CN101056414A - A joint coding method for video data stream source and channel - Google Patents

Abstract

The present invention provides a video data stream information source channel joint decoding method, which uses an LDPC decoder to perform channel decoding on the video communication data stream, and then decodes the channel decoding result through the video decoder to obtain a video image, the LDPC decoder and The video decoder joint iterative decoding, the detailed process is as follows: (1) Firstly, the video decoder decodes the channel decoding result, and uses the video protocol syntax and semantics to analyze the syntax unit during the decoding process to detect whether there is a bit error. In the event of a code error, the error range of the video stream is given, and converted into the probability likelihood ratio of the corresponding error information node; (2) the probability likelihood ratio of the corresponding error information node is substituted into the LDPC decoder, and the LDPC decoder Internal iterative decoding; (3) Repeat the next round of joint iterative decoding from step (1) to the newly obtained channel decoding result until the syntax unit of the channel decoding result conforms to the video protocol.

Description

A joint decoding method of source and channel of video data stream

technical field

The invention belongs to the technical field of wireless video communication, and in particular relates to a video data stream source channel joint decoding method.

Background technique

Video service is one of the most important services on the third-generation and fourth-generation mobile communication platforms. In video communication, the most challenging thing is undoubtedly to realize the real-time transmission of video in the wireless network, which must overcome the large amount of video data, delay jitter, transmission bandwidth fluctuation, burst error packet loss, multipath fading and multipath Address reception and so on. The highly compressed video stream is very sensitive to bit errors during transmission. Once a bit error occurs, it will not only affect the recovery of the bit error data, but also affect the recovery of other related data, resulting in "error diffusion" ( Error Propagation), the recovered signal is beyond recognition, seriously affecting the quality of video communication, and improving the bit error correction capability of the communication system is the key to solving the problem.

At present, the communication system is basically under the guidance of Shannon's theorem, and the communication system is divided into source coding and channel coding to design separately. The main purpose is to decompose the global optimization into several local optimizations. The optimization principles of source coding are compression efficiency and restoration quality, while the optimization principles of channel coding are error correction capability and coding cost. In the design of wireless video communication systems, the local optimization principle of Shannon coding theorem has limitations. It makes ideal assumptions on the source and channel: the source satisfies the stability, asymptotically equivalent division, and experience of all states. , the channel has the characteristics of stationary, independent random distribution and so on. But for video communication, the importance of the data of the source is different, the probability of occurrence is also different, and the tolerance for channel error is different. Moreover, Shannon's theorem is an existence theorem, which requires a long channel coding length to achieve reliable communication, and is not suitable for optimizing wireless video communication systems. It can be seen that it is difficult for the existing technology to provide effective technical means to improve the bit error correction capability of the video communication system, which is also called the robustness of the video communication system to channel bit errors in this field.

Contents of the invention

The purpose of the present invention is to solve the deficiencies of the prior art and provide a video data stream decoding method capable of improving the bit error correction capability of the video communication system. In order to achieve the above object, the present invention adopts joint iterative decoding of LDPC decoder and video decoder, and the detailed process of joint iterative decoding is as follows:

(1) First, during the decoding process of the channel decoding result by the video decoder, use the syntax and semantics of the video protocol to analyze the syntax unit to detect whether there is a bit error, and when a bit error is detected, give the error range of the video stream, and Converted to the probability likelihood ratio of the corresponding error information node;

(2) Substituting the probability-likelihood ratio of the corresponding error information node into the LDPC decoder, and performing iterative decoding inside the LDPC decoder;

(3) Repeat the next round of joint iterative decoding from step (1) to the newly obtained channel decoding result until the syntax unit of the channel decoding result conforms to the video protocol.

Moreover, after the video decoder decodes and obtains the video image, it judges whether the image decoding is correct by analyzing the macroblock image. When the image decoding is wrong, return to the joint iterative decoding of the LDPC decoder and the video decoder, and stop the iteration until the image decoding is correct.

