Image Analysis of Stromatolite Fabric Using a Spatial Frequency Power Spectrum Analysis System

Spatial frequency power spectrum analysis is combined with statistical data processing in a morphological study of stromatolite fabric. Fourteen photographs of stromatolite fabrics from the holotypes of published taxa and 10 line drawings simulating stromatolite fabric were used for the input images; the output data were plotted to make curves of power spectra for each input image, and the data from the 14 stromatolite samples were subjected to Q-mode hierarchical cluster analysis. The diagrams of the power spectrum curves allow similarity-dissimilarity among the analyzed images to be readily perceived by visual inspection. Cluster analysis using different data matrices yielded dendrograms that provided quantitative evaluation of similarity-dissimilarity in fabric features between the 14 stromatolite taxa. The results clearly show the efficiency and utility of this approach in recognition and classification of complex stromatolite fabric patterns, and this method may have a wider application for comparati...