Soft Computing Approaches for Hyperspectral Image Classification

Hyperspectral image classification is one of the most emerging form of image classification. It is able to convey information about an image in a more detailed way as compared to RGB or multispectral data. When spectral measurement is performed using hundreds of narrow contiguous wavelength intervals, the resulting image is called a hyperspectral image. Spectral signature of thousands of materials have been measured in the laboratory and gathered into libraries. Library signatures are used as the basis for identification of materials in Hyperspectral Image (HSI) data. We analyze the spectral signature of the image to extract information. In HSI, each pixel is in fact a high dimensional vector typically containing reflectance measurement from hundreds of continuous narrow band spectral channels (FWHM between 2 and 20) and 400-2500 nm wavelength range. The range of spectrum in HSI data extends beyond the visible range. Hyperspectral data processing comes with many stages such as pre-processing, feature reduction, classification and followed by target detection. Various machine learning and deep learning algorithms have been used to classify HSI data where few of them are Support Vector Machine, Convolutional Neural Network, random forest, SSRN, etc. HSI is being used in variety of fields such as agriculture, mining, food quality, soil types, defense etc.