Remote Sensing Applications and Decision Support

Object-oriented fusion of RADARSAT-2 polarimetric synthetic aperture radar and HJ-1A multispectral data for land-cover classification

[+] Author Affiliations
Yan Xiao, Qigang Jiang, Yuanhua Li, Shu Liu

Jilin University, College of GeoExploration Science and Technology, Changchun 130000, China

Bin Wang

Changchun Institute of Surveying and Mapping, Geographical Information Department, Changchun 130021, China

Can Cui

Dalian Maritime University, Navigation College, Dalian 116000, China

J. Appl. Remote Sens. 10(2), 026021 (May 18, 2016). doi:10.1117/1.JRS.10.026021
History: Received September 24, 2015; Accepted April 22, 2016
Text Size: A A A

Abstract.  The contribution of the integration of optical and polarimetric synthetic aperture radar (PolSAR) data to accurate land-cover classification was investigated. For this purpose, an object-oriented classification methodology that consisted of polarimetric decomposition, hybrid feature selection, and a support vector machine (SVM) was proposed. A RADARSAT-2 Fine Quad-Pol image and an HJ-1A CCD2 multispectral image were used as data sources. First, polarimetric decomposition was implemented for the RADARSAT-2 image. Sixty-one polarimetric parameters were extracted using different polarimetric decomposition methods and then merged with the main diagonal elements (T11, T22, T33) of the coherency matrix to form a multichannel image with 64 layers. Second, the HJ-1A and the multichannel images were divided into numerous image objects by implementing multiresolution segmentation. Third, 1104 features were extracted from the HJ-1A and the multichannel images for each image object. Fourth, the hybrid feature selection method that combined the ReliefF filter approach and the genetic algorithm (GA) wrapper approach (ReliefF–GA) was used. Finally, land-cover classification was performed by an SVM classifier on the basis of the selected features. Five other classification methodologies were conducted for comparison to verify the contribution of optical and PolSAR data integration and to test the superiority of the proposed object-oriented classification methodology. Comparison results show that HJ-1A data, RADARSAT-2 data, polarimetric decomposition, ReliefF–GA, and SVM have a significant contribution by improving land-cover classification accuracy.

Figures in this Article
© 2016 Society of Photo-Optical Instrumentation Engineers

Citation

Yan Xiao ; Qigang Jiang ; Bin Wang ; Yuanhua Li ; Shu Liu, et al.
"Object-oriented fusion of RADARSAT-2 polarimetric synthetic aperture radar and HJ-1A multispectral data for land-cover classification", J. Appl. Remote Sens. 10(2), 026021 (May 18, 2016). ; http://dx.doi.org/10.1117/1.JRS.10.026021


Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.