(depending on satellite and band sensor) and a revisit time of 16 days for most of
Earth’s surface (Irish, 1999; Song et al., 2001). Landsat imagery has been especially
useful in vegetation change detection studies. Typically, annual images obtained
during the period of peak summer greenness are compared based on NDVI or other
vegetation indices calculated from red and NIR reflectance values. Preprocessing of
imagery by the U.S. Geological Survey (USGS) Earth Observation and Science
Center includes georectification and terrain correction based on ground sample points
(where available). Gain values for each sensor are internally calibrated and delivered
to end users as calibrated DN values ranging from 1 to 256. To compare band values
or vegetation index values across dates, the end users must first convert DNs to top-ofatmosphere reflectance values using data on sun angle and Earth–sun distance at the
time of image acquisition (Irish, 1999). Most studies also correct images for
FIGURE 5.3 Data from phenocams showing defoliation of Tamarix by leaf beetles followed
by new leaf production on Dolores River: (a, b) from visual analysis of panchromatic images at
two sites on floodplain; (b, c) scaled NDVI values calculated from red and NIR bands from
multiband cameras at same site.
90
CHANGE DETECTION USING VEGETATION INDICES AND MULTIPLATFORM
Earth’s surface (Irish, 1999; Song et al., 2001). Landsat imagery has been especially
useful in vegetation change detection studies. Typically, annual images obtained
during the period of peak summer greenness are compared based on NDVI or other
vegetation indices calculated from red and NIR reflectance values. Preprocessing of
imagery by the U.S. Geological Survey (USGS) Earth Observation and Science
Center includes georectification and terrain correction based on ground sample points
(where available). Gain values for each sensor are internally calibrated and delivered
to end users as calibrated DN values ranging from 1 to 256. To compare band values
or vegetation index values across dates, the end users must first convert DNs to top-ofatmosphere reflectance values using data on sun angle and Earth–sun distance at the
time of image acquisition (Irish, 1999). Most studies also correct images for
FIGURE 5.3 Data from phenocams showing defoliation of Tamarix by leaf beetles followed
by new leaf production on Dolores River: (a, b) from visual analysis of panchromatic images at
two sites on floodplain; (b, c) scaled NDVI values calculated from red and NIR bands from
multiband cameras at same site.
90
CHANGE DETECTION USING VEGETATION INDICES AND MULTIPLATFORM