[vc_row row_type=”row” text_align=”left” css_animation=””][vc_column][vc_separator type=”normal”][vc_column_text]<\/p>\n
By Josep Bosch Alay Supervised by Dr. Nuno Gr\u00e0cias \/ Dr. Pere Ridao Omnidirectional vision has received increasing interest during the last decade from the computer vision community. A large number of camera models have reached the market to meet the increasing demand for panoramic imagery. However, the use of omnidirectional cameras underwater is still very limited. In this thesis we propose a number of methods to create a reference resource for designing, calibrating and using underwater omnidirectional multi-camera systems.<\/p>\n <\/p>\n <\/p>\n The first problem we address is the design and calibration of omnidirectional cameras for the underwater domain. Among the different imaging system approaches to capturing omnidirectional imagery we chose the use of multi-cameras, due to the higher resolution and quality of the final images obtained. In order to assist the design and insure a proper view coverage, a field-of-view (FOV) simulator has been developed which takes into account the individual FOVs of the cameras, the position and orientation between them and the geometry and relative pose of the waterproof housing. The latter is especially relevant due to the strong image distortions caused by the refraction of the optical rays when travelling through the different media. Nonetheless, once the system is built, a very accurate calibration is required for any metrology or computer vision application. So, a full calibration method is presented for the estimation of both the intrinsic and extrinsic parameters of the cameras and the relative pose of the waterproof housing. This method is able to cope with wide-angle lenses and non-overlapping cameras simultaneously and applicable to both land or water Omnidirectional Multi-camera Systems (OMS).<\/p>\n <\/p>\n <\/p>\n Next, the topic of stitching strategies, to generate omnidirectional panoramas from the individual images, is studied in depth. Stitching strategies have the complex objective of joining the images in a way such that the viewer has the feeling the panoramas were captured from a single location. Conventional approaches either assume that the world is a simple sphere around the camera or use feature-based stitching techniques to align the individual images. However, this leads to artifacts and misalignments in the final panoramas due to parallax effects. This thesis presents a set of new stitching strategies, for both online and offline applications, aiming at processing the images according to available information of the multi-camera system and the environment.<\/p>\n <\/p>\n <\/p>\n Finally, we focus on potential underwater applications. We first explore the promising uses of omnidirectional cameras to create immersive virtual experiences. Then, we demonstrate the capabilities of omnidirectional cameras as complementary sensors for the navigation of underwater robots. Specifically, we present a new tracking system for autonomous underwater vehicles (AUVs) navigating in a close formation. The proposed system, which makes use of active light marker estimates the pose of a target vehicle at short ranges, with high accuracy and execution speed.<\/p>\n <\/p>\n <\/p>\n In order to validate all presented algorithms, two custom omnidirectional cameras were built and several experiments with divers and underwater robots have been carried out to collect the necessary data.<\/p>\n <\/p>\n <\/p>\n Durant l’\u00faltima d\u00e8cada s’ha despertat un inter\u00e8s creixent per a la visi\u00f3 omnidireccional en l’\u00e0mbit de la visi\u00f3 per computador. Podem trobar una gran varietat de c\u00e0meres omnidireccionals al mercat amb l’objectiu de cobrir l\u2019actual demanda d’imatges i v\u00eddeos panor\u00e0mics. Tot i aix\u00f2, l’\u00fas de c\u00e0meres omnidireccionals submarines encara \u00e9s molt limitat. En aquesta tesi doctoral hem creat un recurs de refer\u00e8ncia per al disseny, calibratge i \u00fas de c\u00e0meres omnidireccionals submarines de tipus multi-c\u00e0mera.