Frequency-domain preprocessing and directional correlation-based feature extraction for classification of the buried objects using GPR B-scan data

Yildirim Bahadirlar; Gulay Buyukaksoy Kaplan

doi:10.1117/12.541995

21 September 2004 Frequency-domain preprocessing and directional correlation-based feature extraction for classification of the buried objects using GPR B-scan data

Yildirim Bahadirlar, Gulay Buyukaksoy Kaplan

Author Affiliations +

Proceedings Volume 5415, Detection and Remediation Technologies for Mines and Minelike Targets IX; (2004) https://doi.org/10.1117/12.541995
Event: Defense and Security, 2004, Orlando, Florida, United States

Abstract

A new preprocessing and feature extracting approach for classification of non-metallic buried objects are aimed using GPR B-scan data. A frequency-domain adaptive filter without a reference channel effectively removes the background signal resulting mostly from the discontinuity on the air-to-ground path of the electromagnetic waves. The filter only needs average of the first five A-scans as the reference signal for this elimination, and also serves for masking of the B-scan in the frequency-domain. A preprocessed GPR data with significantly suppressed clutter is then obtained by precisely positioning the Hanning window in the frequency-domain. A directional correlation function defined over a B-scan frame gives distinctive curves of buried objects. The main axis of directional correlation, on which the pivotal correlating pixels and short lines of pixels being correlated are considered, makes an angle to the scanning direction of the B-scan. This form of correlation is applied to the frame from the left-hand and the right-hand side and two over-plotted curves are obtained. Nine measures as features emphasizing directional signatures are extracted from these curves. Nine-element feature vectors are applied to the two-layer Artificial Neural Network and preliminary results over test set are promising to continue to comprehensive training and testing processes.

Citation Download Citation

Yildirim Bahadirlar and Gulay Buyukaksoy Kaplan "Frequency-domain preprocessing and directional correlation-based feature extraction for classification of the buried objects using GPR B-scan data", Proc. SPIE 5415, Detection and Remediation Technologies for Mines and Minelike Targets IX, (21 September 2004); https://doi.org/10.1117/12.541995

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