Acta Geodaetica et Cartographica Sinica ›› 2015, Vol. 44 ›› Issue (7): 753-760.doi: 10.11947/j.AGCS.2015.20130810

Previous Articles     Next Articles

Edge Detection of PolSAR Image Based on Stochastic Distance

WANG Qing, ZENG Qiming, ZHANG Haizhen, JIAO Jian   

  1. Key Laboratory of Spatial Information Integration and Applications of Beijing Municipality, Institute of Remote Sensing and GIS, Peking University, Beijing 100871, China
  • Received:2013-12-30 Revised:2014-10-14 Published:2015-07-28
  • Supported by:

    The National High-tech Research and Development Program of China (863 Program) (No. 2012AA121304);The National Natural Science Foundation of China (No. 41171267)

Abstract:

A new edge detection methodology in PolSAR images is proposed, which is based on stochastic distance in the statistical theory and combined with complex Wishart distribution. Its main principle is inspired from the phenomenon that stochastic distance of two classes separated by an edge is closely related to the edge direction and the contrast of two classes. Simulation experiments are carried out to analyze the performance of the proposed methods. Results prove that methods have better capabilities in edge orientation and edge positioning than common used methods. Then, the proposed detection procedure is tested on a simulated PolSAR image with the complex Wishart distribution, as well as an airborne fully polarimetric SAR image.

Key words: stochastic distance, Wishart distribution, PolSAR, edge detection

CLC Number: