Multi-task encoder with attention mechanism to remove noise from GPR signal

Dongying Li; Zhaolei Yu; Chongpeng Tian; Da Yuan

doi:10.1117/12.2673896

6 April 2023 Multi-task encoder with attention mechanism to remove noise from GPR signal

Dongying Li, Zhaolei Yu, Chongpeng Tian, Da Yuan

Proceedings Volume 12615, International Conference on Signal Processing and Communication Technology (SPCT 2022); 1261505 (2023) https://doi.org/10.1117/12.2673896
Event: International Conference on Signal Processing and Communication Technology (SPCT 2022), 2022, Harbin, China

Abstract

Ground penetrating radar (GPR) signals are non-stationary signals that are easily affected by interference and noise, resulting in reduced exploration capability. This paper proposes a multi-task encoder network model (ADAE) based on deep neural networks, which integrates channel and spatial attention mechanisms. According to ground penetrating radar noise characteristics, the channel and spatial attention mechanism are added to the encoding and decoding stage and the intermediate separation mapping stage, respectively. The composite attention mechanism of channel and space is used to optimize the feature extraction method and improve the characterization ability of features to suppress the influence of noise and irrelevant components from different angles. Experimental results show that the average boost signal to noise ratio of GPR data processed by multi-task encoder is 10.449dB. Furthermore, the results show that the ADAE multi-task encoder has apparent advantages in suppressing the noise of ground-penetrating radar and maintaining the signal amplitude effectively.

Citation Download Citation

Dongying Li, Zhaolei Yu, Chongpeng Tian, and Da Yuan "Multi-task encoder with attention mechanism to remove noise from GPR signal", Proc. SPIE 12615, International Conference on Signal Processing and Communication Technology (SPCT 2022), 1261505 (6 April 2023); https://doi.org/10.1117/12.2673896

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