Biomedical Engineering Reference
In-Depth Information
many insects is adapted to measure polarization
from sky light. The ocelli have wider fields of
view than any of the ommatidia (popularly
known as
facets
or
eyelets
) of the compound eye
of the dragonfly. From the figure it is also clear
that ocelli have an aperture at least an order of
magnitude greater than those of the ommatidia
of the compound eye, suggesting an adaptation
to low-light conditions.
Ocellar stabilization and polarization com-
passing are mediated by aspects of the environ-
ment that are invisible to the human eye. Ocelli
accept input from the ultraviolet region of the
spectrum. The polarization of light is usually
imperceptible to the unaided human eye. The
geometry of insect vision is also outside the
human experience since the compound eye pro-
vides a near-spherical view of the world.
FIGURE 9.2
The panoramic scanning device (top left) pro-
duces a polar-coordinate image in multiple spectra and three
polarization angles. Images are shown from each of the spec-
tral channels sampled by the device. The spectrum of each
channel was centered on the following wavelengths: ultravio-
let (340 nm), blue (400 nm), green (500 nm), red (600 nm), and
infrared (1,000 nm). Each channel was resolved to 16 bits reso-
lution and each image required 60 s to generate.
9.2 STRUCTURE OF THE VISUAL
WORLD OF INSECTS
It is not possible to view the multispectral and
multipolarization world of insects. Understand-
ing the sensors and their limitations as well
as development of flying biomimetic sensors
requires simulators and control system design.
Instrumentation and models of the insect visual
system have been developed to inform the pro-
cess of biomimetic engineering of insect sensory
systems.
To understand and reverse engineer the
insects' view of the static world, we built a device
that approximates the geometry and spectral
response of an insect eye.
Figure 9.2
shows the
multispectral, multipolar, and panoramic scan-
ning device and the output of the seven channels
in polar coordinates. The scanning device pro-
duced a polar-coordinate-system image from
400
×
200 discrete samples over a capture period
of one minute. The polarization pattern distrib-
uted across the sky is revealed in the differences
between the three blue channels. The photodi-
odes in the device and all implementations
described in this chapter were interfaced in
cur-
rent mode
[7]
in order to achieve a linear voltage
representation of light intensity.
Images of the environment were scanned
under different conditions. From these images,
observations about the optical structure of the
environment and quantitative measurements of
the performance of postulated insect-sensor
algorithms could be made.
The polarization pattern of the sky on a clear
day is shown in
Figure 9.3
a. The saturation of
colors indicates the degree of polarization. In
Figure 9.3
b, the ground plane is devoid of pseudo
color apart from one area, which is a body of
water. The water beetle
Notonecta glauca
has been
shown to use the polarized reflections off bodies
of water on the ground plane to navigate
[8]
,
