Domain Adaptation in remote sensing: increasing the portability of land-cover classifiers

Among the types of remote sensing acquisitions, optical images are certainly one of the most widely relied upon data sources for Earth observation. They provide detailed measurements of the electromagnetic radiation reflected or emitted by each pixel in the scene. Through a process termed supervised land-cover classification, this allows to automatically yet accurately distinguish objects at the surface of our planet. In this respect, when producing a land-cover map of the surveyed area, the availability of training examples representative of each thematic class is crucial for the success of the classification procedure. However, in real applications, due to several constraints on the sample collection process, labeled pixels are usually scarce. When analyzing an image for which those key samples are unavailable, a viable solution consists in resorting to the ground truth data of other previously acquired images. This option is attractive but several factors such as atmospheric, ground and acquisition conditions can cause radiometric differences between the images, hindering therefore the transfer of knowledge from one image to another. The goal of this Thesis is to supply remote sensing image analysts with suitable processing techniques to ensure a robust portability of the classification models across different images. The ultimate purpose is to map the land-cover classes over large spatial and temporal extents with minimal ground information. To overcome, or simply quantify, the observed shifts in the statistical distribution of the spectra of the materials, we study four approaches issued from the field of machine learning. First, we propose a strategy to intelligently sample the image of interest to collect the labels only in correspondence of the most useful pixels. This iterative routine is based on a constant evaluation of the pertinence to the new image of the initial training data actually belonging to a different image. Second, an approach to reduce the radiometric differences among the images by projecting the respective pixels in a common new data space is presented. We analyze a kernel-based feature extraction framework suited for such problems, showing that, after this relative normalization, the cross-image generalization abilities of a classifier are highly increased. Third, we test a new data-driven measure of distance between probability distributions to assess the distortions caused by differences in the acquisition geometry affecting series of multi-angle images. Also, we gauge the portability of classification models through the sequences. In both exercises, the efficacy of classic physically- and statistically-based normalization methods is discussed. Finally, we explore a new family of approaches based on sparse representations of the samples to reciprocally convert the data space of two images. The projection function bridging the images allows a synthesis of new pixels with more similar characteristics ultimately facilitating the land-cover mapping across images.

Kernel-based mapping of orographic rainfall enhancement in the Swiss Alps as detected by weatherradar

In this paper, we develop a data-driven methodology to characterize the likelihood of orographic precipitation enhancement using sequences of weather radar images and a digital elevation model (DEM). Geographical locations with topographic characteristics favorable to enforce repeatable and persistent orographic precipitation such as stationary cells, upslope rainfall enhancement, and repeated convective initiation are detected by analyzing the spatial distribution of a set of precipitation cells extracted from radar imagery. Topographic features such as terrain convexity and gradients computed from the DEM at multiple spatial scales as well as velocity fields estimated from sequences of weather radar images are used as explanatory factors to describe the occurrence of localized precipitation enhancement. The latter is represented as a binary process by defining a threshold on the number of cell occurrences at particular locations. Both two-class and one-class support vector machine classifiers are tested to separate the presumed orographic cells from the nonorographic ones in the space of contributing topographic and flow features. Site-based validation is carried out to estimate realistic generalization skills of the obtained spatial prediction models. Due to the high class separability, the decision function of the classifiers can be interpreted as a likelihood or susceptibility of orographic precipitation enhancement. The developed approach can serve as a basis for refining radar-based quantitative precipitation estimates and short-term forecasts or for generating stochastic precipitation ensembles conditioned on the local topography.

