Feature extraction of partial discharge gray-scale images for identification based on multifractal spectrum using fluorescence fiber sensor

Ju Tang; Liang Huang

doi:10.1117/1.OE.53.5.053102

5 May 2014 Feature extraction of partial discharge gray-scale images for identification based on multifractal spectrum using fluorescence fiber sensor

Ju Tang, Liang Huang

Author Affiliations +

Optical Engineering, Vol. 53, Issue 5, 053102 (May 2014). https://doi.org/10.1117/1.OE.53.5.053102

Abstract

This document proposes the utilization of partial discharge (PD) optical signals from a fluorescence optical sensor system for constructing φ-u-n charts and gray-scale images. The fractal characteristics cannot fully characterize the PD gray-scale images, which reduce the PD pattern recognition rates. Multifractal spectrum is used for analyzing the characteristics of gray-scale images, and a new probability calculation method is proposed for computing the multifractal spectrogram of these images. By conducting a series of experiments, the multifractal spectrum is proven effective in describing the variations in the geometric characteristics of the gray-scale images. The main physical features of the multifractal spectrum are extracted and used as pattern recognition features. The backpropagation neural network with an improved conjugate gradient algorithm is used as a classifier in identifying the different PD types, which achieves recognition rates that are <87% . The multifractal spectrum improves the accuracy of the PD pattern recognition unlike other pattern recognition features, such as the box-counting dimension and the information dimension.

Citation Download Citation

Ju Tang and Liang Huang "Feature extraction of partial discharge gray-scale images for identification based on multifractal spectrum using fluorescence fiber sensor," Optical Engineering 53(5), 053102 (5 May 2014). https://doi.org/10.1117/1.OE.53.5.053102

Published: 5 May 2014

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available