Title :
3-D object location determination using line-segments matching
Author :
Shao, Lejun ; Volz, Richard A. ; Walker, Michael W.
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Michigan Univ., Ann Arbor, MI, USA
Abstract :
A real-time high-accuracy object localization system is presented. This system uses a line range sensor to acquire range data from which it extracts line features of an object such as boundary edges for planar surfaces and axes for surfaces of revolution. A combination of Hough transform, filtering, and least-squares optimization techniques is used to guarantee high-precision feature extraction. Line-line matching is then used to compute the transformation parameters. When multiline features are known, a fast optimal algorithm utilizing dual number quaternions is used to carry out the computations
Keywords :
computer vision; computerised picture processing; filtering and prediction theory; least squares approximations; transforms; 3-D object location determination; Hough transform; axes; boundary edges; dual number quaternions; filtering; least-squares optimization; line features extraction; line range sensor; line-line matching; line-segments matching; multiline features; optimal algorithm; planar surfaces; range data acquisition; real-time high-accuracy object localization system; surfaces of revolution; transformation parameters; Cameras; Computer science; Data mining; Feature extraction; Iterative algorithms; Real time systems; Robotic assembly; Sensor phenomena and characterization; Sensor systems; Time measurement;
Conference_Titel :
Industrial Applications of Machine Intelligence and Vision, 1989., International Workshop on
Conference_Location :
Tokyo
DOI :
10.1109/MIV.1989.40540
