Author/Editor		Cruz, Daniela MR
Title		Improvements of mechanical and metallographic techniques based on an image processing system
Type		članek
Source		Image Anal Stereol
Vol. and No.		Letnik 20, št. Suppl 1
Publication year		2001
Volume		str. 239-44
Language		eng
Abstract		Image Processing can be applied in characterization and quality control of materials. This technique can be used to recognize and/or measure shapes, as well as to analyse properties as texture, roughness, colour, and so on. With textured areas, segmentation by threshold is not applicable because does not lead to reproducibility. To process images with these characteristics, we can apply a filter that uses chromatic means of the areas to be identified and their standard deviations; transformation of medley areas in homogeneous ones allows easy quantification. Morphological parameters can be extracted from segmented images with a contour extraction and skeletonisation technique allowing easy statistical studies.
Descriptors		METALLURGY MATERIALS TESTING IMAGE PROCESSING, COMPUTER-ASSISTED