Information Technology Reference
In-Depth Information
spectral ranges including soil electronic maps of the organics content, moisture,
temperature, granulometric composition, and colour. Forming electronic maps of a mineral
fertilizers distribution on fields or virtual maps of planned crop yield using imaging data to
estimate growth conditions and cropping are realized by dosed applying fertilizers in soil [1,
3, 5]. The optimal strategy of the agricultural production can be fast achieved by data
overlapping of electronic virtual maps but also on the basis of current information about
tillage, nutrients carry-over from soil with taken crops, characteristics of used agricultural
units. Then it is possible to control operations of the agricultural machinery, to keep track of
information how much fuel is consumed or whether fertilizers are applied. To produce
electronic maps, we used a point krinning method for the estimation of the distributed
random function in an arbitrary point as the linear combination of its values in initial ones.
A variogram defines a form of the optimal interpolated hypersurface in the space between
reference spots of the sensory control. According to the krinning method, the estimated
value of the soil quality in the known spot p from a set of k neighbouring spots is calculated
as weighed mean measured values in neighbouring spots in the form:
k
W,
(6)
p
i
i
i1
where W
i
- weighting coefficient of an index i of the soil quality in relation to the estimated
spot p from a set of neighbouring spots.
The krinning method provides for solving a set of equations:
k
W γ ξ λγ ξ ,
j
ij
ip
j1
k
(7)
W ,
i
i1
where γ (ξ
i j
), γ (ξ
i p
) - semiv
ario
gram values for the distance ξ
ij
and ξ
i p
between a points i
and estimated points j, p , i1, k; λ - Lagrange factor.
Unknown weighting coefficients W
i
are computed by solving a set of equations (7), but a
value of the controlled variable in the spot p is calculated using the formula (6). The
semivariogram on the area boundary of spots with the different agricultural background in
precision agriculture has the sharp difference in values; therefore, the considered
mathematical model shows the nugget-effect. Having estimated a value of the soil quality in
an agricultural spot q in accord with controlled values k
1
, k
2…
k
m
of appropriate agricultural
backgrounds m, the set of krinning equations for models with the nugget-effect can be
presented as:
k
k
k
1
2
m
W(
)
W(
)
W (
)
(
),
1q
1j
ij
2q
2j
ij
mq
mj
ij
ip
j1
j1
j1
(8)
k
k
k
1
2
m
W
W
W
1,
1q
1i
2q
2i
mq
mi
i1
i1
i1
