Sparse Signal Representation of Ultrasonic Signals for Structural Health Monitoring Applications

Assessment of the integrity of structural components is of great importance for aerospace systems, land and marine transportation, civil infrastructures and other biological and mechanical applications. Guided waves (GWs) based inspections are an attractive mean for structural health monitoring. In this thesis, the study and development of techniques for GW ultrasound signal analysis and compression in the context of non-destructive testing of structures will be presented. In guided wave inspections, it is necessary to address the problem of the dispersion compensation. A signal processing approach based on frequency warping was adopted. Such operator maps the frequencies axis through a function derived by the group velocity of the test material and it is used to remove the dependence on the travelled distance from the acquired signals. Such processing strategy was fruitfully applied for impact location and damage localization tasks in composite and aluminum panels. It has been shown that, basing on this processing tool, low power embedded system for GW structural monitoring can be implemented. Finally, a new procedure based on Compressive Sensing has been developed and applied for data reduction. Such procedure has also a beneficial effect in enhancing the accuracy of structural defects localization. This algorithm uses the convolutive model of the propagation of ultrasonic guided waves which takes advantage of a sparse signal representation in the warped frequency domain. The recovery from the compressed samples is based on an alternating minimization procedure which achieves both an accurate reconstruction of the ultrasonic signal and a precise estimation of waves time of flight. Such information is used to feed hyperbolic or elliptic localization procedures, for accurate impact or damage localization.

Consequences of plant population size for pollinator visitation and plant reproductive success

Habitat loss and fragmentation have a prominent role in determining the size of plant populations, and can affect plant-pollinator interactions. It is hypothesized that in small plant populations the ability to set seeds can be reduced due to limited pollination services, since individuals in small populations can receive less quantity or quality of visits. In this study, I investigated the effect of population size on plant reproductive success and insect visitation in 8 populations of two common species in the island of Lesvos, Greece (Mediterranean Sea), Echium plantagineum and Ballota acetabulosa, and of a rare perennial shrub endemic to north-central Italy, Ononis masquillierii. All the three species depended on insect pollinators for sexual reproduction. For each species, pollen limitation was present in all or nearly all populations, but the relationship between pollen limitation and population size was only present in Ononis masquillierii. However, in Echium plantagineum, significant relationships between both open-pollinated and handcrossed-pollinated seed sets and population size were found, being small populations comparatively less productive than large ones. Additionally, for this species, livestock grazing intensity was greater for small populations and for sparse patches, and had a negative influence on productivity of the remnant plants. Both Echium plantagineum and Ballota acetabulosa attracted a great number of insects, representing a wide spectrum of pollinators, thereby can be considered as generalist species. For Ballota acetabulosa, the most important pollinators were megachilid female bees, and insect diversity didn’t decrease with decreasing plant population size. By contrast, Ononis masquillierii plants generally received few visits, with flowers specialized on small bees (Lasioglossum spp.), representing the most important insect guild. In Echium plantagineum and Ballota acetabulosa, plants in small and large populations received the same amount of visits per flower, and no differences in the number of intraplant visited flowers were detected. On the contrary, large Ononis populations supported higher amounts of pollinators than small ones. At patch level, high Echium flower density was associated with more and higher quality pollinators. My results indicate that small populations were not subject to reduced pollination services than large ones in Echium plantagineum and Ballota acetabulosa, and suggest that grazing and resource limitation could have a major impact on population fitness in Echium plantagineum. The absence of any size effects in these two species can be explained in the light of their high local abundance, wide habitat specificity, and ability to compete with other co-flowering species for pollinators. By contrast, size represents a key characteristic for both pollination and reproduction in Ononis masquillierii populations, as an increase in size could mitigate the negative effects coming from the disadvantageous reproductive traits of the species. Finally, the widespread occurrence of pollen limitation in the three species may be the result of 1) an ongoing weakening or disruption of plantpollinator interactions derived from ecological perturbations, 2) an adaptive equilibrium in response to stochastic processes, and 3) the presence of unfavourable reproductive traits (for Ononis masquillierii).

