In this paper, a useful lossless compression method of hyperspectral remote sensing images is
presented, which combines three-dimension adaptive predictor (3-DAP) and JPEG-LS algorithm. By
using LDPC codes as the error-correction scheme, a compressed image transmission system with high
spectral efficiency is proposed over Rayleigh fading channel. Linear programming (LP) decoding of
LDPC codes is widely concerned for its maximum likelihood features and the maximum likelihood (ML)
decoding for LDPC codes can be relaxed to an LP optimization problem. Therefore, the paper focus on
an efficient algorithm called infeasible primal-dual interior-point (IPDIP) algorithm for solving LP
problem with predictor-corrector technique. This technique decreases the amount of infeasible points and
keeps them closing to the central path. Simulation results show that the lossless compression algorithm
can reconstruct the hyperspectral image completely with higher compression ratio than JPEG-LS's.
Moreover, the image transmission system achieves good bit error rate (BER) performance and good
global convergence properties with less iteration number and time.