PROJECT TITLE :
Nonlinear Multiple Kernel Learning With Multiple-Structure-Element Extended Morphological Profiles for Hyperspectral Image Classification
During this paper, we tend to propose a unique multiple kernel learning (MKL) framework to incorporate each spectral and spatial features for hyperspectral image classification, that is termed multiple-structure-part nonlinear MKL (MultiSE-NMKL). In the proposed framework, multiple structure elements (MultiSEs) are used to come up with extended morphological profiles (EMPs) to gift spatial–spectral information. So as to higher mine interscale and interstructure similarity among EMPs, a nonlinear MKL (NMKL) is introduced to be told an optimal combined kernel from the predefined linear base kernels. We tend to integrate this NMKL with support vector machines (SVMs) and scale back the min–max problem to a straightforward minimization problem. The optimal weight for each kernel matrix is then solved by a projection-primarily based gradient descent algorithm. The benefits of using nonlinear combination of base kernels and multiSE-based mostly EMP are that similarity data generated from the nonlinear interaction of different kernels is absolutely exploited, and the discriminability of the categories of interest is deeply enhanced. Experiments are conducted on three real hyperspectral knowledge sets. The experimental results show that the proposed technique achieves higher performance for hyperspectral image classification, compared with several state-of-the-art algorithms. The MultiSE EMPs will provide abundant higher classification accuracy than employing a single-SE EMP.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here