Detection and Classification of Whale Acoustic Signals

This dissertation focuses on two vital challenges in relation to whale acoustic signals: detection and classification.

In detection, we evaluated the influence of the uncertain ocean environment on the spectrogram-based detector, and derived the likelihood ratio of the proposed Short Time Fourier Transform detector. Experimental results showed that the proposed detector outperforms detectors based on the spectrogram. The proposed detector is more sensitive to environmental changes because it includes phase information.

In classification, our focus is on finding a robust and sparse representation of whale vocalizations. Because whale vocalizations can be modeled as polynomial phase signals, we can represent the whale calls by their polynomial phase coefficients. In this dissertation, we used the Weyl transform to capture chirp rate information, and used a two dimensional feature set to represent whale vocalizations globally. Experimental results showed that our Weyl feature set outperforms chirplet coefficients and MFCC (Mel Frequency Cepstral Coefficients) when applied to our collected data.

Since whale vocalizations can be represented by polynomial phase coefficients, it is plausible that the signals lie on a manifold parameterized by these coefficients. We also studied the intrinsic structure of high dimensional whale data by exploiting its geometry. Experimental results showed that nonlinear mappings such as Laplacian Eigenmap and ISOMAP outperform linear mappings such as PCA and MDS, suggesting that the whale acoustic data is nonlinear.

We also explored deep learning algorithms on whale acoustic data. We built each layer as convolutions with either a PCA filter bank (PCANet) or a DCT filter bank (DCTNet). With the DCT filter bank, each layer has different a time-frequency scale representation, and from this, one can extract different physical information. Experimental results showed that our PCANet and DCTNet achieve high classification rate on the whale vocalization data set. The word error rate of the DCTNet feature is similar to the MFSC in speech recognition tasks, suggesting that the convolutional network is able to reveal acoustic content of speech signals.

Numerical detection of the Gardner transition in a mean-field glass former.

Recent theoretical advances predict the existence, deep into the glass phase, of a novel phase transition, the so-called Gardner transition. This transition is associated with the emergence of a complex free energy landscape composed of many marginally stable sub-basins within a glass metabasin. In this study, we explore several methods to detect numerically the Gardner transition in a simple structural glass former, the infinite-range Mari-Kurchan model. The transition point is robustly located from three independent approaches: (i) the divergence of the characteristic relaxation time, (ii) the divergence of the caging susceptibility, and (iii) the abnormal tail in the probability distribution function of cage order parameters. We show that the numerical results are fully consistent with the theoretical expectation. The methods we propose may also be generalized to more realistic numerical models as well as to experimental systems.

Using Image Processing Methods to Improve the Detection of Buried Explosive Threats in GPR Data

Current state of the art techniques for landmine detection in ground penetrating radar (GPR) utilize statistical methods to identify characteristics of a landmine response. This research makes use of 2-D slices of data in which subsurface landmine responses have hyperbolic shapes. Various methods from the field of visual image processing are adapted to the 2-D GPR data, producing superior landmine detection results. This research goes on to develop a physics-based GPR augmentation method motivated by current advances in visual object detection. This GPR specific augmentation is used to mitigate issues caused by insufficient training sets. This work shows that augmentation improves detection performance under training conditions that are normally very difficult. Finally, this work introduces the use of convolutional neural networks as a method to learn feature extraction parameters. These learned convolutional features outperform hand-designed features in GPR detection tasks. This work presents a number of methods, both borrowed from and motivated by the substantial work in visual image processing. The methods developed and presented in this work show an improvement in overall detection performance and introduce a method to improve the robustness of statistical classification.

