Sensoriamento remoto hiperespectral: avaliação do sensor R95/HSS para a espacialização e caracterização de solos no município de Manaus

Abstract

HSS (Hyperspectral Scanner System) was evaluated for characterization and spatialization of soils in Manaus AM with 3 meter resolution images. The images (37 bands between 0.43 and 2,37 μm) were converted from radiance values to surface reflectance by the application of (FLAASH) Fast Line of Sight Atmospheric Analysis of Spectral Hypercubes based on radiative transfer model MODTRAN 4. To make easy the analysis of the spectral characteristics of the study area and the classification of the main soil types on the site, they were identified by SAM method (Spectral Angle Mapper) before collecting field soil samples. The correlations between soil characteristics and spectral data of the sensor (HSS) and spectral and diffractometric curves obtained in the laboratory were done for 13 samples. Finally, band math technique was used for spatialization of the clay content by the HSS sensor and laboratory data. The results obtained indicate that: (a) the atmospheric correction done with FLAASH application was appropriate, despite of difficulties in totally remove water vapor feature at 0,94 μm and at short wave infrared band, near to 2,0 μm. In general spectral coherence of HSS sensor with the information found in literature and spectrum obtained in the laboratory was confirmed; (b) a good capacity to spectral identification with SAM classifier was observed as the results provided an appropriate pre identification of soil , as well it helped to define sites for field work; (c) The representative curves of soil classes from laboratory data and sensor data are comparable. The reflectance spectrum presented coherence between X- ray (diffraction) data. The correlations between the two environment of data acquisition (laboratory and HSS) are usually high; (d) The results of ITexture index provided the measurement of the clay content of soil.