Analysis of Intelligent Pig Data by Pattern Recognition

To evaluate pipeline condition, we should have several kinds of data for reference. This data bank is the key to analyze the inspection data. Therefore pigging service companies have stored it up and made their own analysis programs with it for decades. This programs and analyzing processes are not released because it is their main technique. Historically, analyzing programs are based on Pattern Recognition (PR) techniques. This techniques are important to intelligent systems for post processor because the science of PR concerns the description or classification of measurements, To perform this process, it needs to extract characteristics from data first. Recently, this science is being combined neural networks which could be tuned to recognize various features. This technical review is concerned with data features and analysis of Geometry and MFL(Magnetic Flux Leakage) survey data. From the understanding typical features, it could be possible to interpret data and evaluate the condition of pipeline