Electron Backscatter Diffraction (EBSD) provides a useful means for characterizing microstructure. However, it can be difficult to obtain index-able diffraction patterns from some samples. This can lead to noisy maps reconstructed from the scan data. Various post-processing methodologies have been developed to improve the scan data generally based on correlating non-indexed or mis-indexed points with the orientations obtained at neighboring points in the scan grid. Two new approaches are introduced (1) a re-scanning approach using local pattern averaging and (2) using the multiple solutions obtained by the triplet indexing method. These methodologies are applied to samples with noise introduced into the patterns artificially and by the operational settings of the EBSD camera. They are also applied to a heavily deformed and a fine-grained sample. In all cases, both techniques provide an improvement in the resulting scan data, the local pattern averaging providing the most improvement of the two. However, the local pattern averaging is most helpful when the noise in the patterns is due to the camera operating conditions as opposed to inherent challenges in the sample itself. A byproduct of this study was insight into the validity of various indexing success rate metrics. A metric based given by the fraction of points with CI values greater than some tolerance value (0.1 in this case) was confirmed to provide an accurate assessment of the indexing success rate.

The Hough transform coupled with triplet indexing [1] has proven very robust in the automated determination of crystallographic orientation from EBSD patterns. However, when the incoming patterns are noisy the ability of these routines to correctly determine the orientation correctly can be compromised. This leads to noisy maps reconstructed from the orientation data, an extreme case is shown in Figure 1(a). A variety of clean-up routines have been developed to find isolated points and assign them orientations generally based on some correlation to the orientations of the surrounding pixels in the scan grid. As the original scan data becomes more and more noisy these clean-up routines [2] tend to produce more and more artifacts. A new approach to re-indexing of the stored patterns has been developed that recovers both non-indexed and mis-indexed points while reducing the number of artifacts. A confirmation of this capability can be observed in Figure 1 (b). This approach is based on averaging the pattern at a scan point with the patterns of the point's neighbors.The details of this re-indexing approach and its limitations will be presented along with results obtained on four different samples: 1) A nickel alloy sample where increasing amounts of artificial noise have been added to the patterns. 2) A nickel alloy sample with varying EBSD camera settings producing patterns with different degrees of noise. 3) A heavily deformed magnesium sample. 4) Fine grained steel wire. Figure 1 shows the results obtained from the first sample with 20 iterations of applied Poisson noise. Figure 2 shows a summary of the results obtained on samples with varying degrees of camera noise. The results show the positive impact of the pattern averaging on re-indexing although not as dramatic as that achieved on the artificially noisy patterns. The averaging of the pattern at a point with the patterns of neighboring points is similar to increasing the size of the interaction volume. This effect is evident on the results obtained on the heavily deformed magnesium sample. On this sample the averaging led to a loss of fine structure in the resulting map. However, by judicious combining of the original results with the results obtained after averaging enables the advantages of the conventional indexing approach and pattern averaged approach to be realized.The results obtained on samples (1) and (2) were compared against a new indexing method based on matching of the experimental patterns with a dictionary of simulated patterns [3] as a baseline. This comparison confirmed the accuracy of the results as well as verified the indexing success metric based on the confidence index [4].