A hyperspectral image spectral unmixing method integrating slic superpixel segmentation

Abstract

The number of pixels in a hyperspectral image is one of the main reasons that affects pixel unmixing (especially end-member extraction). Superpixel segmentation techniques can compose adjacent pixels with similar characteristics into image blocks, and retain useful information for further image processing, thereby significantly reduce the number of pixels involved in endmember extraction and solve the problems above in an effective way. In this paper, endmember extraction method has been modified. Simple Linear Iterative Clustering(SLIC) algorithm is used to segment hyperspectral image pixels into superpixel sets, and conducted experimental comparison of the impact of different dimension reduction methods, different data formats and different algorithms' parameters on hyperspectral image superpixel segmentation results. Furthermore, the impact of SLIC superpixel segmentation results on two typical endmember extraction algorithms are analyzed.