Abstract
Non‐line‐of‐sight
(NLOS)
tracking
has
the
ability
to
detect
and
locate
moving
objects
which
are
hidden
from
direct
line
of
sight.
It
a
number
potential
applications
in
autonomous
driving,
security
search‐and‐rescue
scenarios.
The
current
approaches
for
NLOS
mostly
rely
on
time‐of‐flight
(ToF)
LiDAR,
mmWave
radar
or
RGB
camera.
However,
they
may
either
have
limited
resolution
need
accumulation
position
velocity.
Here
real‐time
comb‐calibrated
frequency‐modulated‐continuous‐wave
(FMCW)
LiDAR
system
is
proposed
demonstrated
high‐resolution
tracking.
A
3D
positioning
(and
velocity)
accuracy
2
mm
s
−1
)
capability
realize
multi‐object
experiment.
Importantly,
FMCW‐LiDAR
approach
can
obtain
velocity
multiple
snapshot
,
thus
resolving
problems
measurement
origin
uncertainty,
false
alarm,
detection
overlap
multi‐target
shows
significant
practical
sensing.
Abstract
Object
detection,
recognition,
and
identification
impose
stringent
requirements
on
the
imaging
resolution
of
light
detection
ranging
(LiDAR)
systems.
The
ability
LiDAR
to
resolve
objects
is
constrained
by
Abbe‐Rayleigh
diffraction
limit
due
wave
nature
light.
Superoscillation,
as
a
far‐field
super‐resolution
technique,
theoretically
allows
for
generation
arbitrarily
small
spots.
However,
its
practical
implementation
primarily
validated
within
microscopic
domain
because
weak
super‐diffractive
information
strong
sidelobe
noise.
Consequently,
application
in
with
waves
remains
insufficiently
explored.
In
this
study,
first
functional
prototype
presented
using
super‐oscillatory
illumination
LiDAR.
By
utilizing
specifically
engineered
field
confocal
illumination,
target
at
distance
20
meters
achieved.
This
work
offers
promising
approach
achieving
active
3D
imaging.
Sensors,
Journal Year:
2024,
Volume and Issue:
24(20), P. 6522 - 6522
Published: Oct. 10, 2024
Non-Line
of
Sight
(NLOS)
imaging
has
gained
attention
for
its
ability
to
detect
and
reconstruct
objects
beyond
the
direct
line
sight,
using
scattered
light,
with
applications
in
surveillance
autonomous
navigation.
This
paper
presents
a
versatile
framework
modeling
temporal
distribution
photon
detections
Time
Flight
(dToF)
Lidar
NLOS
systems.
Our
approach
accurately
accounts
key
factors
such
as
material
reflectivity,
object
distance,
occlusion
by
utilizing
proof-of-principle
simulation
realized
Unreal
Engine.
By
generating
likelihood
distributions
over
time,
we
propose
mechanism
data,
facilitating
optimization
systems
development
novel
reconstruction
algorithms.
The
allows
analysis
individual
components
return
distributions,
yielding
results
consistent
prior
experimental
data
providing
insights
into
effects
extended
surfaces
multi-path
scattering.
We
introduce
an
optimized
secondary
scattering
that
captures
critical
information
reduced
computational
cost.
work
provides
robust
tool
design
improvement
dToF
SPAD
Lidar-based
Abstract
Light
propagation
in
complex
media
results
strong
scattering.
While
wavefront
shaping
(WFS)
enables
the
focusing
of
light
at
depth,
its
speed
is
mainly
limited
by
frame
rate
area‐array
detectors.
The
photodetector
has
been
used
to
achieve
fast
focusing,
but
it
cannot
record
sufficient
speckle
information,
limiting
applications
multi‐point
and
non‐invasive
focusing.
Here,
a
method
for
invasively
or
non‐invasively
through
scattering
proposed
using
1D
signals
as
feedback
shaping.
Experimentally,
not
only
invasive
can
be
realized,
also
maximizing
contrast
linear
fluorescent
signals,
achieved,
suggesting
effectiveness
method.
This
approach
generalized
utilization
array
detectors
WFS
may
hold
interesting
prospects
rapid
within
deep
biological
tissues.
APL Photonics,
Journal Year:
2024,
Volume and Issue:
9(12)
Published: Dec. 1, 2024
Real-time
non-line-of-sight
imaging
is
crucial
for
practical
applications.
Among
existing
methods,
transient
methods
present
the
best
visual
reconstruction
ability.
However,
most
require
a
long
acquisition
time,
thus
failing
to
deal
with
real-time
tasks.
Here,
we
provide
dual
optical
coupling
model
describe
spatiotemporal
propagation
of
photons
in
free
space,
then
propose
an
efficient
non-confocal
transformation
algorithm
and
establish
time-to-space
boundary
migration
model.
Based
on
these,
scan-free
method
proposed.
The
data
speed
can
reach
151
fps,
which
∼7
times
faster
than
current
fastest
method,
while
overall
also
19
fps.
background
stability
brought
by
fast
makes
suitable
dynamic
scenes.
In
addition,
high
robustness
noise
have
capability
outdoor
environments
during
daytime.
To
further
enhance
practicality
this
real-world
scenarios,
exploit
statistical
prior
plug-in-and-play
super-resolution
extract
higher
spatial
resolution
signals,
reducing
detector
array
requirement
from
32
×
8
without
compromising
quality,
device
expense
detectors.
We
propose
a
novel
method
to
reconstruct
non-line-of-sight
(NLOS)
scenes
that
combines
polarization
and
time-of-flight
light
transport
measurements.
Unpolarized
NLOS
imaging
methods
objects
hidden
around
corners
by
inverting
time-gated
indirect
paths
measured
at
visible
relay
surface,
but
fail
scene
features
depending
on
their
position
orientation
with
respect
such
surface.
address
this
limitation
(known
as
the
missing
cone
problem)
capturing
state
of
in
systems
picosecond
time
resolution,
introducing
inversion
leverages
directionality
information
polarized
measurements
reduce
directional
ambiguities
reconstruction.
Our
is
capable
surfaces
inside
space
state-of-the-art
methods,
yielding
fine
reconstruction
details
even
when
using
fraction
points
demonstrate
benefits
our
both
simulated
experimental
scenarios.
