Jules Jaffe - Academia.edu (original) (raw)
Papers by Jules Jaffe
Acoustical Imaging, 2011
ABSTRACT Estimating biophysical parameters of fish populations in situ such as size, orientation,... more ABSTRACT Estimating biophysical parameters of fish populations in situ such as size, orientation, shape, and taxa is a fundamental goal in oceanography. Towards this end, acoustics is a natural choice due to its rapid, non-invasive capabilities. Here, multi-view methods are explored for classification, size and orientation estimation, and 2D image reconstruction for individual fish. Size- and shape-based classification using multi-view data is shown to be accurate (~10% error) using kernel methods and discriminant analysis. For species-based classification in the absence of significant differences in size or shape, multi-view methods offer significant (~40%) reduction in error, but absolute error rates remain high (~20%) due to the lack of discriminant information in acoustic scatter. Length and orientation estimation are investigated using a parameter-based approach with a simple ellipsoidal scattering model. Good accuracy is obtained when the views span the full 360°. When the span is limited to less than 60°, incorporating a prior constraint on possible body shapes can lead to reduced uncertainty in the estimated parameters. Finally, using views that span the full 360°, sparse Bayesian learning coupled with a conventional Radon transform yields accurate two-dimensional, projected images of the fish. KeywordsAcoustics- Fish classification -Scattering-Size estimation- Radon transform - Bayesian learning -Distorted wave Born approximation
To understand the physical and biological dynamics controlling the micro-and fine-scale distribut... more To understand the physical and biological dynamics controlling the micro-and fine-scale distributions of phytoplankton and zooplankton in the ocean. OBJECTIVES To use multispectral optical measurements and three-dimensional acoustic imaging to quantify the microscale and fine-scale vertical patchiness of several taxa of phytoplankton and large zooplankton in situ. APPROACH Our approach was to develop a multispectral in situ imaging system: LUMIS (Low light level Underwater Multispectral Imaging System). Using a sheet of laser light to irradiate a ~1 l sample volume, phytoplankton fluorescence from chlorophyll a and phycoerythrin were imaged with a sensitive CCD camera. In addition to the two fluorescence wavelengths, optical backscatter and water Raman scattering were imaged simultaneously.
Oceanographers are limited by existing technology when observing aquatic microbes that are undist... more Oceanographers are limited by existing technology when observing aquatic microbes that are undisturbed in their natural environment. A Thin Light Sheet Microscope (TLSM (1)), that offers a simple and easy to implement technique for such studies, is described here. The TLSM utilizes conventional microscope optics but replaces the illumination apparatus with a thin (23 micron) sheet of laser light. The light sheet has transverse dimensions of 6 mm by 1 mm , positioned precisely at the depth of field of the microscope's objective lens. The sample volume above and below the light sheet is not exposed to excitation light. The TLSM concentrates light only where excitation illumination is needed, thus maximizing the efficiency of the illumination source and minimizes the background noise level in the image, which maximizes image sharpness. A fluorescent biological marker that connects only to living organisms in the water was added to the test sample. Particles that are not located wit...
We intend to quantify the biological, physical, and chemical dynamics that structure marine plank... more We intend to quantify the biological, physical, and chemical dynamics that structure marine planktonic ecosystems. Observations of the organisms and their environment on the spatial and temporal scales that characterize their interactions, combined with models of the dominant dynamics, will lead to improved understanding of the dynamics, structure, and function of planktonic ecosystems. OBJECTIVES Our objectives in this work are to 1) visualize and quantify herbivorous copepod feeding in the laboratory, and 2) to apply these methods in the field to observe the dynamics of copepod feeding in situ. In particular we intend to test the "feeding sorties" hypothesis vs. the "in situ feeding" hypothesis regarding the location and timing of copepod feeding and vertical migration.
Using a Planar Laser Imaging Fluorometry (PLIF) system mounted on a free-falling platform, we hav... more Using a Planar Laser Imaging Fluorometry (PLIF) system mounted on a free-falling platform, we have observed the spatial distribution patterns and size-frequency spectra of large fluorescent particles in the upper ocean. In drops of the system off San Diego, CA, we found that the size-frequency spectrum was strongly related to the total phytoplankton biomass (measured as chlorophyll a). Differences of the observed spectra from those predicted based on biomass showed regions suggesting aggregation and sinking of particles, systematic and sudden changes in spectra with depth, and associations of anomalously large particles with hydrographic features. The relative spatial distributions of particles in the images showed the particles to be non-randomly distributed on scales of a few cm, except in regions of high concentration (the subsurface chlorophyll maximum). Some images in every profile were found to be streaky, indicating that the water moved while the camera shutter was open. A ca...
