Scheduled System Maintenance:
Some services will be unavailable Sunday, March 29th through Monday, March 30th. We apologize for the inconvenience.
By Topic

An unsupervised classification method for polarimetric SAR images with a projection approach

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

1 Author(s)
Muhtar Qong ; Dept. of Space & Earth Inf. Technol., Kyushu Tokai Univ., Kumamoto, Japan

Two unsupervised training algorithms are presented based on a projection method for polarimetric SAR images. For this, polarimetric features, including the entropy H, angle α, anisotropy A, the largest eigenvalue of the coherence matrix λ1 and coherence γ, are computed. Based on these features three noncorrelated and independent variates, which are the first three principle components are extracted using the principle component analysis. Then, a three-dimensional Cartesian coordinate is constructed using those three resultant variates. In order to reasonably define the initial cluster centers for the unsupervised classification, a projection method is proposed. As a result, a sophisticated two-dimensional feature space is obtained. Two initial cluster center determination algorithms, the histogram and histogram-quadrant methods, based on the above two-dimensional feature space are presented. Unsupervised classifications (with a minimum distance decision rule) are conducted based on the defined initial cluster centers. The classification results show that these approaches have high performance in convergence, speed, data volume and accuracy of training and classifying.

Published in:

Geoscience and Remote Sensing Symposium, 2003. IGARSS '03. Proceedings. 2003 IEEE International  (Volume:7 )

Date of Conference:

21-25 July 2003