Sampling desing on monitoring of multiple- use ofreservoir using cumulative value in oli/Landsat 8 images

Abstract

The remote sensing for studies in water bodies has showed to be of great value for several applications. The determination of the sampling location of field samples is an example, since the satellite images have attributes capable of evaluating the spectral variability of the water surface considering an extensive area. Thus, this paper will define a method of selection of sample elements in Lake Paranoá, located in Brasilia-DF in Brazil, based on variability of images in the visible and infrared spectrum from the Landsat-8 / OLI sensor. The work was based on the selection of images in different seasonal periods that composing a single product called accumulated reflectance. Subsequently an unsupervised classification method was applied, which discriminated five classes with different spectral characteristics. As result, 16 points were obtained, which were the places to obtain limnological variables in places that present different water quality