In this thesis the predictive models the multilayer perceptron and random forest are evaluated to predict cloud top pressure. The dataset used in this thesis contains brightness temperatures, reflectances and other useful variables to determine the cloud top pressure from the Advanced Very High Resolution Radiometer (AVHRR) instrument on the two satellites NOAA-17 and NOAA-18 during the time period 2006-2009. The dataset also contains numerical weather prediction (NWP) variables calculated using mathematical models. In the dataset there are also observed cloud top pressure and cloud top height estimates from the more accurate instrument on the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite. The predicted cloud top pressure is converted into an interpolated cloud top height. The predicted pressure and interpolated height are then evaluated against the more accurate and observed cloud top pressure and cloud top height from the instrument on the satellite CALIPSO. The predictive models have been performed on the data using different sampling strategies to take into account the performance of individual cloud classes prevalent in the data. The multilayer perceptron is performed using both the original response cloud top pressure and a log transformed repsonse to avoid negative values as output which is prevalent when using the original response. Results show that overall the random forest model performs better than the multilayer perceptron in terms of root mean squared error and mean absolute error.

Detta examensarbete undersöker forensisk längdmätning i bild, t ex längduppskattning av människor i bilder rörande brottsmål. Problemen identifieras och några av dagens befintliga längdmätningsmetoder diskuteras. Den metod som bäst uppfyller de i arbetet ställda kraven, d v s snabb handläggning, minimal systeminformation, minimalt arbete på plats och exakthet, har valts ut, anpassats och utvärderats. Metoden bygger på att hitta s k gränspunkter och grundplanets gränslinje i bilden och utifrån en i världen känd referenslängd beräkna den sökta längden. Den bakomliggande teorin presenteras och metoden beskrivs i detalj. Funktioner, algoritmer och ett användargränssnitt har implementerats i beräkningsprogrammet MatLab. Tester har utförts för att validera metodens noggrannhet och parameterberoende. Metoden visar sig ge mycket bra resultat då rätt förutsättningar ges, men har konstaterats vara känslig för variation på gränslinjen. En rad förbättringsförslag presenteras för att utveckla metoden och stabilisera resultatet. Examensarbetet omfattar 20 högskolepoäng och utgör ett obligatoriskt moment i utbildningsprogrammet civilingenjör i datateknik som ges av Linköpings universitet. Arbetet är utfört vid och på uppdrag av Statens kriminaltekniska laboratorium (SKL) i Linköping.

During the last decade the human body detection and tracking has been a very extensive research eld within the computer vision. There are many potential applications of people tracking such as security-monitoring, anthropomorphic analysis or biometrics. In this thesis we present an algorithm and related software to detect human bodies in an indoor environment. It is part of a wider project which aims to estimate the human height. The purposed algorithm performs in real-time to detect people. The algorithm is developed using the free OpenCV library in C++ programming language. As far as this algorithm is rst part of a wider system, our software gives two outputs. The principal one is the coordinates of the detected object. With the coordinates, the aforementioned measuring system will be able to calculate the height by itself. The other output is the video sequence with the detected person bounded by a rectangle, wich provides visual feedback to the user. This software is able to communicate with Matlab Engine. It is important since the subsequent height estimation system works in Matlab®.

