A Monte Carlo Approach for Animated Radiosity Environments

Realistic animation has been always a desired goal in Computer Graphics. The accurate computation of the illumination plays a very important role in the simulation of animated realism. Although many successful physically-based global illumination models, as Ray-tracing or Radiosity methods, have been developed, they are still time-consuming if they are used for the generation of sequences in complex scenarios involving complex light transport phenomena. In this disserta-tion algorithms for accelerating the animation computation process are developed and a new method that computes eeciently animated sequences in radiosity environments is proposed. An important feature of this method, called the Multi-Frame Lighting Method, is that computes all frames of an animated sequence in a single process while the same accuracy as if the frames were computed independently is obtained. Results show a signiicant high speed-up that generally improves previous methods. Algorithms for generating realistic motion blur eeect are included in the method. Our approach is completely based on Monte Carlo techniques, resulting in a good alternative to deterministic approaches for computing accurate realistic animations for moderate complex scenarios. ii iii Acknowledgments There are many people I would like to thank for their help in this work. First of all, to my advisor Xavier Pueyo for his support ooered in all the phases of this work; also thanks to Leo Pini Magalh~ aes for his collaboration in the beginning of this PhD and to my tutor Isabel Navazo for her important comments on the outline of the dissertation. Many thanks to Mateu Sbert for his collaboration in early stages of this work and also for being always interested in discussing results. To all members of the onu-ooce room: Frederic P erez, Ignacio Mart n, Roel Mart nez, Gustavo Patow and Roberto Jim enez for sharing lots of hours of work in a comfortable environment and also for their help with any kind of problems (some of them already thanked with beers). I would also like to thank Albert Mas for using and detecting many bugs in my software and to the rest of the people of the Girona Graphics Group for their support. Many thanks to my wife Susan for her encouragement and for being so patient in all these years, also thanks for correcting my English. I wish to dedecate this work to you and to the rest of my family, Enrique (in memory), Pola, Mariela and Carolina.