We have recently developed and built a novel free-falling platform with a stereoscopic Digital Pa... more We have recently developed and built a novel free-falling platform with a stereoscopic Digital Particle Image Velocimeter (DPIV) to observe and quantify microscale physical and biological structures in the upper ocean. Scattered light from a vertical sheet of laser illumination is imaged from both sides by sensitive CCD cameras. Sequences of images allow two-dimensional maps of three-component velocity to be constructed from cross-correlations between images. The profiler has the potential to provide direct estimates of turbulent kinetic energy dissipation in the near-surface open ocean. Dissipation rates can be estimated using finite-difference approximations for 8 of the 12 velocity gradient terms. Using a filter wheel in front of the camera lenses, scattered light and two wavelengths of fluoresced light can be imaged sequentially, allowing mapping of fluorescent particles to the turbulent structures in the water column. The spatial and temporal resolution of the system is set by ...
We developed and deployed a 2D planar laser imaging fluorometer system mounted on a free-falling ... more We developed and deployed a 2D planar laser imaging fluorometer system mounted on a free-falling vehicle. The vehicle also carried a CTD/fluorometer, an optical nitrate sensor, and a microstructure profiler, to acquire vertical profiles of particle fluorescence and chemical and physical distributions in the field. Our data, gathered 10 km offshore of San Diego, showed large variations in fluorescent particle properties (size-frequency distributions) with depth, with changes over 2-5 m that were larger than changes seen at a single depth over 1000's of kilometers. The phytoplankton were randomly distributed over scales <10 cm, but shoed pronounced gradients over larger scales. Layers -1 m thick of unusual particle properties (e.g., large particles, or long particles) were common, though these layers did not appear as features in the total fluorescence profiles. Such cryptic layers are likely a common occurrence, driven by the interleaving of existing horizontal patches by the ...
LONG-TERM GOALS Our overall goal is to understand the perceptual and mechanistic principles that ... more LONG-TERM GOALS Our overall goal is to understand the perceptual and mechanistic principles that underlay camouflage framed in the context of the animals' environment. In particular, we plan to characterize and understand the perceptual abilities of several species of benthic and pelagic cephalopods (which are unrivaled masters of dynamic camouflage), the aspects of their optical environment that affect their camouflage behavior, the characterization of that behavior, and the molecular mechanisms inside the skin by which those responses are accomplished. OBJECTIVES 1. To fully characterize the spatiotemporal characteristics of the near-surface and shallow benthic underwater light field, including ultraviolet radiation and polarization. 2. To determine the visual abilities of several species of cephalopod and model both the shallow and deep-water world from the animals' points of view. Public reporting burden for the collection of information is estimated to average 1 hour pe...
IEEE Transactions on Acoustics, Speech, and Signal Processing, 1990
ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing, 1988
ABSTRACT A sonar imaging technique which uses specially coded waveforms to project different soun... more ABSTRACT A sonar imaging technique which uses specially coded waveforms to project different sound signals in different directions was previously proposed. In this context, the problem of creating a set of frequency-hopped code words that have mutually small correlation properties has arisen. The authors report on their progress in designing these frequency-hopped signals for multibeam sonar imaging. The special case is considered in which medium spreading effects destroy the phase coherence of the imaging system. A design strategy is discussed which has allowed the design of a set of frequency-hopped code words that exhibit good performance bounds with respect to incoherent processing. The design strategy is a systematic constructions technique which can be used with any size Galois field GF( p ) to generate p -1 codes of length p . Various theorems are presented and examples of the performance of the technique are illustrated for time-bandwidth products of interest in terms of sonar
Ocean Optics XIII, 1997
ABSTRACT This article presents the results of analyzing data which were collected in the sea with... more ABSTRACT This article presents the results of analyzing data which were collected in the sea with a structured illumination/fluorescence imaging system. The system employs all lines of an argon ion laser to create a 2D slice illumination pattern in the blue-green. A sensitive CCD camera was used to measure the fluorescence emission resulting from this stimulation. Under our assumptions, these images are proportional to chlorophyll-a concentration. Inspection of the data reveals a large degree of small scale patchiness with characteristic distances as small as centimeters. In many of the images, the chlorophyll concentration is seen to change by an order of magnitude over several centimeters. Patchiness was characterized via the use of spatial spectral estimation techniques. Preliminary results show that the data follow a -5/3 slope at high wave number, consistent with the hypothesis that the patches were formed by turbulent stirring of larger scale gradients of phytoplankton. In addition, repeated vertical profiles demonstrated that the larger scales were extremely constant, even in the presence of this small scale variability. The technique has provided an original data set with important new information about the small scale structure of pigment in the ocean.
