CN1316829C - Visual frequency flow coding/decoding method based on ripple convertion - Google Patents

Abstract

The present invention discloses a video stream encoding/decoding method based on small wave transformation. The present invention organizes encoding data into code streams according to frames, each frame is composed of three component data of Y, U and V, and the components are respectively arranged on different bit planes according to the quantized precision. The bit plane information of the component data is organized into component information, the information of the three components is organized to form frame information, and thereby, the code streams are finally formed. For the receiving end, the frame information is separated from the received code streams, the component information is separated from the frame information, and the information of various bit planes is separated from the component information and decoded, namely that information of images is restored. The video stream encoding/decoding method based on small wave transformation and a high-efficiency code stream structure ensures that the code streams are freely cut off in the process of transmission, and furthermore, the quality of the restored images is a bit lower but the integral effect of the images can not be affected when the state of a network is poor and the information of some bit planes has errors. The present invention is suitable for frames in an encoding type of intraframe coding to be organized into the code streams and suitable for frames in an encoding type of interframe encoding to be organized into the code streams.

Description

Video flowing coding/decoding method based on wavelet transformation

Technical field

The present invention relates to the coding/decoding method of the video image in Digital Video Processing field, relate in particular to video flowing coding/decoding method based on wavelet transformation.

Background technology

The multimedia technology standard committee set up and propose to be used to encode/standard of compressed audio and video data.H.261/H.263, what this standard comprised the MPEG1/2/4 series standard of moving image and was used for low bit-rate video communication waits standard.

See also Fig. 1, it is the schematic flow sheet of the coding of prior art video flowing.Described coding may further comprise the steps: at first each frame with video flowing carries out format conversion, converts Y to, U, three components of V (step S110); Determine that then each two field picture is I frame, P frame or B frame, and it is carried out intraframe coding or interframe encode (S120); At last, the data behind the coding are organized code stream (S130) according to the code flow structure of standard setting.

Wherein, intraframe coding and interframe encode are two types of video coding.Intraframe coding only utilizes the spatial redundancy in the single image (frame or field), and does not utilize temporal correlation (being the time prediction).And interframe encode uses estimation and motion compensation to reduce time redundancy.Usually, do not use the frame of any interframe encode to be called the I frame, and the frame that interframe encode produces comprise P frame and B frame.The I frame provides code-point for decode procedure, and can provide predicted picture for P frame and B frame.

H.263 wait standard that each frame of video is divided into several macro blocks, each macro block is made up of 4 Y luminance block, 1 Cb chrominance block and 1 Cr chrominance block, and the size of each piece is 8*8.And it is to be the compression that unit carries out frame of video with the macro block.Intraframe coding is the first piece that the brightness and the aberration plane of frame are divided into some 8*8, and each piece carries out dct transform then, uses quantification, entropy coding etc. to obtain coded datas subsequently.Interframe encode is to obtain coded data by estimation, motion compensation, motion vector encoder and residual image coding.Estimation is frame to be divided into several macro blocks put in order pixel motion estimation or half-pix estimation.Macro block after the estimation is divided into piece in interframe block and the frame.Need carry out motion compensation according to its reference frame and present frame for interframe block, and its residual error is carried out dct transform, quantification and entropy coding; For piece in the frame, then directly it is carried out intraframe coding.

Coded data need be organized into code stream and just can be transmitted.To be example H.263, existing code stream institutional framework is described.Code flow structure is divided into image layer (picture layer), gob layer (GOB layer), macroblock layer (macroblock layer) and piece layer (block layer).All comprise a visual head in the data of each frame image in the image layer, followed by the piece group data of this frame, last is the filling position.Each gob layer has comprised a gob layer head, followed by the macro block data of this piece group.The blocks of data that has comprised a macro block head and this macro block in each macroblock layer.That is, in code stream, be that unit organization becomes macro block with the piece, be unit organization becomes the piece group with the macro block, be that unit organization is an image with the piece group.

See also Fig. 2, it is the decoding process schematic diagram of prior art video flowing.Described coding/decoding method may further comprise the steps, and at first, the code stream that receives is obtained frame information (step S210) according to code flow structure; Then, respectively it is carried out the decoding of intraframe decoder or interframe and obtain reducing frame data, to go back original image (step S220); At last, will go back original digital image data and carry out format conversion: be converted to rgb format (step S230) from yuv format.

