|
On Controlling
Light Transport in Poor Visibility Environments
|
|
  
|
|
Abstract
Poor visibility conditions due to
murky water, bad weather, dust and smoke severely impede the performance of
vision systems. Passive methods have been used to restore scene contrast
under moderate visibility by digital postprocessing. However, these methods
are ineffective when the quality of acquired images is poor to begin with.
In this work, we design active lighting and sensing systems for controlling
light transport before image formation, and hence obtain higher quality
data. First, we present a technique of polarized light striping based on
combining polarization imaging and structured light striping. We show that
this technique out-performs different existing illumination and sensing
methodologies. Second, we present a numerical approach for computing the
optimal relative sensor-source position, which results in the best quality
image. Our analysis accounts for the limits imposed by sensor noise.
|
Publication
“On Controlling Light Transport in Poor Visibility Environments”
Mohit Gupta, SG Narasimhan,
YY Schechner
IEEE Computer Vision and Pattern
Recognition (CVPR) 2008
[PDF]
Presentation
|
Pictures (click on thumbnails to enlarge images)
|
 
|
|
Experimental
Setup:
Our experimental setup consisting of a glass tank, filled with moderate
to high concentrations of milk. An LCD projector illuminates the medium with
polarized light. The camera (with a polarizer attached) observes a
contrast chart through the medium.
|
|
|
|
Comparison
of various illumination and sensing techniques:
We compare the performance of various
techniques such as flood-lighting, polarized flood-lighting, light
striping and polarized light-striping. We consider moderate and heavy
scattering conditions. We can notice improvement in contrast using our
technique of polarized light-striping over previous techniques. See full
size image to avoid breaking-up text.
|
|
|
|
Limitations of the high-frequency
illumination based method:
In the presence of moderate to heavy
volumetric scattering, the direct component images have low SNR. The
global image is approximately the same as a flood-lit image, and hence,
suffers from low contrast.
|
|
|
|
Unpolarized vs. Polarized Light Stripe
Scanning:
Using
polarization reduces backscatter, thereby enabling reliable detection of
the intersection of light sheet with the object. Thus, we can improve
image contrast considerably using polarization imaging+light stripe
scanning.
|
|
|
|
What is the optimal sensor-source separation for flood-lighting?
Large separation (60 cms) results in heavy
image noise. On the other hand,
optimal separation (40 cms) results in a high contrast, low noise
image
Both the frames were captured with the same
exposure time.
|
|
|
|
What
is the optimal light stripe scan for the best image quality?
Using
computer simulations, we can design the optimal light stripe scan for the
best image quality. The image quality is quantified in terms of the image
contrast, image SNR, and the gradient across the edge of the stripe resulting
from the intersection of the light sheet with the object.
|
|
|
|
|
