GVMD-NLM: A Hybrid Denoising Method for GNSS Buoy Elevation Time Series Using Optimized VMD and Non-Local Means...

Area of Science:

• Geodesy and Geomatics
• Signal Processing
• Coastal Engineering

Background:

• Global Navigation Satellite System (GNSS) buoys are crucial for real-time elevation monitoring in coastal areas.
• Vertical coordinate time series from GNSS buoys often suffer from complex, non-stationary noise.
• Existing denoising techniques may lack reliability due to empirical parameter settings.

Purpose of the Study:

• To develop an automated and reliable denoising framework for GNSS buoy data.
• To enhance the accuracy of real-time elevation monitoring in coastal waterways.
• To address the limitations of existing denoising methods in handling complex noise.

Main Methods:

• Proposed GVMD-NLM: a hybrid framework combining Variational Mode Decomposition (VMD) optimized by an improved Grey Wolf Optimizer (GWO) with Non-Local Means (NLM) filtering.
• Introduced an adaptive convergence factor decay function (Sigmoid-based) for automatic VMD parameter optimization (K and α).
• Utilized Sample Entropy (SE) to identify effective low-frequency signals and NLM to process high-frequency noise.

Main Results:

• GVMD-NLM demonstrated superior performance over SSA, CEEMDAN, VMD, and GWO-VMD on two datasets from Dongguan Waterway Wharf.
• Achieved significant Root Mean Square Error (RMSE) reductions (up to 28.87%) and Signal-to-Noise Ratio (SNR) improvements (up to 11.13%).
• Denoised signals maintained high fidelity, with correlation coefficients (R) reaching 0.9798.

Conclusions:

• The GVMD-NLM framework offers an objective and automated solution for GNSS data denoising.
• This method provides a more accurate data foundation for waterway hydrodynamics research and water level monitoring.
• The adaptive parameter optimization enhances the reliability and applicability of VMD in noisy time-series data analysis.