DS Ph.D. Qualifier Presentation | Himan Namdari | Machine Learning-Enabled GPR Signal Feature Analysis for Root-Zone Soil Moisture Estimation
2:00 p.m.
DATA SCIENCE
Ph.D. Qualifier Presentation
Himan Namdari
Wednesday, December 13th | 2:00 PM EST
Location: Fuller Labs, Room 311
Zoom: https://wpi.zoom.us/my/hnamdari
Committee:
Advisor: Dr. Reza Zekavat, Physics and Data Science, WPI
Prof. Oren Mangoubi, Mathematical Sciences & Data Science, WPI
Prof. Fabricio Murai, Computer Science and Data Science, WPI
Title:
Machine Learning-Enabled GPR Signal Feature Analysis for Root-Zone Soil Moisture Estimation
Abstract:
In this work, we introduce a novel machine learning (ML) enabled framework for soil moisture estimation at different levels at root-zone using Ground Penetration Radar (GPR) data. The work offers a comprehensive and generalizable approach for soil subsurface characterization at any farm. Our framework comprises three main stages. First, we create a labeled dataset using gprMax simulations. Second, we implement advanced feature extraction techniques, including time-domain analysis, and a feature extraction method based on Peak Extraction. Finally, we demonstrate the application of our framework for soil moisture estimation through a comprehensive feature analysis. We show that our method offers significant improvements over existing techniques, providing a more accurate, efficient, and adaptable solution for soil moisture estimation for different soil compositions. The combination of a robust dataset, sophisticated feature extraction, and thorough analysis positions our framework as a leading approach in the field of GPR-based soil moisture assessment for megafarms.