Land-use classification via ensemble dropout information discriminative extreme learning machine based on deep convolution feature

Tianle Zhang1, Muzhou Hou1, Tao Zhou2, Zhaode Liu2, Weirong Cheng3 and Yangjin Cheng4

  1. School of Mathematics and Statistics, Central South University
    410083 Changsha, China
    houmuzhou@sina.com, csuztl@csu.edu.cn
  2. School of Economics, Guangdong University of Finance and Economics
    510320, Guangzhou, China
    zhoutaoscut@hotmail.com, lzhaode@163.com
  3. School of Mathematics and Science, Peking University
    100871 Beijing, China
    chenweirong@pku.edu.cn
  4. School of Mathematics and Computational ScienceXiangtan University
    411105, Xiangtan, Hunan, China
    yjcheng@xtu.edu.cn

Abstract

Classifying land-use scenes with high quality and accuracy is an important research direction in current hyperspectral remote sensing images, which is conducive to scientific management and utilization of land. An effective classifier and feature extractor can improve classification stability and accuracy. Therefore, based on deep learning technique, a dropout-based ensemble learning method is proposed in this paper, which combines convolutional neural network (CNN) and information discriminating extreme learning machine (IELM). Pre-trained CNN is used to learn effective and robust features, and deep convolution features are fed to the IELM classifier. Then the adoption of dropout technique and ensemble method can improve generalization capabilities and stability. The effectiveness of the proposed algorithm is tested by hyperspectral remote sensing image classification experiments. The experimental results show that the proposed E-CNN-dropIELM has achieved satisfactory results compared to state-of-the-art methods in terms of classification accuracy and stability.

Key words

Extreme learning machine (ELM), Convolutional neural network (CN-N), Dropout, Ensemble, Land-use classification

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS191222010Z

Publication information

Volume 17, Issue 2 (June 2020)
Year of Publication: 2020
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium

Full text

DownloadAvailable in PDF
Portable Document Format

How to cite

Zhang, T., Hou, M., Zhou, T., Liu, Z., Cheng, W., Cheng, Y.: Land-use classification via ensemble dropout information discriminative extreme learning machine based on deep convolution feature. Computer Science and Information Systems, Vol. 17, No. 2, 427–443. (2020), https://doi.org/10.2298/CSIS191222010Z