Orientation estimation of industrial parts for automated inspection

Orientation estimation is an essential step in many machine vision problems. A variety of inspection tasks require the coordinate system of an object model to be aligned with the coordinate system of a set of observations before the actual inspection judgements can be made. In this paper, a procedure is presented to estimate the object's orientation using spectral energy distribution in the Fourier domain. The performance of this technique is determined by conducting simulation experiments on a set of images containing industrial parts under varying conditions of noise.<
>