Above-mentioned video coding/decoding method as can be known, whole cataloged procedure and cataloged procedure all are to be compression and the decompression that unit carries out frame of video with the macro block, cause following defective thus:

First: the code stream that obtains according to above-mentioned video coding mode can not accurately adapt to the bandwidth at networking, can not block code stream arbitrarily;

If current network bandwidth is 1Mb/s, frame per second is 10 frames, and requiring the shared bit number of every two field picture like this is 0.1Mb/s.And the code stream that existing coding method obtains just can not accomplish that every frame only passes 0.1Mb/s.If block code stream in the transmission course, suppose to be provided with in this video flowing 100 macro blocks, if the recipient receives only 80 macro block informations in this video flowing, when then the recipient can only be reduced into two field picture with 80 macro block informations according to existing coding method, obviously can't this width of cloth image of complete demonstration.Therefore, the bandwidth that the code stream that obtains according to above-mentioned video coding mode can not the self adaptation networking can not be blocked code stream arbitrarily.

Second: when the transmission network situation is poor, data on the transmission line may make a mistake, at this moment, and when the code stream that obtains according to existing coding method is sent to the recipient, when the recipient obtains image according to existing coding method, probably mosaic phenomenon can appear.Error bit occurs in this piece group, the macro block of then described error bit back will make a mistake, and will show mistake when causing decoding this piece group, may show that promptly certain delegation partly shows incorrect image, and then the whole image effect of influence.

Summary of the invention

The object of the present invention is to provide a kind of video flowing coding/decoding method based on wavelet transformation, to solve in the prior art, owing to be that the employing piece is that unit organization becomes macro block, with the macro block is that unit organization becomes the piece group, with the piece group is that unit organization is a this code flow structure of image, therefore will encode when afterwards data are organized into code stream according to above-mentioned code flow structure, and can't accurately adapt to the network bandwidth, can not block code stream arbitrarily, when data make a mistake in the image certain delegation partly show technological deficiencies such as incorrect.

For addressing the above problem, the invention discloses a kind of video stream encoding method based on wavelet transformation, comprising:

(1) determines code flow structure, described code flow structure is made up of frame structure, component structure and bit plane structure, described each frame structure comprises frame head and three component structure, each component structure comprises component head and some bit plane structures, described bit plane structure comprises the bit plane head, and this bit plane structure is in order to the bank bit plane information;

(2) each frame of video flowing carries out intraframe coding or interframe encode, and wherein, described intraframe coding comprises carries out wavelet transformation to the view data behind the format transformation, and the coefficient that produces behind the wavelet transformation is carried out the SPIHT coding;

(3) code flow structure that the data after the intraframe coding are set according to step (1) is organized code stream, code stream is organized frame structure according to the frame sequential of video flowing, comprise this frame Y in each frame structure, U, three component information of V comprise in each component information this component are carried out some bit plane information that described SPIHT coding obtains, and, each component structure is in order to preserve one-component information, and each planar structure is in order to preserve a bit plane information.

Also comprise between step (1) and the step (2): if described video flowing is a rgb format, then each frame with video flowing carries out format conversion, converts Y to, U, three components of V.

Step (3) comprises further that also if described frame type of coding is an interframe encode, the motion vector encoder data follow-up that then interframe encode is produced is after described frame head, to organize code stream.

Described frame head comprise at least in order to the frame header field of awareness of identification frame, in order to show frame be the resolution territory of adopting, the frame type territory, frame length territory, frame number territory, image of general video, snap or secret protection, whether adopt arithmetic coding territory, frame type of coding territory and the coding frame per second territory of wishing to adopt the opposite end, keep the territory and be used for after the backward compatible expansion of family release.

Described component head comprises the component leader field of awareness in order to the identification component at least, in order to the component type territory that shows component type, in order to show the high frequency identification field of in wavelet transformation, whether losing high frequency, in order to show the bit plane number field that comprises the bit plane number in the described component, be used to represent wavelet transformation number of plies territory.

Each plane head comprises the bit plane leader field of awareness on discrimination bit plane at least, in order to show the bit-plane data stream length field and the bit plane ID territory of bit-plane data stream length, keep the backward compatible expansion of family release after the territory.

Step (3) also comprises the type of prejudging each frame, if the type of frame is the secret protection type, then comprises a command word that comprises frame head sign, frame type, reserved word and frame length information in the code stream in this frame structure.

