Modern vehicles are very dependent on sensors and especially cameras to analyze different objects and conditions. Single camera systems are frequently used for lane detection and identifying objects in the distance. These systems depend on good conditions to work properly and are easily disrupted by environmental interferences. This project is targeted in developing an image processing algorithm that can detect disturbances applied upon the camera lens. Several focus measure operators are evaluated and compared by expected outcome while still maintaining a satisfying computational time, low resource usage and accuracy with an ERR between 0% - 1%.

The purpose of this research is to develop a process and an algorithm to create a 3D model of the surface a part. This is accomplished using a single camera and a CNC machine as a movable stage. A gradient based focus measure operator written in MATLAB is used to process the images and to generate the surface model. The scopes of this research are image processing and surface model generation as well as verifying part accuracy. The algorithm is able to create a rough surface model of a photographed part, and with careful calibration in a limited number of scenarios has been used in checking part z dimensions.

Dans le cadre de l'acquisition de l'information de profondeur de scènes texturées, un processus d'estimation de la profondeur basé sur la méthode de reconstruction 3D « Shape from Focus » est présenté dans ce manuscrit. Les deux étapes fondamentales de cette approche sont l'acquisition de la séquence d'images de la scène par sectionnement optique et l'évaluation de la netteté locale pour chaque pixel des images acquises. Deux systèmes d'acquisition de cette séquence d'images sont présentés ainsi que les traitements permettant d'exploiter celle-ci pour la suite du processus d'estimation de la profondeur. L'étape d'évaluation de la netteté des pixels passe par la comparaison des différents opérateurs de mesure de netteté. En plus des opérateurs usuels, deux nouveaux opérateurs basés sur les descripteurs généralisés de Fourier sont proposés. Une méthode nouvelle et originale de comparaison est développée et permet une analyse approfondie de la robustesse à différents paramètres des divers opérateurs. Afin de proposer une automatisation du processus de reconstruction, deux méthodes d'évaluation automatique de la netteté sont détaillées. Finalement, le processus complet de reconstruction est appliqué à des scènes agronomiques, mais également à une problématique du domaine de l'analyse de défaillances de circuits intégrés afin d'élargir les domaines d'utilisation / In the context of the acquisition of depth information for textured scenes, a depth estimation process based on a 3D reconstruction method called "shape from focus" is proposed in this thesis. The two crucial steps of this approach are the image sequence acquisition of the scene by optical sectioning and the local sharpness evaluation for each pixel of the acquired images. Two acquisition systems have been developed and are presented as well as different image processing techniques that enable the image exploitation for the depth estimation process. The pixel sharpness evaluation requires comparison of different focus measure operators in order to determine the most appropriate ones. In addition to the usual focus measure operators, two news operators based on generalized Fourier descriptors are presented. A new and original comparison method is developped and provides a further analysis of the robustness to various parameters of the focus measure operators. In order to provide an automatic version of the reconstruction process, two automatic sharpness evaluation methods are detailed. Finally, the whole reconstruction process is applied to agronomic scenes, but also to a problematic in failure analysis domain aiming to expand to other applications

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2018-02-22T13:55:41Z No. of bitstreams: 1 Marcelo Robson de Azevedo Martins da Silva_.pdf: 3067819 bytes, checksum: d7b5c1e064c742114237189fe8cdc3d7 (MD5) / Made available in DSpace on 2018-02-22T13:55:41Z (GMT). No. of bitstreams: 1 Marcelo Robson de Azevedo Martins da Silva_.pdf: 3067819 bytes, checksum: d7b5c1e064c742114237189fe8cdc3d7 (MD5) Previous issue date: 2017-09-27 / Nenhuma / Existem muitas técnicas para reconstrução de objetos tridimensionais em computador, algumas são empregadas em ambientes controlados e outras em ambientes que não necessitam de grande precisão. Shape From Focus é um método bastante conhecido que utiliza uma pilha de fotografias retiradas com diferentes configurações focais para reconstruir um mapa de profundidade bastante preciso. Este método obtém maior estabilidade na reconstrução de objetos muito pequenos ou microscópios, mas recentemente vem sendo utilizado para reconstrução de ambientes. Com isso, o modelo de reconstrução de mapas de profundidade, Shape From Focus, passou a processar maiores quantidades de interferências na pilha de fotografias, como por exemplo, a distorção da lente, o aumento da profundidade de campo, o efeito zoom, entre outros, e também o ruído introduzido pelo ambiente. Este trabalho analisa os efeitos do suporte adaptativo para o janelamento de avaliação do medidor de qualidade de foco do método Shape From Focus. Apesar de diferentes trabalhos sobre este tema utilizarem diversas variações do janelamento de avaliação, o suporte adaptativo pode fornecer uma alternativa para encontrar a estabilidade e confiança na obtenção do mapa de profundidade, limitando o erro introduzido por interferências globais. / There are many techniques for reconstructing three-dimensional objects in a computer, some are used in controlled environments and others in environments that do not require great precision. Shape From Focus is one of the well-known method that uses a stack of cropped photographs with different focal settings to reconstruct a fairly accurate depth map. This method obtains greater stability in the reconstruction of very small objects or microscopes, but has recently been used for reconstruction of environments. As a result, the Shape From Focus reconstruction model began to process greater amounts of interference in the photo stack, such as lens distortion, increased depth of field, zoom effect, among others, as well as noise Introduced by the environment. This work analyzes the effects of the adaptive support for the evaluation window of the focus quality meter of the Shape From Focus method. Although different works on this theme use several variations of the evaluation window, the adaptive support can provide an alternative to find the stability and confidence in obtaining the depth map, limiting the error introduced by global interferences.