# Three-dimensional underwater acoustical imaging Andrea Trucco 1 , Maria Palmese 1 , Andrea Fusiello 2 , and Vittorio Murino 2 1 Department of Biophysical and Electronic Engineering, University of Genova, Italy 2 Department of Computer Science, University of Verona, Italy 1. INTRODUCTION Acoustic imaging is an active research field that aims to study techniques for the formation and processing of images generated by raw signals acquired by an acoustic system [1]. Our purpose is to present a brief survey concerning the generation and processing of acoustic images for underwater applications [2,3], especially focusing on algorithms for three-dimensional (3-D) imaging. Like optical systems, acoustic systems can generate an image by processing the waves backscattered from the objects of a scene. The relative easiness of measuring the time-of-flight of an acoustic signal makes it possible to generate not only acoustic 2- D images similar to optical ones but also range estimates that can be used to produce a real 3-D map. The main advantage of acoustic imaging systems over optical ones is that they properly work also beyond the optical visibility range. Underwater optical vision provides images with finer resolution, but its range is limited to a few tenths of meters in very good water conditions, and the situation is even worse when mud in suspension raised by underwater tasks sharply reduces visibility [2]. Although larger than the range of optical visibility, the range of applicability of a 3-D acoustic imaging system varies according to the specific sensor adopted and the signal frequency characterizing the sensor. Generally, high frequencies are utilized (from about a hundred kilohertz to few megahertz) over a range going from some centimeters to a hundred meters. In general, the scene under investigation is first insonified by an acoustic signal, then the backscattered echoes acquired by the system are processed to create an image of the scene. This process can be performed by two different approaches [3]: