Classification of Airborne Laser Scanning Data in Sandstone Landscapes Abstract This work is concerned with the classification of airborne laser scanning data in sandstone landscapes called "rock cities". Standard filters do not work reliably in such a rugged terrain covered with dense vegetation and in the results the rock formations are smoothed or even removed from the terrain. The method of classification suggested in this work is based on the procedure used in manual filtration. When exploring a sufficiently dense point cloud in 3D, one is able to distinguish rock formations from trees even though their shapes are similar. In contrast to trees, rock pillars are modeled only by points reflected off the surface and therefore they make concave elevations in the ground. Because of penetration of trees, there are points reflected off a treetop, branches, leaves and also ground under the tree. The proposed method segments a point cloud according to local minima in approximated surface and classifies these objects into classes rock, tree, and mix by inner point distribution. Objects in classes tree and mix are then filtered by lasground function from LAStools. The method was tested with merged point cloud consisted of data from the standard airborne laser scanning of the Czech Republic and experimental...

Актуальность и важность исследования автоматизированной системы распознавания эмоций по лицу человека с использованием глубоко разделяемой сверточной нейронной сети во многом определяются использованием преимуществ методов глубокого обучения. Таким образом, для модели с хорошей точностью настройка гиперпараметров является важным аспектом процесса глубокого обучения, а оптимизация поможет в разработке хорошего распознавания эмоций по лицу. Целью диссертации является разработка модели глубокого обучения для распознавания эмоций по лицу с использованием алгоритма сверточной нейронной сети и многоклассовой классификации, а также настройки гиперпараметров с использованием оптимизации ускоренного градиента Нестерова (NAG) для повышения производительности модели глубокого обучения. Целью данной работы является проведение экспериментальных исследований по разработке модели глубокого обучения для определения эмоций человеческого лица на основе черт лица с использованием TensorFlow. Предметом является применение методов машинного обучения для анализа автоматизированной системы распознавания эмоций по лицу человека с использованием DS-CNN. Научная новизна предлагаемой работы заключается в создании нового набора данных по эмоциям лица, который доступен на сайте Kaggle. Во-вторых, для увеличения нелинейности использовались слои глубинной свертки, точечной свертки и глобального среднего пула. И, наконец, алгоритм оптимизации ускоренного градиента Нестерова (NAG) использовался для ускорения обучения и улучшения скорости сходимости. Практическая значимость работы заключается в том, что эта работа позволяет нам разработать модель глубокого обучения с использованием сверточной нейронной сети вместе с многоклассовой классификацией и предлагаемым набором данных, где данные будут предварительно обработаны, а модель DL будет обучена с помощью набора поездов и проверено с помощью тестового набора. Разработанная методология распознает четыре основные эмоции в изображениях людей, такие как счастье, удивление, нейтральность и злость, используя наш предлагаемый набор данных, где эксперимент будет проводиться с предлагаемым набором данных, который доступен на веб-сайте Kaggle. А данные будут оцениваться с помощью статистического анализа с помощью корреляции. Размеченные данные будут предварительно обработаны, а набор данных будет разделен на 3 пары обучающих, проверочных и тестовых наборов различного размера. Модель глубокого обучения будет обучаться с каждым обучающим набором, чтобы найти правильный обучающий размер набора данных, обеспечивающий максимальную точность обучения. Модель глубокого обучения будет проверена с помощью тестового набора для определения наилучшей точности теста, а полученные результаты эксперимента будут проанализированы. / The relevance and importance of the study of automated human facial emotion recognition system using Depthwise Separable Convolutional Neural Network are largely determined by utilizing the benefits of deep learning techniques. So, for a good accuracy model, hyper parameter tuning is an essential aspect of the deep learning process and Optimization will help in developing a good facial emotion recognition. The aim of the thesis is to develop a deep learning model for facial emotion recognition using Convolutional Neural Network algorithm and Multiclass Classification along with Hyper-parameter tuning using Nesterov’s Accelerated Gradient (NAG) Optimization to improve the performance of the deep learning model. The objective of this work is to deal with experimental research to develop a deep learning model to identify the emotion of a human face based on facial features using TensorFlow. The subject is the application of machine learning methods to analyze the automated human facial emotion recognition system using DS-CNN. The scientific novelty of the proposed work is the creation of a new facial emotion dataset which is available on the Kaggle website. Secondly, Depthwise convolutional, pointwise convolutional, and global average pooling layers have been used to increase the nonlinearity. And finally, the Nesterov’s Accelerated Gradient (NAG) optimization algorithm has been used to speed up the training and improve the convergence rate. The practical significance of the work lies in the fact that this work allows us to develop a deep learning model using convolutional neural network along with multiclass classification and proposed dataset where the data will be preprocessed and the DL model will be trained with the train set and validated with the test set. The developed methodology recognizes four basic emotions in images of human beings such as happy, surprise, neutral, and angry using our proposed dataset where experiment will be performed on the proposed dataset which is available on Kaggle website. And the data will be evaluated using statistical analysis with the help of correlation. The labeled data will be pre-processed and the dataset will be split into 3 pairs of training, validation and testing sets of varying sizes. The deep learning model will be trained with each training set to find the proper training size of the dataset which gives highest training accuracy. The deep learning model will be validated with the test set to find the best test accuracy and the obtained results of the experiment will be analyzed.

