Agriculture Reference
In-Depth Information
The correcting effect of averaging the results from several directions is drastic.
With two opposite directions, the reflectance varies only between 24 and 25 %, and
when four equally spaced directions are used, it remains practically constant. It
should be noted that the results are based on an optical model (see legend to
Fig.
9.23
) that does not take into account any other factors than the solar azimuth
angle.
This averaging of the signals from four directions is state of the art in practical
farming when the sensing is based on natural light.
9.4.3
Sensing Nitrogen by Reflectance Based on Artificial
Light
A general experience with the Kiel system in Europe when operating on the basis of
natural illumination is that the solar elevation angle - the complementary angle to
the zenith angle (see Fig.
9.21
) - should not be below 25°. This means that early and
late hours in the day cannot be used. Yet in regions with maritime- or mountainous
influence on the climate, this is just the time of the day when there is less wind. And
high wind speeds can seriously reduce the accuracy of fertilizer-spreading, espe-
cially with the currently dominating centrifugal spreaders.
Artificial illumination can overcome this problem. It also eliminates the noise
associated with any varying irradiance of natural light and with shade effects caused
by the machinery or by trees. It allows operating not only during daytime, but at
dawn and at night as well.
Technically, the spectral effects of the artificial light source need to be sepa-
rated from those of any natural light. This
separation
can be achieved by sub-
tracting the reflectance spectrum of solely natural light from the spectrum that
results, when in addition the artificial illumination is switched on. The artificial
light is applied as
regular flashes
and supplies “On” spectra. In the short inter-
vals between the flashes, the “Off” spectra come from the natural light. The
“Off” spectra are subtracted from the respective adjacent “On” spectra. The thus
obtained spectral signals result only from the artificial illumination and any
effect of varying natural light is eliminated.
In case continuous artificial illumination is applied, separating the effects of the
different light sources would also be possible by
modulating the artificial light
.
However, this separating technique is more complicated.
In the total cost calculation, the higher investment for artificial illumination
might be offset by the capability of treating a larger area because of the longer oper-
ating time per day.
It should be mentioned that the
artificial light source
must adequately supply
the respective wavelengths. Many commonly used light sources do not, as can be
seen in Baille (
1993
) and Lawrence et al. (
2005
). If the wavelengths that are needed
are provided, there is no reason why the nitrogen effect on the reflectance indices
with artificial illumination should be different than with natural light. Commonly
