The problem
Instruments may measure different spectral ranges using different
sensors, resulting in abrupt “jumps” in the reflectance or radiance
data. In such cases, the regions between sensors need to be matched,
i.e. spliced together. Unmatched spectra collected with a 3-sensor
instrument, such as the SVC HR1024, may look like this:
# Path to raw (unmatched) spectra
path_raw = system.file("extdata/svc_raw_and_overlap_matched_serbin/SVC_Files/",
package = "spectrolab")
# Read spectra as reflectance and radiance
reflect_raw = read_spectra(path = path_raw, type = "target_reflectance")
radiance_raw = read_spectra(path = path_raw, type = "target_radiance")
# Sensor overlaps marked with vertical dashed lines
lwd = 0.5
cex = 0.7
oldpar = par(no.readonly = TRUE)
par(mfrow = c(2, 1))
plot(reflect_raw, main = "Reflectance",
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
abline(v = c(990, 1900), col = "red", lty = 2, lwd = lwd)
plot(radiance_raw, main = "Radiance",
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
abline(v = c(990, 1900), col = "red", lty = 2, lwd = lwd)
The solution
Use the function match_sensors to splice the sensor
overlap regions. You must pass the boundary between sensors using the
splice_at argument. It is critical that
you get those bands right (or very close) and every instrument (even
from the same vendor) is different.
You can use plot_interactive zoom into a particular
spectral region and decide what the splice_at values should
be. You can also use the function guess_splice_at to
estimates these bands for you but keep in mind that these are
guesses.
# Spectrolab's guess of what the splice bands are.
# However, you should also visually inspect the spectra to determine what the
# boundaries between sensors are.
splice_bands_guess = guess_splice_at(reflect_raw)
splice_bands_guess
## [1] 986.7333 1901.1000
# Finally, if you know what those sensor bounds should be (say, they're given by
# the manufacturer), just use those numbers instead of spectrolab's guess.
splice_bands = c(990, 1900)
# Match the reflectance and radiance data
reflect_matched = match_sensors(x = reflect_raw, splice_at = splice_bands,
interpolate_wvl = c(5, 1))
radiance_matched = match_sensors(x = radiance_raw, splice_at = splice_bands,
interpolate_wvl = c(5, 1))
lwd = 0.5
cex = 0.7
oldpar = par(no.readonly = TRUE)
par(mfrow = c(2, 1))
plot(reflect_raw, main = "Reflectance",
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
plot(reflect_matched, col = "red", add = TRUE,
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
plot(radiance_raw, main = "Radiance",
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
plot(radiance_matched, col = "red", add = TRUE,
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
And we can check the results from spectrolab’s
match_sensors against SVC’s proprietary matching
algorithm.
path_moc = system.file("extdata/svc_raw_and_overlap_matched_serbin/SVC_Files_moc/",
package = "spectrolab")
reflect_moc = read_spectra(path = path_moc, type = "target_reflectance")
radiance_moc = read_spectra(path = path_moc, type = "target_radiance")
lwd = 0.5
cex = 0.7
oldpar = par(no.readonly = TRUE)
par(mfrow = c(2, 1))
plot(reflect_moc, main = "Reflectance", col = "black",
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
plot(reflect_matched, col = "red", add = TRUE,
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
plot(radiance_moc, main = "Radiance", col = "black",
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
plot(radiance_matched, col = "red", add = TRUE,
lwd = lwd, cex.main = cex, cex.lab = cex, cex.axis = cex)
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.