An Airborne Gravity Gradient Compensation Method Based on Convolutional and Long Short-Term Memory Neural Networks.

Area of Science:

• Geophysics
• Earth Science
• Aerospace Engineering

Background:

• Airborne gravity gradient measurement is crucial for resource exploration but is sensitive to aircraft motion.
• Dynamic environmental factors significantly impact measurement accuracy, necessitating advanced compensation techniques.

Purpose of the Study:

• To develop and evaluate a post-error compensation algorithm for airborne gravity gradient measurements.
• To enhance the accuracy and reliability of gravity detection data collected from dynamic platforms.

Main Methods:

• Proposed a novel compensation algorithm utilizing Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) neural networks (CNN-LSTMs).
• Leveraged CNNs for feature extraction and LSTMs for capturing temporal dependencies in time-series data.
• Applied the algorithm to both simulated and measured airborne gravity gradient data.

Main Results:

• The CNN-LSTM model demonstrated superior performance in learning from coupled time-series data compared to traditional Multi-Layer Perceptrons (MLPs).
• Achieved significant improvements in compensation accuracy for airborne gravity gradient measurements.
• Validated the effectiveness of the proposed method on real-world data from Heilongjiang Province.

Conclusions:

• The CNN-LSTM algorithm offers a robust solution for mitigating motion-induced errors in airborne gravity gradient surveys.
• This advanced compensation technique enhances the precision of geophysical exploration, particularly in resource and hydrocarbon detection.
• The study highlights the potential of deep learning for improving the accuracy of dynamic geophysical measurements.