Title: A coarse-to-fine object recognition method under occlusion for microassembly

Authors: Jian-Ying Zhu, Hua-Ming Wang

Addresses: College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu 210016, PRC. ' College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu 210016, PRC

Abstract: A two-stage object recognition algorithm with the presence of occlusion is presented for microassembly. Coarse localisation determines whether template is in image or not and where it is approximately, and fine localisation gives its accurate position. In coarse localisation, local feature, which is invariant to translation, rotation and occlusion, is used to form signatures. By comparing signature of template with that of image, approximate transformation parameter from template to image is obtained, which is used as initial parameter value for fine localisation. An objective function, which is a function of transformation parameter, is constructed in fine localisation and minimised to realise sub-pixel localisation accuracy. The occluded pixels are not taken into account in objective function, so the localisation accuracy will not be influenced by the occlusion.

Keywords: object recognition; local features; sub-pixel localisation accuracy; objective function; occlusion; microassembly; micromanufacturing.

DOI: 10.1504/IJNM.2006.011380

International Journal of Nanomanufacturing, 2006 Vol.1 No.1, pp.62 - 73

Published online: 29 Nov 2006 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article