a schematic structural diagram of an image rate control apparatus applied to a rate conversion technology according to an embodiment of the present invention;
  • Figure 9 is a block diagram showing the structure of an image rate control device applied to an image coding technique in an embodiment of the present invention. detailed description
  • the embodiment of the invention provides an image rate control method, which can be applied to a technique such as a rate conversion technology and an image coding technology, and the method determines the target bit number of the current image block.
  • the target bit number is the number of bits when the current image block is output at the required output bit rate; among all the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, the actual bit number of the current image block is selected to be the closest.
  • a quantization parameter of the target number of bits wherein a bit number error of each image block is a difference between an actual number of bits of the image block and a corresponding target bit number; and the current image block is quantized according to the determined quantization parameter.
  • the image block suitable for the method may be a macroblock, a stripe or a frame image.
  • the image rate control method provided by the embodiment of the invention not only does not increase the cumulative error of the actual number of bits of all transformed image blocks and the corresponding target number of bits, thereby reducing the image data of multiple image blocks or the entire frame.
  • the cumulative error makes the cumulative error gradually approach zero, and the actual output code rate of each image block and the required output code rate error are small, and the output image quality is high.
  • the image data of Ratel is decompressed by VLD (Variable Length Decoding) module and outputs uncompressed DCT domain image data.
  • the data is inverse quantized by IQ (Inverse Quantization) module, and then passes through Q ( The Quantization, Quantization module re-quantizes and finally compresses the output through the VLC (Variable Length Coding) module to convert the original image rate from Ratel to Rate2.
  • the quantization parameter required for the Q module requantization is determined according to the image rate control method provided by the embodiment of the present invention shown in FIG. 2:
  • the input bit rate determining module determines the original input bit rate Rateel, and sends the determined original input bit rate Rateel to the rate control module, and the VLD module calculates the quantization parameter and the original number of bits of each macroblock of the current frame original image, and the current frame original.
  • the total number of bits of the image and the total number of macroblocks of the original image of the current frame are sent to the rate control module, and the output rate determining module determines the required output rate Rate2 and sends the required output rate Rate2 to the code rate. Control module.
  • the rate control module determines a target number of bits of the current macroblock.
  • the rate control module determines the target bit number of the current macroblock according to the information sent by the input code rate determining module, the VLD module, and the output code rate determining module, where the target bit number is when the current macroblock is output at the required output rate Rate2 The number of bits.
  • the rate control module sends the determined target bit number of the current macroblock to the quantization parameter determining module.
  • the quantization parameter determination module determines the quantization parameter.
  • the quantization parameter determination module selects a quantization parameter that brings the actual number of bits of the current macroblock closest to its target number of bits among all the quantization parameters that do not increase the accumulated bit number error of all the transformed macroblocks.
  • the quantization module and the variable length coding module transform the current macroblock.
  • the quantization module quantizes the current macroblock according to the quantization parameter determined by the quantization parameter determining module, and the variable length coding module performs variable length coding on the current macroblock data quantized by the quantization module, outputs a code stream, and implements an image bit rate from Ratel to Rate2. Transformation.
  • the original input code rate Rateel may also be determined by the rate control module itself without inputting the code rate determining module; or the required output rate rate Rate2 may be directly configured in the rate control module for storage.
  • the required rate rate Rate2 can also be determined by the rate control module itself, without being provided by the output rate determining module, so that the input rate determining module and the output rate determining module can be additionally set to save equipment cost.
  • the rate control module determines the target number of bits of the current macroblock
  • the code rate, the preset current frame original image scale value, and the macroblock of the current frame original image may also be output according to requirements.
  • the total number determines the target number of bits of the current macroblock.
  • the current frame original image scale value is a ratio of the number of target bits of the original frame of the current frame to the number of output bits required per unit time.
  • step S302 further includes:
  • the rate control module calculates the target number of the original image of the current frame according to the original input bit rate Rateel, the required output rate Rate2, and the total number of bits of the original image of the current frame, and the calculation method is: the total number of bits of the original image of the current frame and The product of the output code rate is required to be divided by the original input code rate, and its calculation formula (1) is expressed as:
  • frame_bits is the target number of bits of the original frame of the current frame
  • frame_bits 1 is the total number of bits of the original frame of the current frame
  • Ratel is the original input bit rate
  • Rate2 is the required output bit rate
  • the rate control module calculates the image complexity of the original image of the current frame according to the quantization parameter and the original number of bits of each macroblock of the current frame original image, and the calculation method is: multiplication product of the quantization parameter of each macroblock and the original number of bits And the calculation formula (2) is expressed as:
  • X is the image complexity of the original frame of the current frame
  • mb_no is the total number of macroblocks of the original frame of the current frame
  • Qpi is the quantization parameter of the i-th macroblock
  • mb_bitsi is the original bit of the i-th macroblock number.
  • the rate control module calculates the complexity of the current macroblock according to the quantization parameter of the current macroblock and the original number of bits, and the calculation method is: the product of the quantization parameter of the current macroblock and the original number of bits, and the calculation formula (3) is expressed as :
  • Xj Qpjxmb_bitsj ( 3 ) where xj is the complexity of the current macroblock, Qpj is the quantization parameter of the current macroblock, and mb bitsj is the original number of bits of the current macroblock.
  • the rate control module calculates the target bit number frame_bit of the current frame original image, the image complexity X of the current frame original image, the total number of macroblocks of the current frame original image mb_no, and the complexity of the current macroblock xj.
  • the target number of bits of the current macroblock is calculated as: the product of the complexity of the current macroblock and the target number of bits of the original frame of the current frame, divided by the image complexity of the original frame of the current frame and the total macroblock of the original image of the current frame.
  • the product of the quantity, the formula (4) is expressed as:
  • mb_bits is the target number of bits of the current macroblock.
  • the rate control module calculates the target number of bits of the current macroblock.
  • the rate control module determines the target number of bits of the current macroblock
  • the above steps S401-S404 may not be performed, and the code rate, the preset current frame original image scale value, and the current frame original image may be output according to requirements.
  • the total number of macroblocks, the target number of bits of the current macroblock; its determination formula is expressed as:
  • mb_bits is the target number of bits of the current macroblock
  • B is the number of output bits required per unit time, and its value is equal to the required output bit rate Rate2
  • p is the preset original image ratio of the current frame
  • mb-no is The total number of macroblocks of the original frame of the current frame.
  • the target bit number of each frame image is equal to the required number of output bits per second, that is, when both are 4%, the target number of bits of each frame image is equal.
  • step S304 since the number of quantization parameters is limited, generally 31/51/64 values, and the larger the quantization parameter value, the smaller the actual number of bits of the corresponding current macroblock, so the quantization parameter determination module can A fast search algorithm is used to determine the quantization parameters needed to quantize the current macroblock.
  • the quantization parameter used for quantizing the current macroblock is determined by the dichotomy in the fast search algorithm, the difference between the actual number of bits corresponding to the quantization parameter and the target number of bits of the current macroblock is simply referred to as the bit number error.
  • S304 specifically includes:
  • the quantization parameter determining module arranges all the quantization parameters into a quantization parameter in a certain size order.
  • the quantization parameter determining module selects a first quantization parameter that is in an intermediate position of the quantization parameter arrangement, and calculates an actual number of bits corresponding to the first quantization parameter.
  • the actual number of bits corresponding to the first quantization parameter it can be calculated using the MPEG-2 standard, the MPEG-FOR standard, the H.264 standard, and the like.
  • the quantization parameter determining module calculates a bit number error corresponding to the first quantization parameter.
  • step S505 determining the size relationship between the bit number error and 0, when the bit number error is greater than 0, proceed to step S505;
  • step S507 is continued.
  • bit number error corresponding to the first quantization parameter is greater than 0
  • bit number error corresponding to the second quantization parameter is still greater than 0
  • bit number error corresponding to the second quantization parameter is still greater than 0
  • step S508 when the found quantization parameter corresponds to the bit number error is exactly equal to When 0, step S509 is performed.
  • bit number error corresponding to the first quantization parameter is less than 0, continue to select the second quantization in the middle position in the semi-group quantization parameter arrangement smaller than the first quantization parameter. And calculating a bit number error corresponding to the second quantization parameter, if the bit number error corresponding to the second quantization parameter is still less than 0, continuing to select the quantization parameter in the intermediate position in the quantization parameter arrangement smaller than the second quantization parameter and calculating the The bit number error corresponding to the quantization parameter is sequentially pushed.
  • step S508 is continued.
  • step S509 is performed.
  • step S512 is continued.
  • step S512 When the accumulated bit number error of all the transformed macroblocks is not equal to zero, among the two adjacent quantization parameters, the sign of the error of the corresponding bit number error and the accumulated bit number of all the converted macroblocks are selected.
  • the inverse quantization parameter continues to step S512; when the accumulated bit number error of all the transformed macroblocks is equal to zero, among the two adjacent quantization parameters, a quantization parameter having a smaller absolute value of the corresponding bit number error is selected, Proceed to step S512.
  • a quantization parameter whose corresponding bit number error is less than 0 is selected; when all the transformed macroblocks are accumulated When the bit number error is less than 0, in the above two adjacent quantization parameters, a quantization parameter whose corresponding bit number error is greater than 0 is selected; when the accumulated bit number error of all converted macroblocks is equal to 0, then Adjacent two quantization Among the parameters, a quantization parameter having a smaller absolute value of the corresponding bit number error is selected as a quantization parameter required for quantizing the current macroblock.
  • the quantization parameter determining module sends the selected quantization parameter to the quantization module, and adds the bit number error of the current macro block to the accumulated bit number error of all the transformed macro blocks to obtain a new accumulated bit number error.
  • step S509 a quantization parameter close to the target number of bits, for example, using a golden point method; or calculating a bit number error corresponding to each quantization parameter, and then determining a sign of the bit number error and an accumulated bit number error of all converted macroblocks All the quantization parameters opposite to the sign, and among the determined quantization parameters, the quantization parameter having the smallest absolute value of the corresponding bit number error is selected; or in step S505, the bit number error corresponding to the first quantization parameter is greater than 0 And continuing to select the quantization parameter among all the quantization parameters except the first quantization parameter, and when finding the adjacent two quantization parameter whose corresponding bit number errors are greater than 0 and less than 0 respectively, proceeding to step S508; When the bit number error is exactly equal to the quantization parameter of 0, step S509
