The Algorithm of Pre-Processing a Pulsed Neutron Gamma Ray Logging Signal Taking into Account the Statistical Dependence of the Neighboring Spectrum Elements

Among many geophysical research methods, pulsed neutron gamma ray logging stands out and is one of the most promising, due to the possibility of determining the composition of rocks. However, there are difficulties in interpret ing the results of such research. The gamma ray signal are recorded in the process of pulsed neutron gamma ray logging. The gamma ray spectrums obtained after pre-processing of signal are noisy and have a high fluctuation, which affects the visual analysis of the research results. Visual analysis, on par with automatic, is important when working
with logging data. It helps to prevent the transfer of false results and the work of defective devices. This paper describes an algorithm that allows to reduce fluctuation considering the statistical dependence of neighboring elements of the spectrum to improve the visualization of research results. Since the signals spectrums are random variables, during the consideration of various methods of processing spectrums, it was decided to use smoothing functions, namely the B-spline. After applying different variations of B-spline parameters to gamma ray spectrums, it turned out that a quadratic homogeneous B-spline method with a knots interval – 7 showed best results in achieving the goal of improving visualization. Thus, has been developed algorithm of additional pre-processing that allows reducing fluctuation of gamma ray spectrums, for improve the visualization of the results of pulsed neutron gamma ray logging.

Keywords

logging, pulsed neutron gamma ray logging, processing of experimental signals, B-spline, variance reduction