Geological noise removal in geophysical magnetic survey to detect unexploded ordnance based on image filtering

This paper describes the application of three straightforward image-based filtering methods to remove the geological noise effect which masks unexploded ordnances (UXOs) magnetic signals in geophysical surveys. Three image filters comprising of mean, median and Wiener are used to enhance the location of probable UXOs when they are embedded in a dominant background geological noise. The study area consists of three buried UXOs while a geological dyke structure covers the magnetic anomaly of the desired objects. To provide a better representation of the actual locations of UXOs in the observed magnetic anomaly over this area, all image-based filters could appropriately separate the geological dyke effect from the UXOs. These image filters can be good candidates to remove the geological noise effect in UXO detection when encountering a mixed response of multi-source magnetic anomaly in contaminated territories with UXOs.  An analytic signal map of the separated magnetic anomaly of UXOs was provided to enhance locations of the UXOs in the studied field. Also, a combination of the analytic signal and the Euler deconvolution methods were used to estimate the depth of three buried UXO targets in the study area indicating a high sensitivity of the estimated parameter to the noise level.