Entropy filter for anomaly detection with eddy current remote field sensors

Abstract

Abstract-We consider the problem of extracting a specific feature from a noisy signal generated by a multi-channels Remote Field Eddy Current Sensor. The sensor is installed on a mobile robot whose mission is the detection of anomalous regions in metal pipelines. Given the presence of noise that characterizes the data series, anomaly signals could be masked by noise and therefore difficult to identify in some instances. In order to enhance signal peaks that potentially identify anomalies we consider an entropy filter built on a posteriori probability density functions associated with data series. Thresholds based on the Neyman-Pearson criterion for hypothesis testing are derived. The algorithmic tool is applied to the analysis of data from a portion of pipeline with a set of anomalies introduced at predetermined locations. Critical areas identifying anomalies capture the set of damaged locations, demonstrating the effectiveness of the filter in detection with Remote Field Eddy Current Sensor.

A Model-Free Kullback–Leibler Divergence Filter for Anomaly Detection in Noisy Data Series

“…6, where holes and blind holes are marked by black and red circles, respectively. It is worth mentioning that this test setup is compliant with the standards followed in the gas pipeline nondestructive inspection industry [11].…”

“…An entropy filter has been proposed in [10] to capture the anomaly in a magnetic field. An extension of the entropy filter presented in [11] is proposed in [12] to detect defects on the surface of gas pipelines The filter maps the raw data to a local information entropy space by assigning to each spatial location the entropy calculated from a neighborhood data set. A binary hypothesis test is then applied to partition the space between anomalous and background data series.…”