%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2017/09.04.22.41
%2 sid.inpe.br/sibgrapi/2017/09.04.22.41.48
%T D-KHT: Real time plane detection in depth images
%D 2017
%A Sousa, Eduardo Vera,
%A Velho, Luiz,
%A Fernandes, Leandro Augusto Frata,
%@affiliation Universidade Federal Fluminense
%@affiliation Instituto Nacional de Matemática Pura e Aplicada
%@affiliation Universidade Federal Fluminense
%E Torchelsen, Rafael Piccin,
%E Nascimento, Erickson Rangel do,
%E Panozzo, Daniele,
%E Liu, Zicheng,
%E Farias, Mylène,
%E Viera, Thales,
%E Sacht, Leonardo,
%E Ferreira, Nivan,
%E Comba, João Luiz Dihl,
%E Hirata, Nina,
%E Schiavon Porto, Marcelo,
%E Vital, Creto,
%E Pagot, Christian Azambuja,
%E Petronetto, Fabiano,
%E Clua, Esteban,
%E Cardeal, Flávio,
%B Conference on Graphics, Patterns and Images, 30 (SIBGRAPI)
%C Niterói, RJ, Brazil
%8 17-20 Oct. 2017
%I Sociedade Brasileira de Computação
%J Porto Alegre
%S Proceedings
%K depth images, pattern recognition, Hough transform, real time.
%X The automatic detection of geometric primitives (e.g. planes, spheres, etc.) in depth images provides the basis for solving many computer vision problems. The applications range from robotics to augmented reality. For plane detection, the quality of previous techniques is strongly related to the amount of noise and discontinuities or, in the case of unorganized point clouds, depends on complex structures to organize the points, besides having high computational cost. In this paper, we present a real-time deterministic algorithm for plane detection in depth images. By using an implicit quadtree to cluster approximately coplanar points in the 2.5-D space associated with an efficient Hough Transform voting scheme and a hill climbing strategy to find local maxima, we are able to reach real-time detection. Experiments show that our approach works well even in presence of noise, partial occlusion, and discontinuities.
%@language en
%3 CReady___WTD___SIBGRAPI_2017.pdf