Introduction to BRDF effects
Contact Persons
Keywords
Bidirectional Effects, Goniometry, Spectral Radiometry, BRDF
Introduction
In the past bidirectional reflectance effects of the land surface did not play a major role in global change and ecological analysis although they are assumed to be crucial for multi-temporal studies with varying sun incidence angles. This was mainly due to a lack in bidirectional data either available from remote sensing sensors or acquired in field campaigns. In spite of having bidirectional effects in remote sensing imagery for years especially from sensors with large fields of view, the impact of the bidirectional reflectance distribution function (BRDF) is still not well understood. In the near future a highly increased amount of data from sensors with off-nadir capability will be available above all from NASA's spaceborne sensor MISR. In order to apply this data in land use change and ecologically relevant studies, bidirectional ground reference data must be widely available. Such data is usually stated as Bidirectional Reflectance Factors (BRF) from which the BRDF may be inferred. However, due to the involved illumination and observation geometries (solid angles) spectrodirectional ground-based measurements usually consist only of an approximation and the real target specific BRF needs to be retrieved from respective laboratory or field measurements. A laboratory BRF retrieval for LAGOS measurements is described by Dangel et al. (2005). A field BRF retrieval algorithm can be found in Martonchik et al. (1994) and is applied to dual-view field goniometer measurement (dual-view FIGOS) in Schopfer (2008).
Theory
1. What are BRDF effects?
Most surfaces on earth expose a relationship between the amount of reflected radiance and the directions of irradiance and viewing. According to a specific viewing and irradiance geometry an object appears brighter or fainter. This effect is well known from a persian carpet which exposes its glance only in one preferred direction, or from a soccer field where the mow pattern of the lawn can be observed. But it also affects almost any imagery viewed or illuminated from different directions.
Fig. 1: Bidirectional reflectance effect on a mowed lawn as a result of the mow pattern causing different orientation of the grass leaves.
The BRDF effect is most pronounced in the so-called solar principal plane where the source of illumination, target and sensor are in one plane:
Fig. 2: Bidirectional reflectance effect on a grass lawn, observed under different viewing angles from a FIGOS mounted camera in the solar principal plane. Solar zenith angle is 35°, indicated with red arrows. The view directions are given in blue. The camera is operated in the manual modus keeping aperture, exposure time and focal length constant (k=16, t=1/15, f=135mm).
2. How do BRDF effects influence remote sensing data?
Bidirectional effects affect remotely sensed observations in several ways. As soon as viewing and/or irradiance geometries within a single scene, or a series of scenes, are altered, the spectral reflectance signature of the sensed objects changes according to their object-specific BRDFs. Common cases are e.g.:
- varying irradiance geometry caused by different latitudes in a global data set (e.g. from NOAA-AVHRR)
- varying irradiance geometry caused by seasonal effects in a multitemporal data set (e.g. from LANDSAT-TM)
- varying viewing geometry caused by different viewing angles in a multi-directional data set (e.g. from SPOT or MISR)
- varying viewing geometry caused by large field of view (e.g. Airborne scanner or MODIS)
- varying viewing geometry caused by topography in rugged terrain (e.g. LANDSAT-TM imagery taken over the Alps)
Fig. 3: Bidirectional effects influencing
a satellite imagery
Fig. 4: Hot spot effect in an oblique airphotograph. In the lower right corner, the window of the aircraft shades part of the view. (Image courtesy of Dr. J. Privette, NASA/GSFC).
The bidirectional effect in remote sensing data is most obvious in the hot spot direction where illumination and viewing orientation are identical. Fig. 4 demonstrates an oblique airphotograph showing the hot spot effect over a bare soil area.
Why are ground BRF data needed?
Fig. 5: Bidirectional effects are used for validation of models such as the DART model for forests depicted above (taken from Pinel, et al.,1995)
- Validation of current available BRDF models
- Basic research in order to develop new, more accurate BRDF models
- Investigation of relationships between biogeophysical parameters and BRDF effects
- Calibrating reference panels
- Ground reference data for airborne and spaceborne multidirectional sensors like ASAS, MISR and MODIS
Methods of data acquisition
BRF-data can be acquired in the field under natural illumination conditions or in a controlled laboratory campaign. The acquisition from airborne or spaceborne instrument is technically very challenging and usually only can provide partial bidirectional data sets.
Fig. 6: Acquisition of BRF-data: left: EGO, JRC, Ispra, Italy; center: FIGOS, RSL, Zurich, Switzerland; right: PARABOLA, Biospheric Sciences, NASA/GSFC, Greenbelt, MD, USA.
Instruments to take BRF measurements are being built at various institutes. Some examples:
- The Swiss Field and Laboratory Goniometer (FIGOS/LAGOS) project at RSL.