Reconstruction from image correspondences

A single picture provides a largely incomplete representation of the scene one is looking at. Usually it reproduces only a limited spatial portion of the scene according to the standpoint and the viewing angle, besides it contains only instantaneous information. Thus very little can be understood on the geometrical structure of the scene, the position and orientation of the observer with respect to it remaining also hard to guess. When multiple views, taken from different positions in space and time, observe the same scene, then a much deeper knowledge is potentially achievable. Understanding inter-views relations enables construction of a collective representation by fusing the information contained in every single image. Visual reconstruction methods confront with the formidable, and still unanswered, challenge of delivering a comprehensive representation of structure, motion and appearance of a scene from visual information. Multi-view visual reconstruction deals with the inference of relations among multiple views and the exploitation of revealed connections to attain the best possible representation. This thesis investigates novel methods and applications in the field of visual reconstruction from multiple views. Three main threads of research have been pursued: dense geometric reconstruction, camera pose reconstruction, sparse geometric reconstruction of deformable surfaces. Dense geometric reconstruction aims at delivering the appearance of a scene at every single point. The construction of a large panoramic image from a set of traditional pictures has been extensively studied in the context of image mosaicing techniques. An original algorithm for sequential registration suitable for real-time applications has been conceived. The integration of the algorithm into a visual surveillance system has lead to robust and efficient motion detection with Pan-Tilt-Zoom cameras. Moreover, an evaluation methodology for quantitatively assessing and comparing image mosaicing algorithms has been devised and made available to the community. Camera pose reconstruction deals with the recovery of the camera trajectory across an image sequence. A novel mosaic-based pose reconstruction algorithm has been conceived that exploit image-mosaics and traditional pose estimation algorithms to deliver more accurate estimates. An innovative markerless vision-based human-machine interface has also been proposed, so as to allow a user to interact with a gaming applications by moving a hand held consumer grade camera in unstructured environments. Finally, sparse geometric reconstruction refers to the computation of the coarse geometry of an object at few preset points. In this thesis, an innovative shape reconstruction algorithm for deformable objects has been designed. A cooperation with the Solar Impulse project allowed to deploy the algorithm in a very challenging real-world scenario, i.e. the accurate measurements of airplane wings deformations.

Kernel Methods for Tree Structured Data

Machine learning comprises a series of techniques for automatic extraction of meaningful information from large collections of noisy data. In many real world applications, data is naturally represented in structured form. Since traditional methods in machine learning deal with vectorial information, they require an a priori form of preprocessing. Among all the learning techniques for dealing with structured data, kernel methods are recognized to have a strong theoretical background and to be effective approaches. They do not require an explicit vectorial representation of the data in terms of features, but rely on a measure of similarity between any pair of objects of a domain, the kernel function. Designing fast and good kernel functions is a challenging problem. In the case of tree structured data two issues become relevant: kernel for trees should not be sparse and should be fast to compute. The sparsity problem arises when, given a dataset and a kernel function, most structures of the dataset are completely dissimilar to one another. In those cases the classifier has too few information for making correct predictions on unseen data. In fact, it tends to produce a discriminating function behaving as the nearest neighbour rule. Sparsity is likely to arise for some standard tree kernel functions, such as the subtree and subset tree kernel, when they are applied to datasets with node labels belonging to a large domain. A second drawback of using tree kernels is the time complexity required both in learning and classification phases. Such a complexity can sometimes prevents the kernel application in scenarios involving large amount of data. This thesis proposes three contributions for resolving the above issues of kernel for trees. A first contribution aims at creating kernel functions which adapt to the statistical properties of the dataset, thus reducing its sparsity with respect to traditional tree kernel functions. Specifically, we propose to encode the input trees by an algorithm able to project the data onto a lower dimensional space with the property that similar structures are mapped similarly. By building kernel functions on the lower dimensional representation, we are able to perform inexact matchings between different inputs in the original space. A second contribution is the proposal of a novel kernel function based on the convolution kernel framework. Convolution kernel measures the similarity of two objects in terms of the similarities of their subparts. Most convolution kernels are based on counting the number of shared substructures, partially discarding information about their position in the original structure. The kernel function we propose is, instead, especially focused on this aspect. A third contribution is devoted at reducing the computational burden related to the calculation of a kernel function between a tree and a forest of trees, which is a typical operation in the classification phase and, for some algorithms, also in the learning phase. We propose a general methodology applicable to convolution kernels. Moreover, we show an instantiation of our technique when kernels such as the subtree and subset tree kernels are employed. In those cases, Direct Acyclic Graphs can be used to compactly represent shared substructures in different trees, thus reducing the computational burden and storage requirements.

