Ground Penetrating Radar

Introduction and Goals. System Performance and Current Challenges. GPR Data and Physics from a Signal Processing Perspective. Programmatic Lessons Learned. Pre-Processing: Approaches Tried, Lessons Learned. Pre-Screening: Approaches Tried, Lessons Learned. Inversion and Modeling Approaches (Difference vs. Feature Extraction – End Result is an Image (Inversion) or a Feature Vector (Feature Extraction)). Feature Extraction (Challenges: Time-of-Arrival, Soil Parameter Change Rapidly, Targets Not Hyperbolic, Clutter is Often Hyperbolic, Etc.). Feature Classification (Lots of Supervised Learning Approaches – RVM, SVM, Random Forest, PLSDA, Neural Networks, Log-Linear Models; Problems When Applying Them to GPR Data). Sensor Fusion. Data Visualization and Result Reporting. Open-Source MATLAB Code: GPRTools. Promising Avenues for Future Work. Conclusions and Acknowledgements.