Next Article in Journal
Transparent Fingerprint Sensor System for Large Flat Panel Display
Next Article in Special Issue
A Circular Microstrip Antenna Sensor for Direction Sensitive Strain Evaluation
Previous Article in Journal
Wide-Field Fluorescence Microscopy of Real-Time Bioconjugation Sensing
Article Menu
Issue 1 (January) cover image

Export Article

Open AccessArticle
Sensors 2018, 18(1), 292; https://doi.org/10.3390/s18010292

Multi-Sensor Data Integration Using Deep Learning for Characterization of Defects in Steel Elements

Department of Electrical and Computer Engineering, Faculty of Electrical Engineering, West Pomeranian University of Technology, Szczecin, al. Piastow 17, Szczecin 70-310, Poland
This paper is a continuation of the paper published in Psuj, G.; Biernacki, M.; Kruczynski, K. Application of deep learning procedure to magnetic multi-sensor matrix transducer data for the need of defect characterization in steel elements. In Proceedings of the 18th International Symposium on Electromagnetic Fields in Mechatronics, Electrical and Electronic Engineering (ISEF) Book of Abstracts, Lodz, Poland, 14–16 September 2017.
Received: 9 December 2017 / Revised: 16 January 2018 / Accepted: 16 January 2018 / Published: 19 January 2018
(This article belongs to the Special Issue Small Devices and the High-Tech Society)
View Full-Text   |   Download PDF [5634 KB, uploaded 8 February 2018]   |  

Abstract

Nowadays, there is a strong demand for inspection systems integrating both high sensitivity under various testing conditions and advanced processing allowing automatic identification of the examined object state and detection of threats. This paper presents the possibility of utilization of a magnetic multi-sensor matrix transducer for characterization of defected areas in steel elements and a deep learning based algorithm for integration of data and final identification of the object state. The transducer allows sensing of a magnetic vector in a single location in different directions. Thus, it enables detecting and characterizing any material changes that affect magnetic properties regardless of their orientation in reference to the scanning direction. To assess the general application capability of the system, steel elements with rectangular-shaped artificial defects were used. First, a database was constructed considering numerical and measurements results. A finite element method was used to run a simulation process and provide transducer signal patterns for different defect arrangements. Next, the algorithm integrating responses of the transducer collected in a single position was applied, and a convolutional neural network was used for implementation of the material state evaluation model. Then, validation of the obtained model was carried out. In this paper, the procedure for updating the evaluated local state, referring to the neighboring area results, is presented. Finally, the results and future perspective are discussed. View Full-Text
Keywords: magnetic nondestructive testing; matrix transducer; multi-sensor data integration; large data processing; data aggregation; deep learning; convolutional neural network magnetic nondestructive testing; matrix transducer; multi-sensor data integration; large data processing; data aggregation; deep learning; convolutional neural network
Figures

Figure 1

This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. (CC BY 4.0).

Supplementary material

Share & Cite This Article

MDPI and ACS Style

Psuj, G. Multi-Sensor Data Integration Using Deep Learning for Characterization of Defects in Steel Elements. Sensors 2018, 18, 292.

Show more citation formats Show less citations formats

Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Related Articles

Article Metrics

Article Access Statistics

1

Comments

[Return to top]
Sensors EISSN 1424-8220 Published by MDPI AG, Basel, Switzerland RSS E-Mail Table of Contents Alert
Back to Top