An automatic detection method that uses wavelet analysis and fuzzy logic to characterize elements of Pi geomagnetic irregular pulsations
Masahiro SASAOKA
Abstract
To create an alternative to manual detection, an automatic detection method is developed that uses wavelet analysis and fuzzy logic to characterize elements of Pi geomagnetic irregular pulsations (onset time, period, and amplitude). The values obtained by our new method are compared with the records from manual detection using Memambetsu geomagnetic data for 2012. Although the results show very good agreement overall, the biases due to components of periods less than 10 s in the high-pass filtered data with which we detect Pi by eye might cause amplitude differences between manual and automatic Pi detection. Also, in the manual Pi detection records, the start time of the Pi series does not necessarily agree with the onset time of the Pi.