Advanced Kernel Methods For Remote Sensing Image Classification

Résumé Suite aux recentes avancées technologiques, les archives d'images digitales ont connu une croissance qualitative et quantitative sans précédent. Malgré les énormes possibilités qu'elles offrent, ces avancées posent de nouvelles questions quant au traitement des masses de données saisies. Cette question est à la base de cette Thèse: les problèmes de traitement d'information digitale à très haute résolution spatiale et/ou spectrale y sont considérés en recourant à des approches d'apprentissage statistique, les méthodes à noyau. Cette Thèse étudie des problèmes de classification d'images, c'est à dire de catégorisation de pixels en un nombre réduit de classes refletant les propriétés spectrales et contextuelles des objets qu'elles représentent. L'accent est mis sur l'efficience des algorithmes, ainsi que sur leur simplicité, de manière à augmenter leur potentiel d'implementation pour les utilisateurs. De plus, le défi de cette Thèse est de rester proche des problèmes concrets des utilisateurs d'images satellite sans pour autant perdre de vue l'intéret des méthodes proposées pour le milieu du machine learning dont elles sont issues. En ce sens, ce travail joue la carte de la transdisciplinarité en maintenant un lien fort entre les deux sciences dans tous les développements proposés. Quatre modèles sont proposés: le premier répond au problème de la haute dimensionalité et de la redondance des données par un modèle optimisant les performances en classification en s'adaptant aux particularités de l'image. Ceci est rendu possible par un système de ranking des variables (les bandes) qui est optimisé en même temps que le modèle de base: ce faisant, seules les variables importantes pour résoudre le problème sont utilisées par le classifieur. Le manque d'information étiquétée et l'incertitude quant à sa pertinence pour le problème sont à la source des deux modèles suivants, basés respectivement sur l'apprentissage actif et les méthodes semi-supervisées: le premier permet d'améliorer la qualité d'un ensemble d'entraînement par interaction directe entre l'utilisateur et la machine, alors que le deuxième utilise les pixels non étiquetés pour améliorer la description des données disponibles et la robustesse du modèle. Enfin, le dernier modèle proposé considère la question plus théorique de la structure entre les outputs: l'intègration de cette source d'information, jusqu'à présent jamais considérée en télédétection, ouvre des nouveaux défis de recherche. Advanced kernel methods for remote sensing image classification Devis Tuia Institut de Géomatique et d'Analyse du Risque September 2009 Abstract The technical developments in recent years have brought the quantity and quality of digital information to an unprecedented level, as enormous archives of satellite images are available to the users. However, even if these advances open more and more possibilities in the use of digital imagery, they also rise several problems of storage and treatment. The latter is considered in this Thesis: the processing of very high spatial and spectral resolution images is treated with approaches based on data-driven algorithms relying on kernel methods. In particular, the problem of image classification, i.e. the categorization of the image's pixels into a reduced number of classes reflecting spectral and contextual properties, is studied through the different models presented. The accent is put on algorithmic efficiency and the simplicity of the approaches proposed, to avoid too complex models that would not be used by users. The major challenge of the Thesis is to remain close to concrete remote sensing problems, without losing the methodological interest from the machine learning viewpoint: in this sense, this work aims at building a bridge between the machine learning and remote sensing communities and all the models proposed have been developed keeping in mind the need for such a synergy. Four models are proposed: first, an adaptive model learning the relevant image features has been proposed to solve the problem of high dimensionality and collinearity of the image features. This model provides automatically an accurate classifier and a ranking of the relevance of the single features. The scarcity and unreliability of labeled. information were the common root of the second and third models proposed: when confronted to such problems, the user can either construct the labeled set iteratively by direct interaction with the machine or use the unlabeled data to increase robustness and quality of the description of data. Both solutions have been explored resulting into two methodological contributions, based respectively on active learning and semisupervised learning. Finally, the more theoretical issue of structured outputs has been considered in the last model, which, by integrating outputs similarity into a model, opens new challenges and opportunities for remote sensing image processing.