Riabilitazione protesica ed implanto-protesica di pazienti in età evolutiva affetti da Displasia Ectodermica

Introduction. Ectodermal Dysplasias are a heterogeneous group of inherited disorders characterized by dysplasia of tissues of ectodermal origin (hair, nails, teeth, skins and glands). Clinically, it may be divided into two broad categories: the X-linked hypoidrotic form and the hidrotic form. Hypohidrotic Ectodermal Dysplasia (H.E.D) is characterized by the triad oligo-anodontia, hypotricosis, hypo-anhydrosis (Christ-Siemens-Tourane syndrome). The incidence of HED is about 1/100,000. Mutation in the actodysplasin-A (EDA) and ectodysplasin-A receptor (EDAR) genes are responsible for X-linked and autosomal HED. The clinical features include sparse, fine hair, missing or conical-shaped teeth, decreased sweat and mucous glands, hypoplastic skin, and heat intolerance with exercise or increased ambient temperature. Complete or partial anodontia and malformation of teeth are the most frequent dental findings. Incisors and canines are often conical-shaped while primarily second molars, if present, are mostly affected by taurodontism. Treatment is supportive and includes protection from heat exposure, early prosthetic rehabilitation, skin, hair ear, nose and nail care, and genetic counseling for family planning. The diagnosis of HED in the neonatal and early infancy period may be difficult since sparse hair and absent teeth are normal finding at this age. In childhood the diagnosis is more easily made on the basis of history and clinical examination. Dental abnormalities are the most common complaint. Prosthetic rehabilitation has been recommended as an essential part of the management of HED because is important from functional, esthetic, and psychological standpoint. A team approach that includes input from a pediatric dentist, an orthodontist, a prosthodontist, and an oral and maxillofacial surgeon is necessary for a successful outcome. Conventional prosthodontic rehabilitation in young patient is often difficult because of the anatomical abnormalities of existing teeth and alveolar ridges. The conical shaped teeth and “knife-edge” alveolar ridges result in poor retention and instability of dentures. Moreover, denture must permit jaws expansion and a correct pattern of growth. Materials and Methods. Complete removable dentures were provided to allow for normal physiological development and a corrected masticatory function. Initial maxillary and mandibular impressions were made with smallest stock trays and irreversible hydrocolloid and then final impressions ware made with light-bodied polysulfide rubber base impression material. A base of autopolymerizing resin was constructed and a wax rim was added to the base. The patient’s vertical dimension of occlusion was established by assessing phonetic and esthetic criteria. Preliminary occlusal relations were recorded, and the mandibular cast was mounted on the articulator. Acrylic resin teeth specific for children dentures were selected and mounted. The dentures were tried in and, after proper adjustments, were inserted. The patients were monitored clinically every month to fit prostheses. Cephalometric radiographs were taken every 6 month with the prostheses in place in order to evaluate correct pattern of growth. Cephalometric measurements were realized and used to evaluate the effect of rehabilitation on craniofacial growth. Cephalometric measurements of sound patients were compared with ED patients. After two month expander screws (three-way screw in the upper denture and two-way the lower one)were inserted in each denture in order to permit the expansion of the denture and the jaws growth. Where conical teeth were present, composite crown were realized and luted to improve the esthetic and phonesis. In order to improve retention the placement of endosseous implants was carried out. TC 3D Accuitomo was performed and a resin model of mandibular bone of the patient was realized. At the age of 11 years two implants were inserted into anterior mandible in a child with anodontia. Despite a remarkable multi-dimensional atrophy of the mandibular alveolar process, the insertion of two tapered screw implants (SAMO Smiler, diameter 3.8, length 10 mm). After a submerged healing period of two-three month, the implants were exposed. Implants were connected with an expansion guide that permits mandibular growth and prosthetic retention. The amount of mandibular growth was also evaluate dusing the expansion guide. Results. Early oral rehabilitation improve oral function, phonesis and esthetic, reducing social impairment. Treated patients showed normal cephalometric measurement. Early rehabilitation is able to prevent the prognatissm of the mandibula . The number of teeth was significantly related to several changes in craniofacial morphology. Discussion. In the present study the 5,3% of ED patients showed hypodontia, the l’89,4% di oligodontia, and the 5,3% di anodontia. The cephalometric analysis supports that ED patients showed midface hypoplasia. ED groups showed an increased pogonion to nasion measurement than sound patients, indicative of class III tendency. The present study demonstrated that number of teeth was significantly correlated with deviation of cephalometric measurements from normality. Oligoanodontia is responsible for changing of cephalometric measuraments also on sagittal plane with a class III tendency. Maxillary jaw showed a retrused position related to the presence of hypodontia.

