59217

Análise da correção geométrica por meio de função de refletância bidirecional da superfície em dados de satélite geoestacionário

Favoritar este trabalho

The monitoring of land cover through geostationary satellite sensing images allows greater temporal resolution compared to polar satellites, providing additional information about phenology. However, the bands of red and near infrared, used to obtain much of the environmental index, have anisotropic reflectance characteristics such that the value is dependent of the geometry of observation and vegetation cover. For both are used correction techniques to eliminate this interference, refining the result, however have some limitations. This study evaluated the geometric correction using the BRDF (Bidirectional Reflectance Distribution Function) for obtaining NDVI (Normalized Difference Vegetation Index) daily by data coming from the sensor SEVIRI (Spinning Enhanced Visible and Infrared Imager) of satellite geostationary Meteosat-10, for the state of Goiás. In an evaluation for samples for classes of annual agricultural and pasture NDVI BRDF showed different from maximum, lower and average day NDVI (paired t-test, p <0.001), but is highly correlated (r²> 0.8). The calculation of the BRDF may be an obstacle to the use of high temporal resolution geostationary satellites, because requirement of number of samples per pixel and miscalculation (about 30%); however, there are alternatives that may reduce the effect of these limitations like filtering NDVI and solar azimuth angle values and use BRDF calculation parameters obtained from a previous day