CLASSIFICATION OF CRYSTAL SHAPE USING FOURIER DESCRIPTORS AND MATHEMATICAL MORPHOLOGY

The performances of two image analysis methods for the classification of some randomly selected KCl crystals from a crystallization experime nt into four two-dimensional classes (nearly circular, square, rectang ular and irregular) are compared. The first method uses the first 15 F ourier descriptors of the angular bend as a function of are length of the periphery of the particles, whereas the second method is based on a combination of seven geometrical and morphological parameters of the crystals using a commercially available image analysis system (Visilo g, Noesis, Orsay, France). The feedforward neural network with backpro pagation learning algorithm and discriminant factorial analysis (STATl ab, SLP, Ivry sur Seine, France) were found to classify the crystals w ith similar success.