Bathymetry Measurement by Swath Sonar Systems: The Trade-off Between Accuracy and Resolution

Xavier Lurton
CCOM/JHC Visiting Scholar

Underwater Acoustics Laboratory
Brest, France

Friday, Oct. 21, 2011, 3:00pm
Chase 130

Interferometry is the prevalent processing technique used in swath bathymetry sonars, with two main variants (multibeam echosounders and interferometric side-scan sonars); although the efficiency and accuracy of this technique are well-known and universally admitted, the objective estimation of its performances is not always well understood.

This seminar, after reminding the working principles of bathymetry measurements by swath sonars, proposes to present and discuss the "accuracy-resolution trade-off." This classical dilemma in experimental Physics is nicely illustrated by swath bathymetry sonars: since most often the interferometric measurement is perturbed by intrinsic random fluctuations, the required bathymetry accuracy imposes some kind of spatial averaging of the raw signals, which is detrimental to the resolution of the measurement.

In multibeam echosounders, the interferometric phase ramp (computed inside one beam as a function of time) is fitted with a polynomial, whose intersection at zero-phase gives the instant corresponding to the sounding, the angle being defined by the beam steering value. This decreases the noise level, however at the price of a degradation of the horizontal resolution. This point is usually neglected in performance analysis: hydrographic standards impose a depth measurement accuracy, but fail to specify which horizontal resolution this accuracy is associated with. At the opposite, in interferometric side-scan sonars, the output data is provided as series of individual points giving each a (range, angle) solution at a high sampling rate, hence with potentially an excellent resolution power; unfortunately these point values are submitted to a high level of fluctuations, and are not usable under this form; some averaging is then needed, implying a loss in resolution.

Several solutions are practically possible for reaching an acceptable trade-off; however the issue still deserves some attention from constructors as well as from users and operators. This is illustrated by simulation results and experimental data (case study of EM 3002 on RV Belgica).


Xavier Lurton (b. 1955 in Bordeaux, France) graduated in Physics in 1976 (Universite de Bretagne Occidentale, Brest) and received a Ph.D. in Applied Acoustics in 1979 (Universite du Maine, Le Mans), specializing first in the Physics of brass musical instruments. After spending 2 years of national service as a high-school teacher in Ivory Coast, he was hired by Thomson-Sintra (the French leading manufacturer in the field of military sonar systems - today Thales Underwater Systems) as a R&D engineer, and specialized in underwater propagation modeling and system performance analysis. In 1989 he joined Ifremer (the French government agency for Oceanography) in Brest, where he first participated in various projects of underwater acoustics applied to scientific activities (data transmission, fisheries sonar, ocean tomography…). Along the years he specialized more specifically in seafloor-mapping sonars, both through his own technical research activity (both in physical modeling and in sonar engineering) and through several development projects with sonar manufacturers (Kongsberg, Reson); in this context he has participated in tens of technological trial cruises on research vessels. He has been teaching underwater acoustics for 20 years in several French universities, and consequently committed An Introduction to Underwater Acoustics (Springer) widely based on his own experience as a teacher. He manages the Ifremer's team specialized in underwater acoustics, and has been the PhD advisor of about 15 students. Presently he is for 6 months a visiting scholar at UNH, working on issues related to sonar reflectivity processing, and bathymetry measurement methods.