Military helicopter operations require precise maneuverability characteristics for performance to be determined for the entire helicopter flight envelope. Historically, these maneuverability analyses are combinatorial in nature and involve human-interaction, which hinders their integration into conceptual design. A model formulation that includes the necessary quantitative measures and captures the impact of changing requirements real-time is presented. The formulation is shown to offer a more conservative estimate of maneuverability than traditional energy-based formulations through quantitative analysis of a typical pop-up maneuver. Although the control system design is not directly integrated, two control constraint measures are deemed essential in this work: control deflection rate and trajectory divergence rate. Both of these measures are general enough to be applied to any control architecture, while at the same time enable quantitative trades that relate overall vehicle maneuverability to control system requirements. The dimensionality issues stemming from the immense maneuver space are mitigated through systematic development of a maneuver taxonomy that enables the operational envelope to be decomposed into a minimal set of fundamental maneuvers. The taxonomy approach is applied to a helicopter canonical example that requires maneuverability and design to be assessed simultaneously. The end result is a methodology that enables the impact of design choices on maneuverability to be assessed for the entire helicopter operational envelope, while enabling constraints from control system design to be assessed real-time.

Capillary electrophoresis mass spectrometry (CE-MS) is a versatile instrumental method for metabolomics, which allows for comprehensive metabolite profiling of volume-limited biological specimens in order to better understand the molecular mechanisms associated with chronic diseases, including an alarming epidemic of obesity worldwide. Multiplexed CE separations enable high-throughput metabolite screening with quality assurance to prevent false discoveries when combined with rigorous method validation, robust experimental designs, complementary statistical methods, and high-resolution tandem mass spectrometry (MS/MS) for unknown metabolite identification. In this thesis, multiplexed CE-MS technology is applied for both targeted and untargeted metabolite profiling of various biological fluids, including covalently bound thiol-protein conjugates, as well as free circulating metabolites in serum and plasma, and excreted/bio-transformed compounds in urine due to complex host-gut microflora co-metabolism. This work was applied to characterize aberrant metabolic responses of obese subjects in response to dietary challenges, and measure the benefits of dietary interventions that reduce adiposity without deleterious muscle loss. Chapter 2 presents, a simple, sensitive yet robust analytical protocol to expand metabolome coverage in CE-MS for the discovery of labile protein thiols in human plasma using a rapid chemical derivatization method based on N-tert-butylmaleimide (NTBM). Chapter 3 describes targeted metabolite profiling of serum and plasma to investigate the differential metabolic responses between healthy and unhealthy obese individuals before and after consumption of a standardized high-caloric meal, respectively. Chapter 4 of this thesis describes an untargeted metabolite profiling strategy for urine using multisegment-injection (MSI)-CE-MS for elucidating the effects of protein supplementation following a short-term dietary weight-loss intervention study. This work revealed six urinary metabolites that were classified as top-ranking treatment response biomarkers useful for discriminating between subjects consuming carbohydrate (control), soy, and whey supplemented diets. In summary, this thesis demonstrated the successful implementation of multiplexed CE-MS technology for biomarker discovery in nutritional-based metabolomic studies as required for more effective treatment and prevention of obesity for innovations in public health. / Thesis / Doctor of Philosophy (PhD)

Abstract This thesis focuses on topographic mapping of rock formations with the use of new technologies in a comprehensive manner, from airborne laser scanning (ALS) data acquisition and processing in rocky terrains, followed by their processing to the content of topographic databases and their cartographic processing in maps. The introduction discusses issues of importance for practice, and the relation between topographic mapping of rocks and other fields of human activity. The ALS section describes products for topographic mapping of rocks derived from ALS data, and discusses the specifics of ALS data acquisition and processing in wooded rugged terrain. Existing solutions of this problem are explained and their limitations are identified. Author's own approaches to solving this task are presented as case studies, including three made a further three designed experiments with ALS data processing and evaluation of their results. Recommendation regarding mapping of sandstone landscapes in Czechia have been also addressed. The topographic section describes the current representation of rocks and related objects in the ZABAGED database (Czech national digital topographic database), explains the historical context, analyzes this data and identifies their shortcomings in relation to the ALS. Research...

