The present invention provides a marked
watershed segmentation-based steel
grain boundary automatic extraction method. The method mainly comprises the steps of 1, subjecting a metallographic image to pre-treatment, including
gradation conversion, bilateral filtering and
noise reduction; 2, segmenting the pre-treated
image based on the marked
watershed segmentation
algorithm, including the operations of obtaining a
gradient magnitude image, marking foreground objects, marking background objects, correcting the
gradient magnitude image, and conducting the
watershed segmentation; 3, conducting the automatic extraction of a
grain boundary mainly by using a canny operator. According to the technical scheme of the invention, the marked watershed segmentation
algorithm is applied to the steel crystalline grain
separation process, so that the over-segmentation phenomenon generated by the conventional watershed segmentation
algorithm is greatly improved. The extraction accuracy of a high-precision crystalline boundary is effectively improved. The defect in the prior art that the conventional crystalline boundary is poor in effect can be overcome. Meanwhile, the method is wide in application range, high in accuracy and fast in operation speed.