Color Image Processing for Space Shuttle Main Engine (SSME) Flaw Detection

Health monitoring of the SSME includes manual and robotic inspection of main injector components. Current automated visual tasks include scratch, crack, and erosion detection for oxidizer posts and injector faceplate. This paper discuses extending current visual tasks into the development of color segmentation and clustering algorithms to isolate and extract distinct discoloring flaws. A color image is represented by the combination of the primary colors red, green, and blue (RGB). The processing of the primary colors, individually or in groups, and the transformation into other color domains will be discussed. Finally , the applicability of color processing to the segmentation of color flawed SSME images will be discussed. @ Copyright 1993 by Bryan Everding and William G. W ee. Published by the University of Cincinnati with permission.