ADVANCED ISLANDING DETECTION METHODS FOR SINGLE PHASE GRID CONNECTED PV INVERTERS

Due to the growing concerns associated with fossil fuels, emphasis has been placed on clean and sustainable energy generation. This has resulted in the increase in Photovoltaics (PV) units being integrated into the utility system. The integration of PV units has raised some concerns for utility power systems, including the consequences of failing to detect islanding. Numerous methods for islanding detection have been introduced in literature. They can be categorized into local methods and remote methods. The local methods are categorically divided into passive and active methods. Active methods generally have smaller Non-Detection Zone (NDZ) but the injecting disturbances will slightly degrade the power quality and reliability of the power system. Slip Mode Frequency Shift Islanding Detection Method (SMS IDM) is an active method that uses positive feedback for islanding detection. In this method, the phase angle of the converter is controlled to have a sinusoidal function of the deviation of the Point of Common Coupling (PCC) voltage frequency from the nominal grid frequency. This method has a non-detection zone which means it fails to detect islanding for specific local load conditions. If the SMS IDM employs a different function other than the sinusoidal function for drifting the phase angle of the inverter, its non-detection zone could be smaller. In addition, Advanced Slip Mode Frequency Shift Islanding Detection Method (Advanced SMS IDM), which has been introduced in this thesis, eliminates the non-detection zone of the SMS IDM. In this method the parameters of SMS IDM change based on the local load impedance value. Moreover, the stability of the system is investigated by developing the dynamical equations of the system for two operation modes; grid connected and islanded mode. It is mathematically proven that for some loading conditions the nominal frequency is an unstable point and the operation frequency slides to another stable point, while for other loading conditions the nominal frequency is the only stable point of the system upon islanding occurring. Simulation and experimental results show the accuracy of the proposed methods in detection of islanding and verify the validity of the mathematical analysis.

Air-sea fluxes of CO2 and CH4 from the Penlee Point Atmospheric Observatory on the south-west coast of the UK

We present air–sea fluxes of carbon dioxide (CO2), methane (CH4), momentum, and sensible heat measured by the eddy covariance method from the recently established Penlee Point Atmospheric Observatory (PPAO) on the south-west coast of the United Kingdom. Measurements from the south-westerly direction (open water sector) were made at three different sampling heights (approximately 15, 18, and 27m above mean sea level, a.m.s.l.), each from a different period during 2014–2015. At sampling heights ≥18ma.m.s.l., measured fluxes of momentum and sensible heat demonstrate reasonable (≤ ±20% in the mean) agreement with transfer rates over the open ocean. This confirms the suitability of PPAO for air–sea exchange measurements in shelf regions. Covariance air–sea CO2 fluxes demonstrate high temporal variability. Air-to-sea transport of CO2 declined from spring to summer in both years, coinciding with the breakdown of the spring phytoplankton bloom. We report, to the best of our knowledge, the first successful eddy covariance measurements of CH4 emissions from a marine environment. Higher sea-to-air CH4 fluxes were observed during rising tides (20±3; 38±3; 29±6 μmolem-2 d-1 at 15, 18, 27ma.m.s.l.) than during falling tides (14±2; 22±2; 21±5 μmolem-2 d-1), consistent with an elevated CH4 source from an estuarine outflow driven by local tidal circulation. These fluxes are a few times higher than the predicted CH4 emissions over the open ocean and are significantly lower than estimates from other aquatic CH4 hotspots (e.g. polar regions, freshwater). Finally, we found the detection limit of the air–sea CH4 flux by eddy covariance to be 20 μmolem-2 d-1 over hourly timescales (4 μmolem-2 d-1 over 24 h).

Air-sea fluxes of CO2 and CH4 from the Penlee Point Atmospheric Observatory on the south-west coast of the UK