NUTRIENT RECOMMENDATION MODEL FOR CARROT CROP – FERTICALC CARROT

ABSTRACT The literature on fertilization for carrot growing usually recommends nutrient application rates for yield expectations lower than the yields currently obtained. Moreover, the recommendation only considers the results of soil chemical analysis and does not include effects such as crop residues or variations in yield levels. The aim of this study was to propose a fertilizer recommendation system for carrot cultivation (FERTICALC Carrot) which includes consideration of the nutrient supply by crop residues, variation in intended yield, soil chemical properties, and the growing season (winter or summer). To obtain the data necessary for modeling nutritional requirements, 210 carrot production stands were sampled in the region of Alto Paranaíba, State of Minas Gerais, Brazil. The dry matter content of the roots, the coefficient of biological utilization of nutrients in the roots, and the nutrient harvest index for summer and winter crops were determined for these samples. To model the nutrient supply by the soil, the literature was surveyed in regard to this theme. A modeling system was developed for recommendation of macronutrients and B. For cationic micronutrients, the system only reports crop nutrient export and extraction. The FERTICALC which was developed proved to be efficient for fertilizer recommendation for carrot cultivation. Advantages in relation to official fertilizer recommendation tables are continuous variation of nutrient application rates in accordance with soil properties and in accordance with data regarding the extraction efficiency of modern, higher yielding cultivars.

Learning wind fields with multiple kernels

This paper presents multiple kernel learning (MKL) regression as an exploratory spatial data analysis and modelling tool. The MKL approach is introduced as an extension of support vector regression, where MKL uses dedicated kernels to divide a given task into sub-problems and to treat them separately in an effective way. It provides better interpretability to non-linear robust kernel regression at the cost of a more complex numerical optimization. In particular, we investigate the use of MKL as a tool that allows us to avoid using ad-hoc topographic indices as covariables in statistical models in complex terrains. Instead, MKL learns these relationships from the data in a non-parametric fashion. A study on data simulated from real terrain features confirms the ability of MKL to enhance the interpretability of data-driven models and to aid feature selection without degrading predictive performances. Here we examine the stability of the MKL algorithm with respect to the number of training data samples and to the presence of noise. The results of a real case study are also presented, where MKL is able to exploit a large set of terrain features computed at multiple spatial scales, when predicting mean wind speed in an Alpine region.

Conduites spéculaires dans les démences: les multiples signes du miroir [Mirror behaviors in dementia: the many mirror signs].

Mirror behaviors in advanced dementia are: the mirror sign of Abely and Delmas, where the patient stares at his face (environment-driven behavior of Lhermitte); non recognition of the self in the mirror (autoprosopagnosia and/or delirious auto-Capgras); mirror agnosia of Ramachandran and Binkofski where the patient do not understand the concept of mirror and its use; the psychovisual reflex, or reflex pursuit of the eyes when passively moving a minrror in front of a patient (intact vision); mirror writing (procedural learning). We describe four demented patients with mirror behaviors assessing brain mechanisms of self recognition, social brain and mental and visuo-spatial manipulation of images and objects.

Kernel-based Mapping of Meteorological Fields in Complex Orography

Due to the advances in sensor networks and remote sensing technologies, the acquisition and storage rates of meteorological and climatological data increases every day and ask for novel and efficient processing algorithms. A fundamental problem of data analysis and modeling is the spatial prediction of meteorological variables in complex orography, which serves among others to extended climatological analyses, for the assimilation of data into numerical weather prediction models, for preparing inputs to hydrological models and for real time monitoring and short-term forecasting of weather.In this thesis, a new framework for spatial estimation is proposed by taking advantage of a class of algorithms emerging from the statistical learning theory. Nonparametric kernel-based methods for nonlinear data classification, regression and target detection, known as support vector machines (SVM), are adapted for mapping of meteorological variables in complex orography.With the advent of high resolution digital elevation models, the field of spatial prediction met new horizons. In fact, by exploiting image processing tools along with physical heuristics, an incredible number of terrain features which account for the topographic conditions at multiple spatial scales can be extracted. Such features are highly relevant for the mapping of meteorological variables because they control a considerable part of the spatial variability of meteorological fields in the complex Alpine orography. For instance, patterns of orographic rainfall, wind speed and cold air pools are known to be correlated with particular terrain forms, e.g. convex/concave surfaces and upwind sides of mountain slopes.Kernel-based methods are employed to learn the nonlinear statistical dependence which links the multidimensional space of geographical and topographic explanatory variables to the variable of interest, that is the wind speed as measured at the weather stations or the occurrence of orographic rainfall patterns as extracted from sequences of radar images. Compared to low dimensional models integrating only the geographical coordinates, the proposed framework opens a way to regionalize meteorological variables which are multidimensional in nature and rarely show spatial auto-correlation in the original space making the use of classical geostatistics tangled.The challenges which are explored during the thesis are manifolds. First, the complexity of models is optimized to impose appropriate smoothness properties and reduce the impact of noisy measurements. Secondly, a multiple kernel extension of SVM is considered to select the multiscale features which explain most of the spatial variability of wind speed. Then, SVM target detection methods are implemented to describe the orographic conditions which cause persistent and stationary rainfall patterns. Finally, the optimal splitting of the data is studied to estimate realistic performances and confidence intervals characterizing the uncertainty of predictions.The resulting maps of average wind speeds find applications within renewable resources assessment and opens a route to decrease the temporal scale of analysis to meet hydrological requirements. Furthermore, the maps depicting the susceptibility to orographic rainfall enhancement can be used to improve current radar-based quantitative precipitation estimation and forecasting systems and to generate stochastic ensembles of precipitation fields conditioned upon the orography.