2009 Workshop on Applications of Computer Vision (WACV), 2009
Journal of the Optical Society of America. A, Optics, image science, and vision, 2010
A set of algorithms is proposed to retrieve the size of spherically symmetric particles from the ... more A set of algorithms is proposed to retrieve the size of spherically symmetric particles from the measured intensity of angular scatter data. Of special interest are low-contrast particles whose real part of the index of refraction is between 1.03 and 1.09 and whose size ka is constrained so that pi < or = ka < or = 16pi, where k=2pi/lambda and a is particle radius. Several algorithms are evaluated and compared that are based on either simple matching to the Mie theory predictions or inverse tomography methods. In the tomography methods, a previously proposed algorithm [Opt. Express. 15, 12217 (2007)] was used after estimating the phase of the scattered data or adapted to use intensity-only data. In order to ensure stability, all algorithms' performance was evaluated in the presence of moderate noise. The performance varied as a function of particle size, refractive index, and algorithm. Results suggest that a scattering device that collects only the angular scatter that is...
Proceedings of the 1st ACM international workshop on Underwater networks - WUWNet '06, 2006
ABSTRACT With the increasing sophistication of both manned and unmanned systems for remote ocean ... more ABSTRACT With the increasing sophistication of both manned and unmanned systems for remote ocean exploration, a wealth of knowledge about heretofore-unknown oceanic processes has become available. However, no technologies currently exist to observe organisms and processes without disturbing them, as they move with the natural motion of the oceans. We propose a new class of ocean sensing, whereby free-floating underwater devices operate autonomously and collaborate through an acoustic underwater network between them. This new class of sensing will provide a window into understanding the multifaceted interactions between the ocean's currents, underwater ecosystems and our impact on them. In this paper, we will present the design of our underwater vehicle, which drifts freely with the ocean currents and is equipped with a buoyancy control piston. Results from sea tests illustrate the feasibility of our design, including its depth tracking abilities.
Limnology & Oceanography: Fluids & Environments, 2011
IEEE Embedded Systems Letters, 2012
ABSTRACT In this article we consider the development of acoustic methods which have the potential... more ABSTRACT In this article we consider the development of acoustic methods which have the potential to size marine animals. The proposed technique uses scattered sound in order to invert for both animal size and shape. The technique uses the Distorted Wave Born Approximation (DWBA) in order to model sound scattered from these organisms. The use of the DWBA also provides a valuable context for formulating data analysis techniques in order to invert for parameters of the animal. Although 3-dimensional observations can be obtained from a complete set of views, due to the difficulty of collecting full 3-dimensional scatter, it is useful to simplify the inversion by approximating the animal by a few parameters. Here, the animals are modeled as 3-dimensional ellipsoids. This reduces the complexity of the problem to a determination of the 3 semi axes for the x, y and z dimensions from just a few radial spokes through the 3-dimensional Fourier Transform. In order to test the idea, simulated scatter data is taken from a 3-dimensional model of a marine animal and the resultant data are inverted in order to estimate animal shape
ABSTRACT Positioning is a crucial requirement for mobile underwater systems. Since GPS is not ava... more ABSTRACT Positioning is a crucial requirement for mobile underwater systems. Since GPS is not available underwater the position of vehicles has to be estimated over time. Underwater navigation techniques often rely on acoustic communication with reference beacons that know their location. These beacons in effect act as GPS satellites for underwater vehicles. The beacons may be buoys or vehicles on the surface with access to GPS or they may be deployed on the sea floor at known locations. Since both the beacons and underwater vehicles are battery operated, energy is a key constraint. This is further aggravated by the fact that the energy required to acoustically communicate underwater is high even over moderate distances. Since the acoustic signaling required for tracking vehicles is a recurring cost, we propose to minimize the energy consumption by optimizing the extent of signaling used for localization. Localization techniques that exclusively rely on estimates of time of flight (or time difference of arrivals) require transmissions from beacons to be nearly concurrent. This allows position to be estimated at each point in time based on geometric constraints alone, neglecting vehicle motion between transmissions. Alternatively if vehicles have some knowledge about their motion either from models or from direct measurements of their acceleration and heading as obtained from an on-board inertial measurement unit (IMU), other techniques can take this information into account. The extended Kalman filter, particle filters or factor-graph based Maximum Likelihood (ML) trajectory estimation methods effectively combine IMU measurements with geometric constraints obtained from acoustic time of flight measurements. One of the advantages of such an approach is that beacon transmissions no longer have to be concurrent. In this paper we use the Maximum Likelihood (factor-graph based) tracking method to find the best schedule for beacon transmissions for a fixed signaling rate (or aver- ge power consumption). We explore a number of possible transmission schemes and evaluate their performance as a function of relevant system parameters, specifically, the type of motion measurements available, the accuracy of these measurements and the number of beacons. We also evaluate our proposed schemes on experimental data obtained from sea trials that were conducted off the coast of San Diego. Our experimental and simulation results show that even if coarse information is available about the vehicle's motion, the localization performance can be improved for a fixed signaling rate by appropriately lagging transmissions from reference beacons.