Moreover, when no bit error is detected in step (1), it is further confirmed that the channel decoding result is correct, and the probability-likelihood ratio of the corresponding information bit is increased.

Moreover, the video decoder adopts the MPEG4 video protocol.

The technical scheme of the present invention utilizes the analysis information of the bit stream by the video decoder to participate in the iterative decoding process of the LDPC code, forming a joint iterative decoding of the source channel double decoders. This solution improves the error correction capability of the channel through the high-level syntax information of the video decoder, improves the decoding process of the video code stream by the video decoder, prevents the spread of errors, improves the robustness of video communication to channel errors, and ensures wireless communication. Video Communication Quality.

Description of drawings

FIG. 1 is a diagram of the hierarchical relationship between video data information bits and LDPC channel coding in the present invention.

Fig. 2 is a schematic diagram of the joint iterative decoding process of video streams in the present invention.

Fig. 3 is a schematic diagram of a video decoder checking error of the present invention

Detailed ways

Referring to accompanying drawing 2, the technical solution of the video data stream decoding method provided by the present invention is: adopt LDPC decoder to carry out channel decoding to video data stream, then adopt video decoder to decode channel decoding result to obtain video image, joint iterative decoding The process is, LDPC decoder and video decoder joint iterative decoding, the detailed process of joint iterative decoding is,

(1) First, during the decoding process of the channel decoding result by the video decoder, use the syntax and semantics of the video protocol to analyze the syntax unit to detect whether there is a bit error, and when a bit error is detected, give the error range of the video stream, and Converted to the probability likelihood ratio of the corresponding error information node;

(2) Substituting the probability-likelihood ratio of the corresponding error information node into the LDPC decoder, and performing iterative decoding inside the LDPC decoder;

(3) Repeat the next round of joint iterative decoding from step (1) to the newly obtained channel decoding result until the syntax unit of the channel decoding result conforms to the video protocol.

According to the technical solution provided by the present invention, the basic process of encoding and decoding video data streams is: the video data streams are first processed by video encoding and LDPC encoding before being sent, then transmitted by the modulator, accepted by the demodulator after being transmitted through the channel, and then transmitted through the channel. The joint iterative decoding corrects the bit errors, and then the video image is output by the video decoding.

Key improvement of the present invention is that row LDPC decoder and video decoder joint iterative decoding, adopt this scheme based on following considerations:

Practical variable-length codes are generally prefix codes, that is, any codeword is not a prefix of another codeword. Traditional variable-length code decoders use this rule to decode. While reading in the input bits in sequence, they judge whether the read-in bits match a certain codeword. If they match, they output the corresponding symbols. The prefix decoding operation is very simple, and the memory requirement is also small, so it has been widely used. However, in the case of a bit error, if a certain bit read by the prefix decoder is wrong, it may cause a loss of synchronization between the codecs. At this time, not only the current symbol will be wrongly decoded, but also the following symbols will be misinterpreted, and the result of error propagation is sometimes disastrous. At the same time, because each part of the video communication data has different sensitivity to noise, some data is very sensitive to noise. Therefore, the present invention considers using the prior information from the source and the soft output of the channel decoder to perform soft source decoding. The algorithm based on sequence estimation is used to deal with variable length code decoding. The advantage of this method is not only that the decoding performance can be improved for relatively simple prefix decoding, but also it can accept the difference between the soft input bit sequence and the channel decoder. Joint iterative decoding is realized, thereby greatly improving the performance of the decoder. Therefore, the channel coding of the present invention selects practical LDPC codes.

Considering the mechanism of channel coding and source coding, it can be considered as follows: the video encoder compresses and converts the image sequence into a bit sequence according to the video standard, and the video decoder obtains continuous video images by analyzing the bit sequence and decoding. Therefore, it can be seen that the video decoder understands the video code stream through the relationship between the video information bits and the information bits, and the channel decoder correctly understands the bits through the relationship between the information bits and the error correction bits. got it. The two are different understandings of the same code stream, and the understandings of the two will be different. This difference can be eliminated through iteration to achieve the same understanding result and improve the reliability of the understanding. Therefore, using the high-level syntax information of the video decoder for iterative decoding will hopefully further improve the error correction capability of the channel and break through the performance limit of the current irregular LDPC codes.

