Spectral signatures of the coverage of the Earth, using radiometry field. Phase 1: Region 03 western Honduras

Abstract

The research formed part of a national project whose general purpose is to “Create a spectral signatures catalog nationwide, divided by region (according to National Plan / Country Vision) and characterize land cover according to the FAO classification system, using ASD FieldSpec Handheld with a visible-near infrared (VNIR) range” to reduce errors by atmospheric dispersion and particulate matter as well as ground radiometric contribution. It started from the fact that the main problem is the lack of a spectral library of land cover, the National Project of Spectral Signatures has as general hypothesis that: The spectral signatures of different land cover acquired in field with a spectroradiometer are similar to the orbital spectral signatures of the image corrected by Rayleigh scattering, as long as the image has been acquired at the same observation date. As results, Pine Forest, Thicket, Agriculture, Pasture and Soil spectral signatures were obtained. For dominant vegetable toppings Normalized Difference Vegetation Index was applied to identify the intensity of biomass regarding phenological conditions. At the same time an adjustment to soil conditions was performed to identify the contribution of background reflectance (ground) introduced into the spectral signature of vegetation. In the analysis of the correlation between radiometric measures for image corrected atmospherically and field spectroradiometry, perform with both coverage that provided the most difficult separation, cafetal and thicket, they provided a high correlation coefficient, with a R2 of 0.99 in both cases, which proved the hypothesis. The study concluded that spectral signatures are useful to calibrate images without radiometric corrections.

Revista Ciencias Espaciales, Vol.7(1) 2014, 76-96