Signal Processing and Analysis Techniques for Nuclear Quadrupole Resonance Spectroscopy

Description

Foreword Preface Contents Abbreviations 1 Introduction 1.1 Context of This Work 1.2 Purpose of This Work 1.3 Structure of This Work References 2 Methods and Equipment for Signal Acquisition and Analysis for the Detection of Prohibited Substances 2.1 Signal Detection and Estimation Theory 2.2 Techniques for Detecting Prohibited Substances 2.2.1 Trace Detection Techniques 2.2.2 Bulk Detection Techniques 2.3 Conclusions References 3 Nuclear Quadrupole Resonance Spectroscopy 3.1 The Principle of Nuclear Quadrupole Resonance 3.2 Signal Acquisition Methods Applied in Nuclear Quadrupole Resonance Spectroscopy 3.3 Equipment Used in Nuclear Quadrupole Resonance Spectroscopy 3.4 Conclusions Appendix NQR Data for Different Substances References 4 Signal Processing and Analysis Techniques Applied in Nuclear Quadrupole Resonance 4.1 Classification of the Signal Processing Techniques 4.2 Pre-Processing Techniques 4.2.1 Signal Averaging 4.2.2 Quadrature Detection 4.3 Post-Processing and Analysis Techniques 4.3.1 Single Channel Detection Techniques 4.3.2 Multi-channel Detection Techniques 4.4 Future Research Directions 4.5 Conclusions Appendix Signal Processing and Analysis Techniques Classification Signal Post-Processing and Analysis Techniques Development Timeline Development of the Post-Processing Techniques Distribution of the Post-Processing Techniques According to the Type of Detection [1] References 5 Modeling of Signals Used in Nuclear Quadrupole Resonance Spectroscopy 5.1 Free Induction Decay Models 5.2 Echo Train Models 5.3 Proposal of a New Echo Model 5.4 Noise Models 5.5 Conclusions References 6 Study of the NQR Signal Processing and Analysis Algorithms 6.1 Description of the Algorithms Chosen for the Analysis 6.2 Proposal of a Multi-criteria Detection Algorithm 6.3 Comparative Analysis of the Signal Processing and Signal Analysis Algorithms 6.3.1 Analysis Procedure and Performance Indicators 6.3.2 Implementation of Comparative Analysis 6.3.3 Comparative Analysis Results 6.4 Conclusions Appendix References 7 Analysis of Nuclear Quadrupole Resonance Response Signals 7.1 Description of the Analyzed Signals 7.2 Defining the Signal Features 7.3 Statistical Analysis of Data 7.4 Conclusions Appendix Signal Acquisitions in Different Scanning Scenarios Statistical Analysis of the Data Set References 8 Development of Signal Analysis Algorithms for NQR Detection 8.1 Selection of the Investigated Algorithms 8.2 Implementation of the Signal Analysis Algorithms 8.3 Training and Evaluation of the Signal Analysis Algorithms 8.4 Optimization of the Proposed Algorithm 8.5 Conclusions Appendix Source Code of the Functions Used to Generate the Deep Learning Models Learning Curves for the Developed Algorithms AlexNet Model Optimization Results References 9 Solutions to Improve NQR Detection 9.1 Generalization of the Optimized Algorithm Using Transfer Learning 9.2 Noise Reduction Using Autoencoders 9.3 Conclusions Appendix Source Code of the Functions Used to Generate the Autoencoder Models References 10 Implementation of a Signal Pre-processing, Processing and Analysis System for Nuclear Quadrupole Resonance 10.1 Hardware Implementation of the Detection System 10.2 Software Implementation of the Detection System 10.2.1 Embedded System Firmware 10.2.2 Embedded System Interface 10.2.3 Spincore Module Interface 10.2.4 Database Interface 10.2.5 Interface with External Applications 10.2.6 Application Core 10.2.7 Graphical User Interface 10.2.8 Implementation of the Detection Application 10.3 Evaluation of the Detection System 10.3.1 Evaluation of the Radio Frequency Switch 10.3.2 Detection Limits 10.3.3 Ferrite Core Improvement 10.3.4 Evaluation of the Signal-to-Noise Ratio 10.4 Conclusions References