A kind of based on method for decoding video stream, comprising:

(A) obtain frame information from the code stream that the receives code flow structure by step (1): retrieval frame head sign finds frame head earlier, retrieve the knowledge of component leader again and find the component head, find the bit plane leader to know at last and find the bit plane head, so that obtain the component information of each frame, the bit plane information in each component;

(B) respectively described frame is carried out the decoding of intraframe decoder or interframe and obtain reducing frame data, and then go back original image, described intraframe decoder comprises and carries out the SPIHT decoding earlier, after carry out wavelet transformation, restore frame data.

Step (B) also comprises: if detect when the frame type of coding is interframe encode in the frame structure of code stream of reception, then carry out the interframe decoding by the motion vector encoder data of searching after the frame head, and then reduce the frame data of described frame.

Step (B) also comprises: if detect in the code stream when frame type is the secret protection type in the frame structure of the code stream receive, then obtain to comprise the frame information of frame head sign, frame type and frame length.

Compared with prior art, the present invention has the following advantages: the present invention is organized into code stream with coded data according to frame, and each frame is with Y, U, and three component datas of V are formed, and each component is provided with different bit planes respectively according to the precision that quantizes.The bit plane information organization that is about to each component data becomes component information, and three component information are organized into frame information, form final code stream.For receiving terminal, as long as will isolate frame information in the code stream that receive, isolate component information in each frame information, each component information is isolated each bit plane information, and it is carried out intraframe decoder, can restore frame information.Guaranteed thus in transmission course, can block code stream arbitrarily.And,,, only can be that the picture quality after the reduction has some to descend, but not influence whole image effect even if mistake appears in some bit plane information even network condition is relatively poor.

And it is that the frame of intraframe coding is organized into code stream that the present invention not only is suitable for type of coding, is applicable to that also type of coding is that the frame of interframe encode is organized into code stream.In the present embodiment, the motion vector encoder that interframe encode is produced is placed on after the frame head as special component, and the frame that can finish type of coding and be interframe encode is organized into code stream, further the scope of application of method for improving.

Description of drawings

Fig. 1 is the schematic flow sheet of the coding of prior art video flowing;

Fig. 2 is the schematic flow sheet of the decoding of prior art video flowing;

Fig. 3 is the schematic flow sheet of the coding of video flowing of the present invention;

Fig. 4 is a code flow structure schematic diagram of the present invention;

Fig. 5 is the decoding process schematic diagram of video flowing of the present invention.

Embodiment

Below in conjunction with accompanying drawing, specify the present invention.

See also Fig. 3, it is a kind of video stream encoding method based on wavelet transformation provided by the invention, it is characterized in that, comprising:

Step S310: determine code flow structure, described code flow structure is made up of frame structure, component structure and bit plane structure, described each frame structure comprises frame head and three component structure, each component structure comprises component head and some bit plane structures, described bit plane structure comprises the plane head, and in order to the bank bit plane information.

Step S320: each frame of video flowing carries out intraframe coding or interframe encode, and wherein, described intraframe coding comprises carries out wavelet transformation to the view data behind the format transformation, and the coefficient that produces behind the wavelet transformation is carried out the SPIHT coding;

Step S330: the data after the intraframe coding are organized code stream according to the code flow structure that step S310 sets, code stream is organized frame structure according to the frame sequential of video flowing, comprise this frame Y in each frame structure, U, three component information of V comprise in each component information this component are carried out the various bit plane information that described SPIHT coding obtains, and, each component structure is in order to preserve a component information, and each planar structure is in order to preserve a bit plane information.

See also Fig. 4, it is a kind of code flow structure schematic diagram.Code stream is made up of some frame structures.Each frame structure is made up of a frame head and some components.In Fig. 4, frame head is to be the beginning of frame with SOF territory (frame head sign).Such as SOF is " 0XFFFFFF91 ", then after the recipient receives code stream, detects the beginning that " 0XFFFFFF91 " data in the code stream are a frame information, is convenient to decoding.Adopt Y in the frame in the present invention, U, three components of V promptly comprise three component structure at least in each frame structure, and each component structure is in order to represent a component information.The TOC territory is the sign that component begins.Also defined in the present embodiment in order to show that component tail (BOC territory) that component finishes is with the postamble (BOF territory) in order to show frame structure.We have a look component structure again, and each component structure comprises component head and some bit plane structures.Define the bit plane leader at present embodiment and known (SOF territory), bit plane tail tag knowledge (EOP territory).In fact, it is can not be default that each leader is known (knowing as frame head sign, the knowledge of bit plane leader, bit plane leader), but each tail is default when being identified at definition.But transmission need organize the code flow structure of code stream and the code flow structure of reception decoded bit stream to answer consistent.

