A fuzzy shell clustering approach to recognize hyperbolic signatures in subsurface radar images

被引：59

作者：

Delbò, S ^{[1
]}

Gamba, P

Roccato, D

机构：

[1] Univ Pavia, Dipartimento Elettr, I-27100 Pavia, Italy

[2] CSELT, Ctr Studi E, Turin, Italy

来源：

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING | 2000年 / 38卷 / 03期

关键词：

FCQS term; fuzzy clustering; ground penetrating radar (GPR); image interpretation;

D O I：

10.1109/36.843039

中图分类号：

P3 [地球物理学]; P59 [地球化学];

学科分类号：

0708 ; 070902 ;

摘要：

We propose a pattern recognition approach to analyze subsurface radar images and recognize the hyperbolic signatures produced by targets. After enhancing these signatures using a wavelet denoising procedure, pixels are grouped into significant hyperbolic shapes by fuzzy clustering. The approach also provides a validation measure for each recognized shape using so-called "shell thickness." Excellent experimental results justify the use of this algorithm for automatic interpretation of subsurface radar images.

引用

页码：1447 / 1451

页数：5

共 17 条

[1]

BEZDEK JC, 1981, PATTENR RECOGNITION

[2]

CAROSI S, IN PRESS

[3]

CHANG SG, 1997, P IEEE INT C IM PROC, V1, P604

[4] ELECTROMAGNETIC IMAGING OF UNDERGROUND TARGETS USING CONSTRAINED OPTIMIZATION [J].