Object-based spatial feature for classification of very high resolution remote sensing images

Penglin Zhang, Zhiyong Lv, Wen Zhong Shi

Research output: Journal article publicationJournal articleAcademic researchpeer-review

28 Citations (Scopus)

Abstract

This letter presents a novel spatial feature called object correlative index (OCI) to enhance the classification of very high resolution images. This novel method considers the property of an image object based on spectral similarity to construct a useful OCI to describe the spatial information objectively. Compared with the generic features widely used in image classification, the classification approach based on the OCI spatial feature results in higher classification accuracy than those approaches that only consider spectral features or pixelwise spatial features, such as the pixel shape index and mathematical morphology profiles. Experiments are conducted on QuickBird satellite image and aerial photo data, and results confirm that the proposed method is feasible and effective.
Original languageEnglish
Article number6573351
Pages (from-to)1572-1576
Number of pages5
JournalIEEE Geoscience and Remote Sensing Letters
Volume10
Issue number6
DOIs
Publication statusPublished - 8 Aug 2013

Keywords

  • Classification of very high resolution (VHR) image
  • object correlative index (OCI)
  • spatial feature
  • spectral feature

ASJC Scopus subject areas

  • Geotechnical Engineering and Engineering Geology
  • Electrical and Electronic Engineering

Cite this