A Land Cover Refined Classification Method Based on the Fusion of LiDAR Data and UAV Image

DOI

10.2991/cnci-19.2019.22How to use a DOI?

Keywords

Refined Classification, Airbrone LiDAR, DSM, SVM-DBN.

Abstract

Refined classification is one of the important research directions of remote sensing data classification. With the support of LiDAR data, the fusion of UAV image feature will be an effective mean to expand feature sets and improve classification category differentiation. After expanding the feature sets, there are high dimensional, massive data, small training samples and so on, which are important problems affecting the classification accuracy. Therefore, based on the DSM features obtained from LiDAR, this paper extracts the geometric features of the relevant space according to the neighborhood constraint rules, and proposes a SVM-DBN classification algorithm for small training samples after the fusion spectra, textures and other features. Using the small training samples classification of SVM, the algorithm predicts a large number of inexpensive unmarked samples on the basis of fewer training samples, uses the prediction results as prior-information training DBN and corrects the results of SVM prediction. The classification results show that the method has high overall precision and reliable results, which provides a new idea for the study of land cover refined classification.

Copyright

© 2019, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Lian Liu
AU  - Xing Liu
AU  - Junjiong Shi
AU  - Anran Li
PY  - 2019/05
DA  - 2019/05
TI  - A Land Cover Refined Classification Method Based on the Fusion of LiDAR Data and UAV Image
BT  - Proceedings of the 2019 International Conference on Computer, Network, Communication and Information Systems (CNCI 2019)
PB  - Atlantis Press
SP  - 154
EP  - 162
SN  - 2352-538X
UR  - https://doi.org/10.2991/cnci-19.2019.22
DO  - 10.2991/cnci-19.2019.22
ID  - Liu2019/05
ER  -