John Dubberley

Change detection using Synthetic Aperture Sonar: Preliminary results from the Larvik trial

In April of 2011, FFI led a sea trial near Larvik, Norway on FFIs research vessel the H.U. Sverdrup II with participation by representatives from Canada, United States, and France. One objective of the sea trial was to acquire a data set suitable for examining incoherent and coherent change detection and automated target recognition (ATR) algorithms applied to Synthetic Aperture Sonar (SAS) imagery. The end goal is to produce an automated tool for detecting recently placed objects on the seafloor. To test these algorithms two areas were chosen, one with a comparatively benign seafloor and one with a boulder strewn complex seafloor. Each area was surveyed before and after deployment of objects. The survey time intervals varied from two days to eight days. In this paper we present the trial and show examples of SAS images and change detection of the images.

Herring hydroglyphics in littoral waters of the northern Gulf of Mexico

A large shoal of fish occurring in the vicinity of the 220‐m isobath was observed using a standard 38‐kHz fisheries echosounder and a 1.5‐ to 10‐kHz low‐frequency fish sonar (LFFS) for several days in July 2000. The fish behaved like herring, exhibiting a rapid rise to the sea surface at dawn, formation into schools, and a rapid descent to the sea floor. Schools remained at depth through the day and gradually rose to the sea surface at dusk and then rapidly descended and spread out into a diffuse scattering layer at 75‐ to 125‐m depth. Shifts in resonance frequencies during migration, release of gas bubbles during migration, and strong avoidance of the vessel when maneuvering, all suggest the fish were most likely round herring, Etrumeus teres, which are common at these depths in the NMFS historical trawl survey data. An examination is made of some of the scattering characteristics of the schools and layers of these fish and comparisons of the 38‐kHz data to scattering at 500‐Hz bands from 1.5 to 5 kHz an...

Change detection deconfliction process for sonar clutter items.

When resurveying a seafloor area of interest during change detection operations, an automated method to match found bottom objects with objects detected in a previous survey allows the surveyor to quickly sort new objects from old objects. Here we will demonstrate a software system that accomplishes change detection. The change detection system contains modules for automatic object detection by geospatial bitmap technique, object collocation, feature matching using shadow outlining, scene matching by control point matching, and visualization and filtering capabilities. Emphasis will be placed on the new elements of the system, namely, shadow outlining and optional spatial filtering.

Adaptation and improvement of ASW tactical decision aid design to Mine Warfare tactical decision aid

In this presentation, the adaptation of a tactical decision aid (TDA) designed for Anti Submarine Warfare (ASW) to one used for Mine Warfare is discussed. Previous work at the Naval Research Laboratory established a TDA, namely GRASP for the ASW role. Recently at NRL work has been done to extend the TDA model to Mine Warfare. Specifically this presentation examines search path algorithms for the two TDAs. Comparisons will be shown between various search path algorithms for environmentally influenced mine detection: GRASP, edge covering, and greedy algorithms for limited time searches and searches for assured threshold confidence of detection. Oceanic environmental parameters relevant to designing tactical decision aids for mine detection will be noted. In conclusion this presentation discusses how extensively detailed must the environmental survey be in order to make search path analysis sensitive to the data.

Automated change detection with area matching

When resurveying a geographic area of the seafloor during sidescan change detection operations, an automated method to match bottom objects imaged previously with objects imaged in the resurvey can increase efficiency and accuracy. The geographic position of a new object relative to a historical object is a good indicator of a match. However, due to position error within either survey, there may be more than one spatially‐close object in the new imagery. To complicate matters further, the reflected energy from the new object may be significantly different given a different incidence angle in the resurvey or the partial burial of the object. In addition, the resurveyed object image may be below the threshold set for automatic recognition and falsely eliminated. This presentation will address these problems and suggest possible methods for matching “constellations” of bottom objects by Dijkstras minimum cost ‐ maximum flow algorithm, control point matching, and the data‐association procedure.

Fuzzy Clustering of Oceanographic Sound Speed Profiles for Acoustic Characterization

Historic oceanographic sound speed profiles have traditionally been grouped by area and time period, usually one degree square area and monthly time. After grading the profiles, mean profiles and standard deviations are calculated from the accepted profiles and in the acoustics community they are then used to predict the expected acoustic response of the region. Here the historic profiles in NOAAs World Ocean Database 2005 (WOD2005) will be divided into the same area and time periods, but in subsets with a sufficient number of profiles, fuzzy clustering will be employed on acoustically relevant oceanographic parameters (mixed layer depth, surface temperature, sound speed gradient, etc) to divide the population into multiple clusters. A parabolic equation acoustic transmission model is then applied on the WOD2005 statistical profiles and on the fuzzy cluster populations. Conclusions will be drawn about the suitability of this clustering to capture the variability of acoustic response at a given time and p...

Beam simulation error estimation using the scattering operator

Beam simulation has been used to estimate scattering from large surfaces by calculating the scattering from the individual subsurfaces and then summing over these subsurfaces. Previously physical intuition and comparison to method of moments (MoM) calculations were used to determine beam simulation’s region of validity. Here the MoM operator is used for a direct comparison of beam simulation to MoM. The improved speed of comparison allows a more in‐depth analysis of beam simulation at a lower computational resource cost. Scattering from one‐dimensional random surfaces with Gaussian height distributions will be specifically examined to verify and clarify earlier beam simulation validations. Although one‐dimensional surfaces are an oversimplification for most applications, they are well studied and are likely to give insight into higher dimensional applications.

Modeling the azimuthal dependence of sea surface backscatter during CST‐7

The operator expansion method is applied to realizations of 2‐D roughness measured during CST‐7 in order to probe the azimuthal dependence of the backscatter from time‐evolving, non‐fully developed seas which contain swell from distant surface wave sources. To use wave buoy measured data without extrapolation in frequency beyond the measured surface wave spectra, the measured frequencies examined will be limited to 75 and 150 Hz. The modeling goal will be to determine that if surface scatter was alone responsible for acoustic backscatter from the surface, could there be a measurable azimuthal dependence in the scattering during this experiment? Comparisons to some available measured azimuthal dependence will be shown.

Higher‐order perturbation theory versus the scattering operator expansion method for 3‐D surface scattering

To this point little work available in the literature has examined whether three‐dimensional scattering events can be successfully modeled by a two‐dimensional stimulation. This has mainly been caused by a lack of sufficient computational and modeling power in the past. A contributing reason to the low interest in the validity of this assumption is that there seems to be no compelling reason why the two‐dimensional assumption would fail to correctly approximate the three‐dimensional situation. However, in the interest of completeness, the ramifications of this assumption must be explored. Here two models are used to provide a preliminary look at this issue. Perturbation theory and the scattering operator expansion surface scattering methods will be compared to each other over a variety of canonical and ‘‘realistic’’ surfaces. The strengths and weaknesses of each model for this task will be mentioned. Up to third‐order implementations of these methods will be used.

Operator expansion error estimation by higher‐order terms.

The operator expansion method for quickly solving the Helmholtz boundary value equation was outlined by Milder [J. Acoust. Soc. Am. 89, 529–541 (1991)]. This method was verified by Kaczkowski and Thorsos for impenetrable surface scattering where Kirchhoff and small perturbation theory hold [J. Acoust. Soc. Am. 90, 2258 (A) (1991)]. In this presentation the third‐order and higher terms of the symmetric operator expansion will be derived in order to place error bounds on the integral approximation for second‐order and higher terms of the expansion. These error bounds can then be used to determine the regions of validity for the next lower‐order expansion approximations. Some examples and comparisons to other scattering methods will be shown.