Advances in remote sensing and computing technologies have led to the development of high-resolution hyperspectral sensors capable of acquiring large volumes of spectral data. Hyperspectral imaging (HSI) captures reflectance information in hundreds of contiguous spectral bands, providing detailed spectral and spatial representations of the Earth’s surface. This rich information makes HSI a valuable tool for land-cover classification and environmental monitoring. Notwithstanding the unparalleled information provided by HSI data, their high spectral dimensionality and the limited availability of labeled samples pose significant challenges for image analysis. These issues often result in overfitting, reduced generalization capability, and increased computational burden. In the last decade, Deep Learning emerged as a promising framework for HSI analysis, offering automated feature extraction and the ability to model complex spectral-spatial patterns. Nevertheless, to effectively apply such techniques, it is essential to understand the intrinsic properties of HSI data. In this context, this chapter introduces the fundamentals of hyperspectral imaging, discusses its main challenges, and provides an overview of the transition from traditional analysis methods to data-driven approaches. It sets the groundwork for understanding how the unique characteristics of HSI influence the design and performance of modern processing techniques.