Spectral Mixture

Yosio Edemir Shimabukuro , Flávio Jorge Ponzoni

In the classification of data and application of remote sensing techniques, one of the challenges faced is the classification of non-uniform pixels, with a mixture of components such as soil, vegetation, rock and water. Spectral mixture: linear model and applications deals with the techniques of characterization of these data and quantification of the mixture in the pixels, allowing its use.


The book explains in a didactic way the basic concepts of spectral mixture, digital numbers and orbital sensors, and then presents the technique of linear model of spectral mixture and the generation of fractional images. In addition to a solid theoretical basis, Spectral Mixture brings important examples of practical application, such as in projects for the estimation and monitoring of areas of deforestation and forest fire in the Legal Amazon.

Original title
Mistura espectral
Year of publication

About the authors

Yosio Edemir Shimabukuro

Yosio Edemir Shimabukuro is a forest engineer graduated from Federal Rural University of Rio de Janeiro, Master in Remote Sensing from National Institute For Space Research (INPE) and PhD in Forest Sciences and Remote Sensing from Colorado State University (USA).

Flávio Jorge Ponzoni

Flávio Jorge Ponzoni is a forest engineer graduated from the Federal University of Viçosa, Master in Forest Sciences from the Federal University of Viçosa and PhD in Forest Sciences from the Federal University of Paraná.




Mixing problem in the pixel


The origin of digital numbers


Orbital sensors


Spot Vegetation

Landsat MSS, TM, ETM+, OLI



Linear model of spectral mixture

Mathematical algorithms

Choice of the endmembers


Fraction images

Error images


Application of fractional images

Deforestation monitoring

Mapping burned areas

Detection of selective cutting

Mapping land use and land cover


Final considerations


Works cited

Recommended literature