Skip to main navigation Skip to search Skip to main content

Deep feature extraction and combination for remote sensing image classification based on pre-trained CNN models

  • School of Computer Science and Technology, Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Understanding a scene provided by Very High Resolution (VHR) satellite imagery has become a more and more challenging problem. In this paper, we propose a new method for scene classification based on different pre-trained Deep Features Learning Models (DFLMs). DFLMs are applied simultaneously to extract deep features from the VHR image scene, and then different basic operators are applied for features combination extracted with different pre-trained Convolutional Neural Networks (CNN) models. We conduct experiments on the public UC Merced benchmark dataset, which contains 21 different areal categories with sub-meter resolution. Experimental results demonstrate the effectiveness of the proposed method, as compared to several state-of-the-art methods.

Original languageEnglish
Title of host publicationNinth International Conference on Digital Image Processing, ICDIP 2017
EditorsXudong Jiang, Charles M. Falco
PublisherSPIE
ISBN (Electronic)9781510613041
DOIs
StatePublished - 2017
Event9th International Conference on Digital Image Processing, ICDIP 2017 - Hong Kong, China
Duration: 19 May 201722 May 2017

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume10420
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

Conference9th International Conference on Digital Image Processing, ICDIP 2017
Country/TerritoryChina
CityHong Kong
Period19/05/1722/05/17

Keywords

  • deep feature learning
  • feature combination
  • pre-trained CNN
  • scene classification

Fingerprint

Dive into the research topics of 'Deep feature extraction and combination for remote sensing image classification based on pre-trained CNN models'. Together they form a unique fingerprint.

Cite this