The use of orthomosaicking and reconstruction of three-dimensional (3D) seabed in an inexpensive, asynchronous stereo framework has resulted in a better perception of the underwater scene by using image processing algorithms. The calibration process can be used to detect a rotational, independent way, and localize square corners by using line-fitting and label sorting strategies. The joint use of local and global matching method and imaging enhancement algorithms are considered to identify correct correspondences in severe underwater condition, namely, brightness constancy violation and backscattering. The Iterative Closest Point (ICP) algorithm and a new trend in frame registration, Zhang's epiflow algorithm, under simplified motion model are used. The matching algorithm is based on the underlying assumption of consistent brightness which is regularly violated in underwater environment. Brightness constancy violations are treated by using cross-correlation and image enhancement apart from four frame feature tracking algorithm for ego-motion estimation.
A 3D passive vision system for underwater mosaicking
Leone Alessandro;Diraco Giovanni;Distante Cosimo
2009
Abstract
The use of orthomosaicking and reconstruction of three-dimensional (3D) seabed in an inexpensive, asynchronous stereo framework has resulted in a better perception of the underwater scene by using image processing algorithms. The calibration process can be used to detect a rotational, independent way, and localize square corners by using line-fitting and label sorting strategies. The joint use of local and global matching method and imaging enhancement algorithms are considered to identify correct correspondences in severe underwater condition, namely, brightness constancy violation and backscattering. The Iterative Closest Point (ICP) algorithm and a new trend in frame registration, Zhang's epiflow algorithm, under simplified motion model are used. The matching algorithm is based on the underlying assumption of consistent brightness which is regularly violated in underwater environment. Brightness constancy violations are treated by using cross-correlation and image enhancement apart from four frame feature tracking algorithm for ego-motion estimation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.