Therefore, the present invention proposes to use the information exchange between the LDPC decoder and the video decoder to implement joint iterative decoding, so as to improve the decoding result. In each iteration, the probability-likelihood ratio adjusted by the video decoder is substituted into the LDPC decoder, and each iteration utilizes the result obtained in the previous iteration. After several iterations, the LDPC decoder and the video decoder agree on all bit decisions, which finally improves the error correction capability of the communication system.

The above-mentioned basic solution provided by the present invention is to determine whether to jointly iteratively decode based on the discrimination of the bits. The further solution is based on: after the video decoder decodes and obtains the video image, it judges whether the image decoding is correct by analyzing the macroblock image. When the image decoding error Return to the joint iterative decoding of the LDPC decoder and the video decoder, and stop iterating until the image is decoded correctly. The purpose of adding the judgment of joint iterative decoding based on the evaluation of the image is to be able to judge based on the macroblock image, and provide another way to find errors, which is conducive to further improving the error correction capability of the communication system.

Since the position where the error is detected by the decoder is often not the exact position where the error occurs, the decoder needs to adopt a resynchronization strategy to regain synchronization with the encoder after detecting the error. One of the current video coding standards H.26x and MPEG-x adopts one of the resynchronization strategies is to divide each frame of image into several block groups (GOB or Slice), and the decoder can use the start code of GOB as the resynchronization code. However, in this resynchronization strategy, the distance between resynchronization codes is not equal. For parts of the image with violent movement, the codeword interval is relatively large. Once an error occurs, it takes a long time for the decoder to restore synchronization, and more data is affected. It is beneficial to the further error recovery processing of the decoder. MPEG (Moving Picture Experts Group) is an international standard, the so-called ISO11172. A new video packet-based resynchronization strategy is defined in the MPEG-4 protocol, so that the intervals of the resynchronization codewords in the MPEG-4 video stream are basically kept equal. The strategy is that the encoder divides each frame of image into several video packets. The video packet is composed of complete macroblocks, and its length is determined by a preset threshold. Synchronize. Therefore, the video decoder of the present invention adopts the MPEG4 video protocol. MPEG1 and MPEG2 protocols can also be used during specific implementation.

The video data stream decoding method provided by the present invention is specified below so as to implement:

See Figure 1. LDPC code is a special linear block code. Its check matrix H is a sparse matrix. When it is represented by a Tanner graph, there are M nodes C ₁ , C ₂ ... C _M , each check matrix A node represents a check set of a codeword, called a check node; there are N nodes V ₁ , V ₂ ... V _N in the middle of the graph, and each node represents an information bit of a codeword, called a variable node; There are W nodes I ₁ , I ₂ . . . I _W (in the figure, W is 4), and each node represents a syntax symbol of the video compression code stream. There is a connection edge between the check node corresponding to the "1" element in the check matrix and the variable node, and the present invention refers to the nodes at both ends of an edge as adjacent nodes. The number of edges connected to each node is called the degree of the node. Corresponding to the MPEG4 video protocol, there are connection edges between the nodes I ₁ , I ₂ ... I _W and the variable nodes V ₁ , V ₂ ... V _N , reflecting the many-to-one mapping relationship, which means that the information bits pass through a certain The syntactic elements that make up the video in sequence. For regular LDPC codes, the number of "1"s in each row and column of the check matrix is the same; for irregular codes, the degrees of variable nodes and check nodes are not fixed, they are based on a certain distribution selected. The degree distribution of variable nodes and check nodes can be defined as follows:

$\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&chi;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; \&mu;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; \&chi;</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; v</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}{\mathrm{<span\; class="notranslate">\; \&chi;</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}$

μ(χ) represents the proportion of variable nodes with degree i in the total nodes, υ(χ) represents the proportion of check nodes with degree j in the total nodes, where d _v is the number of maximum variable nodes , d _c is the maximum number of check nodes.