  • step S508 When the parameter is quantized, step S508 is continued; when the quantization parameter corresponding to the bit number error is exactly equal to 0, step S509 is performed.
  • step S508 When the parameter is quantized, step S508 is continued; when the quantization parameter corresponding to the bit number error is exactly equal to 0, step S509 is performed.
  • steps S501-S512 when the quantization parameter corresponding to the bit number error equal to 0 is determined, if the accumulated bit number error of all the converted macroblocks is not equal to 0, the error corresponding to the corresponding bit number may be selected to be equal to 0.
  • the quantization parameter whose quantization parameter is adjacent and whose corresponding bit number error is opposite to the sign of the accumulated bit number error of all converted macroblocks can reduce d, the accumulated bit number error of all converted macroblocks.
  • the current image block may be the first image block of the current frame original image. (ie, steps S508 and S510 are not performed), and the accumulated bit number error of all the converted macroblocks corresponding to the first image block is regarded as 0, so that any one of the current image blocks can be determined according to the method described in step S511. Quantitative parameters.
  • the first embodiment of the present invention only describes the flow of the method for performing rate control on the rate conversion of the macroblock.
  • the method for performing rate control on the rate conversion of other image blocks is the same as the process described in the first embodiment of the present invention. No longer stated.
  • the image rate control method uses the image rate control method to select a quantization parameter for the quantization module, so that the actual output code rate of each macroblock output and the required output code rate error are small, and the error is not increased.
  • the accumulated bit number error of the transformed macroblock is such that the cumulative error of the actual bit number of the plurality of macroblocks or the entire frame image data and the corresponding target bit number approaches zero, so that the output image quality maintains a high quality, and does not
  • the cumulative error increases as the number of transformed macroblocks increases, resulting in a phenomenon in which partial macroblock quantization is large and image quality is poor.
  • the output bit rate of each macroblock output and the required output bit rate error are small, the output bit rate of each macro block is relatively uniform, so a large output buffer is not required, and the output code is required. The difference in rate is the smallest and the requirements are higher. Moreover, the output code stream does not fluctuate within a small time range, and can satisfy the situation where the code rate requirement ratio is high.
  • the image analysis module analyzes the input original image, and analyzes the complexity and current of each macroblock of the original frame original image.
  • the total number of bits of the original image of the frame, the original input code rate Rate of the original frame original image, and the total number of macroblocks of the original image of the original frame are sent to the code rate control module;
  • the output code rate determination module sends the required output code rate Rate2 to Rate control module;
  • the rate control module sends the determined target bit number of the current macroblock to the quantization parameter determining module;
  • the quantization parameter determining module selects the quantization parameter that makes the actual number of bits of the current macroblock closest to the target number of bits among all the quantization parameters that do not increase the accumulated bit number error of all the converted macroblocks according to the foregoing steps S501-S512. And the determined quantization parameter is sent to the quantization module; the quantization module quantizes the current macroblock according to the quantization parameter determined by the quantization parameter determination module, and the variable length coding module performs variable length coding on the current macroblock data quantized by the quantization module, The code stream is output, and the image bit rate is changed from Ratel to Rate2.
  • the original input bit rate Ratel can also be provided by the rate control module itself without the need of the image analysis module; the required output rate rate Rate2 can also be directly configured in the rate control module for storage, or by the rate control module itself. Determining the required output bit rate Rate 2, which is not provided by the output bit rate determining module, so as to avoid additionally setting the output bit rate determining module and saving equipment cost;
  • the rate control module determines the target number of bits of the current macroblock
  • the current macro may also be determined according to the required output code rate, the preset current frame original image scale value, and the total number of macroblocks of the current frame original image.
  • the target number of bits of the block for the specific determination method, see the above formula (5).
  • Another embodiment of the present invention only describes a method flow for performing rate control on a rate conversion of a macroblock, and a method for performing rate control on rate conversion of other image blocks is the same as the process described in the embodiment of the present invention, where No longer stated.
  • the image rate control method uses the image rate control method to select a quantization parameter for the quantization module, so that the actual output code rate of each macroblock output and the required output code rate error are small, and the error is not increased.
  • the cumulative number of bits of the transformed macroblock is incorrect Poor, so that the cumulative error of the actual number of bits of the image data of the plurality of macroblocks or the entire frame and the corresponding target number of bits approaches zero, so that the quality of the output image remains high, and the cumulative error does not follow the transformed macro.
  • the number of blocks increases, the number of macroblocks increases, and the image quality is poor.
  • the output bit rate of each macroblock output and the required output bit rate error are small, the output bit rate of each macro block is relatively uniform, so a large output buffer is not required, and the output code is required. The difference in rate is the smallest and the requirements are higher. Moreover, the output code stream does not fluctuate within a small time range, and can satisfy the situation where the code rate requirement is relatively high.
  • an embodiment of the present invention provides an image rate control apparatus, including: a rate control module 71, configured to determine a target number of bits of a current image block, where the target number of bits is the current image block. Output the number of bits at the bit rate output as required;
  • the quantization parameter determining module 72 is configured to: in all the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, select a quantization parameter that makes the actual number of bits of the current image block closest to the target number of bits, where The bit number error of each image block is the difference between the actual number of bits of the image block and the corresponding target number of bits;
  • the quantization module 73 is configured to quantize the current image block according to the quantization parameter determined by the quantization parameter determination module 72.
  • the image block suitable for the device may be a macroblock, a stripe or a frame image; the quantization parameter determining module 72 may be disposed in the code rate control module 71 or in the quantization module 73.
  • the rate control module 71 determines the target number of bits of the current image block according to the foregoing formula (4) or (5).
  • the quantization parameter determination module 72 further includes:
  • the reference quantization parameter determining sub-module 81 is configured to determine a quantization parameter whose corresponding bit number error is equal to 0, or to determine that two adjacent corresponding bit number errors are greater than 0 and less than
  • the quantization parameter determining sub-module 82 is configured to: when the reference quantization parameter determining sub-module 81 determines the quantization parameter whose corresponding bit number error is equal to 0, select the quantization parameter, or select a corresponding bit number error adjacent to the quantization parameter Positive sign and the accumulated number of bits a quantization parameter opposite to the sign of the error; when the reference quantization parameter determination sub-module 81 determines the two adjacent quantization parameters, selecting a quantization parameter capable of reducing the accumulated bit number error of all the transformed image blocks .
  • the quantization parameter determining sub-module 82 is further configured to: when the reference quantization parameter determining sub-module 81 determines the two adjacent quantization parameters, if the accumulated bit number error of all the transformed image blocks is not equal to zero, Determining, in the two adjacent quantization parameters, a quantization parameter whose sign of the corresponding bit number error is opposite to the sign of the accumulated bit number error; if the accumulated bit number error of all the transformed image blocks is equal to zero, Among the two adjacent quantization parameters, a quantization parameter having a smaller absolute value of the corresponding bit number error is selected.
  • an embodiment of the present invention provides an image rate control apparatus applied to a rate conversion technology, including:
  • variable length decoding module 91 is configured to send the quantization parameter and the original number of bits of each image block of the current frame original image, and the total number of image blocks of the current frame original image to the rate control module 92;
  • the rate control module 92 is configured to determine a target number of bits of the current image block according to the information sent by the variable length decoding module 91, the required output code rate, and the original input code rate of the current frame original image, where the target number of bits is The current image block outputs the number of bits when the code rate is output as required;
  • a quantization parameter determining module 93 configured to: in all the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, select a quantization parameter that makes the actual number of bits of the current image block closest to the target number of bits, where The bit number error of each image block is the difference between the actual number of bits of the image block and the corresponding target number of bits;
  • the quantization module 94 is configured to quantize the current image block according to the quantization parameter determined by the quantization parameter determination module 93.
  • the image rate control device applied to the rate conversion technology may further include:
  • the input code rate determining module 95 is configured to determine an original input code rate of the current frame original image, and send the determined original input code rate to the code rate control module 92;
  • the output code rate determining module 96 is configured to determine a required output code rate of the current frame original image, and send the determined required output code rate to the code rate control module 92.
  • an embodiment of the present invention provides an image rate control apparatus applied to an image coding technology, including:
  • the image analysis module 101 is configured to calculate the complexity of each image block of the current frame original image, the total number of bits of the current frame original image, the original input code rate of the current frame original image, and the total number of image blocks of the current frame original image.
  • the information is sent to the rate control module 102.
  • the rate control module 102 is configured to determine a target number of bits of the current image block according to the information provided by the image analysis module 101 and the required output code rate, where the target number of bits is the current image. The number of bits when the block outputs the bit rate as required;
  • a quantization parameter determining module 103 configured to: in all the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks, select a quantization parameter that makes the actual number of bits of the current image block closest to the target number of bits, where The bit number error of each image block is the difference between the actual number of bits of the image block and the corresponding target number of bits;
  • the quantization module 104 is configured to quantize the current image block according to the quantization parameter determined by the quantization parameter determination module 103.
  • the image rate control apparatus applied to the image coding technology may further include:
  • the output code rate determining module 105 is configured to determine a required output code rate of the original image of the current frame, and send the determined required output code rate to the code rate control module 102.
  • the image rate control method and apparatus therefor provided by the embodiment, by selecting the actual number of bits of the current image block to be closest to the target number of bits, among all the quantization parameters that do not increase the accumulated bit number error of all the transformed image blocks
  • the quantization parameter is used to quantize the current image block according to the quantization parameter, so that not only the actual output code rate of each image block output and the required output code rate error are small, but also the actual number of bits of all the transformed image blocks is not increased.
  • the cumulative error with the corresponding target bit number causes the cumulative error of the plurality of image blocks or the entire frame of image data to approach zero, so that the output image quality maintains a high quality.
  • the present invention can be implemented by hardware, or can be added with necessary general hardware by software.
  • the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.).
  • the instructions include a plurality of instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Description

