Genetic algorithms as applied to line profile reconstruction in the scanning electron microscope

Various procedures based on scanning electron microscope have been applied to the microtopographical study of solid surfaces. This paper presents a reconstruction method of surface morphology with genetic algorithms from secondary electron signals in the scanning electron microscope. In comparison with stereometric and multiple-detector methods, the proposed system requires only conventional secondary electron detectors for a live profile reconstruction on one direction. To deal with statistical fluctuations in the gray level of each pixel on the experimental line scan, the reduced chi-square distribution is sed as the objective function and a scheme for minimizing the number of vertexes in the reconstructed surface profile is adopted. To test the capability of this method, a surface profile is successfully reconstructed from a line scan along the center of a latex substrate.