DocumentCode :
2747645
Title :
Refined Research on MAP Filtering for SAR Image
Author :
Sun, Zengguo ; Han, Chongzhao
Author_Institution :
Inst. of Integrated Autom., Xi´´an Jiaotong Univ.
Volume :
2
fYear :
0
fDate :
0-0 0
Firstpage :
9552
Lastpage :
9556
Abstract :
General form of the MAP filter was deduced and two decisive factors of it were defined. Some MAP equations were obtained in different speckle distribution and prior distribution, and they were analyzed according to the distributing status of real roots between the local averaging value and the observed value. Based on the statistical classification of SAR image, the threshold expressions of MAP filters were developed. In order to completely reflect the advantage and disadvantage of MAP filter, dynamic relation of their de-speckling capability to noise intensity was suggested. De-speckling experiments on SAR image and their quantitative measures demonstrate that MAP filtering performance is influenced directly by speckle intensity and prior distribution. When prior distribution is chosen unreasonable, MAP filtering performance is degraded seriously
Keywords :
filtering theory; image classification; maximum likelihood estimation; radar imaging; statistical distributions; synthetic aperture radar; despeckling; image classification; maximum a posteriori filter; prior distribution; speckle distribution; speckle intensity; statistical classification; synthetic aperture radar image; threshold expression; Automation; Degradation; Differential equations; Electronic mail; Filtering; Filters; Rail to rail inputs; Speckle; Sun; Synthetic aperture radar; Maximum a Posteriori (MAP) filter; prior distribution; speckle; synthetic aperture radar (SAR) image;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress on
Conference_Location :
Dalian
Print_ISBN :
1-4244-0332-4
Type :
conf
DOI :
10.1109/WCICA.2006.1713853
Filename :
1713853
Link To Document :
بازگشت