Field spectroradiometry has progressively evolved in both design and mobility
aspects. Modern measurement devices—called spectroradiometers—are nowadays
available, measuring spectra in a wide spectral range (300–2500 nm) with high
precision and accuracy and are portable and easy to use (Milton et al., 2009). In
addition, mobile platforms such as aerial lift trucks that carry spectroradiometers
above a tree canopy are exploited, for example, MUFSPEM@MED from the
Mediterranean Agronomic Institute of Chania (Manevski et al., 2011, 2012), preceded
by MUFSPEM from the Technical University of Munich, or FIGOS from the Remote
Sensing Laboratories at the University of Zurich (Schopfer et al., 2008). Such
technologies provide repeatable spatial sampling with negligible vegetation disturbance. Nevertheless, at the field scale, reaching the top of high land cover targets such
as trees is a challenging task, as not just the instrument has to be suspended and
operated on a height, but factors that influence the field measurements also have to be
simultaneously taken into consideration. Moreover, the definition of the spectral
signatures has proven difficult due to the variable nature of both the material and the
local environment (e.g., atmosphere, moisture content, or illumination) (Cochrane,
2000; Pfitzner et al., 2006).
The rest of this section is focused on field spectroradiometry. First, the main factors
affecting field spectroradiometric measurements are discussed. Then, the development of spectral libraries and the main types of those libraries depending on the
spectra preprocessing performed are outlined. Finally, an overview of the statistical
approaches employed for vegetation species discrimination from field spectroradiometry data is provided, including relevant case studies.
15.2.1 Factors Affecting Field Spectroradiometric Measurements
15.2.1.1 Instrumentation Factors Consideration of the instrumentation factors
consists of the evaluation of the measurement equipment which includes processes
describing the spectroradiometric behavior with respect to the environment and its
calibration. However, the inherent assumption of the illumination characteristics of
the calibration and the target will always introduce certain error. Therefore, calibration
is used to offset as much as possible from the target the changes in reflectance due to
variations in the sun’s natural illumination, the atmospheric influences, and the
scattering effects between the measurements. It is usually done by the use of a surface
with a known reflectance of almost 100% (white panel) which has diffuse reflectance
properties, thus maintaining constant contrast over a wide range of lighting conditions. The most common is the Spectralon
material, a white reference reflector with a
nominal 99% reflectance manufactured by Labsphere (North Sutton, NH). Periodical
reflectance measurements of the white panel on the field can give direct indication of
how “perfect” a diffuse reflector is during a field campaign. Conditions of high
stability and similarity in its spectral responses as well as low variation of the absolute
spectra of the measured targets during a field campaign allow the influence of the
fluctuations caused by the clear-sky weather to be assumed as negligible (Manevski
et al., 2012).
FIELD SPECTRORADIOMETRY
289
aspects. Modern measurement devices—called spectroradiometers—are nowadays
available, measuring spectra in a wide spectral range (300–2500 nm) with high
precision and accuracy and are portable and easy to use (Milton et al., 2009). In
addition, mobile platforms such as aerial lift trucks that carry spectroradiometers
above a tree canopy are exploited, for example, MUFSPEM@MED from the
Mediterranean Agronomic Institute of Chania (Manevski et al., 2011, 2012), preceded
by MUFSPEM from the Technical University of Munich, or FIGOS from the Remote
Sensing Laboratories at the University of Zurich (Schopfer et al., 2008). Such
technologies provide repeatable spatial sampling with negligible vegetation disturbance. Nevertheless, at the field scale, reaching the top of high land cover targets such
as trees is a challenging task, as not just the instrument has to be suspended and
operated on a height, but factors that influence the field measurements also have to be
simultaneously taken into consideration. Moreover, the definition of the spectral
signatures has proven difficult due to the variable nature of both the material and the
local environment (e.g., atmosphere, moisture content, or illumination) (Cochrane,
2000; Pfitzner et al., 2006).
The rest of this section is focused on field spectroradiometry. First, the main factors
affecting field spectroradiometric measurements are discussed. Then, the development of spectral libraries and the main types of those libraries depending on the
spectra preprocessing performed are outlined. Finally, an overview of the statistical
approaches employed for vegetation species discrimination from field spectroradiometry data is provided, including relevant case studies.
15.2.1 Factors Affecting Field Spectroradiometric Measurements
15.2.1.1 Instrumentation Factors Consideration of the instrumentation factors
consists of the evaluation of the measurement equipment which includes processes
describing the spectroradiometric behavior with respect to the environment and its
calibration. However, the inherent assumption of the illumination characteristics of
the calibration and the target will always introduce certain error. Therefore, calibration
is used to offset as much as possible from the target the changes in reflectance due to
variations in the sun’s natural illumination, the atmospheric influences, and the
scattering effects between the measurements. It is usually done by the use of a surface
with a known reflectance of almost 100% (white panel) which has diffuse reflectance
properties, thus maintaining constant contrast over a wide range of lighting conditions. The most common is the Spectralon
material, a white reference reflector with a
nominal 99% reflectance manufactured by Labsphere (North Sutton, NH). Periodical
reflectance measurements of the white panel on the field can give direct indication of
how “perfect” a diffuse reflector is during a field campaign. Conditions of high
stability and similarity in its spectral responses as well as low variation of the absolute
spectra of the measured targets during a field campaign allow the influence of the
fluctuations caused by the clear-sky weather to be assumed as negligible (Manevski
et al., 2012).
FIELD SPECTRORADIOMETRY
289