A MICROCOMPUTER-BASED SYSTEM FOR QUANTITATIVE PETROGRAPHIC ANALYSIS

An imaging system based on a videocamera and frame grabber is described which is capable of capturing and analyzing composite images. Individual images are captured interactively, this permits manipulation of the illumination to emphasize selected features of interest in sequentially captured images. Data from the sequential images are accumulated to form a synoptic image, which allows analysis to proceed in a manner which emulates the techniques of manual, polarized light microscopy. The effects of rotating a thin section in plane and crossed polarized light can be simulated so that mineral boundaries can be detected across which there is a lack of contrast at some orientations. The imaging system implements algorithms for digital filtering and boundary identification and incorporates facilities for image editing. Mathematical functions are provided for the interpolation of boundaries which are not detected in their entirety, in a way analogous to visual interpretation. The image data are written to 256-color PCX image files which can be manipulated by other software or transmitted electronically. The locations of the boundaries of the features of interest are available as lists of (x, y) coordinates and as chain codes. From these the size, shape, and spatial parameters are computed. In addition, the gray-level and segmented images are used to obtain texture information. The imaging system is illustrated by application to the analysis of grain boundaries, modal composition, and grain shapes in petrographic thin sections. The analytical results are compared with results obtained by traditional petrographic analyses.