一种图像码率控制方法及装置
技术领域
本发明涉及通信技术领域, 特别涉及图像码率控制方法及装置。 背景技术
现有对于压缩的图像信号, 如 mpeg2 , mepg4以及 H.264图像信 号,经常会根据应用场合和传输信道要求的不同来使用相应的图像变 换技术对其进行变换。 比如为适应信道的传输带宽要求, 使用码率变 换方法来调整图像的码率; 为适应终端和信道带宽的要求, 使用图像 格式变换方法和帧率变换方法来调整图像的大小和帧率。所有这些图 像变换过程都需要对输出码率进行码率控制,以最终达到适应信道和 降低带宽的要求。 而且在图像编码技术中, 图像经过编码压缩输出码 流时,也需要对输出码流进行码率控制,以使输出码率达到指定要求。 因此对图像码率的控制是非常重要的。
现有码率变换技术中的图像码率控制装置控制图像码率的大致 工作原理如下所述: 假设 Ratel为输入的图像码率, Rate2为要求输 出的图像码率, 压缩的图像数据经过解压缩后输出非压缩的 DCT ( Discrete Cosine Transformation , 离散余弦变换) i或图像数据, 该 DCT 域图像数据经过反量化处理, 再经过重新量化, 最后被压缩输 出, 最终实现原始图像码率从 Ratel到 Rate2的变换过程。 确定重新 量化所需的量化参数的方法为:
接收反馈的原始的输入图像码率 Ratel、 要求输出的图像码率 Rate2、 以及原始图像的每个宏块的量化参数 (表示为 Qpi )和每个宏 块的原始比特数(表示为 mb— bitsi ); 根据反馈的 Qpi和 mb— bitsi计 算当前一帧图像的复杂度 X和当前一帧图像的每个宏块的复杂度 xi, 然后计算出转换量化参数参考值 Qt: 根据当前一帧图像中已变换宏 块的复杂度与当前一帧图像的复杂度来计算已变换宏块的目标使用 比特数 R— target— used; 将已变换宏块的目标比特数 R— target— used与 已变换宏块的实际使用比特数 R— act— used进行比较,当两者差距大于 定义的误差范围 threshold, 即 R— target— used - R— act— used > threshold 时, 增大量化参数, 得到重新量化所需的量化参数 OUT— Qp, 将该 OUT Qp 送给变换后续的宏块; 当两者差距小于定义的误差范围 threshold, R— target— used— R— act— used < threshold时, 减小量匕参 数, 得到重新量化所需的量化参数 OUT— Qp, 将该 OUT— Qp送给变 换后续的宏块。
在当前宏块变换完成后更新 X— org和 R— act— used的值, 继续按 照上述方法确定重新量化所需的量化参数,继续根据码率控制模块确 定的量化参数变换后续的宏块。
发明人在实现本发明的过程中, 发现现有实现方式存在以下问 题:
现有的图像码率控制方法在确定当前宏块的量化参数时 ,需借助 预先定义的误差范围: 当已变换宏块的目标比特数与实际使用比特数 的差值大于该误差范围时增大量化参数, 当已变换宏块的目标比特数 与实际使用比特数的差值小于该误差范围时减小量化参数,这样得到 的量化参数只是一个预测值,其并不能完全符合量化当前宏块的实际 要求,因此利用该量化参数量化当前宏块后的输出码率就不一定完全 符合要求,如果输出过高或者过低也只能通过控制后续宏块码率使平 均码率达到要求水平,码率控制具有滞后性。当要求输出码率均匀时, 由于每个宏块的输出码率不精确,所以需要釆用多个宏块的平均来平 滑输出, 使输出码率在要求范围之内, 因此需要较大的输出緩冲区。 为了保证平均输出码率, 部分宏块的量化较大, 从而导致图像质量较 差。 并且输出码流在小时间范围内存在波动, 不能完全满足对码率要 求比较高的场合。 发明内容
本发明实施例提供一种图像码率控制方法及其装置,用以实现对 于每一个图像块都选择合适的量化参数 ,不仅使经过码率变换后的每 一个图像块的实际输出码率与要求输出码率的差值较小, 而且可以减 小累计误差。
本发明实施例提供了一种图像码率控制方法, 包括:
确定当前图像块的目标比特数,所述目标比特数为所述当前图像 块按要求输出码率输出时的比特数;
在不增大所有已变换图像块的累计比特数误差的所有量化参数 中,选择使当前图像块的实际比特数最接近所述目标比特数的量化参 数, 其中, 每一个图像块的比特数误差为图像块的实际比特数与对应 目标比特数之间的差值;
根据确定的所述量化参数对当前图像块进行量化。
本发明实施例提供了一种图像码率控制装置, 包括:
码率控制模块, 用于确定当前图像块的目标比特数, 所述目标比 特数为所述当前图像块按要求输出码率输出时的比特数;
量化参数确定模块,用于在不增大所有已变换图像块的累计比特 数误差的所有量化参数中,选择使当前图像块的实际比特数最接近所 述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为图像 块的实际比特数与对应目标比特数之间的差值;
量化模块,用于根据所述量化参数确定模块确定的量化参数对所 述当前图像块进行量化。
本发明实施例提供了一种应用于码率变换技术中的图像码率控 制装置, 包括:
可变长解码模块,用于将当前帧原始图像每个图像块的量化参数 和原始比特数、以及当前帧原始图像的图像块总数量信息发送给所述 码率控制模块;
码率控制模块, 用于根据所述可变长解码模块提供的信息、要求 输出码率、 以及当前帧原始图像的原始输入码率确定当前图像块的目 标比特数,所述目标比特数为所述当前图像块按要求输出码率输出时 的比特数;
量化参数确定模块,用于在不增大所有已变换图像块的累计比特 数误差的所有量化参数中,选择使当前图像块的实际比特数最接近所 述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为图像 块的实际比特数与对应目标比特数之间的差值;
量化模块,用于根据所述量化参数确定模块确定的量化参数对所 述当前图像块进行量化。
本发明实施例提供了一种应用于图像编码技术中的图像码率控 制装置, 包括:
图像分析模块, 用于将当前帧原始图像的每一个图像块的复杂 度、 当前帧原始图像的总比特数、 当前帧原始图像的原始输入码率、 以及当前帧原始图像的图像块总数量信息发送给所述码率控制模块; 码率控制模块, 用于根据所述图像分析模块提供的信息、及要求 输出码率确定当前图像块的目标比特数,所述目标比特数为所述当前 图像块按要求输出码率输出时的比特数;
量化参数确定模块,用于在不增大所有已变换图像块的累计比特 数误差的所有量化参数中,选择使当前图像块的实际比特数最接近所 述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为图像 块的实际比特数与对应目标比特数之间的差值;
量化模块,用于根据所述量化参数确定模块确定的量化参数对所 述当前图像块进行量化。
本发明实施例提供的图像码率控制方法及其装置,通过在不增大 所有已变换图像块的累计比特数误差的所有量化参数中,选择使当前 图像块的实际比特数最接近所述目标比特数的量化参数,根据该量化 参数对当前图像块进行量化,不仅使输出的每一个图像块的实际输出 码率与要求输出码率误差较小,而且不增大所有已变换图像块的实际 比特数与对应目标比特数的累计误差,使多个图像块或整个一帧图像 数据的累计误差趋近于零, 使输出图像质量保持较高的质量。 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面 将对实施例或现有技术描述中所需要使用的附图作一简单地介绍,显 而易见地, 下面描述中的附图仅仅是本发明的一些实施例, 对于本领 域普通技术人员来讲, 在不付出创造性劳动性的前提下, 还可以根据 这些附图获得其它的附图。
图 1为本发明实施例中图像码率控制方法流程示意图; 图 2为本发明实施例中图像码率控制方法另一流程示意图; 意图;
图 4 为本发明实施例中确定量化当前宏块需用的量化参数的方 法流程示意图;
图 5为本发明又一实施例中图像码率控制方法流程示意图; 图 6为本发明实施例中图像码率控制装置结构流程示意图; 图 7为本发明实施例中量化参数确定模块结构流程示意图; 图 8 为本发明实施例中应用于码率变换技术中的图像码率控制 装置结构示意图;
