Bibliographic citations
Garnica, C., (2019). Segmentación de imágenes de satélite en el estudio de bosques húmedos tropicales de la Reserva Nacional Tambopata, Madre de Dios [Tesis, Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/4051
Garnica, C., Segmentación de imágenes de satélite en el estudio de bosques húmedos tropicales de la Reserva Nacional Tambopata, Madre de Dios [Tesis]. : Universidad Nacional Agraria La Molina; 2019. https://hdl.handle.net/20.500.12996/4051
@mastersthesis{renati/243454,
title = "Segmentación de imágenes de satélite en el estudio de bosques húmedos tropicales de la Reserva Nacional Tambopata, Madre de Dios",
author = "Garnica Philipps, Carlos Rodolfo",
publisher = "Universidad Nacional Agraria La Molina",
year = "2019"
}
The purpose of this research is to provide a new methodology in the elaboration of forest maps to evaluate tropical moist forests, based on the segmentation of Landsat images according to physiography. To do this, forest maps have been prepared, classifying digitally Landsat images segmented (depending on the physiography) and without segmentation. The accuracy of each classification was calculated, based on confusion matrices and kappa indexes, and the best classification method was determined, comparing the accuracy of each one. Likewise, Landsat images were interpreted visually, at a work scale of 1:100 000, in order to obtain a reference pattern for the comparisons made. It was concluded that the segmentation of Landsat images allows to obtain forest maps with greater accuracy than the standard processes of digital classification. In the digital classification methods applied to the segmented Landsat images, it was found that the supervised classification is more accurate than the unsupervised classification for the elaboration of tropical forest maps. However, visual interpretation is more accurate than digital classifications for the elaboration of forest maps of the Peruvian Amazonia.
This item is licensed under a Creative Commons License