Enhanced noise reduction for gamma-ray data
The random nature of radioactive decay ensures that noise levels are a significant limitation of the gamma-ray spectrometric method. In the absence of the application of any noise reduction techniques, the noise (fractional standard deviations) on processed airborne gamma-ray spectrometric data can be as high as 10%, 15% and 58% for the estimation of K, eTh and eU concentrations respectively. Fortunately, much of this noise can be removed using one of the spectral noise–reduction methods developed in the mid-1990’s.
There are two methods that are widely used – the NASVD method (Noise Adjusted Singular Value Decomposition) and the MNF method (Maximum Noise Fraction). For large surveys, the implementation of either of these methods can be improved by first sorting the raw spectra into clusters on the basis of similarity in spectral shape. The NASVD or MNF method is then applied to each cluster of spectra in turn. This typically further reduces the random noise by a factor of two. Accuracy can be improved by incorporating spatial information into the analysis. The reduced noise enables smaller radioelement anomalies to be resolved.
GAMMA_Plus™ is a proprietary gamma-ray processing methodology developed by Minty Geophysics for the processing of multichannel gamma-ray spectra for enhanced noise reduction. The method delivers improved noise reduction while maintaining the accuracy of the radiometric signal as follows:
Examples of the application of the method are shown in the figures below. These show a comparison of radioelement estimates derived using the conventional NASVD method (followed by the recommended IAEA processing procedure) and gridded using minimum curvature vs GAMMA_Plus™ followed by GAMMA_Grid™. GAMMA_Plus™ gives improved noise reduction while retaining the radiometric signal.
· a new clustering method is used to sort gamma-ray spectra into clusters based on
their spectral shapes. The new method incorporates spatial information and is optimised for the sorting of gamma-ray spectra;
(data courtesy Dept. of Industry and Resources W.A.)
· the NASVD method is applied to clusters, but also uses spatial information in the
· non-linear filtering is incorporated into the data reduction procedure at more than
one point in the processing chain.