Computer model for predicting underwater color images

Abstract

A computer model to simulate the formation of color underwater images has been developed. The model simulates the appearance of an underwater image as seen by an underwater camera. The images are intended to portray the appearance of a planar reflectance map that has been illuminated with artificial illumination. The input to the model consists of the geometry of the camera, artificial lights, the environmental constants that govern the propagation of light underwater, and the reflectance values of the map. To simulate the appearance of color images the spectrally varying nature of these inputs has been taken into account. This article describes the algorithm and illustrates the images that can be obtained as a function of different water types and camera/light configurations. A specific imaging geometry is used to simulate the appearance of objects that would be seen by a towed underwater imaging platform. The images have been synthesized in two types of ocean water at several distances. The results indicate the important role that frequency dependent scatter and absorption play in the process of underwater image formation.