Validity of the acoustic approximation in full-waveform seismic crosshole tomography

Acoustic waveform inversions are an increasingly popular tool for extracting subsurface information from seismic data. They are computationally much more efficient than elastic inversions. Naturally, an inherent disadvantage is that any elastic effects present in the recorded data are ignored in acoustic inversions. We investigate the extent to which elastic effects influence seismic crosshole data. Our numerical modeling studies reveal that in the presence of high contrast interfaces, at which P-to-S conversions occur, elastic effects can dominate the seismic sections, even for experiments involving pressure sources and pressure receivers. Comparisons of waveform inversion results using a purely acoustic algorithm on synthetic data that is either acoustic or elastic, show that subsurface models comprising small low-to-medium contrast (?30%) structures can be successfully resolved in the acoustic approximation. However, in the presence of extended high-contrast anomalous bodies, P-to-S-conversions may substantially degrade the quality of the tomographic images. In particular, extended low-velocity zones are difficult to image. Likewise, relatively small low-velocity features are unresolved, even when advanced a priori information is included. One option for mitigating elastic effects is data windowing, which suppresses later arriving seismic arrivals, such as shear waves. Our tests of this approach found it to be inappropriate because elastic effects are also included in earlier arriving wavetrains. Furthermore, data windowing removes later arriving P-wave phases that may provide critical constraints on the tomograms. Finally, we investigated the extent to which acoustic inversions of elastic data are useful for time-lapse analyses of high contrast engineered structures, for which accurate reconstruction of the subsurface structure is not as critical as imaging differential changes between sequential experiments. Based on a realistic scenario for monitoring a radioactive waste repository, we demonstrated that acoustic inversions of elastic data yield substantial distortions of the tomograms and also unreliable information on trends in the velocity changes.

Sectoral and occupational employment growth in Spain 1995-2005

Between 1995 and 2005, the Spanish economy grew at an annual average rate higher than 3,5%. Total employment increased by more than 4.9 millions. Most of this growth was in occupations related with university degrees (more than 890,000, 18% of the total employment increase) and vocational qualifications (more than 855,000, 17.5% of the total employment increase). From a sectoral perspective, the main part of this increase took place in “Real estate, renting and business activities” (K sector in NACE rev.1), “Construction” (F sector) and “Health and social sector” (N sector). This paper analyses this employment growth in an Input-output framework, by means of a structural decomposition analysis (SDA). Two kinds of results have been obtained. From a sectoral perspective we decompose employment growth into Labour requirements change, technical change and demand change. From an occupational perspective, we decompose the employment growth in substitutions effect, labour productivity effect and demand effect. The results show that, in aggregated terms, the main part of this growth is attributable to demand growth, with a small technical improvement. But the results also show that this aggregated behaviour hides important sectoral and occupational variation. The purpose of this paper is to contribute to the ongoing debate over productivity growth and what has been called the “growth model” for the Spanish economy.

