Name
Abstract | ||
|---|---|---|
This paper presents a G(p)(0) mixture model-based ship detection method for polarimetric synthetic aperture radar (PolSAR) images. The proposed method is based on the assumption that target pixels can be regarded as a general class and that the PolSAR data contain abundant structure and textural information, which may help distinguish target from azimuth ambiguities and clutter. In the proposed method, a weighted combination of G(p)(0) distributions is used to characterize the PolSAR data to balance the complexity of parameter estimation and modeling accuracy. The proposed method is capable of automatically determining the number of class with an iterative expectation-maximization algorithm incorporating the G(p)(0) distribution. Besides, a prescreening process is integrated to realize computational acceleration. Instead of clustering all pixels in the PolSAR data, only potential target pixels selected in the prescreening stage are clustered. Therefore, fewer class is required to reach convergence due to the removal of most complex background. As a result, better computational efficiency can be achieved. After the clustering, the cluster corresponding to the targets can be distinguished conveniently with the averaged SPAN value of each cluster. The effectiveness and efficiency of the proposed method has been validated by using actual PolSAR datasets and by contrasting the proposed approach with othermethods. Experimental results demonstrate its superiority in improving target detection rate while reducing false alarms caused by clutter and azimuth ambiguities. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1109/JSTARS.2019.2912895 | IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING |
Keywords | DocType | Volume |
Expectation-maximization (EM) algorithm,G(p)(0) distribution,mixture models,polarimetric synthetic aperture radar (PolSAR),target detection | Journal | 12 |
Issue | ISSN | Citations |
SP6.0 | 1939-1404 | 0 |
PageRank | References | Authors |
0.34 | 0 | 5 |