Page 243 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

234 Chapter 8 A review on plant diseases recognition through deep learning

3. Reduction in leaf area index and biomass: It happens due to
pest attacks such as armyworm that can eat maize plant parts
such as leaf and stalk, which leads to significant loss of biomass
and leaf area [35] Moreover, remote sense monitoring assists in
detecting the lack of uncertainty in spectral specificity.
4. Wilting: The loss of firmness in plants is a common symptom
that occurs due to pests and plant diseases. For instance, the
piercing as well as the sucking attitude of pests such as aphid
or beetles causes wilting [36]. In particular, the ruined vascular
system blocks the water flow in the infected situation, which
causes dehydration to all parts of the plant [37].
Remote sensing system can detect symptom changes at
different stages of pest or disease attack.

5.2 Remote sensing systems for monitoring pests
and diseases

A wide variety of remote sensing systems can be adapted to
detect and monitor pests and plant diseases. In the presence of
active and passive radiation, a remote sensing system allows
data acquisition that ranges from gamma to microwave. Such
kind of remote sensing system captures physiological changes
(e.g., water content, pigment content, etc.), structural responses
(e.g., landscape structure, canopy structure), and infection symp-
toms (e.g., pustules, scabs) caused by pests and plant diseases.
The sensing systems can be classified into three kinds based on
pests and plant diseases monitoring:
(a) VIS-SWIR (visible and short-wave infrared)
(b) Fluorescence and thermal sensors
(c) Lidar (light detection and ranging equipment) and SAR
(synthetic aperture radar) systems
These types of sensors are well appropriate for the in-door
examination of pests and plant diseases, especially in fruits and
vegetables.

5.3 Visible and short-wave infrared monitoring
systems
Visible and near-infrared (VIS-NIR) sensor systems capture supe-
rior quality data with relative signal-to-noise ratio. VIS-NIR
sensors are extensively available for different platforms such as
aerial-based, ground-based, and satellite-based that are most
suitable to monitor pests and plant diseases. The wavelength of
SWIR region is 1000e2500 nm, which is more sensitive to detect