The LDPC decoder performs internal decoding. Belief Propagation Algorithm is a probabilistic reasoning algorithm based on graph model for local message passing and updating.

In Fig. 1, the situation where there is a content error in a certain video image is illustrated as an example, the bit error corresponding to the video image content error exists in nodes I ₂ , I ₃ , and then the error range of the video stream is judged accordingly (indicated by a dotted line in the figure), The corresponding error information nodes are nodes V ₅ , V ₆ . . . V _N-3 , V _N-2 . And adjust the probability-likelihood ratio of the corresponding error information node. The above belongs to the process of adjusting according to the video decoding result, and the adjusted probability-likelihood ratio is brought into the LDPC decoder for internal iterative decoding again.

The process of channel decoding performed by an LDPC decoder belongs to the prior art, including the following steps,

Step a: Initialize:

$\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; N</span>}$

Assume that the channel is additive white Gaussian noise channel, where σ ² is the standard deviation of channel noise, q _ij is the external information sent from variable node i to check node j, r _ij is the external information sent from check node j to variable node i external news. R _i is the demodulated information received by the channel, p _i is the initialized log probability likelihood ratio of the corresponding variable node i, for initialization, LLR(r _ij )=0.

Step b: LDPC decoding, including two steps:

Send information from variable nodes to check nodes.

$<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\underset{{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Col</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ]</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span>$

Where Col[i] refers to a non-zero element in column i in the parity check matrix.

From check node to variable node

$\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&Phi;</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\underset{{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; i</span>}}{\underset{{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Row</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; ]</span>}{\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}}\mathrm{<span\; class="notranslate">\; \&Phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$

Wherein Row[j] refers to a non-zero element in row i in the parity check matrix. definition

$\mathrm{<span\; class="notranslate">\; \&Phi;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&chi;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; tanh</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}$

Step c: The channel decoding output is defined as

$\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{{\mathrm{<span\; class="notranslate">\; j</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Col</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ]</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span>$

Then perform hard criterion output:

${\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; u</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\left\{\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; if</span>}& \mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; if</span>}& \mathrm{<span\; class="notranslate">\; LLR</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right.$

Here the hard criterion output is the channel decoding result. where q _i is defined as the final probability likelihood ratio of the iterative output of information node i.

The process in which the video decoder participates in iteration is initiated by the present invention, and the detailed process of checking errors in the embodiment is shown in Figure 3:

First check whether all syntax units conform to the MEPG4 video protocol standard. For some coded video header information including video parameter information: sequence header, picture group header, picture header, it can be directly analyzed according to video syntax and semantics. For example, header information of the MPEG4 protocol: Video Object Header, Video Object Plane Header, User Data Header. Analyze the corresponding video parameter content, and monitor whether errors are found, that is, there are bit errors. In the process of decoding the channel decoding results into images, further decoding is required for the DCT coefficients and motion vectors that contain video content information. In the decoding process, first analyze whether the grammatical units of these variable-length codes conform to the variable-length code table in the video protocol. If there is a discrepancy, there is a bit error and needs to be corrected. In the above inspection process, if a bit error is found, the corresponding wrong information bit is recorded, and converted into the probability likelihood ratio of the corresponding information bit, and then substituted back to the LDPC decoder for corresponding modification. If there is no bit error, you can further confirm that the LDPC decoding is correct, increase the probability likelihood ratio of the corresponding information bit, set the modified coefficient to w, and then substitute the probability likelihood ratio of the entire information bit sequence back to the LDPC decoder , so that the iterative convergence of the LDPC decoder will be better.