<\/em><\/p>\n <\/p>\n <\/p>\n El primer problema que tractem \u00e9s el disseny i el calibratge de c\u00e0meres omnidireccionals per al medi submar\u00ed. D’entre els diferents sistemes capa\u00e7os de capturar imatges omnidireccionals, hem escollit els sistemes multi-c\u00e0mera per l’alta resoluci\u00f3 i qualitat de les imatges obtingudes. Hem dissenyat un simulador de camp de visi\u00f3 per tal d’ajudar en el disseny de sistemes multi-c\u00e0mera i assegurar que s\u00f3n capa\u00e7os de cobrir el camp de visi\u00f3 desitjat. Aquest simulador t\u00e9 en compte les caracter\u00edstiques individuals de les c\u00e0meres, la posici\u00f3 i orientaci\u00f3 entre elles i la posici\u00f3 relativa amb la carcassa submarina. \u00c9s especialment rellevant tenir en compte aquesta \u00faltima degut als importants canvis de direcci\u00f3 dels rajos de llum quan travessen diferents medis. Una vegada el sistema ha estat dissenyat i constru\u00eft, \u00e9s necessari un calibratge molt prec\u00eds per utilitzar-lo en qualsevol aplicaci\u00f3 de metrologia o de visi\u00f3 per computador. Per a resoldre aquest problema, presentem un sistema complet de calibratge que estima els par\u00e0metres intr\u00ednsecs i extr\u00ednsecs de les c\u00e0meres, aix\u00ed com la posici\u00f3 relativa de la carcassa. Aquest m\u00e8tode \u00e9s capa\u00e7 de tractar amb lents de tipus ull-de-peix o gran angular aix\u00ed com sistemes de c\u00e0meres sense solapament i \u00e9s aplicable tant a c\u00e0meres submarines com terrestres.<\/em><\/p>\n <\/p>\n <\/p>\n Despr\u00e9s ens centrem en l’estudi en profunditat de les estrat\u00e8gies de combinaci\u00f3 d’imatges per tal de generar imatges panor\u00e0miques partint de m\u00faltiples imatges capturades per diferents c\u00e0meres. Aquestes estrat\u00e8gies tenen com a objectiu la complexa tasca de crear una imatge panor\u00e0mica tal que l’observador tingui la sensaci\u00f3 que la imatge ha sigut creada des d’un \u00fanic punt de vista. Les estrat\u00e8gies convencionals parteixen de l’assumpci\u00f3 que el m\u00f3n pot ser simplificat com una esfera centrada en el sistema de c\u00e0meres o fan servir t\u00e8cniques de combinaci\u00f3 basades en punts d’especial inter\u00e8s per tal d’alinear correctament les imatges. Malauradament, aquestes estrat\u00e8gies resulten en panor\u00e0miques finals amb alineaments dolents entre imatges i defectes visuals degut als efectes de parallax. Aquesta tesi presenta un conjunt d’estrat\u00e8gies noves per a \u00fas en temps real o post-processat amb l’objectiu de millorar les panor\u00e0miques finals fent \u00fas de la informaci\u00f3 disponible de l’entorn i del sistema multi-c\u00e0mera.<\/em><\/p>\n <\/p>\n <\/p>\n Finalment, ens centrem en les aplicacions potencials de les c\u00e0meres omnidireccionals submarines. Primer, explorem els prometedors usos de les c\u00e0meres per tal de crear experi\u00e8ncies virtuals immersives. Despr\u00e9s, demostrem les capacitats de les c\u00e0meres omnidireccionals com a sensors complementaris per a la navegaci\u00f3 de robots. Concretament, presentem un sistema de seguiment de robots submarins aut\u00f2noms (AUVs) que naveguin en formacions de dist\u00e0ncies curtes. El sistema proposat fa \u00fas de marcadors de llum actius per tal d’estimar la posici\u00f3 relativa del vehicle amb una alta precisi\u00f3 i velocitat d’execuci\u00f3.<\/em><\/p>\n <\/p>\n <\/p>\n
\n<\/strong><\/span><\/p>\n
\n<\/strong>[\/vc_column_text][vc_separator type=”normal”][\/vc_column][\/vc_row][vc_row row_type=”row” text_align=”left” css_animation=””][vc_column][vc_empty_space][vc_column_text]<\/p>\nAbstract<\/strong><\/h3>\n
Resum<\/strong><\/em><\/h3>\n