Share Email Print
cover

Proceedings Paper

Multi-frame restoration of turbulence degraded underwater images
Author(s): Andrey V. Kanaev
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The ability to image underwater is highly desired for scientific and military applications, including optical communications, situational awareness, diver visibility, and mine detection. However, underwater imaging is severely impaired by scattering and optical turbulence associated with refraction index fluctuations. Naturally, the approaches taken to solve the underwater image restoration problem have their origin in atmospheric turbulence compensation algorithms. There is certain similarity between the atmospheric and underwater image degradations but the difference in the scales of refraction index fluctuations in two media brings out the need for significant modifications of atmospheric techniques to be applicable to underwater imagery. Significantly stronger underwater image distortions resulting in large local shifts and warping of the image features require robust tracking using, for example optical flow estimation, even under relatively benign underwater conditions. Comparative performance of multi-frame nonlinear gain “lucky patch” algorithms with variable degree of optical flow technique sophistication is presented for underwater imagery collected in a laboratory tank and in a field exercise. Reliance of image restoration on accuracy of the optical flow algorithm is revealed and one approach to enhance restored image quality using confidence measures of optical flow estimation is proposed.

Paper Details

Date Published: 1 November 2012
PDF: 6 pages
Proc. SPIE 8535, Optics in Atmospheric Propagation and Adaptive Systems XV, 85350N (1 November 2012); doi: 10.1117/12.980433
Show Author Affiliations
Andrey V. Kanaev, U.S. Naval Research Lab. (United States)


Published in SPIE Proceedings Vol. 8535:
Optics in Atmospheric Propagation and Adaptive Systems XV
Karin Stein; John Gonglewski, Editor(s)

© SPIE. Terms of Use
Back to Top