Fast Localization and Characterization of Underground Targets with a Towed Transient Electromagnetic Array System

Area of Science:

• Geophysics
• Electromagnetism
• Signal Processing

Background:

• Traditional algorithms for detecting unexploded ordnance (UXO) using transient electromagnetic (TEM) systems are slow, requiring thousands of iterations.
• Efficient and accurate detection of underground UXO is crucial for safety and land reclamation.

Purpose of the Study:

• To develop a fast inversion algorithm for TEM detection systems to improve the efficiency of UXO detection.
• To accurately estimate the position and characteristics of underground targets.

Main Methods:

• A novel algorithm combining magnetic gradient tensor and singular value decomposition (SVD) was developed.
• Two modes of magnetic gradient tensor were constructed for shallow and deep targets.
• A towed TEM sensor with a 3x3 array of coils was designed and field-tested.

Main Results:

• The magnetic gradient tensor algorithm accurately locates single targets within 2.0 m depth with <8 cm error, and multi-target depths with <12 cm error.
• The SVD algorithm accurately characterizes underground targets, showing improved signal-to-noise ratio for depths >2.0 m compared to traditional methods.
• The proposed method requires approximately 40 ms, a 99% reduction in processing time compared to traditional algorithms.

Conclusions:

• The proposed magnetic gradient tensor and SVD algorithm offers a significant improvement in detection speed and accuracy for UXO.
• This method provides a strong foundation for real-time data processing in TEM detection systems.
• The developed towed TEM system effectively validates the performance of the new algorithm.