Modelling coupled physical-biogeochemical processes in ice-covered oceans

The last decades have seen a large effort of the scientific community to study and understand the physics of sea ice. We currently have a wide - even though still not exhaustive - knowledge of the sea ice dynamics and thermodynamics and of their temporal and spatial variability. Sea ice biogeochemistry is instead largely unknown. Sea ice algae production may account for up to 25% of overall primary production in ice-covered waters of the Southern Ocean. However, the influence of physical factors, such as the location of ice formation, the role of snow cover and light availability on sea ice primary production is poorly understood. There are only sparse localized observations and little knowledge of the functioning of sea ice biogeochemistry at larger scales. Modelling becomes then an auxiliary tool to help qualifying and quantifying the role of sea ice biogeochemistry in the ocean dynamics. In this thesis, a novel approach is used for the modelling and coupling of sea ice biogeochemistry - and in particular its primary production - to sea ice physics. Previous attempts were based on the coupling of rather complex sea ice physical models to empirical or relatively simple biological or biogeochemical models. The focus is moved here to a more biologically-oriented point of view. A simple, however comprehensive, physical model of the sea ice thermodynamics (ESIM) was developed and coupled to a novel sea ice implementation (BFM-SI) of the Biogeochemical Flux Model (BFM). The BFM is a comprehensive model, largely used and validated in the open ocean environment and in regional seas. The physical model has been developed having in mind the biogeochemical properties of sea ice and the physical inputs required to model sea ice biogeochemistry. The central concept of the coupling is the modelling of the Biologically-Active-Layer (BAL), which is the time-varying fraction of sea ice that is continuously connected to the ocean via brines pockets and channels and it acts as rich habitat for many microorganisms. The physical model provides the key physical properties of the BAL (e.g., brines volume, temperature and salinity), and the BFM-SI simulates the physiological and ecological response of the biological community to the physical enviroment. The new biogeochemical model is also coupled to the pelagic BFM through the exchange of organic and inorganic matter at the boundaries between the two systems . This is done by computing the entrapment of matter and gases when sea ice grows and release to the ocean when sea ice melts to ensure mass conservation. The model was tested in different ice-covered regions of the world ocean to test the generality of the parameterizations. The focus was particularly on the regions of landfast ice, where primary production is generally large. The implementation of the BFM in sea ice and the coupling structure in General Circulation Models will add a new component to the latters (and in general to Earth System Models), which will be able to provide adequate estimate of the role and importance of sea ice biogeochemistry in the global carbon cycle.

Spectral estimates and preconditioning for saddle point systems arising from optimization problems

In this thesis, we consider the problem of solving large and sparse linear systems of saddle point type stemming from optimization problems. The focus of the thesis is on iterative methods, and new preconditioning srategies are proposed, along with novel spectral estimtates for the matrices involved.