Thesis (M.S.)--Rochester Institute of Technology, 1983. / Typescript. Includes bibliographical references (leaves 116-118).

Thesis (Ph. D.)--Ohio State University, 2004. / Title from first page of PDF file. Document formatted into pages; contains xiv, 166 p.; also includes graphics (some col.) Includes bibliographical references (p. 156-166).

Thesis (Ph. D.)--University of Wisconsin--Madiosn, 1983. / Typescript. Vita. eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references (leaves 434-442).

Thesis -- Memorial University of Newfoundland. / Typescript. Bibliography : leaves 233-243. Also available online.

Thesis (M.Sc.) - University of KwaZulu-Natal, Pietermaritzburg, 2010. / Full text also available online. Scroll down for electronic link.

Thesis (Ph. D.)--University of Wisconsin--Madison, 1975. / Typescript. Vita. eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references (leaves 198-201).

Os sistemas de estereofotogrametria digital representam um método não radioativo, seguro, preciso e objetivo para quantificar as dimensões dos tecidos moles faciais. Novos equipamentos têm despontado como alternativa viável nas especialidades que tratam do complexo craniofacial. A estereofotogrametria se destaca pela completa ausência de radiação ionizante, alta velocidade de aquisição, grandes ângulos de visualização, navegação virtual da imagem e possibilidade de integração com outros sistemas de imagens. Quando um novo método é proposto, torna-se necessário avaliar a sua confiabilidade na obtenção de resultados. O objetivo deste estudo foi analisar a repetibilidade intraexaminador em medidas faciais lineares e angulares, em um sistema de estereofotogrametria digital utilizado para avaliação da morfologia facial. A amostra final consistiu em 30 indivíduos adultos brasileiros com idade entre 18 e 45 anos, média de 26,71± 6,53 anos. Vinte e cinco mulheres e cinco homens, saudáveis, sem histórico de traumas ou cirurgias faciais e sem anormalidades congênitas, tiveram pontos de referência marcados em suas faces por um mesmo examinador e foram fotografados em dois momentos, com intervalo de uma semana (T1 e T2), pelo sistema de estereofotogrametria digital Vectra M3® (Canfield Scientific, Fairfield, EUA). Trinta e oito pontos foram demarcados em cada captura. Identificou-se um total de 2280 pontos; 11 medidas foram feitas em cada imagem capturada e 330 medições foram realizadas. As imagens obtidas foram medidas em software próprio do sistema. A análise foi realizada a partir de um conjunto de nove medidas angulares (ângulos nasolabial, mentolabial, convexidade facial, convexidade facial total, nasal, maxilar, mandibular, nasofrontal, maxilofacial, e duas medidas lineares (altura facial média AFAM e altura facial inferior - AFAI), propostas para caracterizar a morfologia facial dos indivíduos da amostra. A repetibilidade intraexaminador na aquisição das medidas foi analisada pela Média Absoluta das diferenças (MAD), Magnitude Relativa do Erro (MRE), Erro Técnico de Medição (ETM), Coeficiente de Correlação Intraclasse (ICC) e análise de Bland-Altman. O teste t para amostras pareadas foi utilizado na identificação de erro sistemático entre T1-T2. O coeficiente de correlação de Pearson foi empregado para verificar a associação entre idade/IMC e o erro nas medições realizadas. Resultados: No cálculo do ETM e da MAD, apenas os ângulos nasolabial (2,17°) e mentolabial (2,88°) apresentaram valores maiores que 2 graus (limite clínico estabelecido para o presente estudo). Para a MRE, 9,1% das medidas foram consideradas moderadas (ângulo maxilofacial), 27,3% foram consideradas boas (ângulo nasal, ângulo maxilar, ângulo mandibular), 45,5% muito boas (ângulo nasolabial, ângulo mentolabial, ângulo nasofrontal, AFAM, AFAI) e 18,2% excelentes (ângulo da convexidade facial, ângulo da convexidade facial total.). Os ângulos nasolabial, mentolabial, convexidade facial, convexidade facial total, maxilofacial e nasofrontal apresentaram valores de ICC (R) classificados como excelentes. O ângulo nasal, e as medidas lineares de AFAI e AFAM tiveram um resultado satisfatório. Não ficou