High-Precision Sensor Tuning of Proton Precession Magnetometer by Combining Principal Component Analysis and Singular Value Decomposition

The sensor tuning of a proton precession magnetometer determines the signal-to-noise ratio (SNR) of induced free induction decay (FID) signal, while the quality of the FID influences the measurement accuracy of the geomagnetic field. Even though numerous methods have been proven to be effective in improving the tuning precision, how to efficiently decrease tuning error in a noisy environment is still a crucial challenge. To end this, a new tuning algorithm based on the combination of principal component analysis (PCA) and singular value decomposition (SVD), namely C-PCASVD, was presented in this paper. This novel algorithm aims to suppress the interference of the FID signal. The proposed C-PCASVD can obtain the dominant principal components of FID and noise, respectively. In addition, it is able to identify the corresponding singular values of interference, which could achieve an optimum trade-off between the denoised FID and noise reduction efficiency. The performance of the proposed method is tested on synthetic and field signals from different surveys. Good tuning frequency estimations are obtained at different SNRs. Through comparing the proposed C-PCASVD tuning algorithm with the state-of-the-art methods, including peak detection (PD), auto-correction and FFT (AC-FFT), and secondary tuning based on SVD (ST-SVD), the experimental results demonstrate that the C-PCASVD can significantly improve the tuning precision of proton precession magnetometer in a noisy environment.