OCEANS 2011 IEEE - Spain, 2011
ABSTRACT To measure oceanic radiance and hence, the pelagic environment of marine cephalopods, we... more ABSTRACT To measure oceanic radiance and hence, the pelagic environment of marine cephalopods, we have built a new underwater multiple camera system that consists of a glass sphere that houses 6 machine vision cameras each capable of outputting uncompressed 2 megapixel images at 60 Hz. The cameras are arranged so that their fields of view are all orthogonal, pointing along the positive and negative directions of the principal Cartesian axes (+/- x, +/- y, and +/- z). Taking all of the camera data together, the light field incident on the Omni-directional camera system can be recorded. The system was configured for autonomous operation with batteries, computers, cameras, and ancillary electronics inside the 17” glass sphere. Preliminary data, recorded in both a pelagic and benthic environment, indicate that the data acquired by the system can provide interesting insights into spatial and temporal fluctuations in underwater animal habitats.
Acoustical Imaging, 2011
ABSTRACT Estimating biophysical parameters of fish populations in situ such as size, orientation,... more ABSTRACT Estimating biophysical parameters of fish populations in situ such as size, orientation, shape, and taxa is a fundamental goal in oceanography. Towards this end, acoustics is a natural choice due to its rapid, non-invasive capabilities. Here, multi-view methods are explored for classification, size and orientation estimation, and 2D image reconstruction for individual fish. Size- and shape-based classification using multi-view data is shown to be accurate (~10% error) using kernel methods and discriminant analysis. For species-based classification in the absence of significant differences in size or shape, multi-view methods offer significant (~40%) reduction in error, but absolute error rates remain high (~20%) due to the lack of discriminant information in acoustic scatter. Length and orientation estimation are investigated using a parameter-based approach with a simple ellipsoidal scattering model. Good accuracy is obtained when the views span the full 360°. When the span is limited to less than 60°, incorporating a prior constraint on possible body shapes can lead to reduced uncertainty in the estimated parameters. Finally, using views that span the full 360°, sparse Bayesian learning coupled with a conventional Radon transform yields accurate two-dimensional, projected images of the fish. KeywordsAcoustics- Fish classification -Scattering-Size estimation- Radon transform - Bayesian learning -Distorted wave Born approximation
To understand the physical and biological dynamics controlling the micro-and fine-scale distribut... more To understand the physical and biological dynamics controlling the micro-and fine-scale distributions of phytoplankton and zooplankton in the ocean. OBJECTIVES To use multispectral optical measurements and three-dimensional acoustic imaging to quantify the microscale and fine-scale vertical patchiness of several taxa of phytoplankton and large zooplankton in situ. APPROACH Our approach was to develop a multispectral in situ imaging system: LUMIS (Low light level Underwater Multispectral Imaging System). Using a sheet of laser light to irradiate a ~1 l sample volume, phytoplankton fluorescence from chlorophyll a and phycoerythrin were imaged with a sensitive CCD camera. In addition to the two fluorescence wavelengths, optical backscatter and water Raman scattering were imaged simultaneously.
