Stefan Sandmeier

Radiometric corrections of topographically induced effects on Landsat TM data in an alpine environment

Abstract Four radiometric correction methods for the reduction of slope-aspect effects in a Landsat TM data set are tested in a mountainous test site with regard to their physical soundness and their influence on forest classification, as well as on the visual appearance of the scene. Excellent ground reference information and fine-resolution DEM allowed precise assessment of the applicability of the methods under investigation. The results of the study presented here demonstrate the weakness of the classical cosine correction method for radiometric correction in rugged terrain. The statistical, Minnaert and C-correction approaches, however, yielded an improvement of the forest classification and an impressive reduction of the visual topography effect.

A field goniometer system (FIGOS) for acquisition of hyperspectral BRDF data

A new field goniometer system (FIGOS) is introduced that allows in situ measurements of hyperspectral bidirectional reflectance data under natural illumination conditions. Hyperspectral bidirectional reflectance distribution function (BRDF) data sets taken with FIGOS nominally cover the spectral range between 300 and 2450 nm in 704 bands. Typical targets are small-growing, dense, and homogeneous vegetation canopies, man-made surfaces, and soils. Field BRDF data of a perennial ryegrass surface reveal a strong spectral variability. In the blue and red chlorophyll absorption bands, BRDF effects are strong. Less-pronounced bidirectional reflectance effects are observed in the green and in most of the near-infrared range there surface reflectance is high. An anisotropy index (ANIX), defined as the ratio between the maximum and minimum bidirectional reflectance over the hemisphere, is introduced as a surrogate measurement for the extent of spectral BRDF effects. The ANIX data of the ryegrass surface show a very high correlation with nadir reflectance due to multiple scattering effects. Since canopy geometry, multiple scattering, and BRDF effects are related, these findings may help to derive canopy architecture parameters, such as leaf area index (LAI) or leaf angle distribution (LAD) from remotely sensed hyperspectral BRDF data. Furthermore, they show that normalized difference vegetation index (NDVI) data are strongly biased by the spectral variability of BRDF effects.

The potential of hyperspectral bidirectional reflectance distribution function data for grass canopy characterization

Hyperspectral bidirectional reflectance distribution function (BRDF) data of Konza prairie grassland acquired in the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE) on the ground with two SE-590 instruments and remotely with the airborne advanced solid-state array spectroradiometer (ASAS) are analyzed and compared to BRDF data of dense ryegrass obtained in the laboratory and field with the European goniometric facility (EGO) and the Swiss field-goniometer system (FIGOS). The soil underlying the relatively sparse Konza prairie grass disturbed the spectral BRDF effects of the vegetation components. After a correction of the soil influence based on the bidirectional canopy gap probability, the Konza data from SE-590 and ASAS sensors showed a consistently strong dependence of spectral BRDF effects from nadir reflectance as was observed in the EGO and FIGOS data. BRDF effects were inversely related to reflectance intensities, low reflectances being associated with pronounced BRDF effects and high reflectances with low BRDF effects. This relationship is due to multiple scattering effects and is influenced by the canopy optical properties and architecture parameters such as leaf area index (LAI), leaf angle distribution (LAD), and the gap fraction. The BRDF data of the Konza prairie grass from both ground and aircraft measurements showed a strong relationship between LAI and spectral BRDF variability. Reflectance data with high spectral resolution in the red edge range from 675 to about 900 nm wavelength, acquired from the two viewing directions with maximum and minimum reflectance intensities proved to be useful for deriving vegetation canopy architecture characteristics from hyperspectral BRDF data. BRDF data with high spectral resolution from the airborne ASAS sensor and from planned commercial remote sensing satellites are therefore an ideal testbed for a further exploration of this promising approach.

