|
Step 1. Image Preparation:
The
purpose of this section is to briefly instruct users who are given a Landsat
scene how to prepare the scene for analysis and calculate common
products such as dNBR.
>>> Step 2. Calculating Radiance
When you first
receive your Landsat (or other data source) image, it generally is delivered
as an image of raw digital number (DN) values. You can't apply any spectral
index to this raw data (whether dNBR or any other measure).
The reason you can
not do this is simple. DNBR and other so-called "spectral indices" were
developed to work with measures of how reflective the specific
unburned and burned surfaces are. The DN values do not provide the correct
information. Therefore we have to prepare the data for analysis by
converting the DN values into these
reflective values. This process is called DN to Reflectance to
conversion and has the following steps:
-
Reading the
Data
-
DN to Radiance
-
Radiance to
Reflectance (as measured by the satellite sensor in space)
-
Satellite
Reflectance to Reflectance as if the satellite sensor was positioned
just above the ground
1. Reading
the Data
One of the most
confusing aspect of Landsat images is that there is more than one format
that the data can be delivered to the users as. Some of the different types
include: GeoTiff, Fast format, hdf, and NLAPS. There is also more than one
Landsat sensor , which means that you must be careful how you convert data,
as one set of rules for one image may be different for images derived from
"a different Landsat satellite".
If you are using
advanced remote sensing software packages such as ENVI or Erdas many of
these data formats can be directly read in and are converted on-the-fly by
these programs into both radiance and reflectance as measured by the
satellite sensor in space. However, you will encounter many times when this
can not be done. Therefore, we will cover these main steps in this tutorial
such that you can understand how the data is produced.
|