We present air–sea fluxes of carbon dioxide (CO2), methane (CH4), momentum, and sensible heat measured by the eddy covariance method from the recently established Penlee Point Atmospheric Observatory (PPAO) on the south-west coast of the United Kingdom. Measurements from the south-westerly direction (open water sector) were made at three different sampling heights (approximately 15, 18, and 27m above mean sea level, a.m.s.l.), each from a different period during 2014–2015. At sampling heights ≥18ma.m.s.l., measured fluxes of momentum and sensible heat demonstrate reasonable (≤ ±20% in the mean) agreement with transfer rates over the open ocean. This confirms the suitability of PPAO for air–sea exchange measurements in shelf regions. Covariance air–sea CO2 fluxes demonstrate high temporal variability. Air-to-sea transport of CO2 declined from spring to summer in both years, coinciding with the breakdown of the spring phytoplankton bloom. We report, to the best of our knowledge, the first successful eddy covariance measurements of CH4 emissions from a marine environment. Higher sea-to-air CH4 fluxes were observed during rising tides (20±3; 38±3; 29±6 μmolem-2 d-1 at 15, 18, 27ma.m.s.l.) than during falling tides (14±2; 22±2; 21±5 μmolem-2 d-1), consistent with an elevated CH4 source from an estuarine outflow driven by local tidal circulation. These fluxes are a few times higher than the predicted CH4 emissions over the open ocean and are significantly lower than estimates from other aquatic CH4 hotspots (e.g. polar regions, freshwater). Finally, we found the detection limit of the air–sea CH4 flux by eddy covariance to be 20 μmolem-2 d-1 over hourly timescales (4 μmolem-2 d-1 over 24 h).

Swellable polymer films containing Au nanoparticles for point-of-care therapeutic drug monitoring using surface-enhanced Raman spectroscopy

Large (10 × 10 cm) sheets of surface-enhanced Raman spectroscopy (SERS) active polymer have been prepared by stabilising metal nanoparticle aggregates within dry hydroxyethylcellulose (HEC) films. In these films the aggregates are protected by the polymer matrix during storage but in use they are released when aqueous analyte droplets cause the films to swell to their gel form. The fact that these "Poly-SERS" films can be prepared in bulk but then cut to size and stored in air before use means that they provide a cost effective and convenient method for routine SERS analysis. Here we have tested both Ag and Au Poly-SERS films for use in point-of-care monitoring of therapeutic drugs, using phenytoin as the test compound. Phenytoin in water could readily be detected using Ag Poly-SERS films but dissolving the compound in phosphate buffered saline (PBS) to mimic body fluid samples caused loss of the drug signal due to competition for metal surface sites from Cl^- ions in the buffer solution. However, with Au Poly-SERS films there was no detectable interference from Cl^- and these materials allowed phenytoin to be detected at 1.8 mg L^-1, even in PBS. The target range of detection of phenytoin in therapeutic drug monitoring is 10-20 mg L^-1. With the Au Poly-SERS films, the absolute signal generated by a given concentration of phenytoin was lower for the films than for the parent colloid but the SERS signals were still high enough to be used for therapeutic monitoring, so the cost in sensitivity for moving from simple aqueous colloids to films is not so large that it outweighs the advantages which the films bring for practical applications, in particular their ease of use and long shelf life.

Enzyme-induced Aggregation of Plasmonic Nanoparticles for the Development of Label-free and Ultrasensitive Detection of Bacterial DNA

Sensitive detection of pathogens is critical to ensure the safety of food supplies and to prevent bacterial disease infection and outbreak at the first onset. While conventional techniques such as cell culture, ELISA, PCR, etc. have been used as the predominant detection workhorses, they are however limited by either time-consuming procedure, complicated sample pre-treatment, expensive analysis and operation, or inability to be implemented at point-of-care testing. Here, we present our recently developed assay exploiting enzyme-induced aggregation of plasmonic gold nanoparticles (AuNPs) for label-free and ultrasensitive detection of bacterial DNA. In the experiments, AuNPs are first functionalized with specific, single-stranded RNA probes so that they exhibit high stability in solution even under high electrolytic condition thus exhibiting red color. When bacterial DNA is present in a sample, a DNA-RNA heteroduplex will be formed and subsequently prone to the RNase H cleavage on the RNA probe, allowing the DNA to liberate and hybridize with another RNA strand. This continuously happens until all of the RNA strands are cleaved, leaving the nanoparticles ‘unprotected’. The addition of NaCl will cause the ‘unprotected’ nanoparticles to aggregate, initiating a colour change from red to blue. The reaction is performed in a multi-well plate format, and the distinct colour signal can be discriminated by naked eye or simple optical spectroscopy. As a result, bacterial DNA as low as pM could be unambiguously detected, suggesting that the enzyme-induced aggregation of AuNPs assay is very easy to perform and sensitive, it will significantly benefit to development of fast and ultrasensitive methods that can be used for disease detection and diagnosis.

