Eigen-feature analysis of weighted covariance matrices for LiDAR point cloud classification

Chao Hung Lin, Jyun Yuan Chen, Po Lin Su, Chung Hao Chen

研究成果: Article

34 引文 (Scopus)

摘要

The features used in the separation of different objects are important for successful point cloud classification. Eigen-features from a covariance matrix of a point set with the sample mean are commonly used geometric features that can describe the local geometric characteristics of a point cloud and indicate whether the local geometry is linear, planar, or spherical. However, eigen-features calculated by the principal component analysis of a covariance matrix are sensitive to LiDAR data with inherent noise and incomplete shapes because of the non-robust statistical analysis. To obtain reliable eigen-features from LiDAR data and to improve classification accuracy, we introduce a method of analyzing local geometric characteristics of a point cloud by using a weighted covariance matrix with a geometric median. Each point is assigned a weight to represent its spatial contribution in the weighted principal component analysis and to estimate the geometric median which can be regarded as a localized center of a shape. In the experiments, qualitative and quantitative analyses on airborne LiDAR data and simulated point clouds show a clear improvement of the proposed method compared with the standard eigen-features. The classification accuracy is improved by 1.6-4.5% using a supervised classifier.

原文English
頁(從 - 到)70-79
頁數10
期刊ISPRS Journal of Photogrammetry and Remote Sensing
94
DOIs
出版狀態Published - 2014 八月

指紋

cloud classification
covariance analysis
Covariance matrix
Principal component analysis
matrix
principal component analysis
principal components analysis
Statistical methods
statistical analysis
Classifiers
classifiers
geometry
Geometry
estimates
experiment
Experiments
method

All Science Journal Classification (ASJC) codes

  • Atomic and Molecular Physics, and Optics
  • Engineering (miscellaneous)
  • Computer Science Applications
  • Computers in Earth Sciences

引用此文

@article{cab0a3a29bc044d594f8a6e7c0e0161b,
title = "Eigen-feature analysis of weighted covariance matrices for LiDAR point cloud classification",
abstract = "The features used in the separation of different objects are important for successful point cloud classification. Eigen-features from a covariance matrix of a point set with the sample mean are commonly used geometric features that can describe the local geometric characteristics of a point cloud and indicate whether the local geometry is linear, planar, or spherical. However, eigen-features calculated by the principal component analysis of a covariance matrix are sensitive to LiDAR data with inherent noise and incomplete shapes because of the non-robust statistical analysis. To obtain reliable eigen-features from LiDAR data and to improve classification accuracy, we introduce a method of analyzing local geometric characteristics of a point cloud by using a weighted covariance matrix with a geometric median. Each point is assigned a weight to represent its spatial contribution in the weighted principal component analysis and to estimate the geometric median which can be regarded as a localized center of a shape. In the experiments, qualitative and quantitative analyses on airborne LiDAR data and simulated point clouds show a clear improvement of the proposed method compared with the standard eigen-features. The classification accuracy is improved by 1.6-4.5{\%} using a supervised classifier.",
author = "Lin, {Chao Hung} and Chen, {Jyun Yuan} and Su, {Po Lin} and Chen, {Chung Hao}",
year = "2014",
month = "8",
doi = "10.1016/j.isprsjprs.2014.04.016",
language = "English",
volume = "94",
pages = "70--79",
journal = "ISPRS Journal of Photogrammetry and Remote Sensing",
issn = "0924-2716",
publisher = "Elsevier",

}

Eigen-feature analysis of weighted covariance matrices for LiDAR point cloud classification. / Lin, Chao Hung; Chen, Jyun Yuan; Su, Po Lin; Chen, Chung Hao.

於: ISPRS Journal of Photogrammetry and Remote Sensing, 卷 94, 08.2014, p. 70-79.

研究成果: Article

TY - JOUR

T1 - Eigen-feature analysis of weighted covariance matrices for LiDAR point cloud classification

AU - Lin, Chao Hung

AU - Chen, Jyun Yuan

AU - Su, Po Lin

AU - Chen, Chung Hao

PY - 2014/8

Y1 - 2014/8

N2 - The features used in the separation of different objects are important for successful point cloud classification. Eigen-features from a covariance matrix of a point set with the sample mean are commonly used geometric features that can describe the local geometric characteristics of a point cloud and indicate whether the local geometry is linear, planar, or spherical. However, eigen-features calculated by the principal component analysis of a covariance matrix are sensitive to LiDAR data with inherent noise and incomplete shapes because of the non-robust statistical analysis. To obtain reliable eigen-features from LiDAR data and to improve classification accuracy, we introduce a method of analyzing local geometric characteristics of a point cloud by using a weighted covariance matrix with a geometric median. Each point is assigned a weight to represent its spatial contribution in the weighted principal component analysis and to estimate the geometric median which can be regarded as a localized center of a shape. In the experiments, qualitative and quantitative analyses on airborne LiDAR data and simulated point clouds show a clear improvement of the proposed method compared with the standard eigen-features. The classification accuracy is improved by 1.6-4.5% using a supervised classifier.

AB - The features used in the separation of different objects are important for successful point cloud classification. Eigen-features from a covariance matrix of a point set with the sample mean are commonly used geometric features that can describe the local geometric characteristics of a point cloud and indicate whether the local geometry is linear, planar, or spherical. However, eigen-features calculated by the principal component analysis of a covariance matrix are sensitive to LiDAR data with inherent noise and incomplete shapes because of the non-robust statistical analysis. To obtain reliable eigen-features from LiDAR data and to improve classification accuracy, we introduce a method of analyzing local geometric characteristics of a point cloud by using a weighted covariance matrix with a geometric median. Each point is assigned a weight to represent its spatial contribution in the weighted principal component analysis and to estimate the geometric median which can be regarded as a localized center of a shape. In the experiments, qualitative and quantitative analyses on airborne LiDAR data and simulated point clouds show a clear improvement of the proposed method compared with the standard eigen-features. The classification accuracy is improved by 1.6-4.5% using a supervised classifier.

UR - http://www.scopus.com/inward/record.url?scp=84900799935&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84900799935&partnerID=8YFLogxK

U2 - 10.1016/j.isprsjprs.2014.04.016

DO - 10.1016/j.isprsjprs.2014.04.016

M3 - Article

AN - SCOPUS:84900799935

VL - 94

SP - 70

EP - 79

JO - ISPRS Journal of Photogrammetry and Remote Sensing

JF - ISPRS Journal of Photogrammetry and Remote Sensing

SN - 0924-2716

ER -