Analysis of wear particles using the Radon transform

An image processing technique is presented which automatically selects and characterises only those protrusions on the wear particle most likely to be active in the process of abrasion. The image data are not analyzed directly. Information concerning angularity is extracted from the distribution generated by the transformation of the binary edge map. The method offers an analytical definition of a corner and its apex, a quantitative estimate of angularity, robustness in the presence of missing boundary points, and the potential to run automatically. The theory supporting the methodology is outlined and illustrated using simulated examples. The results of using the technique to process real image data are also presented