A method for determining the effective order of NASVD denoising for the airborne gamma-ray spectrometry data
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
NASVD (Noise Adjusted Singular Value Decomposition) is a widely used and effective method for reducing noise in airborne gamma-ray spectrometry data. This paper proposes a new method, based on the principal component spectral correlation coefficient (PCC), for selecting the optimal number of low-order principal components in NASVD denoising. The PCC method involves accumulating principal components one by one in ascending order using the NASVD denoising algorithm to reconstruct spectra and calculate correlation coefficients between adjacent reconstructed spectra. By comparing these coefficients with the preset threshold 1, the effective order can be determined. Using the low-order principal components determined by the effective order for spectral reconstruction achieves noise reduction of gamma-ray spectrometry data. Finally, this paper verifies the validity and practicability of this method using real measured data. The results demonstrate that the proposed method can accurately and efficiently determine the optimal number of low-order principal components, while also mitigating the impact of human factors on denoising outcomes. This approach can be easily implemented through computer programming and holds significant practical value in reducing noise for gamma-ray spectrometry data.
Abstract
NASVD (Noise Adjusted Singular Value Decomposition) is a widely used and effective method for reducing noise in airborne gamma-ray spectrometry data. This paper proposes a new method, based on the principal component spectral correlation coefficient (PCC), for selecting the optimal number of low-order principal components in NASVD denoising. The PCC method involves accumulating principal components one by one in ascending order using the NASVD denoising algorithm to reconstruct spectra and calculate correlation coefficients between adjacent reconstructed spectra. By comparing these coefficients with the preset threshold 1, the effective order can be determined. Using the low-order principal components determined by the effective order for spectral reconstruction achieves noise reduction of gamma-ray spectrometry data. Finally, this paper verifies the validity and practicability of this method using real measured data. The results demonstrate that the proposed method can accurately and efficiently determine the optimal number of low-order principal components, while also mitigating the impact of human factors on denoising outcomes. This approach can be easily implemented through computer programming and holds significant practical value in reducing noise for gamma-ray spectrometry data.