抄録
This paper proposes a kind of feature extraction based efficient registration algorithm including coarse and fine registrations. During coarse registration, feature points are detected, and feature point links are set up according to the relationship establishment among all detected feature points. Iterative Closest Point (ICP) is used in fine registration to align view pairs after the coarse registration. Experimental results with different real 3D images taken by laser scanner are carried out to compare the convergence and registration error among proposed approaches with classical ICP. The proposed registration approach has high convergence than classical ICP, and can overcome the problems of traditional ICP in low overlapping and bad initial estimate.
