Predicting forest attributes In southeast Alaska using artificial neural networks

Artificial neural network (ANN) methods are used to predict forest characteristics. The data source is the Southeast Alaska (SEAK) Grid Inventory, a ground survey compiled by the USDA Forest Service at several thousand sites. The main objective of this article is to predict characteristics at unsurveyed locations between grid sites. A secondary objective is to evaluate the relative performance of different ANNs. Data from the grid sites are used to train six ANNs: multilayer perceptron, fuzzy ARTMAP, probabilistic, generalized regression, radial basis function, and learning vector quantization. A classification and regression tree method is used for comparison. Topographic variables are used to construct models: latitude and longitude coordinates, elevation, slope, and aspect. The models classify three forest characteristics: crown closure, species land cover, and tree size/structure. Models are constructed using n-fold cross-validation. Predictive accuracy is calculated using a method that accounts for the influence of misclassification as well as measuring correct classifications. The probabilistic and generalized regression networks are found to be the most accurate. The predictions of the ANN models are compared with a classification of the Tongass national forest in southeast Alaska based on the interpretation of satellite imagery and are found to be of similar accuracy.

Modeling of titanium alloys by using artificial neural networks

Titanium alloy exhibits an excellent combination of bio-compatibility, corrosion resistance, strength and toughness. The microstructure of an alloy influences the properties. The microstructures depend mainly on alloying elements, method of production, mechanical, and thermal treatments. The relationships between these variables and final properties of the alloy are complex, non-linear in nature, which is the biggest hurdle in developing proper correlations between them by conventional methods. So, we developed artificial neural networks (ANN) models for solving these complex phenomena in titanium alloys.

In the present work, ANN models were used for the analysis and prediction of the correlation between the process parameters, the alloying elements, microstructural features, beta transus temperature and mechanical properties in titanium alloys. Sensitivity analysis of trained neural network models were studied which resulted a better understanding of relationships between inputs and outputs. The model predictions and the analysis are well in agreement with the experimental results. The simulation results show that the average output-prediction error by models are less than 5% of the prediction range in more than 95% of the cases, which is quite acceptable for all metallurgical purposes.

Two-Stage Orthogonal Least Squares Methods for Neural Network Construction

A number of neural networks can be formulated as the linear-in-the-parameters models. Training such networks can be transformed to a model selection problem where a compact model is selected from all the candidates using subset selection algorithms. Forward selection methods are popular fast subset selection approaches. However, they may only produce suboptimal models and can be trapped into a local minimum. More recently, a two-stage fast recursive algorithm (TSFRA) combining forward selection and backward model refinement has been proposed to improve the compactness and generalization performance of the model. This paper proposes unified two-stage orthogonal least squares methods instead of the fast recursive-based methods. In contrast to the TSFRA, this paper derives a new simplified relationship between the forward and the backward stages to avoid repetitive computations using the inherent orthogonal properties of the least squares methods. Furthermore, a new term exchanging scheme for backward model refinement is introduced to reduce computational demand. Finally, given the error reduction ratio criterion, effective and efficient forward and backward subset selection procedures are proposed. Extensive examples are presented to demonstrate the improved model compactness constructed by the proposed technique in comparison with some popular methods.

Comparison of methods of eutrophication assessment in a coastal lagoon (Ria Formosa)

Dissertação de Mestrado, Gestão da Água e da Costa, Faculdade de Ciências e Tecnologia, Universidade do Algarve, 2010

