Twin and Scratch Detection and Removal in Micrograph Images of Inconel TM 718

Grain size of forged nickel alloy is an important feature for the
 mechanical properties of the material. For fully automatic grain
 size evaluation in images of micrographs it is necessary to detect
 the boundaries of each grain. This grain boundary detection is influenced
 directly by artifacts like scratches and twins. Twins can be seen
 as parallel lines inside one grain, whereas a scratch can be identified
 as a sequence of collinear line segments that can be spread over
 the whole image. Both kinds of artifacts introduce artificial boundaries
 inside grains. To avoid wrong grain size evaluation, it is necessary
 to remove these artifacts prior to the size evaluation process. For
 the generation of boundary images various algorithms have been tested.
 The most stable results were achieved by grayscale reconstruction
 and a subsequent watershed segmentation. A modified line Hough transform
 with a third dimension in the Hough accumulator space, describing
 the distance of the parallel lines, is used to directly detect twins.
 Scratch detection is done by applying the standard line Hough transform
 followed by a rule based segment detection along the found Hough
 lines. The results of these operations give a detection rate of more
 than 90 percent for twins and more than 50 percent for scratches.