Software for interpolation of vegetative growth of yerba mate plants in 3D

The objective of this work was to build mock-ups of complete yerba mate plants in several stages of development, using the InterpolMate software, and to compute photosynthesis on the interpolated structure. The mock-ups of yerba-mate were first built in the VPlants software for three growth stages. Male and female plants grown in two contrasting environments (monoculture and forest understory) were considered. To model the dynamic 3D architecture of yerba-mate plants during the biennial growth interval between two subsequent prunings, data sets of branch development collected in 38 dates were used. The estimated values obtained from the mock-ups, including leaf photosynthesis and sexual dimorphism, are very close to those observed in the field. However, this similarity was limited to reconstructions that included growth units from original data sets. The modeling of growth dynamics enables the estimation of photosynthesis for the entire yerba mate plant, which is not easily measurable in the field. The InterpolMate software is efficient for building yerba mate mock-ups.

A Monte Carlo approach to the inverse problem of diffuse pollution risk in agricultural catchments

The hydrological and biogeochemical processes that operate in catchments influence the ecological quality of freshwater systems through delivery of fine sediment, nutrients and organic matter. Most models that seek to characterise the delivery of diffuse pollutants from land to water are reductionist. The multitude of processes that are parameterised in such models to ensure generic applicability make them complex and difficult to test on available data. Here, we outline an alternative - data-driven - inverse approach. We apply SCIMAP, a parsimonious risk based model that has an explicit treatment of hydrological connectivity. we take a Bayesian approach to the inverse problem of determining the risk that must be assigned to different land uses in a catchment in order to explain the spatial patterns of measured in-stream nutrient concentrations. We apply the model to identify the key sources of nitrogen (N) and phosphorus (P) diffuse pollution risk in eleven UK catchments covering a range of landscapes. The model results show that: 1) some land use generates a consistently high or low risk of diffuse nutrient pollution; but 2) the risks associated with different land uses vary both between catchments and between nutrients; and 3) that the dominant sources of P and N risk in the catchment are often a function of the spatial configuration of land uses. Taken on a case-by-case basis, this type of inverse approach may be used to help prioritise the focus of interventions to reduce diffuse pollution risk for freshwater ecosystems. (C) 2012 Elsevier B.V. All rights reserved.

Experience-based Auditory Predictions Modulate Brain Activity to Silence as Do Real Sounds.

Interactions between stimuli's acoustic features and experience-based internal models of the environment enable listeners to compensate for the disruptions in auditory streams that are regularly encountered in noisy environments. However, whether auditory gaps are filled in predictively or restored a posteriori remains unclear. The current lack of positive statistical evidence that internal models can actually shape brain activity as would real sounds precludes accepting predictive accounts of filling-in phenomenon. We investigated the neurophysiological effects of internal models by testing whether single-trial electrophysiological responses to omitted sounds in a rule-based sequence of tones with varying pitch could be decoded from the responses to real sounds and by analyzing the ERPs to the omissions with data-driven electrical neuroimaging methods. The decoding of the brain responses to different expected, but omitted, tones in both passive and active listening conditions was above chance based on the responses to the real sound in active listening conditions. Topographic ERP analyses and electrical source estimations revealed that, in the absence of any stimulation, experience-based internal models elicit an electrophysiological activity different from noise and that the temporal dynamics of this activity depend on attention. We further found that the expected change in pitch direction of omitted tones modulated the activity of left posterior temporal areas 140-200 msec after the onset of omissions. Collectively, our results indicate that, even in the absence of any stimulation, internal models modulate brain activity as do real sounds, indicating that auditory filling in can be accounted for by predictive activity.

