Shawn F. Johnson

High-Frequency Scattered Envelope Statistics of Patchy Seafloors

The scattering properties of the individual components of seafloors consisting of discrete patches of different materials may have a complicated relationship in terms of their angular response and/or frequency dependence. Consequently, this relationship directly influences the angular and frequency response of scattered envelope probability density functions (pdfs). In this paper, the influence of the relative scattering strength of seafloor patches on scattered envelope statistics will be explored through both a modified form of a recently developed model [Abraham and Lyons, IEEE J. Ocean. Eng., vol. 27, pp. 800-813, 2002] and analysis of experimental data collected off Elba Island, Italy, in May 2003, by the NATO Undersea Research Centre, La Spezia, Italy. Qualitative comparisons of the K-distribution shape parameter (alpha) between that predicted by the model and that measured from data display promising similarities such as the inverse relationship between alpha and bandwidth, the relative difference in values of alpha for the various seafloor types studied, and the dependence on grazing angle. The favorable model/data comparisons show that it is possible to link the scattered envelope distribution to measurable geoacoustic properties, providing the foundation necessary for solving several important problems related to the detection of targets in non-Rayleigh clutter including performance prediction for different systems based on seafloor properties, extrapolation of performance to other system/bandwidths, and optimization of system parameters such as bandwidth to local environment.

Estimating seafloor roughness using synthetic aperture sonar image statistics

In this work, we present a model to predict the impact of intensity scaling caused by random seafloor roughness on SAS image speckle statistics and the possible use of this model to estimate roughness parameters, such as root-mean-square height and slope. The continuous variation in scattering strength produced by roughness (i.e., roughness-induced changes in seafloor slope) is treated as an intensity scaling on image speckle produced by the SAS imaging process. Changes in image statistics caused by roughness are quantified in terms of an effective K-distribution shape parameter. Comparisons between parameter estimates obtained from the scaling model and estimates obtained from high-resolution SAS data collected in experiments off of the Ligurian coast near La Spezia, Italy, are used to illustrate the efficacy of the model. [Work performed under ONR Grants N00014-10-1-0051, N00014-10-1-0047, and N00014-10-1-0151].

Predicting acoustic backscattering from bioturbated seafloors

Seabed roughness is an important property for high-frequency scattering. However, for certain littoral environments, the surficial roughness of the seabed is often continuously being modified by hydrodynamic and biologic forces. For example, ocean surface waves may create orbital ripples on sand seafloors which are characterized by anisotropic roughness, and acoustic scattering from such a seafloor would be strongly aspect-dependent. Yet bottom-dwelling organisms rework the sediment, returning the structured seabed relief to random isotropic roughness, where acoustic scattering is more uniform with respect to the aspect ensonified. In this talk, we will utilize both analytic expressions as well as procedural generation methods to predict the two-dimensional roughness equilibrium power spectra of rippled and bioturbated seafloors, and the resulting impact on acoustic scattering prediction.

A point-based scattering model for the incoherent component of the scattered field

A numerical model for calculation of the incoherent component of the field scattered from random rough surfaces is described. This model is based on the point scattering approach, where the mean scatterer amplitudes are calculated from deterministic models. These amplitudes are then scaled by a complex circular Gaussian random variable to simulate scattering from a surface with minimal coherence length. The resulting simulated fields are shown to agree with theory for the mean field, mean square field, statistical distribution, and the spatial coherence length.

Seafloor measurements using synthetic aperture sonar

The past decade has seen considerable growth in the use of synthetic aperture sonar (SAS) imaging systems in both the civilian and military domains. Although SAS systems are almost always uncalibrated, they can still yield information about the seafloor given an understanding of the mechanisms affecting the statistical properties of the images produced by these systems. This talk will describe our recent efforts to link SAS image statistics to seafloor properties through the use of seafloor scattering models. Sample results from several SAS systems encompassing frequencies ranging from 6 to 300 kHz will be shown.

Acoustic demonstrations for the iPod generation.

