Automatic Feature Matching Based In-Line Inspection Data Alignment (ILIDA) Method for Chinese Terms

In-line Inspection (ILI) is an important reference for pipeline defects growth assessment. To determine the growth of defects, the features acquired through the ILI reports at different times need to be matched. However, the same features of defects in ILI reports described by different detectors are often presented in different Chinese terms, restricting automatic feature matching during the alignment process and making manual matching time consumption. In this paper, an ILI Data Alignment (ILID) method based on automatic feature matching is constructed. In the method, the calculation method denoting Chinese semantic similarity is used to determine the degree of conformity between the various terms in ILI reports and template terms in III A to achieve automatic feature matching for Chinese terms. With reference of English ontology based semantic similarity methods, the depth weight is added as a parameter, and the authoritative semantic test set, Miller & Charles (MC), is used to obtain the weight values of different layers. At last, with the aid of computer and the present research results on procedure of ILI data alignment, the ILIDA method for Chinese terms is achieved. The performance of the improved semantic similarity method is compared with other methods in the MC set, as well as the actual Chinese terms from ILI. Consequently, the improved method achieves the most substantial Pearson correlation coefficient value compared to other methods in MC set and it can distinguish the Chinese terms from the ILI data that other methods are unable to accomplish. Furthermore, the application of ILIDA is verified on a serving pipeline with two times ILI at 2013 and 2015 respectively, which achieves ILI data alignment automatically and accurately compared with manual results.