Agriculture Reference
In-Depth Information
Surrounding Awareness
for Automated
Agricultural Production
12
Francisco Rovira-Más
CONTENTS
12.1 Perception Needs in Agricultural Environments.......................................... 323
12.2 Two-Dimensional Perception for Navigation and Monitoring ..................... 325
12.3 Three-Dimensional Awareness for Navigation and 3-D Mapping ............... 329
12.4 Global Maps from Local Perception ............................................................ 334
12.5 Future Steps in Surrounding Awareness ...................................................... 339
References .............................................................................................................. 340
12.1 PERCEPTION NEEDS IN AGRICULTURAL ENVIRONMENTS
Vehicle automation requires the constant acknowledgment of the neighboring space,
but the specific parameters and properties to be tracked highly depend on the type
of environment in which each vehicle operates, as well as the particular tasks com-
manded to the vehicle in the mission pursued. So, for example, the wind speed is
a key factor to be considered in the automated landing of aerial vehicles but has
negligible influence in the path-planning control algorithm of a farming tractor.
Agricultural applications and environments are quite diverse, although the percep-
tual needs of farm machinery can be grouped into two general sets: sensing for
crop
production
status
and sensing for
navigation and positioning
. The first group
includes mapping of soil properties, detection of water stress in crops, estimation
of plant vigor, monitoring of nitrogen shortage in leaves, and assessment of fruit
maturity or plant diseases. The second set involves all perceptive means—visual and
nonvisual—of positioning the vehicle with respect to surrounding obstacles, as well
as the precise operation of implements such as the automatic maneuver of unload-
ing side pipes in harvesters. The size, power, and weight of farm equipment make
autonomous navigation very complex, not only because machines are oversized and
lanes become tight, but also because of the high levels of reliability required; an
accident caused by an autonomous vehicle will surely jeopardize further research
and innovation in this discipline for years. Yet, there exist plenty of advantages in the
development of
semiautonomous machines
, that is, vehicles that are capable of per-
forming automated tasks under the supervision of a human operator granting safety.
Even the minimum degree of autonomy will require a sound perception system to
323