It has been generally recognised that sick children have an increased risk on malnutrition. An activated immune system requires more nutrients while illness is often associated with a lower absorption and decreased consumption because of lack of appetite. When these increased needs are not balanced with an adequate availability of nutrients, the sick child is at higher risk of developing malnutrition.Médecins Sans Frontières investigated the question whether this process is mitigated by simple short term nutritional supplementation given to sick children alongside medical treatment. Three Randomised Controlled Trials (RCT’s) were conducted. The first, in Democratic Republic of Congo, was a pilot; 180 children with malaria were randomised in 2 arms: 1 group receiving 2 weeks of ready to use therapeutic food (RUTF) and a control group. The children were followed for a period of 4 weeks. Children in the RUTF group showed a higher weight gain in the first 14 days compared to the control group, at day 28 the weight gain in both groups was similar.Thereafter, 2 RCT’s were implemented in Uganda and Nigeria using a similar methodology. Children with malaria, lower respiratory tract infection or diarrhoea (sample size of 2202) were randomised in three groups: supplemented with 2 weeks of RUTF, supplemented with 2 weeks of micronutrient powder (MNP), and not receiving supplementation after each disease episode. The incidence of malnutrition was compared after an observation period of 6 months. The trial in Uganda showed a reduction in malnutrition in the RUITF group with 31%, while in Nigeria, there was no significant reduction in the RUTF group. The MNP group did not show reduction in malnutrition in any site. In the group of moderate malnourished children the RUTF and MNP supplementations were not effective in preventing deterioration to severe malnutrition. However, when the studies were combined the RUTF group showed a lower mortality compared to the MNP group.Multi-variate analysis did not show a reduction of incidence of malnutrition in the supplementation groups. A strong association with morbidity was found. A higher frequency of diarrhoea was associated with an increased incidence of malnutrition. The association with malaria episodes was mixed; it was associated with a higher incidence of malnutrition in Kaabong, but in Goronyo a higher frequency of malaria decreased the incidence of malnutrition. In addition, a more frequent monitoring of the children and treatment of their illnesses was associated with a decreased incidence of malnutrition.The difference in effectiveness of supplementation between the sites can be explained by differences in food security and level of morbidity. It is argued that the fragile food security in Kaabong limits the supply of nutrients, and therefore supplementation with RUTF was effective. In Goronyo the high frequency of morbidity limits convalescence and therefore supplementation was not effective.It is likely that malnutrition is more effectively prevented when several interventions are combined like water and sanitation to prevent diarrhoea, malaria chemoprophylaxis and preventative and curative health and nutrition interventions.This dissertation will present the background, the methods of the trials and the results, followed by a discussion on the implications for programming and research. / Doctorat en Santé Publique / info:eu-repo/semantics/nonPublished

Conselho Nacional de Desenvolvimento Científico e Tecnológico / The study uses dominant trees to describe the tinning, because they are those that possess larger energy inside of a forest settlement. The main objective of the work is to model and to select models to describe the relationship among the potential number of trees for hectare in function of the dominant height, elaborating a system of curves guides how to aid in the determination of the tinning. The municipal and districts where are located the forests are: Cambará do Sul (RS), São Francisco de Paula (RS), Distrito de Tainhas (RS), Distrito dos Kroeff (RS), Bocaína do Sul (SC), Correia Pinto (SC), Curitibanos (SC), Lages (SC), Otacílio Costa (SC), Ponte Alta do Norte (SC) and Santa Cecília (SC). The units temporary sample random were installed using the Method of modified Prodan, or be, 12 trees were measured by unit sample being the center of the unit the dominant tree of the sample. The units were installed at forests in competition from 5 to 40 meters of dominant height. In those units they were measured: the distance of the 11 trees to the central dominant tree; diameter and the height of the 12 trees; the crown diameter, the length of total crown and the exposed crown length to the light of the 2 dominant trees of the sample. The calculated variables were the number of potential trees for hectare, the Index of relative spacing, the crown percentage and the index of competition of Glover and Hool. The modelling made calculations by the statistical package (Statistical Analysis System) version 8.0 and the adjustment of the equations and covariance analysis for the procedure GLM ( General Lineal Models). The equations that best describes the potential number of trees for hectare were Ln(Np) = 10,31285 - 1,28518*(ln(hdom) with a R²aj 0,68 and CV of 4,76% for the class of S'% of 