Citation: Wang, X.; Lyu, H.; Mao, T.;
He, W.; Chen, Q. Point Cloud
Segmentation from iPhone-Based
LiDAR Sensors Using the Tensor
Feature. Appl. Sci. 2022, 12, 1817.
https://doi.org/10.3390/
app12041817
Academic Editor: Nunzio Cennamo
Received: 14 January 2022
Accepted: 7 February 2022
Published: 10 February 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Point Cloud Segmentation from iPhone-Based LiDAR Sensors
Using the Tensor Feature
Xuan Wang
1
, Haiyang Lyu
2
, Tianyi Mao
1,2
, Weiji He
1,
* and Qian Chen
1
1
Jiangsu Key Laboratory of Spectral Imaging & Intelligence Sense (SIIS), Nanjing University of Science and
Technology, Nanjing 210094, China; hengwan210984@njust.edu.cn (X.W.); maoty@njupt.edu.cn (T.M.);
chenqian@njust.edu.cn (Q.C.)
2
Smart Health Big Data Analysis and Location Services Engineering Research Center of Jiangsu Province,
Nanjing University of Posts and Telecommunications, Nanjing 210023, China; hlyu@njupt.edu.cn
* Correspondence: hewj@njust.edu.cn; Tel.: +86-025-85866638
Featured Application: Point cloud segmentation, geometric feature extraction.
Abstract:
With widely used LiDAR sensors included in consumer electronic devices, it is increasingly
convenient to acquire point cloud data, but it is also difficult to segment the point cloud data obtained
from these unprofessional LiDAR devices, due to their low accuracy and high noise. To address the
issue, a point cloud segmentation method using the tensor feature is proposed. The normal vectors
of the point cloud are computed based on initial tensor encoding, which are further encoded into the
tensor of each point. Using the tensor from a nearby point, the tensor of the center point is aggregated
in all dimensions from its neighborhood. Then, the tensor feature in the point is decomposed and
different dimensional shape features are detected, and the point cloud dataset is segmented based on
the clustering of the tensor feature. Using the point cloud dataset acquired from the iPhone-based
LiDAR sensor, experiments were conducted, and results show that both normal vectors and tensors
are computed, then the dataset is successfully segmented.
Keywords: point cloud segmentation; iPhone LiDAR sensor; tensor feature decomposition
1. Introduction
The point cloud is a universal spatial information acquisition format and plays an
important role in indoor and outdoor environment understanding [
1
]. In conventional
point processing methods, the point cloud is usually defined as a spatial location, or as
having textural information, and geometric features are computed from the unstructured
point cloud; then, the data are segmented into different geometric or semantic structures for
further processing, such as in object detection, classification, and scene understanding [
2
,
3
].
In recent years, with different kinds of
Li
ght
D
etection
A
nd
R
anging (LiDAR) sensors
included in consumer electronic devices, such as the iPhone or Kinect, it has become in-
creasingly convenient to acquire point cloud data. However, it also raised many challenges
for data processing, due to the low accuracy and high noise of the point cloud data col-
lected by these unprofessional LiDAR devices. To address these issues, some processing
techniques are applied [
4
–
8
], such as point cloud filter, denoising, normal computation, and
resampling. Then features are extracted from the point cloud, and the data are segmented.
To achieve this goal, many kinds of point cloud segmentation methods were proposed that
extract the feature in different ways, including deep learning-based approaches.
To deal with these kinds of problems in a universal framework, this work proposes
a tensor feature-based point cloud segmentation method, and the point cloud data in an
actual scene is obtained from the iPhone LiDAR sensor. The contributions are as follows:
(1) we conduct the theoretical derivation of the N-d tensor voting, which helps the high-
dimensional normal vector computation and structure decomposition, and design the
Appl. Sci. 2022, 12, 1817. https://doi.org/10.3390/app12041817 https://www.mdpi.com/journal/applsci
