Um descritor robusto e eficiente de pontos de interesse: desenvolvimento e aplicações

Abstract

At the core of a myriad of tasks such as object recognition, tridimensional recon-struction and alignment resides the critical problem of correspondence. Due to the ambiguity in our world and the presence of noise in the data aquisition process, per-forming high quality correspondence is one of themost challenging tasks in roboticsand computer vision. Hence, devising descriptors, which identify the entities to be matched and that are able to correctly and reliably establish pairs of corresponding points is of central importance. In this thesis, we introduce three novel descriptors that efficiently combine appearance and geometrical shape information from RGB-D images, and are largelyinvariant to rotation, illumination changes and scale transformations. For applica-tions that demand speed performance in lieu of a sophisticated and more precise matching process, scale and rotation invariance may be easily disabled. Results of several experiments described here demonstrate that as far as precision and robust-ness are concerned, our descriptors compare favorably to three standard descrip-tors in the literature, namely: SIFT, SURF (for textured images) and Spin-Images(for geometrical shape information). In addition, they outperfom the state-of-the-art CSHOT, which, as well as our descriptors, combines texture and geometry. We use these new descriptors to detect and recognize objects under different illumination conditions to provide semantic information in a mapping task. Fur-thermore, we apply our descriptors for registering multiple indoor textured depth maps, and demonstrate that they are robust and provide reliable results even forsparsely textured and poorly illuminated scenes. In these two pplications we com-pare the performance of our descriptors against the standard ones in the literature and the state-of-the-art. Experimental results show that our descriptors are supe-rior to the others in processing time, memory consumption, recognition rate and alignment quality.