14-20% and Ln(Np) = 10,36304 - 1,16938*(ln(hdom) with a R²aj 0,34 and CV of 5,24% for the class of S'% of 14-20%. The use of the dominant height as independent variable is efficient to describe the guidelines for tinning in the settlements of Pinus taeda. The model of negative exponential regression adjusts with good precision and low mistake the relationship among the potential number of trees for hectare and the dominant height. The variation of the values of potential frequency for hectare for a given dominant height is strongly reduced with the bedding of the data in classes of Index of relative spacing of Hart Becking larger than 14% and smaller than 14% allowing to elaborate frequency curves for hectare for those two situations. The development of the frequency curves stratified by Index of relative spacing influences in the dimension of the diameter of the trees for a same dominant height. / O estudo utiliza árvores dominantes para descrever o desbaste, pois são aquelas que possuem maior vigor dentro de um povoamento florestal. O principal objetivo do trabalho é modelar e selecionar modelos para descrever a relação entre o número potencial de árvores por hectare em função da altura dominante, elaborando um sistema de curvas guias como auxiliar na determinação do desbaste. Os municípios e distritos onde estão localizadas as florestas são: Cambará do Sul (RS), São Francisco de Paula (RS), Distrito de Tainhas (RS), Distrito dos Kroeff (RS), Bocaína do Sul (SC), Correia Pinto (SC), Curitibanos (SC), Lages (SC), Otacílio Costa (SC), Ponte Alta do Norte (SC) e Santa Cecília (SC). As unidades amostrais temporárias foram instaladas aleatoriamente utilizando o Método de Prodan modificado, ou seja, foram medidas 12 árvores por unidade amostral sendo o centro da unidade a árvore dominante da amostra. As unidades foram instaladas em florestas em competição de 5 a 40 metros de altura dominante. Nessas unidades foram medidas: a distância das 11 árvores até a árvore dominante central; o DAP e a altura das 12 árvores; o diâmetro de copa, o comprimento de copa total e o comprimento de copa exposto à luz das 2 árvores dominantes da amostra. As variáveis calculadas foram o número de árvores potenciais por hectare, o Índice de espaçamento relativo, a percentagem de copa e o índice de concorrência de Glover e Hool. A modelagem foi calculada pelo pacote estatístico SAS (Statistical Analysis System) versão 8.0 e o ajuste das equações e análise de covariância pelo procedimento GLM (General Linear Models). As equações que melhor descrevem o número potencial de árvores por hectare foram Ln(Np) = 10,31285 1,28518*(ln(hdom) com um R²aj de 0,68 e CV de 4,76% para a classe de S % de 14 20% e Ln(Np) = 10,36304 1,16938*(ln(hdom) com um R²aj de 0,34 e CV de 5,24% para a classe de S % de 14 20%. O uso da altura dominante como variável independente é eficiente para descrever as diretrizes para desbaste nos povoamentos de Pinus taeda. O modelo de regressão exponencial negativo ajusta com boa precisão e baixo erro a relação entre o número potencial de árvores por hectare e a altura dominante. A variação dos valores de freqüência potencial por hectare para uma dada altura dominante é fortemente reduzida com a estratificação dos dados em classes de Índice de espaçamento relativo de Hart Becking maior do que 14% e menor do que 14% permitindo elaborar curvas de freqüência por hectare para essas duas situações. O desenvolvimento das curvas de freqüência estratificado por Índice de espaçamento relativo influi na dimensão do diâmetro das árvores para uma mesma altura dominante.

The Global Positioning System has all heights referred to the ellipsoid. However, for engineering applications it is necessary to have heights related to the earth gravitational field, which have physical meaning. In order to determine the orthometric height from the geoidal height, which is given at the GPS, it is necessary to know the geoidal undulation. The objective of this paper is to compare an empiric model of local geoidal undulation with a native model present in handheld GPS model GARMIN 12XL and the software call MAPGEO2004 IBGE. A test poligonal had been seted up in the campus of Universidade Regional Integrada do Alto Uruguai e das Missões (URI), in Santiago, RS. The poligonal was leveled and all the poligonal vertices had their orthometric heights determinated. The model for geoidal undulation present in the GARMIN receivers was accessed by a software developed by Geomatics Laboratory at UFSM. Using this program, all geoidal undulations were recorded for each surveyed point. Geoidal undulation with MAPGEO2004 have been obtained from the points geographic coordinates as inputs to the program. Empiric models were developed for the local geoidal undulation, which were compared to the GARMIN 12XL and MAGEO2004 models. The results analyses were conducted with a linear regression between all the three models and the traditional leveling. The results indicate a high correlation has been observed between the empirical model for geoidal undulation and the MAPGEO2004 program. However, there has been no adjust between orthometrics heights of the empirical model and those determined by the handheld receiver / O sistema GPS fornece altitudes referenciadas ao elipsóide. No entanto, para aplicações em