رضا دیانت

The new compressed sensing theory allows the information to be received compactly. In other words it allows the possibility of receiving and compressing the sparse signals simultaneously and optimizationly. The sparse phrase means having low non-zeros values. The with enjoying the spectral and spatial correlation between the bands and pixels of hyperspectral images, compressed sensing theory performs. In this way, with considering one of the two adjacent bands as the reference and the other one as the non-reference band, first the reference bands reconstructed and then for reconstruction of every non-reference bands the adjacent reference band is used. GPSR algorithm is used in reconstructing the reference bands for decoders and for reconstructing the non- reference band another modified GPSR algorithm is used. The adjacent reconstructed reference band is used for creating an initial value for each of the non- reference bands. Using spatial information and blocked image for better image reconstruction the adjacent pixels are used. The simulation results are shown in AVIRIS images the proposed algorithm for the common method of hyperspectral compressed sensing reconstructs all the bands at the same time, this one has a higher compressing operation.