| Recently, with the rapid development of kernel techniques such as computer, communication and network etc., there is an increasing demand for multimedia service. Wireless image and video transmission, an essential component of wireless multimedia, is studied extensively. It is well known that Shannon's separation theorem does not hold in real-world systems, while joint source-channel coding (JSCC) can be significantly superior to it especially for image and video transmission. Considering the property of source and channel, The JSCC combines the source coder and channel coder reasonably to build an integral system in order to obtain significant improvement of system performance. Thus JSCC is of great interest for practical reasons.In this thesis, we present a JSCC scheme to transmit compressed image over wireless fading channels, which is based on unequal error protection (UEP) and adaptive rate allocation. The main contents of this thesis are as follows.①We analyze the characteristics of wireless fading channels for image transmission on them. Simulation has been done to provide the characteristics of some main wireless channel models with different parameters.②This thesis improves the SPIHT algorithm by encoding low-frequency coefficients of a decomposed image with DPCM method and encoding subtree data using virtual-zerotree structure, respectively. Experimental results show the I-SPIHT algorithm can achieve more excellent performance.③In this thesis, a general and effective JSCC scheme is proposed, which provides a 3-grade UEP strategy for I-SPIHT bit streams of different importance and adjusts the rates adaptively between source and channel encoder according to the characteristic of time-varying wireless channels, thus yields good performance and high reliability without adding extra bandwidth. Experimental results of different systems show that the proposed scheme can obviously improve the image quality after transmitting even if the channels are in bad condition. |
PDF Full Text Request