engenharia, necessita-se de altitudes relacionadas com o campo gravitacional terrestre, que possuem ligação com a realidade física. Desse modo, para determinar as altitudes ortométricas a partir das altitudes geométricas, determinadas com o GPS, é preciso, conhecer a ondulação geoidal. O objetivo deste trabalho é comparar o desempenho de um modelo empírico de ondulação geoidal local com o modelo nativo de receptores de navegação modelo Garmin 12XL e o programa MAPGEO2004 - IBGE. Para tanto, foi instalada uma poligonal teste dentro do campus da Universidade Regional Integrada do Alto Uruguai e das Missões (URI), em Santiago, RS. Na mesma, foi executado um nivelamento de precisão para a determinação das altitudes ortométricas dos vértices. O modelo de ondulação geoidal existente nos receptores GPS modelo GARMIN 12XL, foi acessado por meio de um programa desenvolvido no Laboratório de Geomática da Universidade Federal de Santa Maria (UFSM). Por meio deste programa, foram gravadas as ondulações geoidais dadas pelo equipamento para cada ponto rastreado. No programa MAPGEO2004 as ondulações geoidais foram obtidas a partir da entrada das coordenadas curvilíneas latitude e longitude dos pontos de interesse. Foram desenvolvidos modelos empíricos para a ondulação geoidal local, que posteriormente foram comparados com os oriundos do GARMIN 12 XL e do MAOGEO2004. A análise dos resultados foi realizada a partir de uma regressão linear entre cada um dos três modelos e o nivelamento tradicional com nível de precisão. Os resultados levantados apontam uma alta correlação entre o modelo empírico de ondulação geoidal local e o programa MAPGEO2004. No entanto, não houve um bom ajuste entre as altitudes ortométricas do modelo empírico e aquelas determinadas pelo receptor GPS de navegação GARMIN 12XL

Cette thèse s'intéresse aux constantes d'Hermite généralisées associées au groupe linéaire adèlique. A l'image de la théorie de Voronoï classique, on y définit deux propriétés, la perfection et l'eutaxie qui caractérisent les maxima locaux de l'invariant d'Hermite. Des inégalités et liens connus dans le cas classique sont étendus au cas général et fournissent la valeur de la constante dans certains cas. Par une théorie des designs définie pour la variété drapeau et semblable à celle des designs sphériques et grassmaniens, on fournit également de nombreux exemples d'objets atteignant l'extrémum. / This thesis studies generalised Hermite constants associated with the adelic general linear group. Like for the classical Voronoi theory, we define two properties, perfection and eutaxy, which characterise the local maxima of the Hermite invariant. Upper bounds and links known in the classical case are extended to the general case and provide the value of the constant in some cases. Through a theory of designs defined for the flag variety and similar to spherical or grassmanian design, we give also many examples of objects reaching locally the extremum.

Master of Science / Department of Agronomy / Scott A. Staggenborg / Spatial variability in plant available water can be caused by uncontrollable factors such as topography and soil texture as well as controllable factors such as residue management. Research located on the High Plains evaluated the impact of wheat (Triticum aestivum L.) stubble height on snow catch, plant available water at seeding, and optimal corn seeding rates. Treatments consisted of stripper harvest height of 71 cm (28 in.), cut heights of 25 cm (10 in.), and 10 cm (4 in.) Measured snow depths were significantly different among treatments (p<0.0001) with equivalent precipitation of 5.77 (2.27), 3.25 (1.28), and 1.73 cm (0.68 in.) for the stripped, 25 cm, and 10 cm heights respectively. Available soil water at planting increased 24% as stubble height increased from 10 to 71 cm (4 to 28 in) in one year of the study. Two corn hybrids of varying maturity (97 and 108 days) were planted into the stubble treatments at seeding rates ranging from 2.47 to 5.43 plants m[superscript]-2 (10 to 22 000 plants ac[superscript]-1). In the dry year, the long season hybrid responded positively to increasing population in tall stubble and negatively in short stubble. Yield of the short season hybrid increased with increasing stubble height and was mostly unresponsive to population. Grain yields of both hybrids responded positively to increasing plant population in a wet year. Treatments also affected the yield components of yield plant[superscript]-1, kernel weight, and kernels plant[superscript]-1. Managing seeding rates for uncontrollable factors was attempted with small-plot and field scale research across 3 fields in northeast Kansas. A relationship between soil electro-conductivity (EC) and measured water holding capacity values was developed for one study field. This quadratic relationship was significant (p<0.0001) and explained variability in water holding capacity with respect to EC quite well (R[superscript]2=0.6239). Responses from small plots showed that sites differing in population response characteristics could be identified. Field scale data was used to derive a function describing optimal seeding rate with respect to soil EC. In the field under study, optimal seeding rates varied from 3.08 to 8.74 plants m[superscript]-2 (12 500 to 35 375 plants ac-1).

