664
September 2016
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
I
n the last column (July 2016), I discussed a few tricks on how to extract/clip and
save image rasters using the Image Analysis Toolbar in ArcGIS. While using the
tools that I showed last column, you may have noticed that there is a leaf icon
available on the toolbar.
Here at the Southwest Florida Water Management District (SWFWMD), we have
been collecting digital imagery since 2004. The imagery is used for a variety of
purposes, including watershed management planning, regulatory and inspection,
water supply, and habitat health determination. The early years’ data were three-
frequency band false-color infrared (fCIR), while the more recent years’ data are four-
frequency bands; red, green, blue, and infrared (RGBi). For several years, we were
using third-party image analysis software to compute the Normalized Difference
Vegetation Index (NDVI) for our habitat restoration projects. Then we discovered
that the same calculations can be performed with the Image Analysis tool directly
in ArcGIS.
DISCLAIMER: As Remote Sensing scientists, we recognize that the NDVI is defined by
pixel Digital Number values being calibrated for spectral reflectance and not stretched
for color-balancing. The standard 4-band imagery in the SWFWMD database has been
color-balanced and stretched. We recognize that the NDVI values we are calculating
are not actually NDVI in the strictest sense, but for our purposes, the values are
representative of NDVI values.
Using the ArcGIS Image Analysis tools, NDVI can be calculated from a stacked image
(the SWFWMD case) or from two single-band images. An additional convenience of
the Esri NDVI tool is that it will report the NDVI values as either (1) 8-bit values
ranging between 0 – 255 (default option), or (2) 8-bit values ranging between -1 to +1
(Scientific Output option). Details of how the actual calculations are being performed
are available through the Esri Help at the website:
In any case, it is a
good idea to review the image metadata
and determine which band holds the
RED data and which band holds the INFRARED data. The Esri-default is Band 2
= RED and Band 1 = INFRARED, but better to be safe than sorry. Note: If there is
wavelength information encoded in the imagery, the “Use Wavelength” option can be
used; otherwise the band numbers need to be specified.
Once you have determined the correct bands, the tool is actually pretty simple to use.
The screenshots that follow are for the case where you have a stacked, multi-band
image and have determined the band assignments for RED and INFRARED.
Note: The tools will work on raster catalogs and layer files. You still need to know
the band assignments.
GIS
&
Tips Tricks
By Al Karlin, Ph.D.,CMS, GISP
Did you
know
that you
could use
ArcGIS for
Remote
Sensing?
