Article ID Journal Published Year Pages File Type
7123831 Measurement 2016 6 Pages PDF
Abstract
This paper proposes a new approach with digital image processing techniques to determine the number of graphite nodules on samples of nodular cast iron by following the NBR 6913 standard. Counting errors and excessive inspection time are common problems if the procedure is performed without the assistance of computational systems. To overcome these drawbacks, we propose here an algorithm for graphite nodule segmentation based on the Level Set technique. The proposed approach and two other computational methods, Watershed and Region Growing, are compared with the results given by experts using optical microscopy (OM). The results of the proposed method were closer to those of the experts using OM than the other two computational methods. The proposed method presented greater accuracy and faster execution time than the traditional method by visual inspection.
Related Topics
Physical Sciences and Engineering Engineering Control and Systems Engineering
Authors
, , , ,