图 9 本发明实施例中应用于图像编码技术中的图像码率控制装 置结构示意图。 具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方 案进行清楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明一部 分实施例, 而不是全部的实施例。 基于本发明中的实施例, 本领域普 通技术人员在没有作出创造性劳动前提下所获得的所有其它实施例 , 都属于本发明保护的范围。
本发明实施例提出一种图像码率控制方法,可应用于码率变换技 术、 图像编码技术等技术中, 该方法确定当前图像块的目标比特数, 该目标比特数为当前图像块按要求输出码率输出时的比特数;在不增 大所有已变换图像块的累计比特数误差的所有量化参数中,选择使当 前图像块的实际比特数最接近所述目标比特数的量化参数, 其中, 每 一个图像块的比特数误差为图像块的实际比特数与对应目标比特数 之间的差值; 根据确定的量化参数对当前图像块进行量化。 其中, 适 用于该方法的图像块可以是宏块、 条带或者一帧图像。
利用本发明实施例提供的图像码率控制方法,不仅不增大所有已 变换图像块的实际比特数与对应目标比特数的累计误差,进而可以减 小多个图像块或整个一帧图像数据的累计误差,使累计误差逐渐趋近 于零,而且输出的每一个图像块的实际输出码率与要求输出码率误差 较小, 输出图像质量较高。
下面结合附图和具体实施例对本发明实施例提供的图像码率控 制方法进行详细描述:
如图 1所示,为本发明实施例提供的码率控制方法应用于码率变 换技术中时, 对宏块的码率变换进行码率控制的示意图。 输入码率为
Ratel的图像数据经过 VLD ( Variable Length Decoding, 可变长解码) 模块解压缩后输出非压缩的 DCT域图像数据,该数据经过 IQ( Inverse Quantization,反量化)模块进行反量化处理,再经过 Q ( Quantization, 量化)模块重新量化, 最后经过 VLC ( Variable Length Coding, 可 变长编码 )模块压缩输出, 实现原始图像码率从 Ratel到 Rate2的变 换。 其中, Q模块重新量化所需的量化参数按照图 2所示的本发明实 施例提供的图像码率控制方法确定:
S301、向码率控制模块发送确定当前宏块的目标比特数需用的信 息。
输入码率确定模块确定原始输入码率 Ratel , 并将确定的原始输 入码率 Ratel发送给码率控制模块, VLD模块将当前帧原始图像每个 宏块的量化参数和原始比特数、 当前帧原始图像的总比特数、 以及当 前帧原始图像的宏块总数量信息发送给码率控制模块,输出码率确定 模块确定要求输出码率 Rate2 , 并将要求输出码率 Rate2发送给码率 控制模块。
5302、 码率控制模块确定当前宏块的目标比特数。
码率控制模块根据输入码率确定模块、 VLD模块、 及输出码率 确定模块发送来的信息确定当前宏块的目标比特数,该目标比特数为 当前宏块按要求输出码率 Rate2输出时的比特数。
5303、码率控制模块将其确定的当前宏块的目标比特数发送给量 化参数确定模块。
5304、 量化参数确定模块确定量化参数。
量化参数确定模块在不增大所有已变换宏块的累计比特数误差 的所有量化参数中,选择使当前宏块的实际比特数最接近其目标比特 数的量化参数。
5305、 量化模块和可变长编码模块对当前宏块进行变换。
量化模块根据量化参数确定模块确定的量化参数对当前宏块进 行量化,可变长编码模块对量化模块量化后的当前宏块数据进行变长 编码, 输出码流, 实现图像码率从 Ratel到 Rate2的变换。
其中, 在上述步骤 S301 中, 原始输入码率 Ratel也可以由码率 控制模块自己确定而不需输入码率确定模块提供;也可以将要求输出 码率 Rate2直接配置于码率控制模块内存储, 也可以由码率控制模块 自己确定要求输出码率 Rate2, 而不需由输出码率确定模块提供, 这 样可以避免另外设置输入码率确定模块和输出码率确定模块,节省设 备成本。
另外, 在上述步骤 S301 中, 在码率控制模块确定当前宏块的目 标比特数时, 也可以根据要求输出码率、预先设定的当前帧原始图像 比例值、 以及当前帧原始图像的宏块总数量, 确定当前宏块的目标比 特数,上述当前帧原始图像比例值为当前帧原始图像的目标比特数占 每单位时间要求输出比特数的比例值。
并且, VLD模块也可以不将当前帧原始图像的总比特数信息发 送给码率控制模块, 而由码率控制模块根据 VLD模块发送来的当前 帧原始图像每个宏块的原始比特数信息确定当前帧原始图像的总比 特数, 其确定方法为: 当前帧原始图像每个宏块的原始比特数之和。 其中, 如图 3所示, 步骤 S302进一步包括:
5401、 确定当前帧原始图像的目标比特数。
码率控制模块根据原始输入码率 Ratel、 要求输出码率 Rate2和 当前帧原始图像的总比特数, 计算当前帧原始图像的目标比特数, 其 计算方法为: 当前帧原始图像的总比特数和要求输出码率的乘积除以 原始输入码率, 其计算公式( 1 )表示为:
„ , . frame bitslxrate2
frame bits= =
- ratel (1) 其中, frame— bits 为当前帧原始图像的目标比特数, frame— bits 1 为当前帧原始图像的总比特数, Ratel为原始输入码率, Rate2为要求 输出码率。
5402、 确定当前帧原始图像的图像复杂度。
码率控制模块根据当前帧原始图像的每一个宏块的量化参数和 原始比特数, 计算当前帧原始图像的图像复杂度, 其计算方法为: 每 一个宏块的量化参数与原始比特数的乘积之和, 其计算公式(2 )表 示为:
mb_no
X= J (Qpi xmb—bitsi) (2) ι=1
其中, X为当前帧原始图像的图像复杂度, mb— no为当前帧原始 图像的宏块总数量, Qpi为第 i个宏块的量化参数, mb— bitsi为第 i 个宏块的原始比特数。
5403、 确定当前宏块的复杂度。
码率控制模块根据当前宏块的量化参数和原始比特数,计算当前 宏块的复杂度, 其计算方法为: 当前宏块的量化参数与原始比特数的 乘积, 其计算公式(3 )表示为:
xj=Qpjxmb_bitsj (3) 其中, xj 为当前宏块的复杂度, Qpj 为当前宏块的量化参数, mb bitsj为当前宏块的原始比特数。
5404、 确定当前宏块的目标比特数。 码率控制模块根据当前帧原始图像的目标比特数 frame— bits、 当 前帧原始图像的图像复杂度 X、 当前帧原始图像的宏块总数量 mb— no、 及当前宏块的复杂度 xj , 计算当前宏块的目标比特数, 其计 算方法为: 当前宏块的复杂度与当前帧原始图像的目标比特数的乘积 再除以当前帧原始图像的图像复杂度与当前帧原始图像的宏块总数 量的乘积, 其计算公式(4 )表示为:
, , . xix frame bits
mb_bits=― =
_ Xxmb no (4) 其中, mb— bits为当前宏块的目标比特数。
至此, 码率控制模块计算出了当前宏块的目标比特数。
另外, 在码率控制模块确定当前宏块的目标比特数时, 也可以不 执行上述步骤 S401-S404, 而根据要求输出码率、 预先设定的当前帧 原始图像比例值、 以及当前帧原始图像的宏块总数量, 确定当前宏块 的目标比特数; 其确定公式表示为:
mb_bits= B * P (5)
_ mb no
其中, mb— bits为当前宏块的目标比特数, B为每单位时间要 求输出比特数, 其数值等于要求输出码率 Rate2, p为预先设定的当 前帧原始图像比例值; mb— no为当前帧原始图像的宏块总数量。
