Linear Unmixing and Target Detection of Hyperspectral Imagery Using OSP

Proposed technique of dimension reduction, unmixing and target detection is apparent to implement and compute the results in a very fast and efficient manner. Targets Alunite, Buddingtonite, Calcite, Kaolinite, and Muscovite are detected well and have high spectral similarities. To reducing the computational complexity Standard Deviation with respect to correlation distance (STDx - COR) method is used. Number of end-members is enumerating by orthogonal subspace projection (OSP) method. The expectation maximization framework infers the unmixing matrix. Abundance fractions are modeled as a mixture of density functions and it cannot be unmix easily that is why self iteration method is adopted. A set of tests with real hyperspectral data evaluates the performance and illustrates the effectiveness of the proposed method. The experimental results show the effectiveness of the method on hyperspectral data unmixing. Hyperspectral remote sensing is used in a large array of real life applications e.g. Surveillance, Mineralogy, Physics, and Agriculture. The entire work is done by using MATLAB.