Aquisição por Visão Artificial e Medição Automática de Micro Dureza Vickers

O ensaio de dureza, e mais concretamente o ensaio de micro dureza Vickers, é no universo dos ensaios mecânicos um dos mais utilizados quer seja na indústria, no ensino ou na investigação e desenvolvimento de produto no âmbito das ciências dos materiais. Na grande maioria dos casos, a utilização deste ensaio tem como principal aplicação a caracterização ou controlo da qualidade de fabrico de materiais metálicos. Sendo um ensaio de relativa simplicidade de execução, rapidez e com resultados comparáveis e relacionáveis a outras grandezas físicas das propriedades dos materiais. Contudo, e tratando-se de um método de ensaio cuja intervenção humana é importante, na medição da indentação gerada por penetração mecânica através de um sistema ótico, não deixa de exibir algumas debilidades que daí advêm, como sendo o treino dos técnicos e respetivas acuidades visuais, fenómenos de fadiga visual que afetam os resultados ao longo de um turno de trabalho; ora estes fenómenos afetam a repetibilidade e reprodutibilidade dos resultados obtidos no ensaio. O CINFU possui um micro durómetro Vickers, cuja realização dos ensaios depende de um técnico treinado para a execução do mesmo, apresentando todas as debilidades já mencionadas e que o tornou elegível para o estudo e aplicação de uma solução alternativa. Assim, esta dissertação apresenta o desenvolvimento de uma solução alternativa ao método ótico convencional na medição de micro dureza Vickers. Utilizando programação em LabVIEW da National Instruments, juntamente com as ferramentas de visão computacional (NI Vision), o programa começa por solicitar ao técnico a seleção da câmara para aquisição da imagem digital acoplada ao micro durómetro, seleção do método de ensaio (Força de ensaio); posteriormente o programa efetua o tratamento da imagem (aplicação de filtros para eliminação do ruído de fundo da imagem original), segue-se, por indicação do operador, a zona de interesse (ROI) e por sua vez são identificadas automaticamente os vértices da calote e respetivas distâncias das diagonais geradas concluindo, após aceitação das mesmas, com o respetivo cálculo de micro dureza resultante. Para validação dos resultados foram utilizados blocos-padrão de dureza certificada (CRM), cujos resultados foram satisfatórios, tendo-se obtido um elevado nível de exatidão nas medições efetuadas. Por fim, desenvolveu-se uma folha de cálculo em Excel com a determinação da incerteza associada às medições de micro dureza Vickers. Foram então comparados os resultados nas duas metodologias possíveis, pelo método ótico convencional e pela utilização das ferramentas de visão computacional, tendo-se obtido bons resultados com a solução proposta.

Inteligência artificial aplicada na previsão de mercados financeiros

Neste documento, são investigados vários métodos usados na inteligência artificial, com o objetivo de obter previsões precisas da evolução dos mercados financeiros. O uso de ferramentas lineares como os modelos AR, MA, ARMA e GARCH têm muitas limitações, pois torna-se muito difícil adaptá-los às não linearidades dos fenómenos que ocorrem nos mercados. Pelas razões anteriormente referidas, os algoritmos como as redes neuronais dinâmicas (TDNN, NARX e ESN), mostram uma maior capacidade de adaptação a estas não linearidades, pois não fazem qualquer pressuposto sobre as distribuições de probabilidade que caracterizam estes mercados. O facto destas redes neuronais serem dinâmicas, faz com que estas exibam um desempenho superior em relação às redes neuronais estáticas, ou outros algoritmos que não possuem qualquer tipo de memória. Apesar das vantagens reveladas pelas redes neuronais, estas são um sistema do tipo black box, o que torna muito difícil extrair informação dos pesos da rede. Isto significa que estes algoritmos devem ser usados com precaução, pois podem tornar-se instáveis.

Radiothérapie asservie à la respiration en combinaison avec l'utilisation d'un faisceau sans filtre égalisateur

