Fusions, Overlays and Management Zones - Precision in Crop Farming - page 329

However, the variety of sensing techniques that has been developed for getting
information about soil- and crop properties provides many prospects for sensor fusion.
Some feasible and conceivable applications for sensor fusion are listed in Table 13.1 .
The list is oriented on control needs and sensing concepts for farming operations
that were dealt with largely in previous chapters. A concept of sensor fusion can
be, but must not be congruent with data fusion or data merging . The latter terms
stand for the processing of the data into singular signals for the control. The rea-
sonable uses of the data will be different for transient properties like water content
of soils on the one hand or for temporally stable properties like the topography or
organic matter content of soils on the other hand. All properties can be sensed
simultaneously and might also be useful for fused on-the-go control. However,
only the temporally stable properties lend themselves well for delayed use, for
averaging with previous records and hence for mapping. So sensor fusion can
involve data fusion, but it might also be reasonable to complement this technique
with data splitting . And a fused control of a farming operation can be reasonably
based partly on sensing in an on-the-go mode that is supplemented with signals
obtained from maps.
The actual need for sensor fusion must depend on the information variety that
assists in the site-specifi c control of farming operations. The primary sensing for a
site-specifi c objective in Table 13.1 stands for the most important control signals,
which in turn act as substitutes for particular soil and crop properties. Consequently,
the secondary and tertiary sensing are for signals or respective properties that
might be needed to correct or supplement this result. But in some cases, these sec-
ondary and tertiary signals might even be alternatives or surrogates for those from
the primary sensing
The approaches listed go far beyond the presently in practice used technologies,
which are marked in green. Up to now, tertiary sensing is not state of the art. This
can be due to an absence of suitable sensing techniques. As an example, for control-
ling the depth of primary cultivation adequately, sensing the hydromorphic soil
properties can be very reasonable (Sect. 7.2.2.1 ) . But the presently used methods of
sensing electrical conductivities in soils do not allow for this
Several approaches listed for secondary sensing hold for signals that indicate the
water supply in soils or crops. This probably characterizes the special possibilities
and the particular potential of sensor fusion: it makes it possible to include informa-
tion about a transient factor into the control process. This information hardly can
be accounted for within due time if delays due to maps and its processing occur. An
urgent case for including the transient factor water via sensor fusion and secondary
sensing is in-season nitrogen fertilizing. For details to this see Sect. 9.4.10 .
Sensing the electrical conductivity is currently the most frequently used method
of recording soil properties. When this method is used in humid areas, it is mainly
the clay and the water content of the soils that affect the signals. Hence a temporally
stable and a transient property act together. For long-term use, the information that
is derived from electrical conductivity sensing would be more valuable if the effect
of the temporally varying water content could be normalized or taken care of by an
adequate logical mathematical processing. A prerequisite for this is simultaneous
yet separated sensing of electrical conductivity as well as of water.