VLT FORS2 comparative transmission spectroscopy: Detection of Na in the atmosphere of WASP-39b from the ground

We present transmission spectroscopy of the warm Saturn-mass exoplanet WASP-39b made with the Very Large Telescope (VLT) FOcal Reducer and Spectrograph (FORS2) across the wavelength range 411-810nm. The transit depth is measured with a typical precision of 240 parts per million (ppm) in wavelength bins of 10nm on a V = 12.1 magnitude star. We detect the sodium absorption feature (3.2-sigma) and find evidence for potassium. The ground-based transmission spectrum is consistent with Hubble Space Telescope (HST) optical spectroscopy, strengthening the interpretation of WASP-39b having a largely clear atmosphere. Our results demonstrate the great potential of the recently upgraded FORS2 spectrograph for optical transmission spectroscopy, obtaining HST-quality light curves from the ground.

Real-time detection of faces in video streams

[EN]This paper describes a face detection system which goes beyond traditional approaches normally designed for still images. First the video stream context is considered to apply the detector, and therefore, the resulting system is designed taking into consideration a main feature available in a video stream, i.e. temporal coherence. The resulting system builds a feature based model for each detected face, and searches them using various model information in the next frame. The results achieved for video stream processing outperform Rowley-Kanade's and Viola-Jones' solutions providing eye and face data in a reduced time with a notable correct detection rate.

Plane detection from monocular image sequences

AIRES, Kelson R. T. ; ARAÚJO, Hélder J. ; MEDEIROS, Adelardo A. D. . Plane Detection from Monocular Image Sequences. In: VISUALIZATION, IMAGING AND IMAGE PROCESSING, 2008, Palma de Mallorca, Spain. Proceedings..., Palma de Mallorca: VIIP, 2008

Plane Detection Using Affine Homography

AIRES, Kelson R. T.; ARAÚJO, Hélder J.; MEDEIROS, Adelardo A. D. Plane Detection Using Affine Homography. In: CONGRESSO BRASILEIRO DE AUTOMÁTICA, 2008, Juiz de Fora, MG: Anais... do CBA 2008.

Exploring Habitability Markers, Biosignatures, and Their False Positives Using Spectral Models of Terrestrial Exoplanets

Thesis (Ph.D.)--University of Washington, 2016-08

A Reliable Downscaling of ECG Signals for the Detection of T wave Heterogeneity Features

In cardiovascular disease the definition and the detection of the ECG parameters related to repolarization dynamics in post MI patients is still a crucial unmet need. In addition, the use of a 3D sensor in the implantable medical devices would be a crucial mean in the assessment or prediction of Heart Failure status, but the inclusion of such feature is limited by hardware and firmware constraints. The aim of this thesis is the definition of a reliable surrogate of the 500 Hz ECG signal to reach the aforementioned objective. To evaluate the worsening of reliability due to sampling frequency reduction on delineation performance, the signals have been consecutively down sampled by a factor 2, 4, 8 thus obtaining the ECG signals sampled at 250, 125 and 62.5 Hz, respectively. The final goal is the feasibility assessment of the detection of the fiducial points in order to translate those parameters into meaningful clinical parameter for Heart Failure prediction, such as T waves intervals heterogeneity and variability of areas under T waves. An experimental setting for data collection on healthy volunteers has been set up at the Bakken Research Center in Maastricht. A 16 – channel ambulatory system, provided by TMSI, has recorded the standard 12 – Leads ECG, two 3D accelerometers and a respiration sensor. The collection platform has been set up by the TMSI property software Polybench, the data analysis of such signals has been performed with Matlab. The main results of this study show that the 125 Hz sampling rate has demonstrated to be a good candidate for a reliable detection of fiducial points. T wave intervals proved to be consistently stable, even at 62.5 Hz. Further studies would be needed to provide a better comparison between sampling at 250 Hz and 125 Hz for areas under the T waves.