作者: John R. Banta , Geoffrey M. Hargreaves , Kendrick C. Taylor , Joseph R. McCONNELL , Richard B. Alley
DOI:
关键词:
摘要: This study tests novel methods for automatically identifying annual layers in a shallow Antarctic ice core (WDC05Q) using images that were collected with an optical scanner at the US National Ice Core Laboratory. A new method of optimized variance maximization (OVM) modeled density-related changes layer thickness directly from image variance. was done by multi-objective complex (MOCOM) parameter optimization to drive low-pass filtering scheme. The OVM-derived corresponded well results independent glaciochemical interpretation core. Individual cycles brightness then identified OVM apply depth-varying filter and fitting second-order polynomial locally detrended neighborhood. resulting map agreed within 1% overall count interpretation. Agreement on presence specific features 96%. It also shown MOCOM could calibrate image-based match date volcanic marker.