Conventional inline inspection (ILI) tools use odometer wheels to determine the location of identified defects. On top of that, above ground markers (AGMs) are used to confirm and potentially correct for odometer wheel slippage. Free-floating unconventional ILI tools use information from a variety of sensors to accurately locate defects. Accurately identifying joints is a prerequisite for localization and automatic identification of the joints is key for a cost-effective inspection.
This paper focuses on automating the joint identification process with a neural network. This article will describe deep learning strategies for discrete feature identification and segmentation in time series data, how those strategies are increasing data processing efficiency, current accuracy and limitations, and normalization strategies for data from multiple sensors.