Optimization of Direct Slicing Process Using DICOMS for Additive Manufacturing

Abstract

In this article, processing of Computed Tomography (CT) scan data using the principle of direct slicing is discussed. Various segmentation algorithms are compared for fetching the region of interest. Otsu thresholding automates the selection of global threshold value than any intermediate thresholding algorithm and hence it is used for segmentation in presented research work. Further respected contour has been created by selecting the most optimum edge detection technique out of Sobel, Robert, Prewitt and Canny. Canny edge detection produces more sharper edges. Thus, an improvised algorithm of Canny is used which is also relevant for preservation of weak edges. The contour is compared to its Standard Tessellation Language (STL) file, generated in MIMICS software. The coordinates of the two models are superimposed at a given slice height and compared for the deviation of sliced area and perimeter of the two models. Results show that an improvised method of direct slicing with accurate segmentation and edge detection techniques can obtain the precise sections, improve the part’s shape and maintain the outward direction of normal and can therefore overcome the inadequacies of the STL.