The Swiss field-goniometer system (FIGOS)

The reflectance characteristics of most natural objects vary with illumination and viewing geometry, i.e. expose a non-Lambertian behaviour. New sensor systems are capable of viewing targets quasi-simultaneously from nadir and different off-nadir positions. In regard to radiometric corrections the Lambertian assumption has to be overcome by detailed knowledge of the bidirectional reflectance distribution function (BRDF) of targets at the Earths surface. In order to obtain bidirectional reflectance factor (BRF) data of naturally illuminated targets a transportable field-goniometer system (FIGOS) has been developed. It is operated together with a GER-3700 spectroradiometer. The goniometer consists of an azimuth full-circle and a zenith semi-arc of 2 m radius each. It enables one to observe a target from any desired viewing direction. First measurements are taken from a plane meadow under different solar zenith angles over the hemisphere in a resolution of 15/spl deg/ and 30/spl deg/ in zenith and azimuth direction, respectively. A Spectralon panel is measured at the beginning and end of a hemispherical data set. As the position of the Sun and the atmospheric conditions cannot be assumed constant over the measurement period of about 18 minutes, the global solar irradiance is monitored simultaneously to the BRF-data acquisition. The obtained results clearly show the non-Lambertian reflectance characteristics of the meadow.

Analyzing hyperspectral BRDF data of a grass lawn and watercress surface using an empirical model

This paper presents results from analyzing hyperspectral BRDF data of grass lawn and watercress. The intensity of the hotspot as a function of wavelength is determined from fitting an empirical (or rather phenomenological) model to the data. The model consists of a lambertian, bowlshape, hotspot and forward scattering component. An analytical, theoretically derived relation between the hotspot intensity and the lambertian component is given. The intensity of the multiple scattered radiation obtained from this relation can be explained qualitatively with the surface structures of the samples.

A shortwave radiation model for radiometric correction of optical satellite data in rugged terrain

An accurate land-use classification in rugged terrain requires precise correction of atmospheric and slope-aspect effects. This study evaluates the potential of a physical-based correction approach in the mountainous test site Beckenried in Central Switzerland. A Landsat-5 Thematic Mapper (TM) image, a fine-resolution digital elevation model (DEM) and precise ground reference data for several land-use classes serve as the database. Atmospheric correction procedures are based on the CCRS-version of the radiative transfer code 5S. Predefined model data in 5S for atmospheric profiles are compared with actual radiosonde measurements. The approach for the retrieval of atmospherically corrected reflectances is evaluated also in regard to topographic illumination correction. The impact of slope and aspect are taken into account by modifying the irradiance components calculated by 5S for flat terrain to those of inclined terrain using the DEM.<<ETX>>

Comparing hyperspectral BRDF data of grass derived from three different laboratory- and field-goniometer systems

The study compares hyperspectral BRDF data of grass vegetation acquired in the laboratory and field with three different goniometric systems. Reflectance measurements of relatively sparse prairie grass were comparable to measurements taken over a dense grass lawn surface only after correction for the soil influences determined by the canopy gap probability function. Nadir-normalized BRDF data consistently revealed a strong wavelength dependence. In addition, a distinctive relationship between BRDF effects and the canopy reflectance signature was observed in all three grass canopies. These phenomena are due to multiple scattering effects which occur within the grass canopy. Since the canopy structure and multiple scattering effects are ultimately related, the potential for deriving canopy architecture parameters from hyperspectral BRDF data is suggested.

Improving land use classification in rugged terrain using radiometric corrections and a possibility based classification approach

The integration of multi-source ancillary data is one of the most promising techniques for improved classification of remote sensing images. In this research a classification strategy based on possibility theory and fuzzy subsets was applied in order to combine spectral and ancillary information. The ancillary information was used to define expert rules on the geographic context of the different land use classes. The method was tested on a complex, rugged area in Central Switzerland. Special emphasis was laid on the pre-processing of the Landsat TM image. A new physical based radiometric correction was applied to the imagery in order to eliminate the impact of atmosphere and slope-aspect effects. Due to the integration of ancillary information by possibility theory a notable accuracy improvement was achieved in comparison to a maximum-likelihood classifier with nonparametric priors.

Alpine and Subalpine Landuse and Ecosystems Mapping

A multisensor multispectral satellite image data set was used to test geometric and radiometric corrections, necessary for landuse and ecosystems mapping in rugged, alpine and subalpine regions. A traditional maximum likelihood classification of ecosystems is discussed versus a clustering after illumination corrections. A forest mapping experiment demonstrates improvements in classification accuracies through slope aspect and atmosphere corrections.