GPR Antipersonnel Mine Detection: Improved Deconvolution and Time-Frequency Feature Extraction

Timofei Savelyev, Luc Van Kempen, Hichem Sahli

Onderzoeksoutput: Conference paper

Samenvatting

This work deals with the processing of GPR (ground penetrating radar) signals for AP (anti-personnel) mine detection. It focuses on two steps in this processing, namely the deconvolution of the system impulse response, and the extraction of target features for classification.
The received signal is a convolution between the impulse response of the target and that of the measurement system. The latter has to be deconvolved out of the data, in order to receive a clean target signal. The deconvolution is an ill-posed inverse problem, which can be solved with regularization methods. A new algorithm of deconvolution is proposed, based on the iterative nu-method [1] of regularization. Its main advantages are high accuracy, fast convergence and minimal ringing. The algorithm includes an adaptive noise level estimation from a noise deviation defined a priori.
After deconvolution and other pre-processing steps, target recognition, based on the time-frequency features of GPR signals, can be done. State of the art literature [2] shows the investigation of the singular values of the signal’s Wigner distribution (WD) as unique features. However, these features represent the energy of different signal components and as such are dependent on amplitude influencing factors as antenna height, type of soil or target depth. A new feature extraction method is proposed, fully using the energetic and spectral representations of the WD. In order to improve the algorithm’s robustness towards changing real life situations, the differences between the target’s features and those of a reference object, placed in the same conditions, are used as final features for classification.
The developed method has been tested on the data obtained with a ultra-wideband GPR, which has 170 picoseconds effective pulse duration and signal bandwidth 1-4 GHz.
Originele taal-2English
TitelSPIE proceedings Detection and Remediation Technologies for Mines and Minelike Targets VIII; San Diego, USA; April 21-26, 2003.
UitgeverijSPIE, Nondestructive Evaluation and Health Monitoring of Aerospace Materials and Composites II, Vol. 5046, pp. 232-241, San Diego, USA.
Pagina's232-241
Aantal pagina's10
Volume5046
StatusPublished - 21 apr 2003

Bibliografische nota

SPIE, Nondestructive Evaluation and Health Monitoring of Aerospace Materials and Composites II, Vol. 5046, pp. 232-241, San Diego, USA.

Vingerafdruk

Duik in de onderzoeksthema's van 'GPR Antipersonnel Mine Detection: Improved Deconvolution and Time-Frequency Feature Extraction'. Samen vormen ze een unieke vingerafdruk.

Citeer dit