Endmember

In the context of remote sensing, geology, and mixture modeling, an endmember represents a pure, idealized component or constituent within a mixed signal or sample. It is the spectrum or property of a single, homogeneous substance or class, free from the influence of other components. The term is used when dealing with mixtures of different materials, where the goal is to determine the proportions of each endmember present in the mixture.

Endmembers are foundational concepts in spectral mixture analysis (SMA). SMA is used to decompose a remotely sensed pixel's spectrum (a mixture) into the spectra of its constituent endmembers. This allows the estimation of the abundance or proportion of each endmember within that pixel.

Defining endmembers is a crucial step in mixture modeling. They can be derived from spectral libraries (collections of known spectra for various materials), field measurements of pure samples, or through statistical techniques applied directly to the data being analyzed. The accuracy of abundance estimations derived from SMA heavily relies on the quality and representativeness of the chosen endmembers.

In fields other than remote sensing, the concept of an endmember extends to any situation where mixtures of distinct components are analyzed. For instance, in geochemistry, endmembers can represent distinct source rocks or geochemical reservoirs contributing to a mixed sample. In ecology, endmembers might represent pure vegetation types contributing to the overall spectral signature of a landscape.

The process of identifying and selecting appropriate endmembers is often challenging. The ideal endmember represents a pure material, but such purity is rarely achieved in real-world samples. Furthermore, the number of endmembers required to adequately describe a given mixture is not always obvious and can significantly impact the results of the analysis.