Magister Scientiae - MSc / Many Single Nucleotide Polymorphisms (SNPs) found in coding or regulatory regions within the human genome lead to phenotypic differences that make prediction of physical appearance, based on genetic analysis, potentially useful in forensic investigations. Complex traits such as pigmentation can be predicted from the genome sequence, provided that genes with strong effects on the trait exist and are known. Phenotypic traits may also be associated with variations in gene expression due to the presence of SNPs in promoter regions. In this project, the identification of genes associated with these physical traits of potential forensic relevance have been collated from the literature using a text mining platform and hand curation. The SNPs associated with these genes have been acquired from public SNP repositories such as the International HapMap project, dbSNP and Ensembl. Characterization of different population groups based on the SNPs has been performed and the results and data stored in a MySQL database. This database contains SNP genotyping data with respect to physical phenotypic differences of forensic interest. The potential forensicrelevance of the SNP information contained in this database has been verified through in silico SNP analysis aimed at establishing possible relationships between SNP occurrence and phenotype. The software used for this analysis is MATCH™. Data management and access has been enhanced by the use of a functional web-based front-end which enables the users to extract and display SNP information without running complex Structured Query Language (SQL) statements from the command line. This Forensic SNP Phenotype resource can be accessed at http://forensic.sanbi.ac.za/alecia_forensics/Index.html / South Africa

A caprinocultura leiteira é uma atividade que contribui para a melhoria da geração de renda e emprego, em várias localidades do Nordeste do Brasil. O uso de pasto cultivado pode reduzir o efeito da estacionalidade e tornar sustentável a produção de leite de cabra utilizando pastagem. O objetivo deste trabalho foi determinar o efeito de tipos de manejo sobre o potencial de produção de leite de cabra em pasto de capim-tanzânia. O experimento foi conduzido na Embrapa Caprinos e Ovinos no Ceará (3°40´58.42 latitude sul; 40°1650.5 longitude e 79m de altitude). Foram utilizadas 65 cabras Anglo Nubianas, mantidas em pasto de capim-tanzânia manejado sob lotação rotativa, com taxa de lotação variável. Os manejos foram combinações entre alturas residuais do pasto (Altr) e doses de nitrogênio (N). Os manejos foram: intensivo (Altr=32,7 cm e 600 kg de N/ha ano-1); moderado (Altr=47,2 cm e 300 kg N/ha ano-1), leve (Altr=47,3 cm sem N) e extensivo (Altr=32,1 cm sem N). Foram avaliadas características na planta, no animal e no solo. As variáveis analisadas foram: o fluxo de biomassa através da morfogênese e as características estruturais do pasto, comportamento animal, o peso e o escore de condição corporal, produção de leite, as perdas de água e sedimentos e a densidade do solo. As eficiências de uso de nitrogênio (EUN), de água (EUA) e concentrado foram quantificadas. Uma análise econômica foi realizada para determinar a viabilidade econômica dos manejos. O moderado apresentou produtividade de 10.806 kg de leite por ha ano-1 e apresentou maior ERUN (1 kg N produziu 120 kg MS) e do concentrado (0,65kg concentrado por kg de leite). Apesar da densidade do solo ter sido 1,55 g/cm3, apresentou baixa perda de sedimento (72 kg/ha ano-1) e água (0,9%) por erosão. Apesar dos bons índices técnicos o manejo foi economicamente inviável. O manejo leve, também apresentou bons indicadores técnicos, no entanto, a produção de 7.032 kg/ha ano-1 deu prejuízo. O uso de animais mais produtivos e criação de linhas diferenciadas de crédito para investimento e subsídio são opções para estimular produtores a fazer uso destes tipos de manejo. Maiores produtividades (19.442 kg leite/ha ano-1) foram registradas no manejo intensivo, que