A Novel LiDAR Echo Signal Denoising Method Based on the VMD-CPO-IWT Algorithm

Area of Science:

• Signal Processing
• Remote Sensing Technology
• Optimization Algorithms

Background:

• LiDAR echo signals are prone to noise, degrading detection quality and accuracy.
• Optimizing parameters for signal processing techniques like Variational Mode Decomposition (VMD) is challenging.
• Existing denoising methods struggle to effectively suppress diverse noise frequencies in LiDAR signals.

Purpose of the Study:

• To propose a joint denoising method (VMD-CPO-IWT) for improving LiDAR echo signal quality.
• To optimize VMD parameters using the Crested Porcupine Optimizer (CPO) for adaptive decomposition.
• To enhance noise suppression using Improved Wavelet Thresholding (IWT) for better signal fidelity.

Main Methods:

• Variational Mode Decomposition (VMD) for signal decomposition.
• Crested Porcupine Optimizer (CPO) for adaptive optimization of VMD parameters (k, α).
• Wasserstein distance for intrinsic mode function screening and Improved Wavelet Thresholding (IWT) for optimal threshold determination.

Main Results:

• The VMD-CPO-IWT method demonstrated superior performance over WT-db4, EMD-DFA, and VMD-WOA for both simulated and real LiDAR signals.
• Significant improvements in Signal-to-Noise Ratio (SNR) and Root Mean Square Error (RMSE) were achieved.
• For a 25m detection range, SNR improved by 13.64 dB and RMSE reduced by 62.6% for actual LiDAR echo signals.

Conclusions:

• The VMD-CPO-IWT method offers an efficient and practical solution for denoising LiDAR echo signals.
• The parameter-adaptive CPO effectively addresses VMD parameter selection challenges.
• The combined approach provides robust noise suppression across low and high frequencies, enhancing LiDAR data reliability.