Oceanographers are limited by existing technology when observing aquatic microbes that are undist... more Oceanographers are limited by existing technology when observing aquatic microbes that are undisturbed in their natural environment. A Thin Light Sheet Microscope (TLSM (1)), that offers a simple and easy to implement technique for such studies, is described here. The TLSM utilizes conventional microscope optics but replaces the illumination apparatus with a thin (23 micron) sheet of laser light. The light sheet has transverse dimensions of 6 mm by 1 mm , positioned precisely at the depth of field of the microscope's objective lens. The sample volume above and below the light sheet is not exposed to excitation light. The TLSM concentrates light only where excitation illumination is needed, thus maximizing the efficiency of the illumination source and minimizes the background noise level in the image, which maximizes image sharpness. A fluorescent biological marker that connects only to living organisms in the water was added to the test sample. Particles that are not located wit...
We intend to quantify the biological, physical, and chemical dynamics that structure marine plank... more We intend to quantify the biological, physical, and chemical dynamics that structure marine planktonic ecosystems. Observations of the organisms and their environment on the spatial and temporal scales that characterize their interactions, combined with models of the dominant dynamics, will lead to improved understanding of the dynamics, structure, and function of planktonic ecosystems. OBJECTIVES Our objectives in this work are to 1) visualize and quantify herbivorous copepod feeding in the laboratory, and 2) to apply these methods in the field to observe the dynamics of copepod feeding in situ. In particular we intend to test the "feeding sorties" hypothesis vs. the "in situ feeding" hypothesis regarding the location and timing of copepod feeding and vertical migration.
Using a Planar Laser Imaging Fluorometry (PLIF) system mounted on a free-falling platform, we hav... more Using a Planar Laser Imaging Fluorometry (PLIF) system mounted on a free-falling platform, we have observed the spatial distribution patterns and size-frequency spectra of large fluorescent particles in the upper ocean. In drops of the system off San Diego, CA, we found that the size-frequency spectrum was strongly related to the total phytoplankton biomass (measured as chlorophyll a). Differences of the observed spectra from those predicted based on biomass showed regions suggesting aggregation and sinking of particles, systematic and sudden changes in spectra with depth, and associations of anomalously large particles with hydrographic features. The relative spatial distributions of particles in the images showed the particles to be non-randomly distributed on scales of a few cm, except in regions of high concentration (the subsurface chlorophyll maximum). Some images in every profile were found to be streaky, indicating that the water moved while the camera shutter was open. A ca...
We have recently developed and built a novel free-falling platform with a stereoscopic Digital Pa... more We have recently developed and built a novel free-falling platform with a stereoscopic Digital Particle Image Velocimeter (DPIV) to observe and quantify microscale physical and biological structures in the upper ocean. Scattered light from a vertical sheet of laser illumination is imaged from both sides by sensitive CCD cameras. Sequences of images allow two-dimensional maps of three-component velocity to be constructed from cross-correlations between images. The profiler has the potential to provide direct estimates of turbulent kinetic energy dissipation in the near-surface open ocean. Dissipation rates can be estimated using finite-difference approximations for 8 of the 12 velocity gradient terms. Using a filter wheel in front of the camera lenses, scattered light and two wavelengths of fluoresced light can be imaged sequentially, allowing mapping of fluorescent particles to the turbulent structures in the water column. The spatial and temporal resolution of the system is set by ...
We developed and deployed a 2D planar laser imaging fluorometer system mounted on a free-falling ... more We developed and deployed a 2D planar laser imaging fluorometer system mounted on a free-falling vehicle. The vehicle also carried a CTD/fluorometer, an optical nitrate sensor, and a microstructure profiler, to acquire vertical profiles of particle fluorescence and chemical and physical distributions in the field. Our data, gathered 10 km offshore of San Diego, showed large variations in fluorescent particle properties (size-frequency distributions) with depth, with changes over 2-5 m that were larger than changes seen at a single depth over 1000's of kilometers. The phytoplankton were randomly distributed over scales <10 cm, but shoed pronounced gradients over larger scales. Layers -1 m thick of unusual particle properties (e.g., large particles, or long particles) were common, though these layers did not appear as features in the total fluorescence profiles. Such cryptic layers are likely a common occurrence, driven by the interleaving of existing horizontal patches by the ...
