IEEE Robotics & Automation Magazine - December 2023 - 53

a valve to pressurize the chamber while also
observing the water expression at the end of
the exposed stem through a magnifying glass
[Figure 1(b)]. The pressure necessary to force
water out of the stem determines the SWP.
The process measures the capacity of the
cells to retain water by pressurizing the leaf.
The less free water there is in the plant, the
greater the pressure required to cause the leaf
to exude water. When taken in predawn conditions
(i.e., performed before sunrise) and when
plant stomata are closed, the measurement is
at equilibrium with soil moisture conditions.
Subsequent measurements can precisely determine
the water deficit, and, thus, the irrigation
demand, to meet evapotranspiration loss.
A precise measurement of the SWP is essential to assess the
water deficit of the plant and, therefore, to adjust irrigation.
Given that 80% of the managed fresh water in the United
States is consumed by agriculture [21] and that evapotranspiration
model estimates widely diverge [8], even modest
improvements in irrigation practices can have huge impacts,
especially in the semiarid southwestern United States, which
periodically undergoes drought periods while providing a
large fraction of the country's fruits and vegetables.
Considering the aforementioned labor-intensive steps in
measuring SWP using the pressure chamber method, but recognizing
its significance in crop production, some alternative
PHYSICAL
SAMPLING
AND SPECIMEN
ANALYSIS CAN BE
QUITE LABORIOUS
AND OFTEN VARY
AMONG DIFFERENT
TYPES OF CROPS.
methods have been proposed. Some rely on
remote sensing using spectral reflectance or
multispectral imaging to determine water
potential [20], [25]. Their greatest values
are their noninvasive characteristic, much
like the pressure chamber method, and scalability
since these methods are intended as a
faster alternative to the pressure chamber for
mass assessment. Although initially promising,
these methods are highly sensitive to an
intractable external factor, i.e., light variability
due to weather and solar motion. Zhao et
al. [25] mounted multispectral cameras on a
small unmanned aerial vehicle to take highresolution
multispectral images of orchards for
SWP prediction using the canopy Normalized
Difference Vegetation Index but mentioned the high variability
in data collected from different flights within the same
day, due to solar motion. Vila et al. [20] used remote sensors
and spectral reflectance as a proxy for SWP measurements
but had a low correlation coefficient and, thus, concluded that
this method cannot serve as a replacement for the pressure
chamber method. Since the concepts of spectral reflectance
and imaging may not be mature yet for this application, our
work seeks to directly automate the pressure chamber method
for assessing SWP.
Specifically, we are developing a robotic system that can
physically collect and analyze plant specimens in the field
Water
Seal
Air Pressure
Gauge
Sap Drop
(Water)
Leaf
Sample
Compressed Air
Pressure
Chamber
(a)
Leaf
Collection
Pressure
(Air)
Leaf
Leaf
Collection
(c)
(b)
FIGURE 1. (a) The working principle of the pressure chamber. (b) Manual visual inspection is currently performed in the field for SWP
analysis. (c) The concept proposed in this work. A mobile robot autonomously selects multiple measurement locations for sample
collection and retrieves leaves. These are conveyed to a human operator, who uses a pressure chamber retrofitted with our machine
vision-assisted technology to determine the pressure defining the SWP.
DECEMBER 2023 IEEE ROBOTICS & AUTOMATION MAGAZINE
53
Collection

IEEE Robotics & Automation Magazine - December 2023

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - December 2023