La radiothérapie stéréotaxique corporelle (SBRT) est une technique couramment employée pour le traitement de tumeurs aux poumons lorsque la chirurgie n’est pas possible ou refusée par le patient. Une complication de l’utilisation de cette méthode provient du mouvement de la tumeur causé par la respiration. Dans ce contexte, la radiothérapie asservie à la respiration (RGRT) peut être bénéfique. Toutefois, la RGRT augmente le temps de traitement en raison de la plus petite proportion de temps pour laquelle le faisceau est actif. En utilisant un faisceau de photons sans filtre égalisateur (FFF), ce problème peut être compensé par le débit de dose plus élevé d’un faisceau FFF. Ce mémoire traite de la faisabilité d’employer la technique de RGRT en combinaison avec l’utilisation un faisceau FFF sur un accélérateur Synergy S (Elekta, Stockholm, Suède) avec une ceinture pneumatique, le Bellows Belt (Philips, Amsterdam, Pays-Bas), comme dispositif de suivi du signal respiratoire. Un Synergy S a été modifié afin de pouvoir livrer un faisceau 6 MV FFF. Des mesures de profils de dose et de rendements en profondeur ont été acquises en cuve à eau pour différentes tailles de champs. Ces mesures ont été utilisées pour créer un modèle du faisceau 6 MV FFF dans le système de planification de traitement Pinnacle3 de Philips. Les mesures ont été comparées au modèle à l’aide de l’analyse gamma avec un critère de 2%, 2 mm. Par la suite, cinq plans SBRT avec thérapie en arc par modulation volumétrique (VMAT) ont été créés avec le modèle 6 MV du Synergy S, avec et sans filtre. Une comparaison des paramètres dosimétriques a été réalisée entre les plans avec et sans filtre pour évaluer la qualité des plans FFF. Les résultats révèlent qu’il est possible de créer des plans SBRT VMAT avec le faisceau 6 MV FFF du Synergy S qui sont cliniquement acceptables (les crières du Radiation Therapy Oncology Group 0618 sont respectés). Aussi, une interface physique de RGRT a été mise au point pour remplir deux fonctions : lire le signal numérique de la ceinture pneumatique Bellows Belt et envoyer une commande d’irradiation binaire au linac. L’activation/désactivation du faisceau du linac se fait par l’entremise d’un relais électromécanique. L’interface comprend un circuit électronique imprimé fait maison qui fonctionne en tandem avec un Raspberry Pi. Un logiciel de RGRT a été développé pour opérer sur le Raspberry Pi. Celui-ci affiche le signal numérique du Bellows Belt et donne l’option de choisir les limites supérieure et inférieure de la fenêtre d’irradiation, de sorte que lorsque le signal de la ceinture se trouve entre ces limites, le faisceau est actif, et inversement lorsque le signal est hors de ces limites. Le logiciel envoie donc une commande d’irradiation au linac de manière automatique en fonction de l’amplitude du signal respiratoire. Finalement, la comparaison entre la livraison d’un traitement standard sans RGRT avec filtre par rapport à un autre plan standard sans RGRT sans filtre démontre que le temps de traitement en mode FFF est réduit en moyenne de 54.1% pour un arc. De la même manière, la comparaison entre la livraison d’un traitement standard sans RGRT avec filtre par rapport à un plan de RGRT (fenêtre d’irradiation de 75%) sans filtre montre que le temps de traitement de RGRT en mode FFF est réduit en moyenne de 27.3% par arc. Toutefois, il n’a pas été possible de livrer des traitements de RGRT avec une fenêtre de moins de 75%. Le linac ne supporte pas une fréquence d’arrêts élevée.

Feature Extraction Methods Based on Linear Predictive Coding and Wavelet Packet Decomposition for Recognizing Spoken Words in Malayalam

Speech signals are one of the most important means of communication among the human beings. In this paper, a comparative study of two feature extraction techniques are carried out for recognizing speaker independent spoken isolated words. First one is a hybrid approach with Linear Predictive Coding (LPC) and Artificial Neural Networks (ANN) and the second method uses a combination of Wavelet Packet Decomposition (WPD) and Artificial Neural Networks. Voice signals are sampled directly from the microphone and then they are processed using these two techniques for extracting the features. Words from Malayalam, one of the four major Dravidian languages of southern India are chosen for recognition. Training, testing and pattern recognition are performed using Artificial Neural Networks. Back propagation method is used to train the ANN. The proposed method is implemented for 50 speakers uttering 20 isolated words each. Both the methods produce good recognition accuracy. But Wavelet Packet Decomposition is found to be more suitable for recognizing speech because of its multi-resolution characteristics and efficient time frequency localizations