comprovada a relação entre idade/IMC e o erro encontrado entre as medições (diferença absoluta entre T1 e T2). As medidas para AFAM apresentaram um erro sistemático. As medições realizadas foram consideradas precisas dentro dos limites e referências estabelecidas para o presente estudo. Os ângulos nasolabial e mentolabial devem ser interpretados com critério pela variabilidade apresentada. O examinador apresentou boa repetibilidade nas medições realizadas. Os resultados obtidos no presente estudo acrescentam confiabilidade ao uso da estereofotogrametria digital na antropometria facial. Os parâmetros de precisão adequados a cada situação, em clínica ou em pesquisa, devem ser previamente estabelecidos. / Digital stereophotogrammetry systems represent a non-radioactive, safe, and accurate method to quantify the facial soft tissues dimensions. New equipment has emerged as viable alternative in the specialties that deal with the craniofacial complex. Stereophotogrammetry system has been highlighted by the complete absence of ionizing radiation, high-speed acquisition, large viewing angles, virtual image navigation and possibility of integration with others imaging systems. When a new method is proposed, it is necessary to assess its reliability in achieving results. The aim of this study was to analyze the intra-examiner repeatability on linear and angular facial measurements into a digital stereophotogrammetry system used to assess facial morphology. The final sample consisted of 30 Brazilian adults aged between 18 and 45 years, average of 26.71±6,53 years. Five men and twenty-five women, healthy, with no history of trauma or previous facial surgeries and without congenital abnormalities, had landmarks labeled on their faces by the same rater and were photographed in two stages (T1-T2), with an interval of one week, at stereophotogrammetry system digital Vectra M3® (Canfield Scientific, Fairfield, NJ). Thirty-eight points were labeled in each capture. 2280 points were identified; 11 distances were measured in each image, giving a total of 330 measurements. The images were measured by a dedicated software system. The analysis was performed from a set of nine angular measurements (nasolabial, mentolabial, facial convexity, full facial convexity, nasal, maxillary, mandibular, nasofrontal, maxillofacial angles, and two linear measurements (middle facial height - AFAM and lower facial height AFAI). Precision and intra-rater repeatability in the measurements acquisition were evaluated by the Mean Absolute Difference (MAD) Relative Error Magnitude (REM), Technical Error of Measurements (TEM), intra-class correlation coefficient (ICC) and Bland-Altman analysis. Test t for paired samples was used in the systematic error identification between T1-T2. The Pearson correlation coefficient was used to verify the association between age / BMI and measurement errors. Results: to the estimated TEM and MAD only nasolabial angles (2.17°), mentolabial (2.88°) had higher values than 2.0 degrees (clinical limit for this study). For MRE, 9.1% of the measures were considered moderate (maxillofacial angle), 27.3% were considered good (nasal angle, maxillary angle, mandibular angle), 45.5% very good (nasolabial angle, mentolabial angle, nasofrontal angle, AFAM, AFAI) and 18.2% excellent (facial convexity, full facial convexity.). The nasolabial, mentolabial, facial convexity, total facial convexity, maxillofacial and nasofrontal angles showed R values rated as excellent. The nasal angle, and the linear measurements of AFAI and AFAM had a satisfactory outcome. It was not proven the relation between age / body mass index and the error (absolute difference between T1 and T2). Measures to AFAM showed a systematic error. Conclusion: The measurements were considered accurate within the limits and references established for this study. The examiner showed good repeatability in identifying landmarks. The results presented on this study add reliability to the Digital stereophotogrammetry in clinical and research activities, provided there are previously established adequate precision parameters.

