Estimating Multidomain Wall Motion in Polycrystalline Materials

Abstract

This article presents an approach to estimate multidomain wall motion obtained on video from a grain of polycrystalline material. The first step consists of pre-processing the image to reduce noisy artifacts. Then, an object detection technique is applied to the pre-processed images to locate the position on the domain wall. The location data collected from each image are used to display the estimated movement of domain walls from all input video frames. The experimental results allow the improvement of models based on the behavior of the individual movement of each wall.