例如釆用彩色电视制式中的 PAL制输出图像时, 每秒要求输出 25帧图像, 预先设定每秒依序输出的 25帧中, 每一帧图像的目标比 特数占每秒要求输出比特数的比例值分别为 Pl%、 p2% p25%; 假如要求输出码率是 5Mb/s , 当前帧原始图像变换后的图像是当前单 位时间内输出的第 15帧图像, 且 p15%=5%, 当前帧原始图像的宏块 总数量为 1620, 则当前宏块的 p= p15%=5%, B=5Mb, mb_mo=1620, 根据公式 (5)可知, 当前宏块的目标比特数 mb— bits 154.3b。 而当每一 帧图像的目标比特数占每秒要求输出比特数的比例值均相等,即都为 4%时, 则每一帧图像的目标比特数都相等。 这种确定方法可以更加 方便快速地确定当前宏块的目标比特数,进而能更方便快速地确定出 当前宏块的量化参数。 在步骤 S304中, 由于量化参数的个数有限, 一般为 31/51/64个 数值, 且量化参数值越大, 其对应的当前宏块的实际比特数就越小, 故量化参数确定模块可以釆用快速搜索算法来确定量化当前宏块需 用的量化参数。 为方便描述, 当釆用快速搜索算法中的二分法确定量 化当前宏块需用的量化参数时 ,将量化参数对应的实际比特数与当前 宏块的目标比特数的差值简称为比特数误差, 如图 4所示, S304具 体包括:
5501、量化参数确定模块将所有量化参数按一定大小顺序组成量 化参数排列。
5502、量化参数确定模块选择处于量化参数排列中间位置的第一 量化参数, 并计算第一量化参数对应的实际比特数。
计算第一量化参数对应的实际比特数时, 可以釆用 MPEG-2 标 准、 MPEG-FOR标准、 H.264标准等来计算。
S503、 量化参数确定模块计算第一量化参数对应的比特数误差。
5504、确定该比特数误差与 0的大小关系, 当该比特数误差大于 0时, 继续执行步骤 S505;
当该比特数误差小于 0时, 继续执行步骤 S506;
当该比特数误差等于 0时, 继续执行步骤 S507。
5505、 当第一量化参数对应的比特数误差大于 0时, 继续在大于 第一量化参数的半组量化参数排列中选择处于中间位置的第二量化 参数并计算第二量化参数的比特数误差,如果第二量化参数对应的比 特数误差依然大于 0 , 则继续在大于第二量化参数的量化参数排列中 选择处于中间位置的量化参数并计算该量化参数对应的比特数误差, 依次例推, 当找到相邻的两个量化参数, 一个量化参数对应的比特数 误差大于 0、 另一个量化参数对应的比特数误差小于 0时, 继续执行 步骤 S508; 当找到的量化参数对应的比特数误差恰好等于 0时, 执 行步骤 S509。
5506、 当第一量化参数对应的比特数误差小于 0时, 继续在小于 第一量化参数的半组量化参数排列中选择处于中间位置的第二量化 参数并计算第二量化参数对应的比特数误差,如果第二量化参数对应 的比特数误差依然小于 0 , 则继续在小于第二量化参数的量化参数排 列中选择处于中间位置的量化参数并计算该量化参数对应的比特数 误差, 依次例推, 当找到相邻的两个量化参数, 一个量化参数对应的 比特数误差大于 0、 另一个量化参数对应的比特数误差小于 0时, 继 续执行步骤 S508; 当找到的量化参数对应的比特数误差恰好等于 0 时, 执行步骤 S509。
S507、 当第一量化参数对应的比特数误差等于 0时, 选择第一量 化参数为量化当前宏块需用的量化参数, 继续执行步骤 S512。
S508、 判断当前宏块是否是当前帧原始图像的第一个宏块; 如果是, 则执行步骤 S510;
否则, 继续执行步骤 S511。
5509、选择对应比特数误差等于 0的该量化参数, 继续执行步骤 S512。
5510、在上述相邻的两个量化参数中, 选择对应的比特数误差的 绝对值较小的量化参数, 继续执行步骤 S512。
5511、在上述相邻的两个量化参数中, 选择能减小所有已变换宏 块的累计比特数误差的量化参数, 其选择方法为:
当所有已变换宏块的累计比特数误差不等于零时,在上述相邻的 两个量化参数中,选择对应比特数误差的正负号与所有已变换宏块的 累计比特数误差的正负号相反的量化参数, 继续执行步骤 S512; 当所有已变换宏块的累计比特数误差等于零时,在上述相邻的两 个量化参数中, 选择对应的比特数误差的绝对值较小的量化参数, 继 续执行步骤 S512。
例如, 当所有已变换宏块的累计比特数误差为大于 0时, 则在上 述相邻的两个量化参数中, 选择对应比特数误差为小于 0 的量化参 数; 当所有已变换宏块的累计比特数误差为小于 0时, 则在上述相邻 的两个量化参数中, 选择对应比特数误差为大于 0的量化参数; 当所 有已变换宏块的累计比特数误差等于 0时,则在上述相邻的两个量化 参数中,选择对应的比特数误差的绝对值较小的量化参数作为量化当 前宏块需用的量化参数。
S512、量化参数确定模块将选择的量化参数发送给量化模块,将 当前宏块的比特数误差与所有已变换宏块的累计比特数误差相加,得 到新的累计比特数误差。
其中, 除步骤 S501— S512中描述的二分法之外, 也可以利用其 它方法在不增大所有已变换宏块的累计比特数误差的所有量化参数 中, 选择使当前宏块的实际比特数最接近所述目标比特数的量化参 数, 例如釆用黄金分割点法; 或者计算每一个量化参数对应的比特数 误差,然后确定比特数误差的正负号与所有已变换宏块的累计比特数 误差的正负号相反的所有量化参数, 再在确定的这些量化参数中, 选 择对应比特数误差的绝对值最小的量化参数; 或者在步骤 S505中, 当第一量化参数对应的比特数误差大于 0时,继续在除第一量化参数 之外的所有量化参数中选择量化参数, 当找到相邻的两个对应比特数 误差分别大于 0和小于 0的量化参数时, 继续执行步骤 S508; 当找 到对应比特数误差恰好等于 0的量化参数时, 执行步骤 S509; 同样 的在步骤 S506中, 当第一量化参数对应的比特数误差小于 0时, 继 续在除第一量化参数之外的所有量化参数中选择量化参数, 当找到相 邻的两个对应比特数误差分别大于 0和小于 0的量化参数时,继续执 行步骤 S508; 当找到对应比特数误差恰好等于 0的量化参数时, 执 行步骤 S509。 当然还会有很多其它基于本发明实施例量化参数确定 思想的方法, 这里不再——详述。
另外, 在步骤 S501— S512中, 当确定了对应比特数误差等于 0 的量化参数时, 如果所有已变换宏块的累计比特数误差不等于 0 , 则 也可以选择与该对应比特数误差等于 0的量化参数相邻的、且对应比 特数误差与所有已变换宏块的累计比特数误差的正负号相反的量化 参数, 这样便可以减 d、所有已变换宏块的累计比特数误差。
并且, 在步骤 S501— S512中, 在确定了上述相邻的两个量化参 数后,也可以不管当前图像块是否为当前帧原始图像的第一个图像块 (即不执行步骤 S508和 S510 ), 而将第一个图像块对应的所有已变 换宏块的累计比特数误差看成 0, 这样便可以按照步骤 S511 所述的 方法确定任何一个当前图像块对应的量化参数。
本发明实施例一仅说明了对宏块的码率变换进行码率控制的方 法流程,对其它图像块的码率变换进行码率控制的方法流程与本发明 实施例一所述流程一样, 这里不再叙述。
本发明实施例利用本发明实施例提供的图像码率控制方法为量 化模块选择量化参数,不仅使输出的每一个宏块的实际输出码率与要 求输出码率误差较小, 而且不增大所有已变换宏块的累计比特数误 差,使多个宏块或整个一帧图像数据的实际比特数与对应目标比特数 的累计误差趋近于零, 使输出图像质量保持较高的质量, 不会使累计 误差随着已变换宏块数目的增多而增大进而出现部分宏块量化较大、 图像质量较差的现象。 而且, 由于输出的每一个宏块的实际输出码率 与要求输出码率误差较小, 每一个宏块的输出码率较均匀, 因此不需 要较大的输出緩冲区, 且与要求输出码率的差值最小, 符合要求度较 高。 并且输出码流在小时间范围内不存在波动, 能满足对码率要求比 较高的场合。
如图 5所示,为本发明实施例提供的码率控制方法应用于图像编 码技术中时, 对宏块的码率变换进行码率控制的示意图。
在图像编码技术中应用本发明实施例提供的方法控制图像码率 时, 由图像分析模块对输入的原始图像进行分析, 并将分析得到的当 前帧原始图像的每一个宏块的复杂度、 当前帧原始图像的总比特数、 当前帧原始图像的原始输入码率 Ratel、 以及当前帧原始图像的宏块 总数量信息发送给码率控制模块;输出码率确定模块将要求输出码率 Rate2发送给码率控制模块;