Electrical Design: A Problem for Artificial Intelligence Research

This report outlines the problem of intelligent failure recovery in a problem-solver for electrical design. We want our problem solver to learn as much as it can from its mistakes. Thus we cast the engineering design process on terms of Problem Solving by Debugging Almost-Right Plans, a paradigm for automatic problem solving based on the belief that creation and removal of "bugs" is an unavoidable part of the process of solving a complex problem. The process of localization and removal of bugs called for by the PSBDARP theory requires an approach to engineering analysis in which every result has a justification which describes the exact set of assumptions it depends upon. We have developed a program based on Analysis by Propagation of Constraints which can explain the basis of its deductions. In addition to being useful to a PSBDARP designer, these justifications are used in Dependency-Directed Backtracking to limit the combinatorial search in the analysis routines. Although the research we will describe is explicitly about electrical circuits, we believe that similar principles and methods are employed by other kinds of engineers, including computer programmers.

Comparing methods for dimensionality reduction when data are density functions

Functional Data Analysis (FDA) deals with samples where a whole function is observed for each individual. A particular case of FDA is when the observed functions are density functions, that are also an example of infinite dimensional compositional data. In this work we compare several methods for dimensionality reduction for this particular type of data: functional principal components analysis (PCA) with or without a previous data transformation and multidimensional scaling (MDS) for diferent inter-densities distances, one of them taking into account the compositional nature of density functions. The difeerent methods are applied to both artificial and real data (households income distributions)

Case based reasoning as an extension of fault dictionary methods for linear electronic analog circuits diagnosis

El test de circuits és una fase del procés de producció que cada vegada pren més importància quan es desenvolupa un nou producte. Les tècniques de test i diagnosi per a circuits digitals han estat desenvolupades i automatitzades amb èxit, mentre que aquest no és encara el cas dels circuits analògics. D'entre tots els mètodes proposats per diagnosticar circuits analògics els més utilitzats són els diccionaris de falles. En aquesta tesi se'n descriuen alguns, tot analitzant-ne els seus avantatges i inconvenients. Durant aquests últims anys, les tècniques d'Intel·ligència Artificial han esdevingut un dels camps de recerca més importants per a la diagnosi de falles. Aquesta tesi desenvolupa dues d'aquestes tècniques per tal de cobrir algunes de les mancances que presenten els diccionaris de falles. La primera proposta es basa en construir un sistema fuzzy com a eina per identificar. Els resultats obtinguts son força bons, ja que s'aconsegueix localitzar la falla en un elevat tant percent dels casos. Per altra banda, el percentatge d'encerts no és prou bo quan a més a més s'intenta esbrinar la desviació. Com que els diccionaris de falles es poden veure com una aproximació simplificada al Raonament Basat en Casos (CBR), la segona proposta fa una extensió dels diccionaris de falles cap a un sistema CBR. El propòsit no és donar una solució general del problema sinó contribuir amb una nova metodologia. Aquesta consisteix en millorar la diagnosis dels diccionaris de falles mitjançant l'addició i l'adaptació dels nous casos per tal d'esdevenir un sistema de Raonament Basat en Casos. Es descriu l'estructura de la base de casos així com les tasques d'extracció, de reutilització, de revisió i de retenció, fent èmfasi al procés d'aprenentatge. En el transcurs del text s'utilitzen diversos circuits per mostrar exemples dels mètodes de test descrits, però en particular el filtre biquadràtic és l'utilitzat per provar les metodologies plantejades, ja que és un dels benchmarks proposats en el context dels circuits analògics. Les falles considerades son paramètriques, permanents, independents i simples, encara que la metodologia pot ser fàcilment extrapolable per a la diagnosi de falles múltiples i catastròfiques. El mètode es centra en el test dels components passius, encara que també es podria extendre per a falles en els actius.

