Texture characterization of atmospheric fine particles by fractional Brownian motion analysis

This study describes some preliminary results of a new approach which seeks to develop a method by which the surface structure of individual airborne fine particles can be defined and classified by a specific parameter, Hurst coefficient (H), which is directly related to the fractal dimension. The proposed method uses the fractional Brownian motion analysis to parameterize the microscopic images recorded from five particles with various components (Mugica et al., J. Aerosol. Sci., 33 (2002) 91–102). On the microscopic scale, the particle was segmented from its background via detection of H values of the selected regions of images. Furthermore, the central texture of the above five particles was parameterized by a set of H values. The range of H values was between 0.27 and 0.56, according to the fluctuation of the particle surface. Finally, some future applications of texture characterization of fine particles, which are currently under development, are briefly discussed.