Agriculture Reference
In-Depth Information
Tree row plane
Original
canopy boundary
Smoothed
canopy boundary
Through hole
Sensor
L
ij
+1
LL
ij
+1
Shallow gap
L
ij
ε
LL
ij
Deep gap
β
H
s
D
Ground
D
max
FIGURE 7.41
Illustration of original and smoothed canopy boundary. (From Wei, J.,
Salyani, M.,
Trans. ASAE
, 48, 1595-1601, 2005. With permission.)
outer (smoothed) canopy boundary and the tree row plane (both are shown in Figure
7.41). Then they defined foliage density (
D
f
) as the ratio of foliage volume to tree
canopy volume.
Swanson et al. (2010) built a linear and kernel regression model to estimate the
citrus yield based on canopy volume, density, and fruit counts. The laser scanner was
mounted on a vehicle. The vehicle traveled along the tree row. The 3-D point clouds
were collected using multiple scans. Then 3-D point clouds were mapped into a 3-D
grid composed of voxels. Figure 7.42 shows one of the voxelized volume slices. All
of the voxels' volume were summed together for each slice to form the total canopy
volume.
FIGURE 7.42
Voxelized volume slice. (From Swanson, M. et al.,
A multi-modal sys-
tem for yield prediction in citrus trees,
ASABE Annual Intl. Meeting
, 1009474, 2010. With
permission.)
