Radiographic Least Squares Fitting Technique Accurately Measures Dimensions and X-Ray Attenuation

Abstract

In support of stockpile stewardship and other important nondestructive test (NDT) applications, we seek improved methods for rapid evaluation of materials to detect degradation, warping, and shrinkage. Typically, such tests involve manual measurements of dimensions on radiographs. We seek to speed the process and reduce the costs of performing NDT by analyzing radiographic data using a least-square fitting technique for rapid evaluation of industrial parts. In 1985, Whitman, Hanson, and Mueller have demonstrated a least-square fitting technique that very accurately locates the edges of cylindrically symmetrical objects in radiographs. [1] To test the feasibility of applying this technique to a large number of parts, we examine whether an automated least squares algorithm can be routinely used for measuring the dimensions and attenuations of materials in two nested cylinders. The proposed technique involves making digital radiographs of the cylinders and analyzing the images. In our preliminary study, however, we use computer simulations of radiographs.