La thermographie infrarouge est une méthode largement employée pour la caractérisation des propriétés thermophysiques des matériaux. L’avènement des diodes laser pratiques, peu onéreuses et aux multiples caractéristiques, étendent les possibilités métrologiques des caméras infrarouges et mettent à disposition un ensemble de nouveaux outils puissants pour la caractérisation thermique et le contrôle non desturctif. Cependant, un lot de nouvelles difficultés doit être surmonté, comme le traitement d’une grande quantité de données bruitées et la faible sensibilité de ces données aux paramètres recherchés. Cela oblige de revisiter les méthodes de traitement du signal existantes, d’adopter de nouveaux outils mathématiques sophistiqués pour la compression de données et le traitement d’informations pertinentes. Les nouvelles stratégies consistent à utiliser des transformations orthogonales du signal comme outils de compression préalable de données, de réduction et maîtrise du bruit de mesure. L’analyse de sensibilité, basée sur l’étude locale des corrélations entre les dérivées partielles du signal expérimental, complète ces nouvelles approches. L'analogie avec la théorie dans l'espace de Fourier a permis d'apporter de nouveaux éléments de réponse pour mieux cerner la «physique» des approches modales.La réponse au point source impulsionnel a été revisitée de manière numérique et expérimentale. En utilisant la séparabilité des champs de température nous avons proposé une nouvelle méthode d'inversion basée sur une double décomposition en valeurs singulières du signal expérimental. Cette méthode par rapport aux précédentes, permet de tenir compte de la diffusion bi ou tridimensionnelle et offre ainsi une meilleure exploitation du contenu spatial des images infrarouges. Des exemples numériques et expérimentaux nous ont permis de valider dans une première approche cette nouvelle méthode d'estimation pour la caractérisation de diffusivités thermiques longitudinales. Des applications dans le domaine du contrôle non destructif des matériaux sont également proposées. Une ancienne problématique qui consiste à retrouver les champs de température initiaux à partir de données bruitées a été abordée sous un nouveau jour. La nécessité de connaitre les diffusivités thermiques du matériau orthotrope et la prise en compte des transferts souvent tridimensionnels sont complexes à gérer. L'application de la double décomposition en valeurs singulières a permis d'obtenir des résultats intéressants compte tenu de la simplicité de la méthode. En effet, les méthodes modales sont basées sur des approches statistiques de traitement d'une grande quantité de données, censément plus robustes quant au bruit de mesure, comme cela a pu être observé. / Infrared thermography is a widely used method for characterization of thermophysical properties of materials. The advent of the laser diodes, which are handy, inexpensive, with a broad spectrum of characteristics, extend metrological possibilities of infrared cameras and provide a combination of new powerful tools for thermal characterization and non destructive evaluation. However, this new dynamic has also brought numerous difficulties that must be overcome, such as high volume noisy data processing and low sensitivity to estimated parameters of such data. This requires revisiting the existing methods of signal processing, adopting new sophisticated mathematical tools for data compression and processing of relevant information.New strategies consist in using orthogonal transforms of the signal as a prior data compression tools, which allow noise reduction and control over it. Correlation analysis, based on the local cerrelation study between partial derivatives of the experimental signal, completes these new strategies. A theoretical analogy in Fourier space has been performed in order to better understand the «physical» meaning of modal approaches.The response to the instantaneous point source of heat, has been revisited both numerically and experimentally. By using separable temperature fields, a new inversion technique based on a double singular value decomposition of experimental signal has been introduced. In comparison with previous methods, it takes into account two or three-dimensional heat diffusion and therefore offers a better exploitation of the spatial content of infrared images. Numerical and experimental examples have allowed us to validate in the first approach our new estimation method of longitudinal thermal diffusivities. Non destructive testing applications based on the new technique have also been introduced.An old issue, which consists in determining the initial temperature field from noisy data, has been approached in a new light. The necessity to know the thermal diffusivities of an orthotropic medium and the need to take into account often three-dimensional heat transfer, are complicated issues. The implementation of the double singular value decomposition allowed us to achieve interesting results according to its ease of use. Indeed, modal approaches are statistical methods based on high volume data processing, supposedly robust as to the measurement noise.