Application of Active Contours on Metallurgical Microstructures

Employing computer algorithms for finding and outlining the boundaries between phases or grain boundaries (referred to as edge detection) is a widely used technique employed as an intermediate step in microstructure analysis. Having an outlined region of interest enables the user to extract data about the region or use it in the reconstruction of three-dimensional models. Because traditional edge detection relies on a user-selected threshold value, the results often are subjective. Furthermore, traditional edge detection frequently results in outlines that are incomplete, requiring additional processing steps, such as edge linking and spur removal. Active contouring is an edge-detection-based technique that typically yields results superior to that of a traditional edge detector. High noise tolerance and built-in flexibilities of active contours make the technique desirable to use across a broad range of applications. Although initially used for meteorological applications, several uses of active contours are presented for metallurgical microstructural data obtained using optical and electron microscopies to demonstrate robustness and the range of applications that can employ active contours.