Networked multiband digital cameras (phenocams) (Richardson et al., 2007)
deployed over individual shrubs documented the phenology of defoliation with
high temporal and spatial resolution but over a very limited sample of shrubs.
Then Landsat TM and MODIS satellite images were used to quantify the effects
of defoliation on riparian green plant cover and ET from 2000 to 2010, and ET
measured in years before beetle release and years after release were subjected to an
analysis of variance (ANOVA) to test the hypothesis that beetles reduced riparian ET
(Nagler et al., 2012).
5.2.3 Phenocams Combine High Spatial and Temporal Resolution with
Limited Field of View
Phenocams are inexpensive digital cameras mounted over vegetation. They automatically acquire images on a frequent basis (every 15 min to daily or longer), which are
transmitted by an onsite packet radio station to a computer at a data processing site.
Phenocams and other tower-mounted, remote sensors have become practical continuous-monitoring tools due to advances in computers and sensor technologies and
improvements in the Internet in the past 10 years (e.g., Richardson et al., 2007).
Details of their use have been reviewed in Rundel et al. (2009), Zerger et al. (2010),
and Sonnentag et al. (2012). In vegetation studies they are especially useful in
determining the exact timing of phenological events such as leaf-out, peak greenness,
flowering, and maturation of either natural stands of plants or agricultural crops.
(Similar cameras are used to monitor the movement of animals.) Phenocam sites have
been organized into state, regional, national, and international networks, for example,
the USA National Phenology Network, to monitor effects of climate change on
vegetation growth and flowering cycles.
The cameras provided multiband imagery with red, green, blue, and NIR bands
which are combined to compute NDVI values:
NDVI =
NIR − red
NIR + red
(5.1)
Unlike satellite sensors, band values from digital cameras are rarely calibrated to
reflectance values, and digital number (DN) values vary widely across different
cameras viewing the same scene. Many phenocams have only red, green, and blue
bands available: greenness from these images is computed from these bands by
algorithms such as the “excess greenness” index (Sonnentag et al., 2012):
ExG =
2 green
red + green + blue
(5.2)
Either NDVI or ExG or simple visual interpretation of images can be used for analyses
of phonological phenomenon.
In our work with saltcedar leaf beetles, we used phenocams to track the defoliation
process from its onset in early summer through the regreening of shrubs in late
86
CHANGE DETECTION USING VEGETATION INDICES AND MULTIPLATFORM
deployed over individual shrubs documented the phenology of defoliation with
high temporal and spatial resolution but over a very limited sample of shrubs.
Then Landsat TM and MODIS satellite images were used to quantify the effects
of defoliation on riparian green plant cover and ET from 2000 to 2010, and ET
measured in years before beetle release and years after release were subjected to an
analysis of variance (ANOVA) to test the hypothesis that beetles reduced riparian ET
(Nagler et al., 2012).
5.2.3 Phenocams Combine High Spatial and Temporal Resolution with
Limited Field of View
Phenocams are inexpensive digital cameras mounted over vegetation. They automatically acquire images on a frequent basis (every 15 min to daily or longer), which are
transmitted by an onsite packet radio station to a computer at a data processing site.
Phenocams and other tower-mounted, remote sensors have become practical continuous-monitoring tools due to advances in computers and sensor technologies and
improvements in the Internet in the past 10 years (e.g., Richardson et al., 2007).
Details of their use have been reviewed in Rundel et al. (2009), Zerger et al. (2010),
and Sonnentag et al. (2012). In vegetation studies they are especially useful in
determining the exact timing of phenological events such as leaf-out, peak greenness,
flowering, and maturation of either natural stands of plants or agricultural crops.
(Similar cameras are used to monitor the movement of animals.) Phenocam sites have
been organized into state, regional, national, and international networks, for example,
the USA National Phenology Network, to monitor effects of climate change on
vegetation growth and flowering cycles.
The cameras provided multiband imagery with red, green, blue, and NIR bands
which are combined to compute NDVI values:
NDVI =
NIR − red
NIR + red
(5.1)
Unlike satellite sensors, band values from digital cameras are rarely calibrated to
reflectance values, and digital number (DN) values vary widely across different
cameras viewing the same scene. Many phenocams have only red, green, and blue
bands available: greenness from these images is computed from these bands by
algorithms such as the “excess greenness” index (Sonnentag et al., 2012):
ExG =
2 green
red + green + blue
(5.2)
Either NDVI or ExG or simple visual interpretation of images can be used for analyses
of phonological phenomenon.
In our work with saltcedar leaf beetles, we used phenocams to track the defoliation
process from its onset in early summer through the regreening of shrubs in late
86
CHANGE DETECTION USING VEGETATION INDICES AND MULTIPLATFORM