Improved MCA–TV algorithm for interference hyperspectral image decomposition

• MCA is adopted to for interference hyperspectral image decomposition.
• Using adaptive threshold to improve the efficiency of the traditional MCA.
• TV algorithm is improved due to the characteristic of the interference fringes.

The technology of interference hyperspectral imaging, which can get the spectral and spatial information of the observed targets, is a very powerful technology in the field of remote sensing. Due to the special imaging principle, there are many position-fixed interference fringes in each frame of the interference hyperspectral image (IHI) data. This characteristic will affect the result of compressed sensing theory and traditional compression algorithms used on IHI data. According to this characteristic of the IHI data, morphological component analysis (MCA) is adopted to separate the interference fringes layers and the background layers of the LSMIS (Large Spatially Modulated Interference Spectral Image) data, and an improved MCA and Total Variation (TV) combined algorithm is proposed in this paper. An update mode of the threshold in traditional MCA is proposed, and the traditional TV algorithm is also improved according to the unidirectional characteristic of the interference fringes in IHI data. The experimental results prove that the proposed improved MCA–TV (IMT) algorithm can get better results than the traditional MCA, and also can meet the convergence conditions much faster than the traditional MCA.