Non-binary Tail-Biting LDPC Convolutional Code Encoding for Image Transmission

In current image transmission, data channel encoding is needed to consider noise and bit-error protection in order to enhance and guarantee image-data quality. This paper focuses on a problem of image transmission with high bit error rate (BER) and additive white Gaussian noise (AWGN) channel, and proposes a method of non-binary tail-biting low-density parity-check convolutional (NB TB-LDPCC) code encoding for image transmission. In this method, image data compressed by JPEG are used to construct finite fields in non-binary, then construct a base matrix in column weight two, and finally create tail biting LDPC convolution code for modulation in the next process in image transmission. The performance of the proposed method is evaluated by 30 images on USC-SIPI Image Database, and the comparison results with conventional method reveal BER is improved at most 1 dB, and PSNR is increased 2.01 dB approximately.