Optical remote sensing involves acquisition and analysis of optical data - electromagnetic radiation captured by the sensing modality after reflecting off an area of interest on ground.  Optical image acquisition modalities have come a long way - from gray-scale photogrammetric images to hyperspectral images. The advances in imaging hardware over recent decades have enabled availability of high spatial, spectral and temporal resolution imagery to the remote sensing analyst. These advances have created unique challenges for researchers in the remote sensing community working on algorithms for representation, exploitation and analysis of such data. Early optical remote sensing systems relied on multispectral sensors, which are characterized by a small number of wide spectral bands. Although multispectral sensors are still employed by analysts, in recent years, the remote sensing community has seen a steady shift to hyperspectral sensors, which are characterized by hundreds of fine resolution co-registered spectral bands, as the dominant optical sensing technology. Such data has the potential to reveal the underlying phenomenology as described by spectral characteristics accurately. This "extension" from multispectral to hyperspectral imaging does not imply that the signal processing and exploitation techniques can be simply scaled up to accommodate the extra dimensions in the data. This book presents state-of-the-art signal processing and exploitation algorithms that address three key challenges within the context of modern optical remote sensing: (1) Representation and visualization of high dimensional data for efficient and reliable transmission, storage and interpretation; (2) Statistical pattern classification for robust land-cover-classification, target recognition and pixel unmixing; (3) Fusion of multi-sensor data to effectively exploit multiple sources of information for analysis.