Orientador: Aluir Porfírio Dal Poz / Banca: Mauricio Galo / Banca: Fernando Marques de Almeida Nogueira / Resumo: Nesta pesquisa é proposta uma metodologia automática para a extração de segmentos de rodovia em imagens aéreas ou orbitais de diferentes resoluções (baixa, média e alta resolução). A metodologia é baseada no conceito generalizado de linhas em imagens digitais, pelo qual as linhas podem ser descritas por feições simétricas entre duas bordas paralelas. No caso específico de imagens de baixa resolução, onde as rodovias manifestam-se como entidades de 1 ou 2 pixels de espessura, as linhas coincidem com as próprias rodovias. Neste caso, a metodologia baseou-se nas seguintes etapas: Realce da malha viária; Aquisição dos limiares de histerese; Detecção de linhas com seleção automática de escala, utilizando o conceito de bordas paralelas; Limiarização por comprimento e Extração das linhas poligonais representando segmentos de rodovia através de algoritmos de poligonização. Em imagens de média e alta resolução as rodovias manifestam-se como faixas estreitas e alongadas e, conseqüentemente, o alvo de extração torna-se o eixo de simetria de cada rodovia. Neste caso, não é necessária apenas a primeira etapa da metodologia proposta para imagens de baixa resolução. Os resultados obtidos na avaliação experimental mostram, de forma geral, que o método proposto atendeu todos os critérios estabelecidos para a extração eficiente de segmentos de rodovias em cenas rurais, proporcionando resultados satisfatórios, de maneira inteiramente automática. / Abstract: In this research an automatic methodology is proposed for road segment extraction in aerial or satellite imagery in different resolutions (low, medium and high resolution). The methodology is based on the generalized concept of lines in digital images, by which lines can be described as center lines between two parallel edges. In specific case of low resolution images, where roads manifest as features of 1 or 2 pixels wide, lines representations coincide with the own roads. In this case, the methodology is based on the following steps: Road network enhancement; Acquisition of the thresholds of histerese; Road detection using the concept of parallel edges; Threshold by the line length criteria; and Polygon extraction representing road segments by polygonization algorithms. In medium and high resolution images roads manifest as narrow and elongated strips and, consequently, the extraction goal becomes the symmetric axis of each road. In this case, it is not necessary only the first step of the methodology proposed for images of low resolution. The results obtained in the experimental estimation indicate, in general form, that the proposed method met all the criteria established for the efficient extraction of road segments in rural scenes, providing satisfactory results, in a completely automatic way. / Mestre

Ice plays an important role in hydraulic processes of rivers in cold regions such as Canada. The formation, progression, recession and breakup of river ice cover known as river ice processes affect river hydraulics, sediment transport characteristics as well as river morphology. Ice jamming and break up are responsible of winter flash floods, river bed modification and bank scour. River ice cover monitoring using terrestrial images from cameras installed on the shores can help monitor and understand river ice processes. In this study, the benefits of terrestrial monitoring of river ice using a camera installed on the shore are evaluated. A time-lapse camera system was installed during three consecutive winters at two locations on the shores of the Lower Nelson River, in Northern Manitoba and programmed to take an image of the river ice cover approximatively every hour. An image analysis algorithm was then developed to automatically extract quantitative characteristics of the river ice cover from the captured images. The developed algorithm consists of four main steps: preprocessing, image registration, georectification and river ice detection. The contributions of this thesis include the development of a novel approach for performing georectification while accounting for a fluctuating water surface elevation, and the use of categorization approach and a locally adaptive image thresholding technique for target detection. The developed algorithm was able to detect and quantify important river ice cover characteristics such as the area covered by ice, border ice progression and ablation rate, and river ice break up processes with an acceptable accuracy.

Quantitative measures of stand density and stocking are very important because only with full knowledge of the growing stock can a forest be managed efficiently. Stocking is concerned with fraction of area occupied with trees. Density is related to the degree of crowding within the area occupied. These quantitative values are determined by parameters that could be measured on the ground and on aerial photographs. The methods used to estimate height, crown width, and crown closure from aerial photographs are thoroughly described. The writer also describes the stocking and density conditions under which trees grow, with illustrations by both theoretical models and actual sample plot crown models. Forty-seven sets of ground and photo-measurements were taken and analysed by simple and multiple regression methods. A comparison of photo and ground values was then made to evaluate the usefulness of aerial photographs for density and stocking measurements. The correlation of the ratio of height (Ht) to crown width (CW) from the ground and photo data to age, site index, crown closure, basal area, adjusted basal area, crowding factor and adjusted crowding factor were also studied. Eight assumptions regarding normality of density and full stocking were made so that the interrelationship between the individual density and stocking measurements could be studied more effectively. It is concluded that Ht/CW ratios are measurable from aerial photographs and shown that they are useful as a measure of stand density and stocking. / Forestry, Faculty of / Graduate