Shape preferred orientation of object populations: automatic analysis of digitized images

An automatic technique is presented for systematic shape preferred orientation (SPO) analysis of crystal fabrics in rocks. This technique is based on digitization of an image with a CCD camera and separation of an object population from its matrix. The image is analysed globally using a rotating grid that is superimposed on the square pixel grid of the computer screen. The parametric method is based on counting intercepts in all directions. The interceps rose provides information on both the orientation and strength of the SPO. Problems of hardware-related counting noise anisotropy are avoided using an original method of filtering the intercept function with a linear, sliding mask on the counting grid. The technique was tested both on single objects of different shapes and on different populations of identical objects. A very high degree of accuracy was obtained for orientation measurements. Fabric shape parameters are introduced which are specific to this intercept technique; they compare favourably with shape parameters used in axial fabric analysis.