By Ty Kennedy-Bowdoin, Processing Platform Product Manager
SkySat-1 collects imagery using 5 channels: blue, green, red, near infrared (NIR), and panchromatic. Everyone is familiar with the first three channels because the human eye is sensitive to this range of the electromagnetic spectrum and our brains have evolved to interpret this information intuitively. For this reason, cameras and satellite data are typical viewed in this “true-color” scheme that we see in platforms like Google Maps.
Figure: MSAVI Results of a Center Pivot Irrigated Field in Saudi Arabia. Captured by SkySat-1 on 12/26/13.
SkySat-1 collects imagery using 5 channels: blue, green, red, near infrared (NIR), and panchromatic. Everyone is familiar with the first three channels because the human eye is sensitive to this range of the electromagnetic spectrum and our brains have evolved to interpret this information intuitively. For this reason, cameras and satellite data are typical viewed in this “true-color” scheme that we see in platforms like Google Maps.
The NIR channel (740 - 900 nm) is designed to capture
light in a range just beyond the visible spectrum (390 - 700 nm). To
visualize a “color infrared” image we typically map the NIR channel to
red, while assigning the red to the green and the green to the blue
channels. This allows us to visualize the near infrared channel as red,
so materials that reflect well in this wavelength range appear very
red.
One of the more interesting characteristics of the NIR
channel is that lush vegetation reflects very strongly relative to other
materials or woody vegetation. Healthy vegetation generates more
chlorophyll in the leaves, which reflects well in the NIR, while less
healthy leaves are much less reflective. Many interesting metrics depict
vegetation health based on ratios of the red to NIR channel.
One of the most common vegetation health metrics is the
Normalized Difference Vegetation Index (NDVI) which works well in
consistently vegetated areas. The Modified Soil Adjusted Vegetation
Index (MSAVI) takes this metric one step further by correcting for the
amount of exposed soil in each pixel in agricultural areas where
vegetation is surrounded by exposed soil.
The example in Figure 1 illustrates a color infrared
rendering of some center pivot irrigated fields in Saudi Arabia. Figure 2
illustrates the MSAVI results from this same dataset. I have rendered
the colors to depict healthy vegetation with hot colors and less healthy
vegetation in cool colors. While we could qualitatively determine that
one of the center pivot irrigated fields was well watered and the others
appear to be fallow in the color infrared image, we now can quantify
this with metrics that can be compared throughout the seasons or years
and correlated with the factors measured on the ground like water usage,
fertilization, seed varieties, crop yield, etc.
Mineral deposits rich in Iron Oxide are another common
source of reflective NIR values in satellite images. This can be very
interesting to geologists searching for assemblages of minerals
indicative of higher concentrations of valuable metals.
This is only the beginning of what we can do multispectral imagery! Interested in learning more? Feel free to reach out to me at ty@skybox.com
This is only the beginning of what we can do multispectral imagery! Interested in learning more? Feel free to reach out to me at ty@skybox.com
Figure: MSAVI Results of a Center Pivot Irrigated Field in Saudi Arabia. Captured by SkySat-1 on 12/26/13.