LONG-TERM GOALS Our overall goal is to understand the perceptual and mechanistic principles that ... more LONG-TERM GOALS Our overall goal is to understand the perceptual and mechanistic principles that underlay camouflage framed in the context of the animals' environment. In particular, we plan to characterize and understand the perceptual abilities of several species of benthic and pelagic cephalopods (which are unrivaled masters of dynamic camouflage), the aspects of their optical environment that affect their camouflage behavior, the characterization of that behavior, and the molecular mechanisms inside the skin by which those responses are accomplished. OBJECTIVES 1. To fully characterize the spatiotemporal characteristics of the near-surface and shallow benthic underwater light field, including ultraviolet radiation and polarization. 2. To determine the visual abilities of several species of cephalopod and model both the shallow and deep-water world from the animals' points of view. Public reporting burden for the collection of information is estimated to average 1 hour pe...
IEEE Transactions on Acoustics, Speech, and Signal Processing, 1990
ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing, 1988
ABSTRACT A sonar imaging technique which uses specially coded waveforms to project different soun... more ABSTRACT A sonar imaging technique which uses specially coded waveforms to project different sound signals in different directions was previously proposed. In this context, the problem of creating a set of frequency-hopped code words that have mutually small correlation properties has arisen. The authors report on their progress in designing these frequency-hopped signals for multibeam sonar imaging. The special case is considered in which medium spreading effects destroy the phase coherence of the imaging system. A design strategy is discussed which has allowed the design of a set of frequency-hopped code words that exhibit good performance bounds with respect to incoherent processing. The design strategy is a systematic constructions technique which can be used with any size Galois field GF( p ) to generate p -1 codes of length p . Various theorems are presented and examples of the performance of the technique are illustrated for time-bandwidth products of interest in terms of sonar
Ocean Optics XIII, 1997
ABSTRACT This article presents the results of analyzing data which were collected in the sea with... more ABSTRACT This article presents the results of analyzing data which were collected in the sea with a structured illumination/fluorescence imaging system. The system employs all lines of an argon ion laser to create a 2D slice illumination pattern in the blue-green. A sensitive CCD camera was used to measure the fluorescence emission resulting from this stimulation. Under our assumptions, these images are proportional to chlorophyll-a concentration. Inspection of the data reveals a large degree of small scale patchiness with characteristic distances as small as centimeters. In many of the images, the chlorophyll concentration is seen to change by an order of magnitude over several centimeters. Patchiness was characterized via the use of spatial spectral estimation techniques. Preliminary results show that the data follow a -5/3 slope at high wave number, consistent with the hypothesis that the patches were formed by turbulent stirring of larger scale gradients of phytoplankton. In addition, repeated vertical profiles demonstrated that the larger scales were extremely constant, even in the presence of this small scale variability. The technique has provided an original data set with important new information about the small scale structure of pigment in the ocean.
2009 Workshop on Applications of Computer Vision (WACV), 2009
Journal of the Optical Society of America. A, Optics, image science, and vision, 2010
A set of algorithms is proposed to retrieve the size of spherically symmetric particles from the ... more A set of algorithms is proposed to retrieve the size of spherically symmetric particles from the measured intensity of angular scatter data. Of special interest are low-contrast particles whose real part of the index of refraction is between 1.03 and 1.09 and whose size ka is constrained so that pi < or = ka < or = 16pi, where k=2pi/lambda and a is particle radius. Several algorithms are evaluated and compared that are based on either simple matching to the Mie theory predictions or inverse tomography methods. In the tomography methods, a previously proposed algorithm [Opt. Express. 15, 12217 (2007)] was used after estimating the phase of the scattered data or adapted to use intensity-only data. In order to ensure stability, all algorithms' performance was evaluated in the presence of moderate noise. The performance varied as a function of particle size, refractive index, and algorithm. Results suggest that a scattering device that collects only the angular scatter that is...
Proceedings of the 1st ACM international workshop on Underwater networks - WUWNet '06, 2006
ABSTRACT With the increasing sophistication of both manned and unmanned systems for remote ocean ... more ABSTRACT With the increasing sophistication of both manned and unmanned systems for remote ocean exploration, a wealth of knowledge about heretofore-unknown oceanic processes has become available. However, no technologies currently exist to observe organisms and processes without disturbing them, as they move with the natural motion of the oceans. We propose a new class of ocean sensing, whereby free-floating underwater devices operate autonomously and collaborate through an acoustic underwater network between them. This new class of sensing will provide a window into understanding the multifaceted interactions between the ocean's currents, underwater ecosystems and our impact on them. In this paper, we will present the design of our underwater vehicle, which drifts freely with the ocean currents and is equipped with a buoyancy control piston. Results from sea tests illustrate the feasibility of our design, including its depth tracking abilities.
