Does flicker underwater disturb underwater photography?

Technion researchers find a way to utilize the “light field” to reconstruct the 3D structure of an underwater photo

Researchers from the Technion’s Faculty of Electrical Engineering have found a way to utilize the phenomenon of underwater flicker to reconstruct the three dimensional structure of a photograph. Natural lighting underwater significantly varies with time and space. The researchers explain that the reason for this is that the way in which surface waves scatter the light penetrating the water changes all the time. The “light field” created underwater as a result of this is called flicker. In previous research, flicker was considered an undesirable phenomenon for the quality of the photos. Now, the Technion researchers have found that the phenomenon actually can be useful and can be used to coordinate stereoscopic imaging (in stereoscopic imaging, two cameras are used to photograph the same scene, just like a pair of eyes. Using coordination between the two cameras, it is possible to reconstruct the three dimensional structure of the pictured being photographed).

In research from 2004 by Prof. Yoav Schechner and Nir Karpel from the Technion’s Faculty of Electrical Engineering, the pair proposed a way to reduce the phenomenon – processing video pictures after photographing in order to create more uniform pictures that look as if they were photographed under stable lighting conditions. In their latest research, Prof. Schechner and graduate student Yohay Swirski show that flicker can be useful for underwater computerized vision.

“The phenomenon of underwater flicker creates a unique time/space signature for every point in the scene,” explain the Technion researchers. “This signature can be used to identify and coordinate two or more photographs, which depict the same scene photographed at the same time. By coordinating a large number of points between the different photographs, it is possible to reconstruct the three dimensional structure of the photographed scene precisely and densely.”

Without the phenomenon of flicker, the problem of coordinating stereoscopic imaging is considered one of the complicated problems of computerized vision, especially in areas in which there is no texture or spatial details in the scene – such as partial walls or sandy bottoms. Flicker enables solving the coordination problem with relatively simple tools and in a very accurate way using spatial and time information of the “field of light”.

The method was investigated and proven in various tests both in a swimming pool and in the Mediterranean Sea under various light and visibility conditions and presented at the International Conference on Computerized Vision (ICCV) in 2009 in Japan.