Edge detection and measurement of nose radii of cutting tool inserts from scanned 2-D images

摘要

Flatbed scanners have significant potential in metrological applications as a low-cost alternative for acquiring high-resolution digital images. This work proposes an improved edge detection algorithm applied to the 2-D images of cutting inserts captured using the common flatbed scanner. The internal lighting provided in the scanner is used as the source of illumination. A localized edge detection method is proposed using morphological processing and the gradient of the gray level profile across the edge is used to locate the edge pixels of the insert. The edge pixels are then used in the nose radius determination using least squares circle fitting. A comparison of the proposed method with the existing global edge detection illustrates the effectiveness of the proposed method. Radius measurement using the simple least squares circle fitting performed up to sub-pixel accuracy illustrates the effectiveness of the measurement using the edge pixels generated using the proposed algorithm. A maximum absolute deviation of about 5% was obtained in the radii values determined using the scanner compared to those measured using the Alicona InfiniteFocus 3-D Metrology System, as well as the nominal radii specified by the manufacturer of the cutting tool inserts.