Then, in order to be able to distinguish based on the macroblock image, it is necessary to restore the corresponding macroblock image, by analyzing the continuity between the macroblock images, the continuity of the macroblock edge, the continuity of the average gray level of the macroblock, etc., Determine whether the image decoding is correct. This analysis process can be carried out by software, and the present invention is called a decoded image evaluator. If the image decoding is incorrect, it is necessary to determine the corresponding error information bits through analysis, and the corresponding probability likelihood ratio needs to be adjusted, and the LDPC decoder is replaced to make corresponding modifications, that is, the joint iteration of the LDPC decoder and the video decoder is performed again decoding. Since the error correction judgment by image parameters is less reliable than the judgment result by video grammar, the corresponding adjustment value is smaller than the grammar error correction adjustment, and the modified coefficient is set to v, where v<w.

The MPEG4 video protocol standard includes multiple reference modes, and the present embodiment provides the image evaluation process of the I frame and the P frame: for the I frame, after detecting the video parameter content, check the synchronization mark and decode the video packet header, and then decode the MCBPC (macroblock type and chroma coding mode) to obtain the parameter information of the macroblock, analyze the MCBPC, and monitor whether any errors are found; then decode the six block information, analyze the variable-length coding, and monitor whether any errors are found; finally, judge based on the macroblock image, and reconstruct the decoded Image content, verify image content, monitor for errors. After an error is found, record the corresponding wrong information bit, convert it into the probability likelihood ratio of the corresponding information bit, and substitute it back to the LDPC decoder for corresponding modification. The P frame contains motion characteristics, so it is necessary to decode the motion vector while decoding the 6 block information, analyze the motion vector value, and monitor whether errors are found. Generally, B frames are not used in real-time communication, and the evaluation method of B frames should be consistent with that of P frames, which will not be described in detail.

As an iterative algorithm, joint iterative decoding needs to design iteration conditions:

According to the judgment information of the video decoder on channel decoding and the judgment result of the decoded image evaluator, it is decided whether iterative decoding is required; if iteration is required, the probability likelihood ratio adjusted by the video decoder is substituted into the LDPC decoder for new The internal iterative decoding of the channel realizes a joint iterative decoding; if no iterative decoding is required, the joint iterative decoder of the channel (that is, the combination of the LDPC decoder and the video decoder) performs a hard criterion and sends the decoding result to Go to the upper layer protocol stack to complete the decoding process of the channel. The specific algorithm implementation in the joint iterative decoding process can adopt DSP (Digital Signal Processing) technology.

Claims (5)

1. joint coding method for video data stream source and channel, at first adopt ldpc decoder that the video communication data flow is carried out channel decoding, obtain video image by video decoder decodes channel decoding result then, it is characterized in that: ldpc decoder and Video Decoder associating iterative decoding, the detailed process of associating iterative decoding is

(1) at first by the channel decoding result being carried out in the decode procedure at Video Decoder, utilize video protocols grammer, semantic analysis syntactic units, detect whether error code is arranged, when detecting error code, provide the video flowing fault domain, and be converted to the probability likelihood ratio of corresponding information node of makeing mistakes;

(2), carry out the inner iterative decoding of ldpc decoder with the probability likelihood ratio substitution ldpc decoder of corresponding information node of makeing mistakes;

(3) the channel decoding result that newly obtains is repeated to begin to carry out next round associating iterative decoding by step (1), meet video protocols up to channel decoding result's syntactic units.

2. video data stream decoding method according to claim 1, it is characterized in that: after video decoder decodes obtains video image, judge by analyzing macroblock image whether picture decoding is correct, when the picture decoding mistake, return and carry out ldpc decoder and Video Decoder associating iterative decoding, when picture decoding is correct, stop iteration.

3. video data stream decoding method as claimed in claim 1 or 2, it is characterized in that: step detects when not having error code in (1), and further the acknowledgement channel decode results is correct, heightens the probability likelihood ratio of corresponding information bit.

4. video data stream decoding method as claimed in claim 1 or 2, it is characterized in that: Video Decoder adopts MPEG4 video protocols standard.

5. state the video data stream decoding method as claim 3, it is characterized in that: Video Decoder adopts MPEG4 video protocols standard.

Images