Wireless Image Transmission, Joint Source Channel Coding

With the progress in the fields of computer, telecommunication, networking and multimedia, etc., image communication, an essential component of multimedia communications, is studied extensively. Based on Shannon's separation principle, the separation approach of the source and channel coding often leads to low efficiency especially for image and video transmission. Real-world systems benefit through joint design of source and channel coders, given knowledge of the channel. This thesis proposes a joint source-channel coding scheme for image transmission over error-prone wireless fading channel, which is based on unequal error protection and rate allocation. The main contribution of this thesis is:1) This approach combines a simple but yet elegant concept that joint source-channel rate allocation with UEP to realize a general and effective JSCC system. The method provides unequal protection for improved wavelet zerotree-based SPIHT bit streams with different importance, and adaptively adjusts the source and channel coding rates according to the time-varying characteristic of wireless channel, thus yields both a good performance and a high reliability without adding extra bandwidth. The efficacy, feasible and error-resilient property of the proposed progressive image transmission system has been demonstrated through the comparison of three different transmission systems over Rayleigh fading channel. And especially, the simulation of the proposed scheme shows that it can obviously improve the image quality compared to the EEP and UEP schemes proposed previously when the channel is in bad conditions. So the scheme can be applied to the fluctuant channels with large range.2) The SPIHT algorithm based on the wavelet transform benefits high performance on image compression. But the resulted code cannot resist bit errors. Even one bit error can damage the whole rebuilt image. Here, the error resilience performance of SPIHT algorithm is analyzed and the virtual-subtree algorithm is proposed to resolve this question.3) A new rate compatible punctured Turbo (RCPT) code is to provide UEP for the progressive compressed data according to error sensitiveness, which takes advantage of the superior performance of Turbo code and the property of the rate compatible punctured code. This channel coding based on source utilizes the property of compressed data which generated in order of importance, and improves thelimitation to error sensitivity. The simulation results also show that the method can get better balance between bit error rate and code length, and the image can get better anti-noise property.