Research Of Joint Source And Channel Coding In The Key Technologies

As the main carrier of information in the modern society, image’s distinguishing featureis the large volume of data. There are two problems when using the digital images:（1） itneeds great memory space;（2） channel capacity is high when it is transmitted. The key pointsolving two problems is to promote the development of the application in digital imaging.This paper studies basic knowledge of the joint source and channel coding, and proposes anew joint source channel coding algorithm combined with wavelet transform andself-organizing map feature algorithm （SOM）, it’s very effective when images are transmittedin harsh environment.The main research contents are as follows:1: This paper elaborates the subject of research’s background and significance, and givesthe generate process the source coding theory and channel coding theory, analyzes the presentstate of the domestic and foreign of the joint source coding、image compression coding basedon wavelet transform and vector quantization algorithm.2: This paper also introduces the basic framework and basic principles of the joint sourcechannel coding, describes the system model of joint coding and the current jointsource-channel coding techniques, and analyzes the image and image coding performanceevaluation methods.3: This paper introduces the detailed theory of source coding, including wavelettransform、compressive sensing theory and vector quantization coding. Describing the basictheory of these methods, including definitions, characteristics, key technologies. It showsthese thee methods’ superiority through simulation experiment.4: This paper introduces the basic knowledge of the channel coding, the channel codingmethods and basic theory, including the characteristics of channel coding, focusing on the RSerror correction code, it’s effectiveness through simulation analysis.5: This paper proposes a joint source coding method based on Daubechies9/7wavelettransform and SOFM algorithm, firstly the image is decomposed by wavelet transformation;then using the cross-band classification vector strategy to construct a new vector, vectorquantization using SOFM algorithm and use the RS code for error correction to the differentcharacteristics of the sub-band wavelet coefficients; finally, it puts the processed data streamthrough the bad channel (BER=10^-2). This method’s experiment results show that the imagecan still be accepted after through the harsh environment, PSNR values have been improvedand the image restoration quality has been much better.