Hyperspectral image classification using support vector machines

Abstract

In this paper, a novel approach for hyperspectral image classification technique is presented using principal component analysis (PCA), bidimensional empirical mode decomposition (BEMD) and support vector machines (SVM). In this process, using PCA feature extraction technique on Hyperspectral Dataset, the first principal component is extracted. This component is supplied as input to BEMD algorithm, which divides the component into four parts, the first three parts represents intrensic mode functions (IMF) and last part shows the residue. These BIMFs and residue image is further taken as input to the SVM for classification. The results of experiments on two popular datasets of hyperspectral remote sensing scenes represent that the proposed-model offers a competitive analyticalperformance in comparison to some established methods.

Unsupervised Classification of Hyperspectral Images using PCA and K-Means

“…This can be overcome by dimensionality reduction techniques e.g. Principal component analysis (PCA), Discrete wavelet transform and Independent component analysis (ICA) [1]. The first few principal components of PCA result in 70 percent correct classification rate [6].…”

“…SVM, a supervised learning method, is effective in high dimensional spaces and is accurate for classification. It requires a smaller number of training samples and labels [1]. However, SVM is computationally expensive.…”

“…first, data dimensionality was reduced, followed by K-means clustering. PCA is a standard dimensionality reduction algorithm for hyperspectral images [12] that aims to remove the correlation among bands [1].…”

“…A. Importing libraries in Python to load and process data  numpy: for multidimensional arrays and matrics [14] and preserves most of the spectral information in a compact number of principal components [1]. It is based on the fact that neighboring bands of hyperspectral images are highly correlated and often convey almost the same information.…”

“…Remote Sensing is the data collection of earth's surface, by measuring the reflected signal at various wavelength bands to form a spectral signature from objects at long distances, via either of the two techniques: infrared sensing or optical remote sensing (via satellite's optical sensor) [1]. Based on the number of bands, satellite images are classified into: panchromatic, multispectral and hyperspectral images.…”