Below be example with a kind of code flow structure of the present invention, the structure of frame head, component head and bit plane head is described.

See also table 1, the information in the various territories of forming frame head is described.

Table 1

SOF	Ptype	RSV	LEN	FrameID	BTOV	BOAC	BPTYPE	BframeRate	Ppara	TUFF
											32	2	6	16	8	3	1	1	5	20	X

The meaning in various territories, and shared figure place, span see Table 2.

Table 2

The territory	Meaning	Length (bit)	Value
				SOF	Sign	32	0xFFFFFF91
Ptype	The snap sign	2	The fast beat of data 11-secret protection of the general video 01-of 00-order
				RSV	Keep, in order to the expansion of code stream in the future	6	000000
LEN	Frame length	16	0～65535bytes
				FrameID	Frame number	8	0～255
BTOV	Image resolution ratio	3	000-160×128 001-320×256

				010-176×144 011-352×288 100-640×512 101-720×480 110-352×240 111-RSV
					BOAC		Option of Arithmetic Coding	1	1-on 0-off
BPTYPE		Frame type	1	0--intraframe coding 1--interframe encode
					BframeRate		The opposite end frame per second	5	00001-1F/s 00010-2F/s 00011-3F/s 00100-4F/s 00101-5F/s 00110-6F/s 00111-7F/s 01000-8F/s ... .11110-30F/s (informing the frame per second that opposite end this end subscriber of encoder is wished with this command word)
Interframe encode parameter Ppara (not using in the actual code)	Vlen	Vector code stream length	16
					Tdef	Haffman code table version identifier	2	00-H263 code table 01-H264 code table
	Bsize	Macroblock size	2	00--4*4 01--8*8 10--16*16
					TUFF		Filler is used for byte-aligned	X	0.....0

Promptly; each frame head comprise at least in order to the frame header field of awareness (SOF) of identification frame, in order to show frame be general video, snap or secret protection frame type territory (Ptype), keep the resolution territory (BTOV) of territory, frame length territory (LEN), frame number territory (FrameID), image adopting, the coding frame per second territory (BframeRate) of whether adopting arithmetic coding territory, frame type of coding territory (BPTYPE) and wishing to adopt the opposite end, filler.

See also table 3, it is the structural representation table of component.

Table 3

TOC	bCPT	bLOW	bQP	bBPN	BREV	RSV
							16	2	3	1	4	3	3

The meaning in various territories, shared figure place, span see Table 4.

Table 4

The territory	Meaning	Length (bit)	Value
				TOC	Sign	16	0xFFA1
bCPT	Components Type	2	00-Y 01-U 10-V 11-Vector
				bLOW	Level of Wavelet	3	000-1 001-2 010-3 011-4 100-5 101-6 110-7 111-8
bQP	Quantization Param	1 (originally 3bits)	0--does not lose high frequency 1--and loses high frequency
				bBPN	Bit Plane Num	4	0～15
bREV	bFlagOrder	3 (originally 4bits)	1-encode_y1 2-encode_y2

			...8-decode_uv2
				RSV	Keep	3	000

That is: each component head comprises the component leader field of awareness (TOC) in order to the identification component at least, in order to the component type territory (bCPT) that shows component type, in order to show the high frequency identification field (bQP) of in wavelet transformation, whether losing high frequency, in order to show the bit plane number field (bBPN) that comprises the bit plane number in the described component, the number of plies of wavelet coding.

The bit plane header structure is as shown in table 5, uses when component type is YUV.

Table 5

SOP	uLEN	BID	RSV
				16	14	4	6

The meaning in each territory, shared figure place and span see Table 6.

Table 6

Sign	Meaning	Length (bit)	Value
				SOP	Sign	16	0xFFD9
ULEN	Bit-plane data stream length	14	0～5000byte
				BID	Bit plane ID number	4	0～15
RSV	Keep	6

That is: each plane head comprise the bit plane leader field of awareness (SOP) on discrimination bit plane at least, in order to show the bit-plane data stream length field (ULEN) and the bit plane ID territory (BID) of bit-plane data stream length.

