Self-similar and self-affine characteristics of microstructural images of HSLA steel

Abstract

This paper reports a procedure based on discrete Fourier transformation adopted for verifying statistical self-affinity and self-similarity in microstructural images of high strength low alloy steel. The procedure is first applied on synthetic images based on fractional Brownian motion to show the degree of self-similarity in them and finally, applied on microstructural images of three different magnifications. Effects of length scales on power spectral estimates of the microstructural images captured by scanning electron microscopy have been discussed. Importance of appropriate image resolutions and characteristic length for fractal based quantification of microstructural images is highlighted