码率控制模块根据图像分析模块和输出码率确定模块发送来的 信息, 通过公式(1 )确定当前帧原始图像的目标比特数; 通过以下 公式(6 )计算当前帧原始图像的图像复杂度: X=∑ xi 其中, X为当前帧原始图像的图像复杂度, mb— no为当前帧原始 图像的宏块总数量; 通过公式(4 )或( 5 )确定当前宏块的目标比特 数;
码率控制模块将确定的当前宏块的目标比特数发送给量化参数 确定模块;
量化参数确定模块按前述步骤 S501— S512, 在不增大所有已变 换宏块的累计比特数误差的所有量化参数中,选择使当前宏块的实际 比特数最接近所述目标比特数的量化参数,并将确定的量化参数发送 给量化模块;量化模块根据量化参数确定模块确定的量化参数对当前 宏块进行量化,可变长编码模块对量化模块量化后的当前宏块数据进 行变长编码, 输出码流, 实现图像码率从 Ratel到 Rate2的变换。
至此, 完成了对当前宏块的变换。
其中, 原始输入码率 Ratel也可以由码率控制模块自己统计而不 需图像分析模块提供; 也可以将要求输出码率 Rate2直接配置于码率 控制模块内存储,也可以由码率控制模块自己确定要求输出码率 Rate 2, 而不需由输出码率确定模块提供, 这样可以避免另外设置输出码 率确定模块, 节省设备成本;
另外, 在码率控制模块确定当前宏块的目标比特数时, 也可以根 据要求输出码率、预先设定的当前帧原始图像比例值、 以及当前帧原 始图像的宏块总数量, 确定当前宏块的目标比特数; 其具体确定方法 见前述公式( 5 )。
本发明又一实施例仅说明了对宏块的码率变换进行码率控制的 方法流程,对其它图像块的码率变换进行码率控制的方法流程与本发 明实施例所述流程一样, 这里不再叙述。
本发明实施例利用本发明实施例提供的图像码率控制方法为量 化模块选择量化参数,不仅使输出的每一个宏块的实际输出码率与要 求输出码率误差较小, 而且不增大所有已变换宏块的累计比特数误 差,使多个宏块或整个一帧图像数据的实际比特数与对应目标比特数 的累计误差趋近于零, 使输出图像质量保持较高的质量, 不会使累计 误差随着已变换宏块数目的增多而增大进而出现部分宏块量化较大、 图像质量较差的现象。 而且, 由于输出的每一个宏块的实际输出码率 与要求输出码率误差较小, 每一个宏块的输出码率较均匀, 因此不需 要较大的输出緩冲区, 且与要求输出码率的差值最小, 符合要求度较 高。 并且输出码流在小时间范围内不存在波动, 能满足对码率要求比 较高的场合。
如图 6所示,本发明实施例提供了一种图像码率控制装置,包括: 码率控制模块 71 , 用于确定当前图像块的目标比特数, 所述目 标比特数为所述当前图像块按要求输出码率输出时的比特数;
量化参数确定模块 72 , 用于在不增大所有已变换图像块的累计 比特数误差的所有量化参数中,选择使当前图像块的实际比特数最接 近所述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为 图像块的实际比特数与对应目标比特数之间的差值;
量化模块 73 , 用于根据量化参数确定模块 72确定的量化参数对 所述当前图像块进行量化。
其中, 适用于该装置的图像块可以是宏块、 条带或者一帧图像; 量化参数确定模块 72可以设置于码率控制模块 71内,或者设置于量 化模块 73内。
其中, 码率控制模块 71根据前述公式(4 )或 (5 )确定所述当 前图像块的目标比特数。
如图 7所示, 量化参数确定模块 72进一步包括:
参考量化参数确定子模块 81 , 用于确定对应比特数误差等于 0 的量化参数,或者确定相邻的两个对应比特数误差分别大于 0和小于
0的量化参数;
量化参数确定子模块 82 , 用于在参考量化参数确定子模块 81确 定了对应比特数误差等于 0的量化参数时, 选择该量化参数, 或者选 择与该量化参数相邻的、对应比特数误差的正负号与所述累计比特数 误差的正负号相反的量化参数; 当参考量化参数确定子模块 81确定 了所述相邻的两个量化参数时,选择能减小所述所有已变换图像块的 累计比特数误差的量化参数。
其中, 量化参数确定子模块 82还用于在参考量化参数确定子模 块 81确定了所述相邻的两个量化参数时, 如果所述所有已变换图像 块的累计比特数误差不等于零,在所述相邻的两个量化参数中选择对 应比特数误差的正负号与所述累计比特数误差的正负号相反的量化 参数; 如果所述所有已变换图像块的累计比特数误差等于零, 在所述 相邻的两个量化参数中选择对应比特数误差的绝对值较小的量化参 数。
如图 8所示,本发明实施例提供一种应用于码率变换技术中的图 像码率控制装置, 包括:
可变长解码模块 91 , 用于将当前帧原始图像每个图像块的量化 参数和原始比特数、以及当前帧原始图像的图像块总数量信息发送给 码率控制模块 92;
码率控制模块 92, 用于根据可变长解码模块 91发送来的信息、 要求输出码率、以及当前帧原始图像的原始输入码率确定当前图像块 的目标比特数,所述目标比特数为所述当前图像块按要求输出码率输 出时的比特数;
量化参数确定模块 93 , 用于在不增大所有已变换图像块的累计 比特数误差的所有量化参数中,选择使当前图像块的实际比特数最接 近所述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为 图像块的实际比特数与对应目标比特数之间的差值;
量化模块 94 , 用于根据量化参数确定模块 93确定的量化参数对 所述当前图像块进行量化。
其中,上述应用于码率变换技术中的图像码率控制装置还可以包 括:
输入码率确定模块 95 , 用于确定当前帧原始图像的原始输入码 率, 并将确定的原始输入码率发送给码率控制模块 92; 输出码率确定模块 96 , 用于确定当前帧原始图像的要求输出码 率, 并将确定的要求输出码率发送给码率控制模块 92。
参阅图 9所示,本发明实施例提供一种应用于图像编码技术中的 图像码率控制装置, 包括:
图像分析模块 101 , 用于将当前帧原始图像的每一个图像块的复 杂度、当前帧原始图像的总比特数、当前帧原始图像的原始输入码率、 以及当前帧原始图像的图像块总数量信息发送给码率控制模块 102; 码率控制模块 102 , 用于根据图像分析模块 101提供的信息、 及 要求输出码率确定当前图像块的目标比特数,所述目标比特数为所述 当前图像块按要求输出码率输出时的比特数;
量化参数确定模块 103 , 用于在不增大所有已变换图像块的累计 比特数误差的所有量化参数中,选择使当前图像块的实际比特数最接 近所述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为 图像块的实际比特数与对应目标比特数之间的差值;
量化模块 104 , 用于根据量化参数确定模块 103确定的量化参数 对所述当前图像块进行量化。
其中,上述应用于图像编码技术中的图像码率控制装置还可以包 括:
输出码率确定模块 105 , 用于确定当前帧原始图像的要求输出码 率, 并将确定的要求输出码率发送给码率控制模块 102。
实施例提供的图像码率控制方法及其装置,通过在不增大所有已 变换图像块的累计比特数误差的所有量化参数中,选择使当前图像块 的实际比特数最接近所述目标比特数的量化参数,根据该量化参数对 当前图像块进行量化,不仅使输出的每一个图像块的实际输出码率与 要求输出码率误差较小,而且不增大所有已变换图像块的实际比特数 与对应目标比特数的累计误差,使多个图像块或整个一帧图像数据的 累计误差趋近于零, 使输出图像质量保持较高的质量。
通过以上的实施方式的描述,本领域的技术人员可以清楚地了解 到本发明可以通过硬件实现,也可以可借助软件加必要的通用硬件平 台的方式来实现基于这样的理解,本发明的技术方案可以以软件产品 的形式体现出来, 该软件产品可以存储在一个非易失性存储介质(可 以是 CD-ROM, U盘, 移动硬盘等) 中, 包括若干指令用以使得一 台计算机设备(可以是个人计算机, 服务器, 或者网络设备等)执行 本发明各个实施例所述的方法。
以上公开的仅为本发明的几个具体实施例, 但是, 本发明并非局 限于此,任何本领域的技术人员能思之的变化都应落入本发明的保护 范围。

