Detection of magnetic UXO anomalies in noisy environments
Successful application of the magnetic method to unexploded ordnance (UXO) clearance in large areas depends on the ability to automatically detect targets by picking the anomalies that are caused by UXO-like objects. Efficient and reliable first-order detection of potential anomalies lays the foundation for achieving accurate and cost-effective discrimination in subsequent analysis. Towards this goal, we have developed an approach to automatic anomaly detection using the concept of structural index (SI), in which any anomaly having an SI close to 3 is identified as a potential target. To fully develop it for application in geologically noisy environments, we examine different methods for pre-processing data and post-processing detection results to minimize the influence of geology and other magnetic noise. We have examined the use of iterative Wiener and wavelet filtering techniques to extract the residual UXO anomaly. We have also developed an amplitude analysis technique based on approximate source strengths to winnow false alarms statistically. In this presentation, we introduce the basics of the detection algorithm, discuss the effect of pre-processing, and illustrate the importance of post-processing when carrying out anomaly detection in noisy environments.
UXO anomaly detection signal separation amplitude analysis
Yaoguo Li Kristofer Davis Richard Krahenbuhl Todd Meglich
Center for Gravity, Electrical, and Magnetic Studies, Colorado School of Mines
国际会议
成都
英文
734-740
2010-06-14（万方平台首次上网日期，不代表论文的发表时间）