Structured Light in Scattering Media

Virtually all structured light methods assume that the scene and the sources are immersed in pure air and that light is neither scattered nor absorbed. Recently, however, structured lighting has found growing application in underwater and aerial imaging, where scattering effects cannot be ignored. In this project, we conduct a comprehensive analysis of two representative methods - light stripe range scanning and photometric stereo - in the presence of scattering. For both methods, we derive physical models for the appearances of a surface immersed in a scattering medium. Based on these models, we present results on (a) the condition for object detectability in light striping and (b) the number of sources required for photometric stereo. In both cases, we demonstrate that while traditional methods fail when scattering is significant, our methods accurately recover the scene (depths, normals, albedos) as well as the properties of the medium. These results are in turn used to restore the appearances of scenes as if they were captured in clear air. Although we have focused on light striping and photometric stereo, our approach can also be extended to other methods such as grid coding, gated and active polarization imaging.


"Structured Light in Scattering Media"
S.G. Narasimhan, S.K. Nayar, B. Sun and S.J. Koppal
IEEE International Conference on Computer Vision (ICCV),
Oct, 2005.

"Structured Light Methods for Underwater Imaging: Light Stripe Scanning and Photometric Stereo"
S.G. Narasimhan and S.K. Nayar
IEEE/MTS Oceans,
September 2005.


Experimental Setup:
This picture shows the apparatus used in our light striping and photometric stereo experiments. The tank is made of transparent anti-reflection coated glass and contains the scattering medium (for example, dilute milk) and the scene. In the case of light stripe range finding, a projector (Infocus LP120) is used sweep a light plane across the scene. A video camera (Canon XL-1S) views the scene as well as the medium with the effects of scattering. In the case of photometric stereo, the same tank setup is illuminated by 11 light sources and is observed by a digital still camera (not shown in image).
Light Stripe Range Finding in Scattering Media:
This picture shows the results of applying our scene and medium recovery algorithms to two scenes. The detection of the object intersections and the 3D reconstruction obtained under different densities of scattering (dilutions of milk) compare well with the results obtained in clear air. Despite the strong effects of scattering, we are able to remove them completely to restore the original scene contrast. A comparison with the floodlit images shows that simply using bright sources does not enhance visibility in scattering media, and that new structured lighting methods that are designed to focus light on the scene to alleviate blurring and backscattering must be used.
Photometric Stereo in Scattering Media:
This picture shows the results of applying our photometric stereo algorithm that is adapted to work in scattering media. Scattering effects cause significant decrease in scene contrast and color and hence, applying the traditional algorithm directly on these images results in poor performance. Our modified method requires 5 images (instead of the usual 3) to correctly recover not only the surface normals and albedos, but also a depth map of the scene, which is not possible in the original method.
Detecting and reconstructing mirrors:
This picture illustrates the concept of reconstructing surfaces irrespective of their reflectances. The light plane is visible due to scattering until it hits a mirror where a dark vertical stripe is formed. This dark stripe can then be used for reconstruction.


(Video Result Playlist)
ICCV 2005 Video (use Apple Quicktime 6.0):
This video is a compilation of the main results of this project (30 MB). (With narration)
Light Stripe Range Scanning:
This video shows the recovery of the 3D structures of objects immersed in a scattering medium using the light striping method developed for scattering media (23 MB). (With narration)
Photometric Stereo:
This video shows the recovery of surface normals, depths and albedos of scene points using the photometric stereo algorithm developed for scattering media (6 MB). (With narration).