foi o único que apresentou viabilidade econômica. Apesar da maior EUA (2 kg leite/mm de água), sua dependência de insumos externos (600 kg N/ha ano-1 e 0,89 kg/kg de leite) aumenta o risco de alterações de mercado no preço de insumos o que pode tornar inviável o uso intensivo. O extensivo foi o pior desempenho (1kg concentrado/kg de leite), registrou-se perda no vigor das plantas, queda na densidade de perfilhos ao longo do ano (de 200 para 150 perfilhos/m2) e maior perda dágua por escorrimento superficial (3%) sendo o único dos manejos não indicado. Sistemas de produção sustentáveis de leite de cabra em pastagem cultivada usando os manejos moderado, leve ou intensivo dependem do desenvolvimento de métodos mais eficientes de controle de verminose. Políticas públicas incluindo linhas de crédito e subsídios são necessários para produzir leite de cabra em pastagem cultivada. / The goat milk production is an activity that contributes to improve income generation and employment in several locations in northeast of Brazil (3°40´58.42 latitude south; 40°1650.5 longitude and 79m of elevation). The use of cultivated pasture can reduce the effect of seasonality making the dairy goat production more sustainable using pasture. The objective of this study was to determine the effect of types of management on the potential production of dairy goats in Tanzania grass pasture. The experiment was conducted at Embrapa Goats and Sheep in Ceará. Anglo Nubian goats (65) were kept in Guineagrass pasture managed under rotational stocking with variable stocking rate. The managements were combinations of pasture residual heights (Altr) and nitrogen doses (N). The managements were: intensive (ALTr = 32.7 cm and 600 kg N / ha year-1), moderate (ALTr = 47.2 cm and 300 kg N / ha year-1), light (ALTr = 47.3 without N cm) and extensive (ALTr = 32.1 cm without N). Plant, animal and soil characteristics were evaluated. The variables analyzed were: the flow of biomass through the morphogenetics and structural characteristics of the pasture, animal behavior, weight and body condition score, milk production, losses of water and sediment and soil density. Nitrogen (UEN), Water (WUE) and concentrate use efficiency were quantified. An economic analysis was performed to determine the economic viability of each management. Moderate treatment productivity was 10,806 kg of milk per ha year-1 and UEN (1 kg of N produced 120 kg of DM) and concentrate use efficiency (0.65 kg concentrate per kg of milk) were higher. Although soil density was 1.55 g/cm3, this treatment showed losses of sediment (72 kg / ha year-1) and water (0.9%) by erosion. Despite the good technical indexes moderate one was not economically different sustainable. Light management, as the moderate one, had good technical indicators, however, the production of 7032 kg / ha year-1 determined a negative economically viable treatment. The use of more productive animals and the possibility to create differentiated lines of credit for investment and subsidies are alternatives to encourage producers to use this kind of management. The highest yield (19,442 kg milk / ha year-1) was observed in the intensive management, which was the only economically viable treatment. Despite the higher WUE (2 kg milk/mm of water), its dependence on external inputs (600 kg N/ha year-1 and 0.89 kg/kg of milk) increases the risk of changes in market prices of inputs making the intensive system risky. Extensive management had the worst performance (1kg concentrate/kg of milk), there was loss in plant vigor, decrease in tiller density throughout the year (from 200 to 150 tillers/m2) and greater water loss by leaching (3%) and is not recommended. Sustainable goat milk production systems using moderate, light or intensive pasture management depends on the development of more efficient methods of worms control. Public policies including lines of credit and subsidies are necessary to goats milk production on cultivated pastures in Northeast area of Brazil.