Long-term functional improvements in the 2-year treatment of schizophrenia outpatients with olanzapine long-acting injection

BACKGROUND: Little is known about the long-term changes in the functioning of schizophrenia patients receiving maintenance therapy with olanzapine long-acting injection (LAI), and whether observed changes differ from those seen with oral olanzapine. METHODS: This study describes changes in the levels of functioning among outpatients with schizophrenia treated with olanzapine-LAI compared with oral olanzapine over 2 years. This was a secondary analysis of data from a multicenter, randomized, open-label, 2-year study comparing the long-term treatment effectiveness of monthly olanzapine-LAI (405 mg/4 weeks; n=264) with daily oral olanzapine (10 mg/day; n=260). Levels of functioning were assessed with the Heinrichs-Carpenter Quality of Life Scale. Functional status was also classified as 'good', 'moderate', or 'poor', using a previous data-driven approach. Changes in functional levels were assessed with McNemar's test and comparisons between olanzapine-LAI and oral olanzapine employed the Student's t-test. RESULTS: Over the 2-year study, the patients treated with olanzapine-LAI improved their level of functioning (per Quality of Life total score) from 64.0-70.8 (P<0.001). Patients on oral olanzapine also increased their level of functioning from 62.1-70.1 (P<0.001). At baseline, 19.2% of the olanzapine-LAI-treated patients had a 'good' level of functioning, which increased to 27.5% (P<0.05). The figures for oral olanzapine were 14.2% and 24.5%, respectively (P<0.001). Results did not significantly differ between olanzapine-LAI and oral olanzapine. CONCLUSION: In this 2-year, open-label, randomized study of olanzapine-LAI, outpatients with schizophrenia maintained or improved their favorable baseline level of functioning over time. Results did not significantly differ between olanzapine-LAI and oral olanzapine.

Separation of Gas Bubbles from Circulation Waters and Pulp Suspensions in Open Channel Flow

The objective of this thesis was to study the removal of gases from paper mill circulation waters experimentally and to provide data for CFD modeling. Flow and bubble size measurements were carried out in a laboratory scale open gas separation channel. Particle Image Velocimetry (PIV) technique was used to measure the gas and liquid flow fields, while bubble size measurements were conducted using digital imaging technique with back light illumination. Samples of paper machine waters as well as a model solution were used for the experiments. The PIV results show that the gas bubbles near the feed position have the tendency to escape from the circulation channel at a faster rate than those bubbles which are further away from the feed position. This was due to an increased rate of bubble coalescence as a result of the relatively larger bubbles near the feed position. Moreover, a close similarity between the measured slip velocities of the paper mill waters and that of literature values was obtained. It was found that due to dilution of paper mill waters, the observed average bubble size was considerably large as compared to the average bubble sizes in real industrial pulp suspension and circulation waters. Among the studied solutions, the model solution has the highest average drag coefficient value due to its relatively high viscosity. The results were compared to a 2D steady sate CFD simulation model. A standard Euler-Euler k-ε turbulence model was used in the simulations. The channel free surface was modeled as a degassing boundary. From the drag models used in the simulations, the Grace drag model gave velocity fields closest to the experimental values. In general, the results obtained from experiments and CFD simulations are in good qualitative agreement.

Georeferenced database generation with the purpose of hydrologic molding in reservoirs of the hydrographic basin of Jaguaribe river in the state of Ceará, Brazil

The edafoclimatic conditions of the Brazilian semiarid region favor the water loss by surface runoff. The state of Ceará, almost completely covered by semiarid, has developed public policies for the construction of dams in order to attend the varied water demand. Several hydrological models were developed to support decisive processes in the complex management of reservoirs. This study aimed to establish a methodology for obtaining a georeferenced database suitable for use as input data in hydrological modeling in the semiarid of Ceará. It was used images of Landsat satellite and SRTM Mission, and soil maps of the state of Ceará. The Landsat images allowed the determination of the land cover and the SRTM Mission images, the automatic delineation of hydrographic basins. The soil type was obtained through the soil map. The database was obtained for Jaguaribe River hydrographic basin, in the state of Ceará, and is applicable to hydrological modeling based on the Curve Number method for estimating the surface runoff.