Today’s “iPod generation” can gain a unique and interactive exposure to physics by leveraging their musical interests, and their iPod, to explore the science of sound. Inside certain models reside a mini‐computer and a suite of acoustic and vibration sensors, including loudspeakers, microphones, and accelerometers. Combined with free or inexpensive software applications or “apps,” these highly portable devices can offer a tangible means to help students understand many physics, audio, and acoustics concepts that they can explore both inside and outside the classroom. We will demonstrate several apps that are currently available, for example, a seismometer, signal generator, noise level meter, spectrum analyzer, oscilloscope, and various musical instruments. Demonstrations utilizing an iPod Touch are guaranteed to excite students of all ages.

Characterization and modeling of synthetic aperture sonar imagery.

The characterization and modeling of synthetic aperture sonar (SAS) image statistics are of importance for developing target‐on‐background detection and classification algorithms and for developing specialized filters for speckle noise reduction. A simple model is presented to predict the impact of amplitude scaling caused by seafloor ripples on SAS image speckle statistics. The continuous variation in scattering strength produced by ripples (i.e., ripple‐induced changes in seafloor slope) is treated as a deterministic amplitude scaling on image speckle produced by the SAS imaging process. Changes in image statistics caused by ripples are quantified in terms of an effective K‐distribution shape parameter. Agreement between shape parameter estimated from the scaling model and from SAS data collected in experiments off of Panama City, FL and off of the Ligurian coast of near La Spezia, Italy illustrates the efficacy of the model. [Work supported by ONR Grant Nos. N00014‐04‐1‐0013 and N00014‐06‐1‐0245.]

Simulation of rippled‐sand sonar imagery with visual and statistical verification.

Synthetic aperture sonar (SAS) imagery is often characterized by a decidedly non‐Rayleigh pixel amplitude distribution, owing to its inherent high‐resolution combined with speckle induced by the coherent image formation process. Recent work has shown increasing the resolution cell size by limiting the image bandwidth typically produces images with statistics tending toward a Rayleigh distribution. A caveat to this generalization is that the trend toward Rayleigh is dependent on orientation of the SAS system to strongly correlated sea‐floor structures such as sand‐ripples. A method has been developed to simulate rippled‐sand sonar imagery at high‐frequencies (i.e., on the order of 100 kHz), which accounts for nonsymmetric ripple shape, sediment acoustic properties, sonar to ripple orientation, system resolution, and coherent imaging induced speckle. This numerical simulation method is computationally inexpensive and compares well both visually and statistically with collected data over a wide range of orie...

Statistics of synthetic aperture sonar image resolution degradation.

Synthetic aperture sonar (SAS) image statistics can often be characterized by a probability density function with a heavier tail than the expected Rayleigh distribution. The K‐distribution shape parameter can be used as a metric of non‐Rayleighness, with physical ties to both seafloor properties and sonar parameters. Recent results have shown that increasing the resolution cell size, or degrading the resolution of the image, produces images with statistics tending toward Rayleigh (i.e., a higher K‐distribution shape parameter). In general, a doubling of the resolution cell area results in a doubling of the K‐distribution shape parameter. A caveat to this generalization is the orientation of the sonar system to any features that may exist on the seafloor (i.e., sand ripples). In such a situation, image statistics may continue to be significantly non‐Rayleigh for certain orientations in spite of resolution degradation. SAS images of seafloors with various bottom types and feature orientations collected with...

Effect of Resolution on Synthetic Aperture Sonar Image Statistics

Synthetic Aperture Sonar (SAS) images often exhibit large inter‐pixel intensity fluctuations owing to the small resolution scales of such systems, with the reverberation envelope probability density potentially being significantly heavier‐tailed than the Rayleigh distribution traditionally assumed for lower‐resolution systems. The K‐distribution has been shown to well describe sonar reverberation over a wide range of resolutions, with the shape parameter being proportional to the resolution cell size or inversely proportional to the bandwidth. The resolution cell size can be increased by restricting the bandwidth of the data in post processing, which in turn reduces the tails of the reverberation envelope for a higher matched K‐distribution shape parameter. Broadband acoustic data suitable for SAS processing collected as part of the SAX04 experiment has been analyzed, and results concerning image resolution and statistics of such data will be presented.