Claims

权利要求
1、 一种图像码率控制方法, 其特征在于, 包括:
确定当前图像块的目标比特数,所述目标比特数为所述当前图像 块按要求输出码率输出时的比特数;
在不增大所有已变换图像块的累计比特数误差的所有量化参数 中,选择使当前图像块的实际比特数最接近所述目标比特数的量化参 数, 其中, 每一个图像块的比特数误差为图像块的实际比特数与对应 目标比特数之间的差值;
根据确定的所述量化参数对当前图像块进行量化。
2、 如权利要求 1所述的方法, 其特征在于, 所述图像块或者为 宏块、 或者为条带、 或者为一帧图像。
3、 如权利要求 1所述的方法, 其特征在于, 根据如下公式确定 所述当前图像块的目标比特数:
, , . xix frame bits
mb_bits=― =
~~ Xxmb no
其中, 所述 mb— bits 为所述当前图像块的目标比特数, 所述 xj 为所述当前图像块的复杂度, 所述 frame— bits为当前帧原始图像的目 标比特数, 所述 X为当前帧原始图像的图像复杂度, 所述 mb— no为 所述当前帧原始图像的图像块总数量。
4、 如权利要求 1所述的方法, 其特征在于, 根据如下公式确定 所述当前图像块的目标比特数:
. . . B * p
mb_bits= ―
_ mb no
其中, 所述 mb— bits为所述当前图像块的目标比特数, 所述 B为 每单位时间要求输出比特数, 其数值等于所述要求输出码率, 所述 p 为预先设定的当前帧原始图像比例值,所述比例值为当前帧原始图像 的目标比特数占每单位时间要求输出比特数的比例值, 所述 mb— no 为所述当前帧原始图像的图像块总数量。
5、 如权利要求 1所述的方法, 其特征在于, 所述在不增大所有 已变换图像块的累计比特数误差的所有量化参数中,选择使当前图像 块的实际比特数最接近所述目标比特数的量化参数, 进一步包括: 当所有量化参数中存在对应比特数误差等于 0的量化参数,则选 择该量化参数, 或者选择与该量化参数相邻的、对应比特数误差的正 负号与所述累计比特数误差的正负号相反的量化参数; 否则
当所有量化参数中存在相邻的两个对应比特数误差分别大于 0和 小于 0的量化参数,则选择能减小所述所有已变换图像块的累计比特 数误差的量化参数。
6、 如权利要求 5所述的方法, 其特征在于, 当所有量化参数中 存在相邻的两个对应比特数误差分别大于 0和小于 0的量化参数时, 所述能减小所有已变换图像块的累计比特数误差的量化参数的选择 方法为:
如果所述所有已变换图像块的累计比特数误差不等于零,在所述 相邻的两个量化参数中选择对应比特数误差的正负号与所述累计比 特数误差的正负号相反的量化参数;
如果所述所有已变换图像块的累计比特数误差等于零, 在所述相 邻的两个量化参数中选择对应比特数误差的绝对值较小的量化参数。
7、 一种图像码率控制装置, 其特征在于, 包括:
码率控制模块, 用于确定当前图像块的目标比特数, 所述目标比 特数为所述当前图像块按要求输出码率输出时的比特数;
量化参数确定模块,用于在不增大所有已变换图像块的累计比特 数误差的所有量化参数中,选择使当前图像块的实际比特数最接近所 述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为图像 块的实际比特数与对应目标比特数之间的差值;
量化模块,用于根据所述量化参数确定模块确定的量化参数对所 述当前图像块进行量化。
8、 如权利要求 7所述的装置, 其特征在于, 所述图像块或者为 宏块、 或者为条带、 或者为一帧图像。
9、 如权利要求 7所述的装置, 其特征在于, 所述码率控制模块 确定当前图像块的目标比特数时,根据如下公式确定所述当前图像块 的目标比特数:
, , . xix frame bits
mb_bits=― =
~~ Xxmb no
其中, 所述 mb— bits 为所述当前图像块的目标比特数, 所述 xj 为所述当前图像块的复杂度, 所述 frame— bits为当前帧原始图像的目 标比特数, 所述 X为当前帧原始图像的图像复杂度, 所述 mb— no为 所述当前帧原始图像的图像块总数量。
10、 如权利要求 7所述的装置, 其特征在于, 所述码率控制模块 确定当前图像块的目标比特数时,根据如下公式确定所述当前图像块 的目标比特数:
. . . B * p
mb_bits= ―
_ mb no
其中, 所述 mb— bits为所述当前图像块的目标比特数, 所述 B为 每单位时间要求输出比特数, 其数值等于所述要求输出码率, 所述 p 为预先设定的当前帧原始图像的目标比特数占每单位时间要求输出 比特数的比例值, 所述 mb— no为所述当前帧原始图像的图像块总数 量。
11、 如权利要求 7所述的装置, 其特征在于, 所述量化参数确定 模块进一步包括:
参考量化参数确定子模块,用于确定对应比特数误差等于 0的量 化参数,或者确定相邻的两个对应比特数误差分别大于 0和小于 0的 量化参数;
量化参数确定子模块, 用于在所述参考量化参数确定子模块确定 了对应比特数误差等于 0的量化参数时, 选择该量化参数, 或者选择 与该量化参数相邻的、对应比特数误差的正负号与所述累计比特数误 差的正负号相反的量化参数; 当所述参考量化参数确定子模块确定了 所述相邻的两个量化参数时,选择能减小所述所有已变换图像块的累 计比特数误差的量化参数。
12、 如权利要求 11 所述的装置, 其特征在于, 所述量化参数确 定子模块还用于在所述参考量化参数确定子模块确定了所述相邻的 两个量化参数时,如果所述所有已变换图像块的累计比特数误差不等 于零,在所述相邻的两个量化参数中选择对应比特数误差的正负号与 所述累计比特数误差的正负号相反的量化参数;如果所述所有已变换 图像块的累计比特数误差等于零,在所述相邻的两个量化参数中选择 对应比特数误差的绝对值较小的量化参数。
13、 一种应用于码率变换技术中的图像码率控制装置, 其特征在 于, 包括:
可变长解码模块,用于将当前帧原始图像每个图像块的量化参数 和原始比特数、以及当前帧原始图像的图像块总数量信息发送给所述 码率控制模块;
码率控制模块, 用于根据所述可变长解码模块提供的信息、要求 输出码率、 以及当前帧原始图像的原始输入码率确定当前图像块的目 标比特数,所述目标比特数为所述当前图像块按要求输出码率输出时 的比特数;
量化参数确定模块,用于在不增大所有已变换图像块的累计比特 数误差的所有量化参数中,选择使当前图像块的实际比特数最接近所 述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为图像 块的实际比特数与对应目标比特数之间的差值;
量化模块,用于根据所述量化参数确定模块确定的量化参数对所 述当前图像块进行量化。
14、 如权利要求 13所述的装置, 其特征在于, 所述装置进一步 包括:
输入码率确定模块,用于确定所述当前帧原始图像的原始输入码 率, 并将确定的原始输入码率发送给所述码率控制模块;
输出码率确定模块,用于确定所述当前帧原始图像的要求输出码 率, 并将确定的要求输出码率发送给所述码率控制模块。
15、 一种应用于图像编码技术中的图像码率控制装置, 其特征在 于, 包括: 图像分析模块, 用于将当前帧原始图像的每一个图像块的复杂 度、 当前帧原始图像的总比特数、 当前帧原始图像的原始输入码率、 以及当前帧原始图像的图像块总数量信息发送给所述码率控制模块; 码率控制模块, 用于根据所述图像分析模块提供的信息、 及要求 输出码率确定当前图像块的目标比特数,所述目标比特数为所述当前 图像块按要求输出码率输出时的比特数;
量化参数确定模块,用于在不增大所有已变换图像块的累计比特 数误差的所有量化参数中,选择使当前图像块的实际比特数最接近所 述目标比特数的量化参数, 其中, 每一个图像块的比特数误差为图像 块的实际比特数与对应目标比特数之间的差值;
量化模块,用于根据所述量化参数确定模块确定的量化参数对所 述当前图像块进行量化。
16、 如权利要求 15所述的装置, 其特征在于, 所述装置进一步 包括:
输出码率确定模块,用于确定所述当前帧原始图像的要求输出码 率, 并将确定的要求输出码率发送给所述码率控制模块。
PCT/CN2008/072415 2007-09-29 2008-09-18 A control method and device for image code rate WO2009043264A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200710161580.2 2007-09-29
CN 200710161580 CN101127909B (zh) 2007-09-29 2007-09-29 一种图像码率控制方法及装置

