Introduction

The book presents methods of dimensionality reduction for remotely sensed hyperspectral imagery. Hence, for better understanding the problem domain and the associated dataset, this chapter introduces the background of remote sensing (RS), which includes a brief discussion on electromagnetic spectrum, atmospheric transmission window, atmospheric scattering, surface reflection, reflectance curve and RS data characteristics. This chapter also includes an overview of hyperspectral remote sensing that describes the representation of hyperspectral image and its differences with multispectral image. Airborne/space-borne imaging spectrometers capture hyperspectral imagery with different spectral ranges and resolutions. The discussion on hyperspectral sensors provides a detailed description of the available hyperspectral data acquired by different sensors. This chapter also introduces the techniques of digital image processing which is an integral part of any digital image analysis task. A brief note on machine learning is also presented in this chapter to provide the concept of supervised and unsupervised learning. These concepts are useful to understand the image classification and clustering tasks that are usually performed on remote sensing imagery.