New Methods for Triangulation-based Shape Acquisition using Laser Scanners

Tradicionalment, la reproducció del mon real se'ns ha mostrat a traves d'imatges planes. Aquestes imatges se solien materialitzar mitjançant pintures sobre tela o be amb dibuixos. Avui, per sort, encara podem veure pintures fetes a ma, tot i que la majoria d'imatges s'adquireixen mitjançant càmeres, i es mostren directament a una audiència, com en el cinema, la televisió o exposicions de fotografies, o be son processades per un sistema computeritzat per tal d'obtenir un resultat en particular. Aquests processaments s'apliquen en camps com en el control de qualitat industrial o be en la recerca mes puntera en intel·ligència artificial. Aplicant algorismes de processament de nivell mitja es poden obtenir imatges 3D a partir d'imatges 2D, utilitzant tècniques ben conegudes anomenades Shape From X, on X es el mètode per obtenir la tercera dimensió, i varia en funció de la tècnica que s'utilitza a tal nalitat. Tot i que l'evolució cap a la càmera 3D va començar en els 90, cal que les tècniques per obtenir les formes tridimensionals siguin mes i mes acurades. Les aplicacions dels escàners 3D han augmentat considerablement en els darrers anys, especialment en camps com el lleure, diagnosi/cirurgia assistida, robòtica, etc. Una de les tècniques mes utilitzades per obtenir informació 3D d'una escena, es la triangulació, i mes concretament, la utilització d'escàners laser tridimensionals. Des de la seva aparició formal en publicacions científiques al 1971 [SS71], hi ha hagut contribucions per solucionar problemes inherents com ara la disminució d'oclusions, millora de la precisió, velocitat d'adquisició, descripció de la forma, etc. Tots i cadascun dels mètodes per obtenir punts 3D d'una escena te associat un procés de calibració, i aquest procés juga un paper decisiu en el rendiment d'un dispositiu d'adquisició tridimensional. La nalitat d'aquesta tesi es la d'abordar el problema de l'adquisició de forma 3D, des d'un punt de vista total, reportant un estat de l'art sobre escàners laser basats en triangulació, provant el funcionament i rendiment de diferents sistemes, i fent aportacions per millorar la precisió en la detecció del feix laser, especialment en condicions adverses, i solucionant el problema de la calibració a partir de mètodes geomètrics projectius.

Intelligent methods for load forecasting

Load forecasting is an important task in the management of a power utility. The most recent developments in forecasting involve the use of artificial intelligence techniques, which offer powerful modelling capabilities. This paper discusses these techniques and provides a review of their application to load forecasting.

Drivers of increased soil respiration in a poplar coppice exposed to elevated CO2

Background and Aims. The response of soil respiration (SR) to elevated CO2 is driven by a number of processes and feedbacks. This work aims to i) detect the effect of elevated CO2 on soil respiration during the second rotation of a short rotation forest, at two levels of N availability; and ii) identify the main drivers behind any changes in soil respiration. Methods. A poplar plantation (POP-EUROFACE) was grown for two rotations of three years under elevated CO2 maintained by a FACE (Free Air CO2 Enrichment) technique. Root biomass, litter production and soil respiration were followed for two consecutive years after coppice. Results. In the plantation, the stimulation of fine root and litter production under elevated CO2 observed at the beginning of the rotation declined over time. Soil respiration (SR) was continuously stimulated by elevated CO2, with a much larger enhancement during the growing (up to 111 %) than in the dormant season (40 %). The SR increase at first appeared to be due to the increase in fine root biomass, but at the end of the 2nd rotation was supported by litter decomposition and the availability of labile C. Soil respiration increase under elevated CO2 was not affected by N availability. Conclusions. The stimulation of SR by elevated CO2 was sustained by the decomposition of above and belowground litter and by the greater availability of easily decomposable substrates into the soil. C losses through SR were greater in the last year of the plantation due to a lack of effect of elevated CO2 on C allocation to roots, reducing the potential for C accumulation.