Publications (1)

Publication Number Publication Date
WO2009043264A1 true WO2009043264A1 (en) 2009-04-09

Family

ID=39095805

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2008/072415 WO2009043264A1 (en) 2007-09-29 2008-09-18 A control method and device for image code rate

Country Status (2)

Country Link
CN (1) CN101127909B (zh)
WO (1) WO2009043264A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112887725A (zh) * 2021-01-19 2021-06-01 珠海全志科技股份有限公司 参考帧缓存数据压缩方法、计算机装置及计算机可读存储介质
CN114513664A (zh) * 2022-04-18 2022-05-17 鹏城实验室 视频帧编码方法、装置、智能终端及计算机可读存储介质

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101127909B (zh) * 2007-09-29 2010-07-07 华为技术有限公司 一种图像码率控制方法及装置
EP2101503A1 (en) * 2008-03-11 2009-09-16 British Telecommunications Public Limited Company Video coding
CN101552917B (zh) * 2008-03-31 2011-01-05 深圳市融创天下科技发展有限公司 一种视频压缩码率控制方法
JP6120490B2 (ja) 2011-11-07 2017-04-26 キヤノン株式会社 画像符号化装置、画像符号化方法及びプログラム、画像復号装置、画像復号方法及びプログラム
CN106331704B (zh) 2015-07-07 2019-10-22 杭州海康威视数字技术股份有限公司 一种视频码率控制方法及视频编码装置
CN105323556B (zh) * 2015-11-11 2019-05-17 深圳市阿格斯科技有限公司 监控摄像机影像多通道传送方法及装置
WO2018068239A1 (zh) 2016-10-12 2018-04-19 深圳市大疆创新科技有限公司 码率控制的方法、计算机系统和装置
CN106961604B (zh) * 2017-03-07 2018-06-15 腾讯科技(深圳)有限公司 视频编码的码率分配、编码单元码率分配方法及装置
WO2019127136A1 (zh) * 2017-12-27 2019-07-04 深圳市大疆创新科技有限公司 码率控制的方法与编码装置
CN110324615A (zh) * 2018-03-30 2019-10-11 华为技术有限公司 一种码率分配方法及装置
US11551077B2 (en) * 2018-06-13 2023-01-10 International Business Machines Corporation Statistics-aware weight quantization
WO2020150992A1 (zh) * 2019-01-25 2020-07-30 深圳市大疆创新科技有限公司 码率分配的方法与装置
CN112040246B (zh) * 2020-08-27 2022-03-04 西安迪威码半导体有限公司 一种低时延低复杂度的固定码率控制方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5990957A (en) * 1996-09-09 1999-11-23 Samsung Electronics Co., Ltd. Video signal bit amount control using adaptive quantization
EP1182888A2 (en) * 1996-04-25 2002-02-27 Matsushita Electric Industrial Co., Ltd. Video coding with bit rate dependent quantisation
CN1437401A (zh) * 2002-12-23 2003-08-20 乐金电子(沈阳)有限公司 影像转换编码装置
US20030202580A1 (en) * 2002-04-18 2003-10-30 Samsung Electronics Co., Ltd. Apparatus and method for controlling variable bit rate in real time
CN1738423A (zh) * 2005-08-26 2006-02-22 华中科技大学 视频编码比特率的控制方法
CN1941915A (zh) * 2005-09-26 2007-04-04 株式会社东芝 视频编码方法及装置
CN101102494A (zh) * 2007-07-23 2008-01-09 武汉大学 一种视频转码中宏块级码率控制方法
CN101127909A (zh) * 2007-09-29 2008-02-20 华为技术有限公司 一种图像码率控制方法及装置

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1182888A2 (en) * 1996-04-25 2002-02-27 Matsushita Electric Industrial Co., Ltd. Video coding with bit rate dependent quantisation
US5990957A (en) * 1996-09-09 1999-11-23 Samsung Electronics Co., Ltd. Video signal bit amount control using adaptive quantization
US20030202580A1 (en) * 2002-04-18 2003-10-30 Samsung Electronics Co., Ltd. Apparatus and method for controlling variable bit rate in real time
CN1437401A (zh) * 2002-12-23 2003-08-20 乐金电子(沈阳)有限公司 影像转换编码装置
CN1738423A (zh) * 2005-08-26 2006-02-22 华中科技大学 视频编码比特率的控制方法
CN1941915A (zh) * 2005-09-26 2007-04-04 株式会社东芝 视频编码方法及装置
CN101102494A (zh) * 2007-07-23 2008-01-09 武汉大学 一种视频转码中宏块级码率控制方法
CN101127909A (zh) * 2007-09-29 2008-02-20 华为技术有限公司 一种图像码率控制方法及装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112887725A (zh) * 2021-01-19 2021-06-01 珠海全志科技股份有限公司 参考帧缓存数据压缩方法、计算机装置及计算机可读存储介质
CN114513664A (zh) * 2022-04-18 2022-05-17 鹏城实验室 视频帧编码方法、装置、智能终端及计算机可读存储介质

Also Published As

Publication number Publication date
CN101127909A (zh) 2008-02-20
CN101127909B (zh) 2010-07-07

Similar Documents

Publication Publication Date Title
WO2009043264A1 (en) A control method and device for image code rate
JP5410549B2 (ja) 事前計算動き情報を用いたビデオ・エンコード
KR100850705B1 (ko) 시공간적 복잡도를 고려한 적응적 동영상 부호화 방법 및그 장치
JP4555257B2 (ja) 画像符号化装置
JP4987080B2 (ja) 映像符号化装置および方法、映像符号化プログラムおよびそのプログラムを記録した記録媒体
JPH1098726A (ja) 量子化活動度を用いてビット発生量を制御する映像符号化方法及び装置
CN101331773B (zh) 使用速率失真特性进行速率受控视频编码的装置和方法
US7451080B2 (en) Controlling apparatus and method for bit rate
JP4227067B2 (ja) 動画像符号化方法と装置及びプログラム
US8964836B2 (en) Method and apparatus for spatial predictive encoding and/or decoding of video data
JP2015515812A (ja) ビデオ符号化のための量子化係数を提供するための機器および方法
MX2013000355A (es) Dispositivo de procesamiento de imagenes, metodo de procesamiento de imagenes, y programa.
JP2008011431A (ja) 画像符号化装置
US20060088095A1 (en) Image-encoding controlling apparatus for using a table reflecting statistical frequency of quantization parameter selection and method thereof
KR20040007818A (ko) 동영상 부호화를 위한 dct연산량 조절 방법 및 그 장치
RU2587412C2 (ru) Управление скоростью передачи видео на основе гистограммы коэффициентов преобразования
KR20070077955A (ko) 인트라 예측 모드 결정 방법 및 장치
KR20040069445A (ko) 낮은 메모리 대역폭을 갖는 동영상 압축 장치와 그 방법
JP4407249B2 (ja) データ処理装置およびその方法と符号化装置、その方法及びプログラム
KR20180081512A (ko) 인코딩 및 디코딩 방법 및 대응하는 디바이스들
JP2009010608A (ja) 符号化処理装置
JPH09289452A (ja) 変換符号化システムの変換係数選択方法及びその装置
CN112995667B (zh) 一种增强型R-Lambda码率控制方法
US20050141608A1 (en) Pipeline-type operation method for a video processing apparatus and bit rate control method using the same
JP4857152B2 (ja) 直交変換・量子化装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 08800909

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 08800909

Country of ref document: EP

Kind code of ref document: A1