Robots are expected to operate autonomously in dynamic environments.
Understanding the underlying dynamic characteristics of objects is a key
enabler for achieving this goal. In this paper, we propose a method for
pointwise semantic classification of 3D LiDAR data into three classes:
non-movable, movable and dynamic. We concentrate on understanding these
specific semantics because they characterize important information required for
an autonomous system. Non-movable points in the scene belong to unchanging
segments of the environment, whereas the remaining classes corresponds to the
changing parts of the scene. The difference between the movable and dynamic
class is their motion state. The dynamic points can be perceived as moving,
whereas movable objects can move, but are perceived as static. To learn the
distinction between movable and non-movable points in the environment, we
introduce an approach based on deep neural network and for detecting the
dynamic points, we estimate pointwise motion. We propose a Bayes filter
framework for combining the learned semantic cues with the motion cues to infer
the required semantic classification. In extensive experiments, we compare our
approach with other methods on a standard benchmark dataset and report
competitive results in comparison to the existing state-of-the-art.
Furthermore, we show an improvement in the classification of points by
combining the semantic cues retrieved from the neural network with the motion
cues.

本文提出了一种基于深度神经网络和贝叶斯滤波框架的方法，用于将三维 LiDAR 数据点语义分为三个类别：不可移动、可移动和动态。结果表明，本文方法与其他方法相比在标准基准数据集上具有竞争力，并通过将神经网络检测的语义线索与运动线索结合来提高点的分类精度。