Limnology & Oceanography: Fluids & Environments, 2011
IEEE Embedded Systems Letters, 2012
ABSTRACT In this article we consider the development of acoustic methods which have the potential... more ABSTRACT In this article we consider the development of acoustic methods which have the potential to size marine animals. The proposed technique uses scattered sound in order to invert for both animal size and shape. The technique uses the Distorted Wave Born Approximation (DWBA) in order to model sound scattered from these organisms. The use of the DWBA also provides a valuable context for formulating data analysis techniques in order to invert for parameters of the animal. Although 3-dimensional observations can be obtained from a complete set of views, due to the difficulty of collecting full 3-dimensional scatter, it is useful to simplify the inversion by approximating the animal by a few parameters. Here, the animals are modeled as 3-dimensional ellipsoids. This reduces the complexity of the problem to a determination of the 3 semi axes for the x, y and z dimensions from just a few radial spokes through the 3-dimensional Fourier Transform. In order to test the idea, simulated scatter data is taken from a 3-dimensional model of a marine animal and the resultant data are inverted in order to estimate animal shape
ABSTRACT Positioning is a crucial requirement for mobile underwater systems. Since GPS is not ava... more ABSTRACT Positioning is a crucial requirement for mobile underwater systems. Since GPS is not available underwater the position of vehicles has to be estimated over time. Underwater navigation techniques often rely on acoustic communication with reference beacons that know their location. These beacons in effect act as GPS satellites for underwater vehicles. The beacons may be buoys or vehicles on the surface with access to GPS or they may be deployed on the sea floor at known locations. Since both the beacons and underwater vehicles are battery operated, energy is a key constraint. This is further aggravated by the fact that the energy required to acoustically communicate underwater is high even over moderate distances. Since the acoustic signaling required for tracking vehicles is a recurring cost, we propose to minimize the energy consumption by optimizing the extent of signaling used for localization. Localization techniques that exclusively rely on estimates of time of flight (or time difference of arrivals) require transmissions from beacons to be nearly concurrent. This allows position to be estimated at each point in time based on geometric constraints alone, neglecting vehicle motion between transmissions. Alternatively if vehicles have some knowledge about their motion either from models or from direct measurements of their acceleration and heading as obtained from an on-board inertial measurement unit (IMU), other techniques can take this information into account. The extended Kalman filter, particle filters or factor-graph based Maximum Likelihood (ML) trajectory estimation methods effectively combine IMU measurements with geometric constraints obtained from acoustic time of flight measurements. One of the advantages of such an approach is that beacon transmissions no longer have to be concurrent. In this paper we use the Maximum Likelihood (factor-graph based) tracking method to find the best schedule for beacon transmissions for a fixed signaling rate (or aver- ge power consumption). We explore a number of possible transmission schemes and evaluate their performance as a function of relevant system parameters, specifically, the type of motion measurements available, the accuracy of these measurements and the number of beacons. We also evaluate our proposed schemes on experimental data obtained from sea trials that were conducted off the coast of San Diego. Our experimental and simulation results show that even if coarse information is available about the vehicle's motion, the localization performance can be improved for a fixed signaling rate by appropriately lagging transmissions from reference beacons.
OCEANS 2011 IEEE - Spain, 2011
ABSTRACT To measure oceanic radiance and hence, the pelagic environment of marine cephalopods, we... more ABSTRACT To measure oceanic radiance and hence, the pelagic environment of marine cephalopods, we have built a new underwater multiple camera system that consists of a glass sphere that houses 6 machine vision cameras each capable of outputting uncompressed 2 megapixel images at 60 Hz. The cameras are arranged so that their fields of view are all orthogonal, pointing along the positive and negative directions of the principal Cartesian axes (+/- x, +/- y, and +/- z). Taking all of the camera data together, the light field incident on the Omni-directional camera system can be recorded. The system was configured for autonomous operation with batteries, computers, cameras, and ancillary electronics inside the 17” glass sphere. Preliminary data, recorded in both a pelagic and benthic environment, indicate that the data acquired by the system can provide interesting insights into spatial and temporal fluctuations in underwater animal habitats.