IMAGING OF FLUCTUATING DOMAIN SHAPES - METHODS OF IMAGE-ANALYSIS AND THEIR IMPLEMENTATION IN A PERSONAL COMPUTING ENVIRONMENT

Methods for the analysis of digital images have been implemented on a personal computer, augmented with a 32-bit floating-point array processor located in its backplane. They are applied to derive a quantitative description of domain shapes and patterns, as well as dynamic shape fluctuations, based on series of images recorded on video disk or tape. This description relies on a variety of algorithms, in part developed in the context of optical character recognition, to extract and encode domain boundaries, and to evaluate moments and power spectra of momentary boundary configurations. Time-series analysis of principal axes directions and of selected, active Fourier modes in terms of suitable correlation functions permits the study of dynamic phenomena. Examples are discussed to illustrate the scope and versatility of the analytical procedures whose wide applicability in laboratory environments is facilitated by their implementation on a personal computer platform.