Advertisement

Final Considerations

  • Yosio Edemir Shimabukuro
  • Flávio Jorge Ponzoni
Chapter
Part of the Springer Remote Sensing/Photogrammetry book series (SPRINGERREMO)

Abstract

This chapter presents some final thoughts about the book.

Keywords

LSMM objectives Future applications 

The spectral mixture can be linear and nonlinear. The linear model was discussed because of the facility of implementation with very satisfactory results.

The linear spectral mixture model is a technique of data transformation of remote sensing data, i.e., converts the spectral information into physical proportion information of the components (endmembers) within the pixel. This information of proportion of the components is represented in new images called fraction images. In that way, the linear spectral mixture model is a data reduction technique, and in addition it enhances the information of these components within the image pixel. It is not a thematic classifier, but provides useful information of fraction images for a variety of applications in several areas.

In general, these endmembers are vegetation, soil, and shade/water elements present on the ground. The vegetation fraction image presents similar information of vegetation indices such as NDVI, SAVI, and EVI, highlighting the vegetation cover areas, while the soil fraction image highlights the areas without vegetation cover, and the shade/water fraction image highlights the water bodies and the burned areas.

The soil and shade/water fraction images were important for automating the PRODES project, which was done through the digital PRODES project, providing the estimate of deforested areas and the map of spatial distribution of these areas.

Hopefully, at the end of this book, we have contributed to the provision of useful information to the deepest reflections for those who intend to use the fraction images derived from linear spectral mixture model in the development of their works.

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Yosio Edemir Shimabukuro
    • 1
  • Flávio Jorge Ponzoni
    • 1
  1. 1.Remote Sensing DivisionNational Institute for Space ResearchSão José dos CamposBrazil

Personalised recommendations