Global registration of point clouds aims to find an optimal alignment of a sequence of 2D or 3D point sets.
It's a vital but challenging task in numerous geometric applications given the existence of noise and
sequential observations that are only partially-overlapped. Iterative fitting algorithm represented by
Iterative Closest Point(ICP) is the most widely adopted method to solve such registration problems,
yet it suffers from its reliance on a close initial distance between target point sets.
In this paper, we present a novel approach that takes advantage of current deep learning techniques for
unsupervised learning of global registration from a temporal sequence of point clouds. Our key novelty is
that we introduce a deep Spatio-Temporal REPresentation (STREP)
feature, which describes the geometric essence of both temporal and spatial relationship of the sequence of
point clouds acquired with sensors in an unknown environment. In contrast to the previous practice that
treats each time step (pair-wise registration) individually, our unsupervised model starts with optimizing
a sequence of latent STREP feature, which is then decoded to a temporally and spatially continuous sequence
of geometric transformations to globally align multiple point clouds. We have evaluated our proposed
approach over both simulated 2D and real 3D datasets and the experimental results demonstrate that our
method can beat other techniques by taking into account the temporal information in deep feature learning.
|
