Title: Automatic Restoration of Underwater Monocular Sequences of Images

Authors: Paulo Drews-Jr, Erickson R. Nascimento, Mario F. M. Campos, Alberto Elfes

Abstract: Underwater environments present a considerable challenge for computer vision, as water is a scattering medium with substantial light absorption characteristics made even more severe if the water is turbid. This poses significant problems for visual underwater navigation, object detection, tracking and recognition. Previous works tackle the problem by using unreliable priors or expensive and complex devices. This paper adopts a physical underwater light attenuation model which is used to enhance the quality of visual features of the images and the applicability of traditional computer vision techniques for underwater images. The proposed method simultaneously estimates the attenuation parameter of the medium and the depth map of the scene to compute the image radiance, i.e. reducing the effect of the medium in the images. Our approach is based on an novel optical flow method, which is capable of deal with scattering media, and a new technique that robustly estimates the medium parameter. Combined with structure-from-motion techniques, the depth map is estimated and a model-based restoration is performed. The method was tested both in simulation and with real sequences of images. The experimental images were acquired by a camera on a remotely operated vehicle (ROV) in naturally lit shallow seawater. The results show that the proposed technique allows substantial restoration of the images, thereby improving the ability to identify and match features, which in turn is essential for other computer vision algorithms such as object detection and tracking and autonomous navigation .