La dernière décennie a vu l'exploitation d’application d’inspection automatique dans plusieurs domaines grâce à l’avancé des capteurs de vision et des méthodes d’analyse de texture et de segmentation d’images. Cependant, la nature difficile des images de chaussées (fortement texturée), la petite taille des défauts (fissures) conduisent au constat que l’inspection dans ce domaine est réalisée manuellement. Chaque année, en France, des opérateurs doivent visualiser des milliers de kilomètres d'images de route pour y relever des dégradations. Cette façon de faire est couteuse, lente et a un résultat plutôt subjectif. L’objectif de ce travail de thèse est de développer une méthode permettant la détection et la classification des fissures automatiquement sur ces images de chaussées. Le coeur de la thèse est une nouvelle méthode de segmentation, la Free Form Anisotropy (FFA). D'une part, elle permet de prendre en compte simultanément les attributs concernant la forme et l’intensité des pixels d’une fissure pour la détection. D’autre part, une nouvelle modélisation est utilisée en recherchant des chemins minima dans des graphes (images) afin de trouver la forme de la fissure dès qu'elle est présente dans l’image. Après la segmentation, l’extraction et la classification de défauts sont réalisées par une transformée de Hough et par le calcul de l’orientation locale des pixels. Les résultats expérimentaux ont été obtenus à partir de plusieurs bases d’images et compares avec des méthodes existantes. / These last decades have seen application of automatic inspection in many fields thanks to advanced vision sensors and image analysis methods. However, the difficult nature of pavement images, the small size of defects (cracks) lead to the fact that inspection in this area is done mostly manually. Each year in France, operator must view images of thousands kilometers of roads to detect these degradations. This method is expensive, slow and has a rather subjective result. The objective of this thesis is to develop a method for the detection and the classification of cracks on these pavement images automatically. In this thesis, a new method of segmentation has been developed: the Free Form Anisotropy (FFA). On one hand, this method allows to take into account both the features concerning form and intensity of cracks, for the detection. On the other hand, a new model is used to search minimum paths in graphs (images). This minimum path follows crack form when crack is present. After segmentation, extraction and classification of defects are performed by the Standard Hough Transform and by calculating local orientation of pixels. Experimental results have been obtained from different image databases and compared with other existing methods.

Submitted by Luciana Ferreira (lucgeral@gmail.com) on 2014-10-30T14:24:27Z No. of bitstreams: 2 Dissertação - Roussian Di Ramos Alves Gaioso - 2014.pdf: 6127790 bytes, checksum: 9990f791c0f9abaee7e3e03e4cdc8ee4 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-10-30T14:29:29Z (GMT) No. of bitstreams: 2 Dissertação - Roussian Di Ramos Alves Gaioso - 2014.pdf: 6127790 bytes, checksum: 9990f791c0f9abaee7e3e03e4cdc8ee4 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-10-30T14:29:29Z (GMT). No. of bitstreams: 2 Dissertação - Roussian Di Ramos Alves Gaioso - 2014.pdf: 6127790 bytes, checksum: 9990f791c0f9abaee7e3e03e4cdc8ee4 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2014-08-08 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / This paper presents a Graphics Processing Unit (GPU) based parallels implementations for the All Pairs Shortest Paths and Transitive Closure problems in graph. The implementations are based on the main sequential algorithms and takes full advantage of the highly multithreaded architecture of current manycore GPUs. Our solutions reduces the communication between CPU and GPU, improves the Streaming Multiprocessors (SMs) utilization, and makes intensive use of coalesced memory access to optimize graph data access. The advantages of the proposed implementations are demonstrated for several graphs randomly generated using the widely known graph library GTgraph. Graphs containing thousands of vertices and different edges densities, varying from sparse to complete graphs, were generated and used in the experiments. Our results confirm that GPU implementations can be competitive even for graph algorithms whose memory accesses and work distribution are both irregular and data-dependent. Keywords / Este trabalho apresenta implementações paralelas baseadas em Graphics Processing Unit (GPU) para os problemas da identificação dos caminhos mínimos entre todos os pares de vértices e do fecho transitivo em um grafo. As implementações são baseadas nos principais algoritmos sequenciais e tiram o máximo proveito da arquitetura multithreaded das GPUs atuais. Nossa solução reduz a comunicação entre a Central Processing Unit (CPU) e a GPU, melhora a utilização dos Streaming Multiprocessors (SMs) e faz um uso intensivo de acesso aglutinado em memória para otimizar o acesso de dados do grafo. As vantagens dessas implementações propostas são demonstradas por vários grafos gerados aleatoriamente utilizando a ferramenta GTgraph. Grafos contendo milhares de vértices foram gerados e utilizados nos experimentos. Nossos resultados confirmam que implementações baseadas em GPU podem ser viáveis mesmo para algoritmos de grafos cujo acessos à memória e distribuição de trabalho são irregulares e causam dependência de dados.