P. Boccardo : Environmental Impact Analysis Using Remotely Sensed Data
151
by Landsat's MSS and T.M. sensors), or using different images (such as imagery
in the microwave spectral band, digital conversions of aerial photographs in the
near infrared spectral band, etc.). The two different methods of integration,
although extremely different, present several common characteristics that are
analyzed in the following paragraphs.
2
DIGITAL IMAGE PROCESSING TECHNIQUE
2.1 Data acquisition
Data acquisition essentially depends on two factors: scale and spectral bands. Scale
is intrinsically related to the acquisition system (sensor) geometrical resolution and
allows one to obtain a quantitative measure of the minimum ground details that
can be identified. IFOV (Instantaneous Field Of View) is not always that which can
be detected on the ground; in fact, some features (such as highways, railroads,
mirrors and comer reflectors) though smaller than the minimum sensor geometric
resolution, register a consistent radiant flux on the sensor (a radiant flux that is
considerably different from that of the surrounding area), leaving a "trace" of their
presence on the digital image.
Nowadays, the geometric resolution of the satellite acquisition systems does not
allow large scale applications, being suitable for medium scale surveys (1:25.000,
1 :50.000). Several new sensors are due to be launched in the near future, thus
improving geometric resolution to a few meters (bib Washington).
To overcome resolution shortcomings, one can often use aerial imagery, acquired
both in a direct digital form (scanner mounted on board of airborne platform) and
hardcopy scanned (aerial photograph).
Spectral resolution (the possibility of acquiring signals in different bands of the
electromagnetic spectrum) is particularly important in the feature recognition
process and in the classification of environmental phenomena. It is obvious that
the greater the spectral resolution, the greater the possibility of extracting
information from the data. This extraction is also related to the radiometric
resolution concept, that is the capacity to (quantify) the radiant flux that comes
from the ground in a certain number of radiometric classes, with the subsequent
storage of these classes in digital data fraction (bits).
2.2 Data pre-processing
The data available must be previously corrected in order to eliminate two types of
distortion: radiometric and geometric distortion.
Radiometric distortion refers to the scanners' calibration instability and to the
interactions between the electromagnetic energy, the atmosphere and the ground.
Geometric distortions refer to both the relative position of the acquisition platform
with respect to the atmosphere and the ground (terrestrial rotation and curvature
effect, reflection and refraction phenomena due to the different atmospheric layer
151
by Landsat's MSS and T.M. sensors), or using different images (such as imagery
in the microwave spectral band, digital conversions of aerial photographs in the
near infrared spectral band, etc.). The two different methods of integration,
although extremely different, present several common characteristics that are
analyzed in the following paragraphs.
2
DIGITAL IMAGE PROCESSING TECHNIQUE
2.1 Data acquisition
Data acquisition essentially depends on two factors: scale and spectral bands. Scale
is intrinsically related to the acquisition system (sensor) geometrical resolution and
allows one to obtain a quantitative measure of the minimum ground details that
can be identified. IFOV (Instantaneous Field Of View) is not always that which can
be detected on the ground; in fact, some features (such as highways, railroads,
mirrors and comer reflectors) though smaller than the minimum sensor geometric
resolution, register a consistent radiant flux on the sensor (a radiant flux that is
considerably different from that of the surrounding area), leaving a "trace" of their
presence on the digital image.
Nowadays, the geometric resolution of the satellite acquisition systems does not
allow large scale applications, being suitable for medium scale surveys (1:25.000,
1 :50.000). Several new sensors are due to be launched in the near future, thus
improving geometric resolution to a few meters (bib Washington).
To overcome resolution shortcomings, one can often use aerial imagery, acquired
both in a direct digital form (scanner mounted on board of airborne platform) and
hardcopy scanned (aerial photograph).
Spectral resolution (the possibility of acquiring signals in different bands of the
electromagnetic spectrum) is particularly important in the feature recognition
process and in the classification of environmental phenomena. It is obvious that
the greater the spectral resolution, the greater the possibility of extracting
information from the data. This extraction is also related to the radiometric
resolution concept, that is the capacity to (quantify) the radiant flux that comes
from the ground in a certain number of radiometric classes, with the subsequent
storage of these classes in digital data fraction (bits).
2.2 Data pre-processing
The data available must be previously corrected in order to eliminate two types of
distortion: radiometric and geometric distortion.
Radiometric distortion refers to the scanners' calibration instability and to the
interactions between the electromagnetic energy, the atmosphere and the ground.
Geometric distortions refer to both the relative position of the acquisition platform
with respect to the atmosphere and the ground (terrestrial rotation and curvature
effect, reflection and refraction phenomena due to the different atmospheric layer