| Abstract Scope |
Within EBSD analysis it is known that pattern quality is critical for successfully indexing patterns. Perhaps not as well understood, pattern quality strongly effects the accuracy of the orientation measurement, regardless of indexing method. The non-local pattern averaging and re-indexing (NLPAR) algorithm was developed to significantly improve pattern quality by averaging pattern signals that contain similar information, preventing signal mixing across high angle grain boundaries. It was less clear what effect this might have on low-angle boundaries. Here, using pattern simulation combined with traditional Radon indexing and Dictionary Indexing, we will show that this edge preserving feature of NLPAR is robust to small angle boundaries with misorientations less than 1°, and in some cases as small as 0.1°. Additionally, we will show that the increase in pattern quality using NLPAR improves indexing accuracy enough that measurements of kernel average misorientation are significantly improved in nearly data collection cases. |