To adopt above-mentioned disclosed code flow structure is example, specifies the coding/decoding process of video flowing of the present invention.

If the form of view data is RGB, then need to carry out the format conversion yuv format, the mutual conversion of RGB and YUV can be passed through floating-point operation, also can change mutually by integer arithmetic, and is open already in the prior art, do not repeat them here.

Determine the type of each frame then: I frame, P frame or B frame.And the I frame carried out intraframe coding, P frame or B frame are carried out interframe encode.Wherein, intraframe coding comprises: earlier the view data of described yuv format is carried out wavelet transformation, and the coefficient that produces behind the wavelet transformation is encoded according to carrying out SPIHT, so that the data behind the coding can generate code stream, and then output to channel.

Wavelet transformation can adopt two-dimentional bi-orthogonal wavelet transformation, at this moment, can adopt following formula to carry out direct transform during coding:

$s_l^{\left(0\right)}=s_{21}$

$d_l^{\left(0\right)}=s_{2l+1}$

$d_l^{\left(1\right)}=d_l^{\left(0\right)}+\mathrm{\α}(s_l^{\left(0\right)}+s_{l+1}^{\left(0\right)})$

$s_l^{\left(1\right)}=s_l^{\left(0\right)}+\mathrm{\β}(d_l^{\left(1\right)}+d_{l-1}^{\left(1\right)})$

$s_l=k\·s_l^{\left(1\right)}$

$d_l=d_l^{\left(1\right)}/k$

S wherein _l ⁽⁰⁾Be the even samples of one dimension image information S, d _l ⁽⁰⁾Be the odd samples of one dimension picture signal S, s _lBe the even samples after the conversion, d _lBe the Room numerical example after the conversion, α, β are the coefficient of little conversion, and k is the stretching coefficient, wherein α=-0.5, β=0.25, k=1.414.

Each component according to little volume conversion carries out the SPIHT coding.It is the important step of coding that the importance of subclass detects, and in reality was carried out, important information was stored in three orderly tabulations, LIS: non-important set tabulation, LIP: non-important pixel tabulation, LSP: important pixel tabulation.In all tabulations, each enters the mouth all by a coordinates logo, and LIP and LSP represent single pixel ordered list, LIS represent to gather D (i, j) or L (i, ordered list j).For distinguish the set D (i, j) with the set L (i j), is two types with the sets definition among the LIS, type A represent to gather D (i, j), type B represent to gather L (i, j).

In sequencer procedure, pixel among the LIP (the important pixel of process right and wrong in the past) is detected, everyly become important pixel and just be moved among the LSP, set among the same ordered list LIS also detects, when a set is found to be important, this set is just shifted out from the LIS tabulation so, the row set of going forward side by side is divided, the new subclass more than an element that division obtains is put into ordered list LIS again, and the set of having only single-element is added among ordered list LIP or the LSP according to the importance of this element.Pixel among the LSP is continued to do the refinement process to be handled.

Following encryption algorithm uses the set division methods to realize sequencer procedure.

Encryption algorithm:

1 initialization:

Output dominant bit plane exponent number n=[log2 (max (i, j) | Ci, j|})], establishing LSP is empty set, with coordinate (i, j) Join among the set LIP, simultaneously with wherein those have offspring's coordinate to put into LIS as the category-A node.

2 sequencer procedures:

2.1 (i j) encodes to each element among the LIP

2.1.1 output Sn (i, j);

If 2.1.2 Sn (i, j)=1, move so (i, j) in LSP, and output Ci, the symbol of j.

2.2 (i j) encodes to each element among the LIS

2.2.1 if the category-A node, so

Output Sn (D (i, j));

If Sn (D (i, j))=1, so

To any (k, 1)

Output Sn (k, 1);

If Sn (k, 1)=1 joins (k, 1) among the LSP so, and output Ck, 1 symbol;

If Sn (k, 1)=0 joins (k, 1) among the LIP so.

If Ln (i, j) ≠ 0, so as the category-B node will (i j) moves on to the LIS afterbody, then go to2.2.2;

If Ln (i, j)=0, deletion of node from LIS (i, j).

2.2.2 if category-B type node, so

Output Sn (L (i, j));

If Sn (L (i, j))=1, so

As the category-A node with (k, 1) (i j) moves on to the LIS afterbody;

Deletion of node from LIS (i, j).

