Applied Analysis
Inc.
630 Boston Road
Suite 201
Billerica, MA 01821
USA
PH: 978-663-6828
FAX: 978-663-6389
|
|
|
|
 |
|
|
24. What is Material Pixel Fraction?
|
|
Material Pixel Fraction is the fraction of
the pixels spatial area covered by a particular material. IMAGINE
Subpixel Classifier determines the fractional contribution to the
total pixel spectral radiance due to reflected light from a given
material. In general, this is the equivalent to the material pixel
fraction.
|
|
|
|
25. What are the output classes?
|
| Internally,
IMAGINE Subpixel Classifier computes the material pixel fraction detected
for each pixel in floating point form as a fraction from 0.0 to 1.0.
The output of the classification process is a 4-bit single-plane image
so the internal floating point values must be converted to fraction
classes which represent ranges of fractions. You can choose either 2, 4,
or 8 output classes which correspond to the following fraction ranges:
|
|
Class
|
2 Classes
|
4 Classes
|
8 Classes
|
| 1
|
0.20 0.59 |
0.20 0.39 |
0.20 0.29 |
| 2
|
0.60 1.00 |
0.40 0.59 |
0.30 0.39 |
| 3
|
|
0.60 0.79 |
0.40 0.49 |
| 4
|
|
0.80 1.00 |
0.50 0.59 |
| 5
|
|
|
0.60 0.69 |
| 6
|
|
|
0.70 0.79 |
| 7
|
|
|
0.80 0.89 |
| 8
|
|
|
0.90 1.00 |
|
|
|
|
26. What is classification tolerance?
|
| Classification tolerance
is a user-selected parameter which you can use to control detections
and false alarms. If you increase the tolerance value, the spectral
filters are enlarged and more detections are made. This is used to
increase the number of valid detections along with possible increased
false alarms. If you decrease the tolerance, the spectral filter is
decreased and the number of detections decreases. Generally the number
of false detections decreased as well. You can adjust the tolerance
value to best suit your application and sometimes to compensate for
signature deficiencies. |
|
|
|
27. Can I use more than one signature?
|
| IMAGINE Subpixel Classifier
allows you to combine two or more signatures into a multi-signature
file. The classification output for a multi-signature contains multiple
layers, one for each individual signature plus one for the combined
result. The combined signatures can be considered part of the same
family or separate families. Individual signatures in the same family
do not compete with each other during classification. This means that
each signature is treated independently when computing its classification
output layer. Signatures in separate families compete at classification
time which means that only the best matching signature is assigned
a classification fraction until no detections are possible from the
resulting residuals. In both cases, the combined classification layer
shows combined detections. |
|
|
|
|
|
|
|
AAI's Alaskan Oil Spill Assessment Using Satellite Imagery. (more...)
|
|
Demo Web-based mapping using SVG (more)
|
|
