Full text loading...
-
Classification of Arrival-time Picks for Microseismic Event Localization
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 79th EAGE Conference and Exhibition 2017, Jun 2017, Volume 2017, p.1 - 5
Abstract
In practice, passive surface microseismic data often has low SNR, so arrival times picked on single traces are unreliable. False picks that are far from the true arrival times are common which can severely degrade the accuracy of subsequent estimation of an event moveout curve which leads to a poor event location estimate. Our previous work utilized a RANdom SAmpling Consensus (RANSAC) based scheme to improve arrival time estimates from first-break picking. In this paper, we present an improved RANSAC curve fitting scheme to approximate any parametric moveout curve, e.g., hyperbolic, which can also efficiently classify the picks distorted by noise into true and false picks. The proposed method is a significant extension of last year’s method with respect to its accuracy, stability, and efficiency. 2-D arrays with hyperboloid moveout surfaces are now addressed in a robust manner. Results are validated with a large-scale dataset from passive earthquake monitoring.