3. thinning process: to each node among the LSP, except nearest sequencer procedure, output factor | Ci, n the bit of j|.

4. quantize to proofread and correct: the 2nd step of n=n-1 go to.

The bit plane information of each component of each frame that SPIHT coding back obtains is used to organize code stream.

Mainly comprise estimation, motion compensation, motion vector encoder and residual image coding for interframe encode, wherein, the predictive frame that estimation obtains, according to its reference frame and present frame, carry out motion compensation, and its residual image carried out wavelet transformation, use the SPIHT coding then, so that coded data is formed code stream.Be coded in when organizing code stream for the moving dynamic vector of interframe encode, can regard special component structure as, be placed on frame head after.

No matter be interframe encode and intraframe coding, after each frame of video flowing is encoded, when forming code stream according to the frame sequential of video flowing.Whether the type of at first judging described frame is secret protection, if the information of directly importing SOF, Ptype, RSV, LEN territory in frame head gets final product, otherwise, insert the bit plane information of frame head information, three component information and each component.

The method for decoding video stream corresponding of having a talk again with above-mentioned coding.See also Fig. 5, it is the flow chart of video decode.It comprises:

At first carry out step S410: obtain frame information from the code stream that receives by the code flow structure that coding adopts: retrieval frame head sign finds frame head earlier, retrieve the knowledge of component leader again and find the component head, find the bit plane leader to know at last and find the bit plane head, so that obtain the component information of each frame, the bit plane information in each component;

Carry out step S420 again: respectively described frame is carried out the decoding of intraframe decoder or interframe and obtain reducing frame data, and then go back original image, described intraframe decoder comprises and carries out the SPIHT decoding earlier, after carry out wavelet transformation, restore frame data.

Intraframe decoder comprises the coefficient of elder generation by the wavelet transformation of SPIHT decoding acquisition, and then carries out wavelet inverse transformation, at last image transitions is become rgb format.

The SPIHT decoding algorithm:

1 initialization:

Input dominant bit plane exponent number n=[log2 (max (i, j) | Ci, j|})], establishing LSP is empty set, with coordinate (i, j)

Join among the set LIP, simultaneously those are had offspring's coordinate to put into LIS as the category-A node.

2 sequencer procedures:

2.1 (i j) decodes to each element among the LIP

2.1.1 input Sn (i, j);

If 2.1.2 Sn (i, j)=1, move so (i, j) in LSP, and input Ci, the symbol of j.

2.2 (i j) decodes to each element among the LIS

2.2.1 if the category-A node, so

Input Sn (D (i, j));

If Sn (D (i, j))=1, so

To any (k, 1)

Input Sn (k, 1);

If Sn (k, 1)=1 joins (k, 1) among the LSP so, and input Ck, 1 symbol;

If Sn (k, 1)=0 joins (k, 1) among the LIP so.

If Ln (i, j) ≠ 0, so as the category-B node will (i j) moves on among the LIS, then go to 2.2.2;

If Ln (i, j)=0, deletion of node from LIS (i, j).

2.2.2 if category-B type node, so

Input Sn (L (i, j));

If Sn (L (i, j))=1, so

As the category-A node with (k, 1)

(i j) moves on among the LIS;

Deletion of node from LIS (i, j).

3. thinning process: to each node among the LSP, except nearest sequencer procedure, input coefficient | Ci, n the bit of j|.

4. quantize to proofread and correct: n=n-1 go to 2.

Inverse transformation process:

${d_l}^{\left(1\right)}=\mathrm{\κ}\·d_l$

$s_l^{\left(1\right)}=s_l/\mathrm{\κ}$

$s_l^{\left(0\right)}=s_l^{\left(1\right)}-\mathrm{\β}(d_l^{\left(1\right)}+d_{l-1}^{\left(1\right)})$

$d_l^{\left(0\right)}=d_l^{\left(1\right)}-\mathrm{\α}(s_l^{\left(0\right)}+s_{l+1}^{\left(0\right)})$

$x_{21+1}=d_l^{\left(0\right)}$

$s_{2l}=s_l^{\left(0\right)}$

The coefficient of Cai Yonging: a=-0.5 wherein, b=0.25, κ=1.414.

If detect when the frame type of coding is interframe encode in the frame structure of code stream of reception, then carry out the interframe decoding, and then reduce the frame data of described frame by the motion vector encoder data of searching after the frame head.

