TY - CONF A1 - Heideklang, René A1 - Shokouhi, Parisa T1 - Quantitative multi-modal NDT data analysis T2 - QNDE conference 2013 - Review of progress in quantitative nondestructive evaluation N2 - A single NDT technique is often not adequate to provide assessments about the integrity of test objects with the required coverage or accuracy. In such situations, it is often resorted to multi-modal testing, where complementary and overlapping information from different NDT techniques are combined for a more comprehensive evaluation. Multimodal material and defect characterization is an interesting task which involves several diverse fields of research, including signal and image processing, statistics and data mining. The fusion of different modalities may improve quantitative nondestructive evaluation by effectively exploiting the augmented set of multi-sensor information about the material. It is the redundant information in particular, whose quantification is expected to lead to increased reliability and robustness of the inspection results. There are different systematic approaches to data fusion, each with its specific advantages and drawbacks. In our contribution, these will be discussed in the context of nondestructive materials testing. A practical study adopting a high-level scheme for the fusion of Eddy Current, GMR and Thermography measurements on a reference metallic specimen with built-in grooves will be presented. Results show that fusion is able to outperform the best single sensor regarding detection specificity, while retaining the same level of sensitivity. T2 - Quantitative nondestructive Evaluation conference 2013 (QNDE) CY - Baltimore, MD, USA DA - 2013-07-21 KW - Data fusion KW - Multi-sensor KW - Reliability PY - 2013 SP - 1 EP - 5(?) AN - OPUS4-30058 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -