Hyperspectral image classification using FPCA-based kernel extreme learning machine

In this paper, the capabilities of functional data feature extraction technique are combined with the advantages of kernel extreme learning machine (KELM), to develop an effective hyperspectral image (HSI) classification method. In the proposed method, the hyperspectral pixels are firstly represented by functions. Each pixel in the HSI is processed from the perspective of function rather than high-dimensional vector. These functional representations are transformed to a lower dimensionality feature space using functional principal components analysis (FPCA). And then the obtained lower dimensional representations are processed by a multiclass KELM classifier. Experimental results on two HSI datasets show that the proposed method provides a relatively promising performance compared with other methods.