When if frame type is the secret protection type in the frame structure of the code stream that detection receives in the code stream, then obtain to comprise the frame information of frame head sign, frame type and frame length.

More than disclosed only be several specific embodiment of the present invention, but those skilled in the art can think variation all should drop in protection scope of the present invention.

Claims (10)

1, a kind of video stream encoding method based on wavelet transformation is characterized in that, comprising:

(1) determines code flow structure, described code flow structure is made up of frame structure, component structure and bit plane structure, described each frame structure comprises frame head and three component structure, each component structure comprises component head and some bit plane structures, described bit plane structure comprises the bit plane head, and this bit plane structure is in order to the bank bit plane information;

(2) each frame of video flowing carries out intraframe coding or interframe encode, and wherein, described intraframe coding comprises carries out wavelet transformation to the view data behind the format transformation, and the coefficient that produces behind the wavelet transformation is carried out the SPIHT coding;

(3) code flow structure that the data after the intraframe coding are set according to step (1) is organized code stream, code stream is organized frame structure according to the frame sequential of video flowing, comprise this frame Y in each frame structure, U, three component information of V comprise in each component information this component are carried out some bit plane information that described SPIHT coding obtains, and, each component structure is in order to preserve one-component information, and each planar structure is in order to preserve a bit plane information.

2, the video stream encoding method based on wavelet transformation as claimed in claim 1 is characterized in that, also comprises between step (1) and the step (2): if described video flowing is a rgb format, then each frame with video flowing carries out format conversion, convert Y to, U, three components of V.

3, the video stream encoding method based on wavelet transformation as claimed in claim 1, it is characterized in that step (3) also further comprises, if described frame type of coding is an interframe encode, the motion vector encoder data follow-up that then interframe encode is produced is after described frame head, to organize code stream.

4, as claim 1 or 3 described video stream encoding methods based on wavelet transformation; it is characterized in that, described frame head comprise at least in order to the frame header field of awareness of identification frame, in order to show frame be the branch pigtail rate territory of adopting, the frame type territory, frame length territory, frame number territory, image of general video, snap or secret protection, whether adopt arithmetic coding territory, frame type of coding territory and the coding frame per second territory of wishing to adopt the opposite end, keep the territory and be used for after the backward compatible expansion of family release.

5, the video stream encoding method based on wavelet transformation as claimed in claim 1, it is characterized in that, described component head comprises the component leader field of awareness in order to the identification component at least, in order to the component type territory that shows component type, in order to show the high frequency identification field of in wavelet transformation, whether losing high frequency, in order to show the bit plane number field that comprises the bit plane number in the described component, be used to represent wavelet transformation number of plies territory.

6, the video stream encoding method based on wavelet transformation as claimed in claim 1, it is characterized in that, each plane head comprises the bit plane leader field of awareness on discrimination bit plane at least, in order to show the bit-plane data stream length field and the bit plane ID territory of bit-plane data stream length, keep the backward compatible expansion of family release after the territory.

7, the video stream encoding method based on wavelet transformation as claimed in claim 1; it is characterized in that; step (3) also comprises the type of prejudging each frame; if the type of frame is the secret protection type, then comprise a command word that comprises frame head sign, frame type, reserved word and frame length information in the code stream in this frame structure.

8, a kind of method for decoding video stream based on claim 1 is characterized in that, comprising:

(A) obtain frame information from the code stream that the receives code flow structure by the step the claim 1 (1): retrieval frame head sign finds frame head earlier, retrieve the knowledge of component leader again and find the component head, find the bit plane leader to know at last and find the bit plane head, so that obtain the component information of each frame, the bit plane information in each component;

(B) respectively described frame is carried out the decoding of intraframe decoder or interframe and obtain reducing frame data, and then go back original image, described intraframe decoder comprises and carries out the SPIHT decoding earlier, after carry out wavelet transformation, restore frame data.

9, method for decoding video stream as claimed in claim 8, it is characterized in that, comprise: step (B) also comprises: if detect when the frame type of coding is interframe encode in the frame structure of code stream of reception, then carry out the interframe decoding, and then reduce the frame data of described frame by the motion vector encoder data of searching after the frame head.

10, method for decoding video stream as claimed in claim 8; it is characterized in that; comprise: step (B) also comprises: if detect in the code stream when frame type is the secret protection type in the frame structure of the code stream receive, then obtain to comprise the frame information of frame head sign, frame type and frame length.

Images