FishEye-RR: unwrapping and object recognition in omnidirectional images

• Grant number: 11/18821-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: December 01, 2011
• End date: July 31, 2012
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Anna Helena Reali Costa
• Grantee: Camila de Andrade Scatolini
• Host Institution: Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil

Abstract

The use of omnidirectional images to extract attributes and form object-based maps is an important topic in mobile robotics and computer vision since they allow the capture of the scene around the observer with a wide field of view. However, this advantage comes with the difficulty in processing distorted images. The study of how best to rectify the image obtained and thus be able, for example, to get a faster or more effective method for real-time applications or even apply existing classifiers that showed good results with conventional images is necessary as the development of interactive robots represent a clear challenge in future research. The objective of this study is to define the model of the wide-angle lens (popularly known as fish-eye) existing at the "Laboratório de Técnicas Inteligentes" (LTI) to use it in the rectification of the images obtained in an autonomous navigation system and integrate them to an effective object classifier, already tested with conventional images. To this end, it is necessary to calibrate the camera and study methods of correction of images formed from wide-angle lenses, and study the classifier that will be integrated.(AU)